Safety Culture and Safety Management Within The Norwegian-Controlled Shipping Industry State of Art, Interrelationships, and in Uencing Factors

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/234056603
Safety culture and safety management within the Norwegian-controlled

shipping industry State of art, interrelationships, and influencing factors
Thesis · October 2011
CITATIONS READS
21 1,125
1 author:
Helle Oltedal
Western Norway University of Applied Sciences (HVL)
25 PUBLICATIONS 150 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Safety culture and safety management within the Norwegian-controlled shipping industry View project
RISKOP Managing Risk in Offshore Operations View project
All content following this page was uploaded by Helle Oltedal on 16 May 2014.
The user has requested enhancement of the downloaded file.

Safety culture and safety management
within the Norwegian-controlled
shipping industry
State of art, interrelationships, and influencing factors
By
Helle A. Oltedal
Thesis submitted in fulfillment of

The requirements for the degree of
PHILOSOPHIAE DOCTOR
(PhD)
Faculty of Social Sciences
2011
University of Stavanger
N-4036 Stavanger
NORWAY
www.uis.no
©2011 Helle A. Oltedal
ISBN: 978-82-7644-464-3
ISSN: 1890-1387
ii
Preface
My interest in the shipping industry started while working on a project in

which I calculated the risk for ships making contact with high-rise
constructions located shore side. When searching for background information,
I realized that very little research concerning safety related to human factors
and seafarers’ welfare existed. I found evidence of seafarers working under
unacceptable working conditions, sailing on “rust buckets,” jeopardizing their
own safety as well as that of the vessels—in some cases afraid to come
forward out of fear of losing their jobs. At the same time, working at sea was
ranked as one of the most dangerous occupations in the world. This situation
triggered my interest in the field and the research presented in this thesis.
I need to thank several people for their contribution during the process of
completing this PhD. First of all, I would like to thank my supervisor Ole A.
Engen for his encouragement, help, and support throughout this process. I
would also like to thank my colleague Doctor David McArthur for helping me
with statistical problems as well as useful comments and suggestions while
writing the articles. Dean A. Rune Johansen has also been a great support
during my research. Thanks to his engagement in the maritime industry, I was
able to receive funding for five more years of research—something that I
appreciate tremendously. I would like to mention Captain Vigleik Storesund,
who has been a great help in enabling me to understand the maritime industry
and life at sea. Last, but not least, Chief Officer Johanne Marie Trovåg,
Professor Knud Knudsen, and Professor Preben Hempel Lindøe all made
tremendous efforts in reading through my final work.
iii
Furthermore, I would like to thank the Norwegian Shipowners’ Association
for helping me get in touch with their members. Although they cannot all be
mentioned by name, I wish to thank the shipping companies who volunteered
to participate and all the seafarers who—despite a hectic working situation—
found time to contribute with their views. In addition several companies have
enabled me to sail on their vessels and in other ways participate in their daily
operations.
I would also like to thank Dr. Nick Bailey and Professor Helen Sampson at
the Seafarers International Research Centre (SIRC) in Cardiff, Wales, for
making my research stay there possible.
My apologies to all of you who have not been mentioned. I cannot mention
you all, but no one has been forgotten. During my research, I have gained
valuable and useful contributions from many people of different nationalities,
including academics, practitioners, seafarers, and shore personnel working in
all parts of the shipping industry.
Finally, I would like to thank my family for their encouragement and support.
I also want to apologize to my two young children who have had a mum
preoccupied with research for such a long time. I have spent long periods
away from home and, when staying at home, I have been mentally absent. I
hope I can now make up for the lost time. It would not have been possible to
complete this thesis without the support of my incredibly patient and
supportive husband Mikal. Thus, I dedicate this thesis to them—Mikal and
our two exceptional and wonderful children, Jørgen and Elene.
Haugesund, April 2011, Helle A. Oltedal
iv
Summary
This research focuses attention on safety challenges within the Norwegian

shipping industry. A status picture of the shipboard safety culture and the
interrelationships with safety management and organizational factors is given.
Three research questions are explored: (1) What characterizes safety culture
and safety management within the shipping industry? (2) What is the
relationship between safety culture and safety performance within the
shipping industry? (3) What characterizes shipping companies’ application of
the safety management concept? In order to explore these research questions,
four aims were defined to guide this work: (1) to outline and discuss the
application of safety culture and safety management within merchant
shipping; (2) to outline and discuss relevant theories of safety culture and
safety management and analyze the relationship between safety culture and
safety management; (3) to support the use of a methodological framework for
the assessment of safety culture in relation to safety management; and (4) to
assess safety culture within merchant shipping and analyze the relationship
with safety management and actual performance. The research questions are
further examined and specified in six journal articles.
The thesis is divided into two main parts. Part I includes the overall
framework in relation to research aims. Part II presents the six journal articles.
In part I, chapter 1, a general introduction and a status picture of risk, safety
management, and safety culture within the shipping industry are presented,
which gives reason for the research aims and questions introduced in the
chapter. Chapter 2 outlines the safety responsibilities within the industry at
the international, national, and company levels. Emphasis is placed on the
v
International Safety Management (ISM) Code, which provides the minimum
standards and guidelines for operational safety management. Chapter 3
provides theoretical clarification and framing with regard to safety culture and
safety management. This chapter also introduces a general working model
used in the studies of safety culture and safety management in this thesis.
Chapter 4 presents the methodological approach. The thesis builds upon a
mixed method approach where both qualitative and quantitative techniques
are used. The main results are briefly summarized in Chapter 5, followed by a
discussion in Chapter 6 and concluding remarks in Chapter 7. The concluding
remarks concern study limitations, implications, and suggestions for future
research.
The thesis draws upon theory from both the socio-anthropological and
organizational psychological directions. In accordance with the organizational
psychological perspective, a survey was carried out. A safety culture
questionnaire developed by Studio Apertura, a constituent centre of The
Norwegian University of Science and Technology (NTNU), in collaboration
with the Norwegian DNV and the research institution SINTEF was used. In
total, 1,574 questionnaires were distributed to 83 tanker and bulk/dry cargo
carriers, with 1,262 being returned from 76 of the vessels. The vessels were
initially randomly selected from the Norwegian Shipowners’ Association
member list, but as participation was voluntary, some withdrawal occurred.
Statistical analysis involves descriptive statistics, factor analysis, regression
analysis, and structural equation modeling. The statistical survey results were
complemented by qualitative data obtained through document studies, case
studies including two tanker companies and two bulk/dry cargo companies,
vi
interviews, participating observations and field studies at sea, and
participation in other maritime forums.
The study results indicate several deficiencies in all parts of a traditional

safety management system defined as: (1) the reporting and collection of
experience data from the vessel; (2) data processing, summarizing, and
analysis; (3) the development of safety measures; and (4) implementation.
The underreporting of experience data is found to be a problem, resulting in
limitations related to the data-processing process. Regarding the development
of safety measures, it is found that the industry emphasizes the development
of standardized safety measures in the form of procedures and checklists.
Organizational root causes related to company policies (e.g., crewing policy)
is to a lesser degree identified and addressed.
The most prominently identified organizational influential factors are the

shipping companies crewing policy, which includes rotation systems, crew
stability, and contract conditions, and shipboard management. The
companies’ orientation toward local management, which includes leadership
training, educational, and other managerial support, are also essential. The
shore part of the organization is identified as the driving force for
development and change in the shipboard safety culture. Thus, safety
campaigns should to a larger degree include and be directed toward shore
personnel.
vii
List of articles in thesis:
Article 1
Oltedal, H. A., & Engen, O. A. (2009). Local management and its impact on
safety culture and safety within Norwegian shipping. In S. Martorell, C.
Guedes Soares & J. Barnett (Eds.), Safety, Reliability and Risk Analysis:
Theory, Methods and Applications (pp. 1423-1430). London: Taylor &
Francis Group.
Article 2
Oltedal, H. & Wadsworth, E. (2010). Risk perception in the Norwegian

shipping industry and identification of influencing factors. Maritime Policy &
Management, 37(6), 601-623.
Article 3
Oltedal, H. A. (2010). The use of safety management systems within the

Norwegian tanker industry—Do they really improve safety? In R. Bris, C.
Guedes Soares, & S. Martorell (Eds.), Reliability, Risk and Safety: Theory
and Applications (pp. 2355-2362). London: Taylor & Francis Group.
viii
Article 4
Oltedal, H. A., & Engen, O. A. (2010). Tanker versus dry cargo—The use of
safety management systems within Norwegian dry cargo shipping. In J.M.
Ale, I.A. Papazoglou, & E. Zio (Eds.), Reliability, Risk and Safety (pp. 2118-
2125). London: Taylor & Francis Group.
Article 5
Oltedal, H. & McArthur, D. (2010). Reporting practices in merchant shipping,

and the identification of influencing factors. Safety Science, 49(2), 331-338.
Article 6
Oltedal, H. A., & Engen, O. A. (2010). Safety Management in Shipping—

Making Sense of limited Success. Safety Science Monitor, submitted.
ix
Contents
Preface ............................................................................................................ iii
Summary ..........................................................................................................v
List of articles in thesis: ............................................................................... viii
Contents ............................................................................................................x
Part I .................................................................................................................1
1 General background and introduction ..................................................1
1.1 Safety culture and safety management within shipping.................... 2
1.2 Research aims and research questions .............................................. 5
1.3 Aims of articles ................................................................................. 7
2 Safety responsibilities in maritime industry ........................................11
2.1 The International Safety Management (ISM) Code ........................ 13
2.2 Maritime administrations and responsibilities ................................ 17
The flag state ........................................................................................... 18
The Port State Control ............................................................................ 20
The company and crew management ...................................................... 23
2.3 When the regulatory framework and safety management fails ....... 24
2.4 Safety responsibilities in the maritime industry—a summary ........ 27
x
3 Safety culture and safety management in theory and practice ..........29
3.1 Safety culture as an organizational integrated concept ................... 33
3.2 Organizations, management and cultural change ........................... 37
3.3 Organizational culture and safety management .............................. 41
The Man Made Disaster model ............................................................... 43
Normal Accident Theory ........................................................................ 45
High Reliability Organizations Theory ................................................... 47
Managing risk and safety culture ............................................................ 49
The theory of Practical Drift – why organizations fails .......................... 51
3.4 Safety culture and measurable outcome variables .......................... 53
4 Research methodology ...........................................................................56
4.1 Quantitative research and questionnaire survey.............................. 58
Questionnaire development .................................................................... 60
Survey sample and respondents’ demographics ..................................... 61
Validity and reliability through theoretical conceptualization ................ 62
Validity and reliability through factor analysis....................................... 66
Validity and reliability through scale analysis ........................................ 68
Causal relationship through structural equation modeling ..................... 70
xi
4.2 Qualitative research design ............................................................. 71
Document study ...................................................................................... 71
Case studies............................................................................................. 73
Interviews................................................................................................ 73
Participatory observation ........................................................................ 74
Participation in maritime forums ............................................................ 75
4.3 Applied methods and statistics in articles ....................................... 76
5 Research results .....................................................................................78
5.1 Summary and results of article 1 .................................................... 78
5.7 Causal relationships between components of safety culture ........... 91
5.8 Summing up and presentations of main conclusions ...................... 98
6 Discussion ...............................................................................................99
7 Concluding remarks ............................................................................107
xii
7.1 Methodological limitations ........................................................... 107
7.2 Theoretical limitations .................................................................. 109
7.3 Future research .............................................................................. 109
7.4 Final remarks ................................................................................ 110
8 References .............................................................................................112
Part II ............................................................................................................124
Article 1
Article 2
Article 3
Article 4
Article 5
Article 6
xiii
General background and introduction
Part I
1 General background and introduction
“Never before have so few done so much for so many.” When opening the
Year of the Seafarer in 2010, these brave words—a quotation paraphrased
from one of Winston Churchill’s most famous speeches—were similarly
strikingly declared by Efthimios E. Mitropoulos, Secretary General of the
International Maritime Organization (IMO). Few people seem to understand
the importance of seafarers and shipping in our society. The worldwide
population of seafarers serving in international trade is estimated to be
approximately 1,187,000 people from virtually every nationality. Worldwide,
about 50,000 ships carry about 90% of the world’s trade; thus, these more
than one million seafarers are transporting goods for the benefit for the
world’s population of almost 7 billion. The seafarers’ and shipping industry’s
global importance is commonly highlighted by the phrase “without shipping,
half the world would starve and the other half would freeze.”
The current research is conducted in light of safety challenges within

merchant shipping. During the first five years of the previous decade (i.e.,
2000 to 2005), an average of 18 ships collided, grounded, or caught fire every
single day, and two vessels were sinking every day (Gregory & Shanahan,
2010). Merchant shipping and seafaring are traditionally perceived as a risky
industry—a risk partly induced by its situational characteristics. Work at sea
is demanding as both work and leisure time happens within a small group, at
the same place, for a long period of time and with few possibilities to interact
with the surrounding world. The seafarer’s only alternative whereabouts when
1
at sea is the sea itself, where harsh conditions prevail. Seafarers and their
vessels are constantly exposed to forces beyond their control, such as storms,
freak waves, and strong currents. Being far from port most of the time, the
seafarer must handle critical situations with little or no support from others,
with only their own competence and expertise to rely on.
1.1 Safety culture and safety management within

shipping
Human error is associated with the vast majority of accidents and incidents
within shipping. An estimated 75% to 96% of marine casualties are caused—
at least in part—by some form of human error (Anderson, 2003; Rothblum,
2000; Wagenaar & Groeneweg, 1987). However, within recent safety
management theories, human error is not seen as a cause of accidents and
incidents, but rather as something shaped and provoked by upstream
organizational factors. Thus, human error is not an explanation per se, but
something that needs further explaining (Hollnagel, 2004; Reason, 2001).
Possible explanatory factors may be related to seafarers’ cognitive system
(e.g., human information processing, training, motivation, and fear), social
system (e.g., social pressure, role, and life stress), and situational system (e.g.,
physical stress, environmental stress, and ergonomic aspects), which are all
assumed to be mutually interdependent (Wagenaar & Groeneweg, 1987). It is
also widely accepted that individual factors are inextricably linked to
organizational factors and decisions (Hollnagel, 2004; Reason, 2001; Schager,
2008).
2
Safety at sea is regulated by the UN’s agency for maritime affairs, the
International Maritime Organization (IMO). From the IMO perspective, safety
management and human error are closely intertwined with the industry’s
definition and application of the safety culture concept, regulated through the
International Safety Management (ISM) Code (IMO, 2010a; Lappalainen,
2008; Mitroussi, 2003). The ISM Code, which became mandatory for all
merchant vessels from July 1998 to July 2002, formally introduced the idea of
safety culture in shipping:
The application of the ISM Code should support and encourage the
development of a safety culture in shipping. Success factors for the
development of a safety culture are, inter alia, commitment, values
and beliefs (IMO, 2010a, p. 35).
However, despite the implementation of the ISM Code, recent statistics

indicate that losses are continuing to increase, resulting in a heavy loss of life
and serious damage to the environment (Soma, 2010). The statistics in Figure
1 illustrate the frequency of navigational accidents (collisions, contacts, and
wrecked/stranded vessels) from 1993 to 2009.
3
Figure 1: Navigational accident frequency in relation to the world fleet size,

1993-2009 (Source: Lloyds’ Fairplay, 2010) Fleet size in number of crude oil
tankers over 100,000 dwt, chemical tankers over 10,000 dwt, containers over
20,000 dwt, RoRo cargo over 10,000 dwt, bulk over 50,000 dwt.
As shown in Figure 1, the frequency of serious navigational accidents has

increased significantly since 2002. It is also interesting to note that, since the
first introduction of the code in 1998, none of the subsequent years show
lower accident frequency than before the code was introduced. This statistical
trend raises a fundamental question: Why do we have such an increase in the
accident rates despite the introduction of the ISM Code, emphasis on safety
culture, and lower tolerance for non-conformities? Three possible
explanations have been put forth:
1. Shipping companies’ implementation of the ISM Code and

understanding of safety management are inadequate. The IMO
assessment of the effectiveness of the ISM Code (IMO, 2005)
indicates that implementation has resulted in more administrative
4
work, procedures, checklists, and other means in order to control

human behavior. However, is safety best ensured by controlling and
restricting human behavior?
2. The ISM Code’s underlying theoretical rationale of linear causality is

inadequate. Is it possible to prevent future accidents by learning from
past events? Are there any causal links between near misses, minor
incidents, and major accidents? When dealing with future events
evolving in an unforeseen and complex pattern, are other rationalities
more adequate?
3. The ISM Code’s assumption of a relationship between safety culture

and actual safety performance and outcome is inadequate. What is
organizational safety culture and what determines its relations to
safety management, organizational practices, and safety performance?
This thesis provides an account of these three possible explanations.

Empirical data are collected from the Norwegian controlled liquid and dry
cargo shipping industry for this purpose. The data are analyzed and discussed
in light of theory on safety management and safety culture. The seafarers’
perspective and their operative experiences are emphasized. The analyses and
discussion will be consistent with the scope of the research aims and
questions, as formulated in the following sections.
1.2 Research aims and research questions
Although shipping is known to be a risky industry, surprisingly little research

has been done within this area. In recent years, a few articles and doctoral
5
theses on maritime safety culture and climate have been published (e.g.,
Antonsen & Norges teknisk-naturvitenskapelige universitet. Institutt for
sosiologi og statsvitenskap, 2009; Christophersen, 2009; Ek, 2006; Håvold &
Norges teknisk-naturvitenskapelige universitet. Institutt for industriell
økonomi og teknologiledelse, 2007; Lamvik, 2002; Østreng, 2007). In light of
the discussed situation, it is important to get a better understanding of what
characterizes safety culture within shipping and how shipboard safety culture
relates to safety management and human error. Thus, the following four
research aims have been developed in order to provide direction for this
thesis:
1. To outline and discuss the application of safety culture and safety

management within merchant shipping.
2. To outline and discuss relevant theories of safety culture and safety
management and analyze the relationship between safety culture and
safety management.
3. To give reason for a methodological framework for assessment of
safety culture in relation to safety management.
4. To assess safety culture within merchant shipping and analyze the
relationship with safety management and actual performance.
In order to pursue these aims, a questionnaire survey was carried out within
merchant shipping along with field studies, case studies, interviews, and other
qualitative methods. The population is defined within the Norwegian
controlled shipping industry as liquid tankers and dry cargo carriers above
500 gross ton. Norwegian controlled is defined as vessels owned by
6
Norwegian parties where the owners’ safety management department is

located in Norway. In the study, the seafarers’ perspective is emphasized.
Based on the four research aims previously described, three research

questions were developed for the purpose of the thesis:
1. What characterizes safety culture and safety management within the

shipping industry?
2. What is the relationship between safety culture and safety performance
within the shipping industry?
3. What characterizes shipping companies’ application of the safety
management concept?
1.3 Aims of articles
The thesis includes six separate studies with their own main aims. All aims
for each study are discussed in this section.
1. Local management and its impact on safety culture and safety within
Norwegian shipping:
The first study uses survey data collected in 2006. The data cover all sailing
personnel on 76 Norwegian controlled liquid tankers and dry cargo carriers.
The aims of the study are to:
 Explore and analyze the shipboard characteristics of safety culture;
 Identify which factors affect the shipboard safety culture; and
 Get results in order to set direction for further studies.
7
2. Risk perception in the Norwegian shipping industry and

identification of influencing factors:
This second article makes use of the same survey data. In this study, the
catering personnel and captains are excluded. Risk perception is used as an
indicator for shipboard safety. The aims of the study are to:
 Assess the relationship between risk perception and dimensions of

safety culture; and
 Explore the influence of organizational structural variables.
3. The use of safety management systems within the Norwegian tanker

industry and whether they really improve safety:
The third study involves both quantitative survey data and qualitative data.
The sub-sample dry cargo carriers are excluded. The data and analyses are
organized in accordance with the sub-components and information flow of a
traditional safety management system. The aims of the study are to:
 Describe the status of safety management within the liquid tanker

sector; and
 Identify organizational structural factors that influence the safety

management performance.
8
4. Tanker versus dry cargo regarding the use of safety management

systems within Norwegian dry cargo shipping:
The fourth study is a follow-up of the third, and follows a similar structure
related to the sub-components and information flow of a traditional safety
management system. The study includes both quantitative survey data and
qualitative data. The sub-sample liquid cargo carriers are excluded. The aims
of the study are to:
 Describe the status of safety management within the dry cargo sector;
 Identify organizational structural factors that influence the safety

management performance; and
 Compare current situations between the two sectors—namely, dry and

liquid cargo carriers.
5. Reporting practices in merchant shipping and the identification of

influencing factors:
The fifth study involves quantitative survey data. Both the third and fourth
study identified underreporting of experience data as a substantial problem.
The reporting of experience data is regarded as a main cornerstone in a safety
management system. Thus, the aims of the article are to:
 Assess the relationship between reporting practices and the dimensions

of safety culture;
 Explore the influence of local management; and
9
 Further explore differences between the dry and liquid cargo carrier
sectors.
6. Safety management in shipping and making sense of limited success:
The sixth study involves both quantitative survey data and qualitative data. As
the previous studies (i.e., three through five) point to a substantial weakness in
current safety management practices, the aim of this study is to:
 Explore and identify reasons for the gaps between safety ambitions
inherent in traditional safety management systems and operational
practices.
10
Safety responsibilities in maritime industry
2 Safety responsibilities in maritime industry
This section is, in accordance with specified research aim 1, formulated as

follows:
1. To outline and discuss the application of safety culture and safety

management within merchant shipping.
The development of international trade and shipping in today’s globalized

market has to a large degree determined the regulative structure of the
industry. The international regulative system is of high importance for the
safety of ships and crew sailing the seven seas, as every shipping company is
required to relate to this during daily operations. In order to gain proper
understanding of safety management within shipping, knowledge of the most
important laws and the international regulative framework is necessary. Thus,
some of the historical mainlines and the present situation related to safety
management and the regulatory system will be further presented. An overview
of the international regulatory system, maritime administration, and
conventions (conventions in bold) related to safety management are shown in
Figure 2 (next page).
11
Maritime Administrations
(Flag and Costal States)
United Nations (UN)
United Nations Convention on the

Law of the Sea (UNCLOS)
International Maritime Organization

(IMO)
Convention for the Safety of Life The International Safety

at Sea (SOLAS) Management (ISM) Code
Figure 2: The maritime international regulatory system related to safety

management
As shown in Figure 2, the maritime administrations’ (flag and coastal states)

safety responsibilities are determined by the UN through the Convention on
the Law of the Seas (UNCLOS). Although UNCLOS sets the broad regulative
framework, the task of developing and maintaining workable regulations on
ship safety within this framework is delegated to the UN agency the IMO,
which is now responsible for 35 international conventions and agreements.
For the purpose of this thesis, the most relevant is the Convention for the
Safety of Life at Sea (SOLAS) chapter IX, management for the safe operation
of ships, and the guidelines for SOLAS IX—namely, the ISM Code.
12
2.1 The International Safety Management (ISM) Code
The ISM Code became mandatory for all merchant vessels above 500 gross
tons in two waves, depending upon type of vessel—namely, July 1, 1998, and
July 1, 2002. Until the adoption of the ISM Code, IMO had attempted to
improve shipping safety largely by improving the hardware of shipping (e.g.,
the construction of ships and their equipment). By comparison, the ISM Code
focuses on the way shipping companies are managed. The ISM Code is the
first to provide regulations and guidelines to promote the development of
sound management and operating practices in order to ensure crew safety and
avoid damage to the environment. The shipping industry is known to have a
reactive approach toward safety as the process of regulating the activity has
evolved primarily as a response to maritime disasters. Development of the
ISM Code was also based upon a growing recognition that loss of life at sea
and environmental pollution are influenced by the way in which companies
manage their fleets. Table 1 (next page) summarizes some of the accidents’
precursory to the ISM Code (Anderson, 2003).
13
Table 1.
Accidents Precursory to the ISM Code
1987 Herald of Free Enterprise capsized off Zeebrugge; 190 people lost their
lives.
1987 Donna Paz ferry collided with a tanker in the Philippines; an estimated 4,386
people were killed.
1989 Exxon Valdes ran aground off the coast of Alaska, spilling 37,000 tons of oil
and causing extensive environmental damage.
1990 Scandinavian Star caught fire; 158 people lost their lives.
1991 Agip Abruzzo, with 80,000 tonnes of light crude on board, was in a collision
with the ro-ro ferry Moby Prince off Livorno, Italy. Fire and pollution
occurred, and 143 people died.
1991 Have experienced fire and explosion off Genova, spilling 50,000 tons of
crude oil; 6 people were killed.
1991 The Egyptian ferry Salem Express struck a reef and sank; 470 people were
killed.
1991 Aegean Sea broke in two off La Coruna, Spain, spilling about 74,000 tons of
crude oil; extensive pollution occurred.
1993 Braer driven onto the Shetland Island, carrying about 84,700 light crude oil;
extensive pollution occurred.
1994 Estonia ro-ro passenger ferry sank after the bow door fell off during heavy
weather at sea; 852 people lost their lives.
A common factor appearing in these accidents was human error, which could
be traced back to poor safety management and organizational practice. By
introducing the ISM Code, IMO intended to adopt a proactive approach
14
toward safety, where future accidents should be prevented by learning from

and reflecting upon previous mistakes and experiences.
The ISM Code requires shipping companies to develop, implement, and

maintain a safety management system, which includes the following
functional requirements: (1) a safety policy; (2) instructions and procedures to
ensure safe operations of ships in compliance with relevant international and
flag state legislation; (3) defined levels of authority and lines of
communication between and amongst shore and shipboard personnel; (4)
procedures for reporting accidents and non-conformities with the provision of
the ISM Code; (5) procedures to prepare for and respond to emergency
situations; and (6) procedures for internal audits and management reviews. In
the code’s guidelines, emphasis is placed on near-miss reporting and how to
create an organizational atmosphere in which people are willing to report
accidents and non-conformities by developing a just culture. The concept of a
just culture is also known to be a fundamental element in James Reason’s
theory of safety culture and safety management (Reason, 2001). Moreover, in
order to achieve the development of an organizational safety culture, IMO
identifies three key elements: (1) recognizing that accidents are preventable
by following correct procedures and establishing best practices; (2) constantly
thinking about safety; and (3) seeking continuous improvement. IMO’s
approach and perspective to safety culture is apparently instrumental, where
safety culture is seen as something that may be engineered by an
organization’s structures and control systems in order to produce desired
behavioral norms and accompanying safety outcomes.
Safety management, as described in the ISM Code, is founded on a linear

causality, in which future events are attempted, predicted, and prevented by
15
analyzing past operational experiences. Thus, a critical system requirement is

reliability and accuracy of input data—the experience, near miss, and accident
reports ; as long as the input is reliable, the overall system presupposes the
possibility of developing efficient standardized measures in order to control
operational safety (Kjellen, 2000). One underlying assumption is that serious
injuries and accidents may be prevented by learning from and reflecting upon
incidents with no injury or damage. This idea is frequently illustrated as a near
miss-accident pyramid. However, previous research does not support this
theory (Anderson, 2003). Moreover, IMO recognized that near misses are
underreported (IMO, 2007c), and the input system requirement is not met.
This also provides a reason to question the underlying theory of linear
causality, especially if near misses, small-scale accidents, and more serious
events have the same causal chain (Rundmo, 1996).
In 2005, IMO provided a report assessing the impact and effectiveness of

implementation of the ISM Code (IMO, 2005). Based on the data collected,
IMO concluded that—when the ISM Code and safety cultural development is
embraced as a positive step—tangible positive benefits are evident. It was also
recognized that ISM Code compliance could be made easier through a
reduction in the administrative processes by, inter alia, the reduction of
paperwork, increased reporting of operational experience data, and greater
involvement of seafarers in the development of ISM manuals, the procedural
system, and checklists. In the industry, it seems to be a common
misconception that the ISM Code requires large quantities of paperwork and
administration to function and that ticking boxes and checklists would replace
good training and seamanship (Anderson, 2003).
16
Although the national maritime government is responsible for implementation

of the ISM Code, the coastal state is responsible for enforcement of the code,
and each shipping company has the primary responsibility for safe operations.
However, these responsibilities are challenging as shipping today has
become—more than ever before—a globalized industry. For example, a
vessel may have owners in one state, be registered in a second state, be
chartered by a company from a third state, and be transporting goods whose
owners belong to a fourth state. To make it even more complicated, the vessel
is sailing between ports in different states and is manned with a multinational
and culturally diverse crew, who are managed by a company in yet another
state. These sector-related circumstances have resulted in specific challenges
with regard to the administration and enforcement of international regulations,
as outlined in the following section.
2.2 Maritime administrations and responsibilities
A maritime administration may have two different roles: a flag state and a
coastal state. The coastal states’ responsibility for the enforcement of
international regulations is done through inspections and Port State Control
(PSC) of vessels entering their own coastal territorial waters, regardless of
which flag the vessel is flying (Stopford, 2009). The coastal state may be the
same as the flag state, but this is far from always the rule. Any ship owner is
free to register a vessel in any of the world’s flag states.1 The term flag state
1
The definition of a flag state is not straightforward. A myriad of descriptions of flag states have evolved,
including traditional maritime nation, embedded maritime nation, national flag, classis register, open
register, opportunist register, international open register, international register, closed register, second
register, dependent territory register, offshore register, and flag of convenience (Mansell & SpringerLink,
17
(or administration) is used to refer to a country that maintains a vessel’s

registry. The flag state has the overall responsibility for ensuring compliance
with international regulations. This responsibility encompasses the operation
of the ship, the physical status of the ship, the activities of the ship owners,
and the working conditions of the seafarers. The flag administration, in the
first instance, underwrites the safe operation of those ships under its flag.
The flag state
Each flag state may have a national register, second register, and/or open
register. A national register is reserved to vessels with national ownership.
Second registers, which are additional to national registers, are mostly open
registers. In an open register, ships owned by foreign entities may register.
The creation of second registers is a response to intensified competition in the
market for ship registration. In the early 1980s, the shipping market
experienced a severe depression. Since the late 1980s, a number of states have
created second registers in addition to their first national register in order to
provide some or all of the advantages of an open register as a result of the
economic crisis. A common motivation for establishing such second registers
are to attract shipowners or prevent shipowners from flagging out by
providing other or more relaxed application of the international IMO
regulations (Alderton, 2004). Income in the form of tonnage taxation fee is
also a motivation for some nations to establish a register when they do not
necessarily have the means, will, or competence to meet their responsibilities
as a flag state. According to Alderton (2004), low barriers to entry into the
2009). In this thesis, flag state refers to an open international register where any shipping owner is free to
register a merchant vessel.
18
flag market exist, with minimal start-up cost or time being required. This
situation has led to competition among some maritime administrations,
which—in order to encourage registration of vessels under their flag—permits
less bureaucratic control along with relaxed requirements.
International regulations adopted by IMO intend to provide a harmonized set

of rules for the industry. The previously described situation has resulted in
variations among maritime administrations in performance and application of
the international regulations (Alderton, 2004). Alderton (2004) distinguishes
between three types of administrations:
(1) Regulatory efficient states in which the state seeks to regulate the full
extent of maritime operations.
(2) Regulatory inefficient states; the main distinction between this

category and the first lies in the treatment of labor issues.
(3) Unregulated states, in which the regulatory environment within these

registers is almost non-existent.
Many of today’s safety-related criticisms are related to ship registration and to

which flag the vessels fly. Although substandard shipping is mostly associated
with regulatory inefficient and unregulated states, even administrations
regarded as being regulatory efficient may have defective performance. An
audit of the Norwegian Maritime Directorate, which is regarded as being
regulatory efficient, revealed that the administration does not have, inter alia,
adequate operational control with its own working procedures and that the
administrated regulations are not comprehended in unison, which may result
in misinterpretations and erroneous decisions (Riksrevisjonen, 2010).
19
As long as the flag states benefit from running open registries and shipowners
can benefit from it, the situation will most likely never change. Moreover, the
lack of flag state control, as evident in several countries, has made PSC even
more important.
The Port State Control
In the wake of some major maritime disasters in the European area (e.g.,
accidents with the Erika in 2000 and Prestige in 2002, which both occurred
after the implementation of the ISM Code), it was realized that the PSC and
ParisMOU2 could and should take a more determined stance against
substandard shipping in order to ensure better enforcement of the international
regulations.
The Prestige was a Greek-operated oil tanker, officially registered in the

Bahamas, but with a Liberian corporation registered as the owner. The
2
The ParisMOU (Memorandum of Understanding) on Port State Control is the official document in which
the 27 participating maritime authorities agree to implement a harmonized system of PSC. The MOU
consists of a the main body in which the authorities agree on: 1) their commitments and the relevant
international conventions, 2) the inspection procedures and the investigation of operational procedures, 3)
the exchange of information, and 4) the structure of the organization and amendment procedures. The
current member states of the ParisMOU region are, in alphabetical order, Belgium, Bulgaria, Canada,
Croatia, Cyprus, Denmark, Estonia, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Latvia,
Lithuania, Malta, Netherlands, Norway, Poland, Portugal, Romania, Russian Federation, Slovenia, Spain,
Sweden and The United Kingdom. Following the foundation built by the ParisMOU, several other regional
MOUs have been signed, including the Tokyo MOU (Pacific Ocean), Acuerdo Latino or Acuerdo de Viña
del Mar (South and Central America), the Caribbean MOU, the Mediterranean MOU, the Indian Ocean
MOU, the Abuja MOU (West and Central Atlantic Africa), the Black Sea MOU, and the Riyadh MOU
(Persian Gulf). In this thesis, due to its geographic area, references will be made to the ParisMOU
(http://www.parismou.org/).
20
ownership of the Prestige was unclear, and it was difficult to establish

responsibility for the accident, which resulted in the spill of more than 60,000
tons of heavy fuel oil. Prior to the accident, the Prestige set sail without being
properly inspected, although a previous captain had complained about
numerous structural deficiencies.
The oil tanker Erika also experienced structural failure. Although the
structural failures were visible, the vessel was found to be seaworthy by the
classification society. The Erika was sailing under a Maltese flag and
chartered by a shipping company registered in the Bahamas on behalf of a
French oil company. With regard to the Erika, it was also difficult to establish
responsibilities (CPEM, 1999). The Erika accident resulted in a spill of about
19,800 tons of heavy fuel oil.
Although both accidents involved structural failures, they can also be

characterized as stemming from a non-functional ISM system. In a functional
safety management system, such structural failures should have been detected
and handled appropriately by the shipping company. In the case of Prestige,
the captain had even notified the company about structural deficiencies that
had not been handled properly. In the aftermath of these accidents, the
response from ParisMOU came in form of developing a harmonized vessel
detention policy, guidelines for operational PSC, and others (ParisMOU,
2007). One of the strategies was naming and shaming. Today, all inspection
results and detentions, with detailed information about the company, vessel,
and flag, are registered in a public database (at parismou.org). Information
about banned vessels and “rust buckets” are also made public. A more recent
initiative includes a list for the performance of flag states (MARISEC, 2008;
MARISEC, 2006; Winchester, Alderton, & Seafarers International Research
Centre, 2003). On this list, each flag state is evaluated and ranked based on
21
their performance on certain aspects, as PSC records ships flying their flags
along with the implementation and enforcement of important international
treaties, such as the ISM Code. Flag state performance is then ranked and
placed on a black list (poor performance), grey list (mediocre performance),
and white list (good performance). The black, grey, and white lists for 2009
included a total number of 82 flags, 24 on the black list, 19 on the grey list,
and 39 on the white list (ParisMOU. 2010).
Despite ParisMOU’s intention to make shipping safer, the current inspection

system has inherent weaknesses. ParisMOU’s target is to inspect 25% of all
vessels calling port. The ships are selected based on criteria as previous
inspections reports from the MOU region. A vessel flying a poor performance
black-listed flag is more likely to be selected for an inspection than others.
These criteria for the selection of vessels are understandable; however, they
disregard the fact the some of the ships flying a poor performance flag are
owned by companies that take their responsibilities for the safe operation and
crew welfare seriously. As a result, PSC resources may be used inefficiently.
It is widely known that vessels in bad shape continue to be operational,
without getting caught by the inspection net (Corbett, 2009; Tradewinds,
2007). In order to improve the current system, the ParisMOU introduced a
new inspection system in January 2011, whereby each ship is ranked as high,
standard, or low risk; this will determine the frequency of inspections. These
changes intend to prevent low-risk vessels from being overly inspected in
order to release resources for more frequent inspections of high-risk vessels.
With the implementation of this new system, it remains to be seen if it will
capture those vessels that deliberately avoid inspections.
22
The company and crew management
Although the overall responsibility for ensuring compliance with international
regulations belongs to the flag state, each shipping company has the primary
responsibility for the safety of their ships and crews. The ISM Code requires
shipping companies to develop, implement, and maintain a safety
management system for this purpose. However, safety management is not
only a system property; system efficiency is determined by its human
interrelationships. On board the vessels, the crew is the ultimate asset to
ensure safety at sea, which is dependent on their experience and competence.
The recession of the 1980s brought about several structural changes apart
from flagging out—namely, the establishment of manning and crewing
agencies. Crew management normally involves finding, organizing, paying,
and training crews. In order to survive financially, some companies turned to
managing ships for other owners as a means of utilizing spare management
capacity. Others found it necessary to turn to crew managers in order to hire
cheaper crew in other and unfamiliar parts of the world. This has resulted in
ships being crewed by mixed nationalities working under different contracts
and employment terms. The sudden switch to employing seafarers from
nations without maritime traditions is claimed to result in a reduction in
standards of competence, except from those relatively unusual cases where
the shipowners invest in training.
In addition, within crew management, there is variability in performance. At

one end are those who have become seriously involved in training, some with
their own training facilities and with established systems of testing crew
competence (Alderton, 2004). On the other end are those who do not—nor
intend to—perform such control over crew competence, which is quite a
23
problem, as fraudulent certificates of competence are an issue within the

industry (IMO STW 41/4, 3, 2009). As pointed out by Anderson (2003), one
might question if crew from manning agents takes the companies’ (safety)
goals and objectives to heart due to a lack of ownership and short
employment. When crew management is carried out by an external party, the
shipowner will lose control over assessing and ensuring qualifications,
training, and competence. The shipowner then depends on third-party
qualifications, thoroughness, and follow-up when providing crew.
2.3 When the regulatory framework and safety

management fails
Major accidents are valuable sources of information about the regulatory
framework, organizational practices, and cultures and in which way these
impact safety. In order to illustrate how the regulatory system may fail, the
explosion and sinking of the chemical tanker Bow Mariner is further outlined.
The Bow Mariner was one out of four3 chemical tankers that exploded during
a six-month period between December 2003 and June 2004. The Bow
Mariner case is interesting for several reasons. First, the accident occurred
after the introduction of the ISM Code. Second, the accident investigation of
Bow Mariner indicated that safety culture and poor safety management are
explanatory factors of the accidents. Finally, what happened aboard the Bow
Mariner is a driving force for further amendments of the international
regulations, concerning shipboard leadership and managerial skills (IMO,
2007a; IMO, 2007b). Based on this situation, the Bow Mariner is seen as a
3
The other three were the tankers Sun Venus, Panama Serena, and NCC Mekka.
24
case suitable for the purpose of understanding and exemplification. All factual
information derives from the official accident investigation report (United
States Coast Guard, 2005).
On February 28, 2004, the chemical tanker Bow Mariner exploded and sank
in the seas outside Virginia, United States, causing the death of 21
crewmembers. The Bow Mariner, owned by the Norwegian company Odfjell
Tankers, was flying a Singaporean flag, was operated by the Greek company
Ceres, and was manned by Greek officers and Filipino crew. The vessel had a
valid Safety Management Certificate (SMC). As part of the investigation, the
vessel’s inspection history for a five-year period before the explosion were
reviewed and found to be unremarkable. However, during an internal audit in
June 2003, 25 observations were recorded, including one pertaining to the
failure to complete an enclosed space entry permit and another for failure to
record training. These latter non-conformities were also present during the
accident and pointed to as possible influencing factors to what happened. The
accident investigators point to numerous indications that the ISM Code
requirements were not fully implemented or functional aboard the vessel,
despite apparent documentation of full compliance with the code. These ISM
non-conformities were found to contribute to the accident. Amongst others,
no crew familiarization with the vessel was conducted.
It is also explicitly stated that the shipboard social culture and safety culture
contributed to the occurrence of the accident. This included poor shipboard
management. Aboard the Bow Mariner, the Greek captain had—in
accordance with the company policy—full authority over all personnel. Such
full authority is not unusual aboard a seagoing vessel. However, at the Bow
Mariner the distinctions between Greek and Filipino nationality were
25
remarkable. Filipino officers did not take their meals in the officers’ mess, and
the Filipino crew were given almost no responsibility and were closely
supervised in every task. The Filipino crew were simply doing what they were
ordered to do. As a result, they gained little knowledge about important
aspects of their jobs. The lack of technical knowledge and fear of the Greek
senior officers provide an explanation as to why the Filipino crew did not
question the masters’ unsafe order to open all the empty tanks, which was a
significant breach of normal safe practices for such ships. If the tanks had
remained closed, the explosions would not have occurred. However, as stated
in the company policy, the captain’s orders should never be questioned, and
the failure to obey orders was a reason for disciplinary actions. Investigations
of the accident leave no question that such fear of the ship management or
senior officers can lead to a shipboard culture where safety takes a backseat to
preserving one’s employment. Interviews with crew from another Ceres
vessel indicate that this poor culture was the general rule in the entire Ceres
company.
With the case of Bow Mariner, it is evident that the ISM Code did not
generate the intended outcome—namely, safe operations and a good safety
culture—as a result of reasons not necessarily related to shortcomings in the
ISM Code itself. This situation can also be related to the lack of ability to
reveal the onboard conditions, which are related to coastal administration and
inspections. Despite documentation and certification confirming full
compliance with the code, the accident investigators pointed to numerous
indicators that the code was neither fully implemented nor functional aboard
the vessel. The onboard situation is created by organizational factors (e.g.,
crewing and shipboard management policies). The accident investigators also
regarded commercial pressure as a contributory factor to what happened.
26
2.4 Safety responsibilities in the maritime industry—

a summary
Thus far, the safety responsibilities in the maritime industry have been shortly
outlined and analyzed. The situation is summarized in the following:
 With the UN delegation of authority, IMO is responsible for

developing and maintaining workable safety regulations and laws
regulating ship safety.
 The maritime administrations encompass flag state and coastal state.

The coastal state is responsible for enforcing maritime regulations
while the flag state is responsible for ensuring compliance with
international regulations.
 The shipping companies have the primary responsibility for the safe
operations of ships and crew safety. Safety management is regulated
through the ISM Code, which is developed by IMO.
 The crew is the ultimate asset for ensuring safety at sea. Shipping
companies may ensure safe operations by investing in crew training
and competence and ensuring that crew experience is made use of in
the company safety management system.
When ensuring operational safety, the crew relates to and is influenced by all
these actors and levels of authority. National and international legislations
represent minimum standards. Beyond the minimum standards, each shipping
company determines the crew’s working conditions. On each vessel, the
framework given by the shipping company is moderated by the ship’s
27
management. Safety culture and safety management in theory and practice are
further elaborated in the following chapter.
28
Safety culture and safety management in theory and practise
3 Safety culture and safety management in

theory and practice
This section focused on specified research aims 2 and 3, formulated as
follows:
2. To outline and discuss relevant theories of safety culture and safety

management and analyze the relationship between safety culture and
safety management.
3. To give reason for a methodological framework for assessment of

safety culture in relation to safety management.
The concept of safety culture as a term and an explanatory factor in an

accident investigation was first used by the International Atomic Energy
Agency (IAEA) International Nuclear Safety Advisory Group (INSAG)
following the Chernobyl accident that occurred on April 26, 1986 (IAEA,
1991). The Chernobyl accident occurred while a test was being performed on
a turbine generator during a normal, scheduled shutdown of one of the
reactors. At the time, written test procedures were unsatisfactory from a safety
point of view. In addition, serious violations of basic operating safety rules
were present, as the operators deliberately withdrew most control rods from
the core and switched off important safety systems (The United Nations
Scientific Committee on the Effects of Atomic Radiation, 1988). Both safety
management and the interrelationship with the human factors and human error
were brought into the safety culture concept and safe operations. Safety
culture was defined as both attitudinal as well as structural, relating to both
the organizational framework and structures along with the attitude of
29
employees at all levels in responding to and benefitting from the framework

(IAEA, 1991).
More recently, several diverse definitions of the safety culture concept have
abounded in the safety research and organizational literature (Guldenmund,
2000; Sorensen, 2002; Wiegmann, Zang, von Thaden, Sharma, & Mitchell,
2002a). In general, all the conceptual definitions can be placed in two broad
categories: the socio-anthropological and the organizational psychology
perspective (Wiegmann et al., 2002b). One difference between these
perspectives concerns the conceptual definition, which is also reflected in
methodology. From the socio-anthropological perspective, it is argued that a
superficial research model of culture should be avoided in order to build
cultural research on a deeper, more complex anthropological model. From an
anthropological perspective, the practice of ethnography and fieldwork,
qualitative in-depth studies with data deriving from interviews, observations,
and/or participation is commonly accepted as appropriate research methods.
Within this scientific direction, culture is described in text with an emphasis
on the organizational member’s subjective interpretation and sense making.
From the organizational psychology perspective, it is argued that culture can
be described with a limited number of dimensions, usually sought through
large organization-wide questionnaire surveys. From this latter perspective,
the culture concept is assumed to express itself through an organizational
climate—a set of perceptually based psychological attributes (Guldenmund,
2000).
Another important difference between the two directions concerns their view
toward cultural change. The socio-anthropological direction considers
organizational culture to be an “evolved construct” deeply rooted in history,
30
collectively held, and sufficiently complex to resist any attempt at direct

manipulation (Mearns & Flin, 1999; Wiegmann et al., 2002b). In contrast, the
organizational psychologists regard culture as changeable and tend to focus on
its functional significance and the means by which it may be manipulated to
improve productivity and safety (Wiegmann et al., 2002b). The organizational
psychology perspective provides a conceptual bridge between safety culture,
safety behavior, and organizational safety management systems, with the aim
of controlling, guiding, or directing first-line operators’ attitude and behavior
toward safe operations.
The concept of safety culture—and climate—has over time been a theme of

heated discussion, with little theoretical consensus emerging on the
ontological, epistemological, and methodological questions relating to the
subject. The main differences in these questions seem to be: (1) What is the
scope of safety culture and the relationship between culture and climate? (2)
How does the concept relate to other organizational aspects and outcome? (3)
Which methods are most suitable for measurement? (Peterson, Ashkanasy, &
Wilderom, 2000). These fundamental questions have already been elaborated
upon by many researchers (e.g., Antonsen & Norges teknisk-
naturvitenskapelige universitet. Institutt for sosiologi og statsvitenskap, 2009;
Cooper, 2000; Glendon & Stanton, 2000; Guldenmund, 2000; Håvold &
Nesset, 2009; Olsen, 2009; Sorensen, 2002; Tharaldsen, 2011; Wiegmann et
al., 2002a; Wiegmann et al., 2002c; Zhang, Wiegmann, von Thaden, Sharma,
& Mitchell, 2002). Based on his review, Guldenmund (2000) pointed out that
most of the characteristics given to culture equally apply to climate, and
within recent research it is more commonly accepted that climate is a
reflection of an underlying culture. Hale (2000) even proposed that one should
stop talking about safety culture completely and instead talk about
31
(organizational) cultural influences on safety (Hale, 2000). In order to grasp

as many facets as possible of the safety culture concept, a multi-method
approach is needed. As the safety culture-climate debate seems to be settling
down and has already been thoroughly discussed by many, the concept will in
this thesis only be touched upon in brief.
A general working model used in the studies of safety culture and safety
management in this thesis is shown in Figure 3.
Figure 3: The general working model used in the studies of safety culture and
safety management
With reference to Figure 3, organizational culture/climate is seen as an

integrated concept subject to change by organizational management practices
and structures. It is assumed that both organizational cultural and managerial
features influence safety, which is defined as safety culture. As such,
organizational safety culture is perceived as a concept with integrated parts
from organizational management system practices and organizational
culture/climate. The organizational safety culture is assumed to reflect the
status of safety in the organization. Organizational safety, or safety culture, is
assessed using two measurement outcome variables: risk perception and the
32
state of the safety management system (SMS). The various concepts and
relationships shown in Figure 3 are further elaborated upon in the following
three sub-chapters.
3.1 Safety culture as an organizational integrated

concept
Within the field of organizational safety, the climate concept was first
introduced by Zohar in 1980 (Zohar, 1980). In more recent publications,
Zohar relates safety climate to an overall organizational climate made of
shared perceptions among employees concerning the procedures, practices,
and kinds of behavior that are rewarded and supported with regard to a
specific strategic focus. When the strategic focus involves the performance of
high-risk operations, the resultant shared perceptions define safety climate
(Zohar, 2010). Although climate and the underlying culture may have a
particular referent as safety, they embrace and are influenced by more than a
single unit or function in the organization. An organization has multiple goals
and multiple policies that are all manifested in organizational behavior and
practices. The different goals are often in competing conflict, like profit and
safety (Hollnagel, 2004; Hollnagel, 2009). For example, the crew may be
expected to cut corners and work faster without getting crossing prevailing
rules and regulations or jeopardizing safety. The safety climate concept
integrates perceptions toward the organization’s total contexts as a regulatory
framework and competitors as well as internal matters as finance, marketing,
human resources, control systems, safety management systems, and so on.
Consequently, when measuring climate with a particular reference, it is
important to embrace the organization in a wider sense in order to reveal
conflicting areas and the priority of importance. True priorities at work (e.g.,
33
efficiency versus safety) have been shown to provide the strongest prediction
of actual behavior (Zohar, 2008).
The theoretical roots of the safety culture discussion can be traced back to
Barry Turner and the introduction of the manmade disaster model (Pidgeon &
O'Leary, 2000; Turner, 1978). In the manmade disaster model, an accident is
defined not by its physical impact, but in sociological terms, as a significant
disruption or collapse of the existing cultural beliefs and norms regarding
hazard. These cultural beliefs and norms are assumed to be formally laid
down in rules and procedures or more tacitly taken for granted and embedded
within working practices. This is also related to managerial and organizational
practices.
Andrew Pettigrew (1979), whose background is in anthropology and

sociology, relates the cultural concept to the everyday tasks and objectives in
organizations as a product of social processes connecting the past, present,
and future. In many ways, Pettigrew’s definition encompasses Turner’s
definition. According to Pettigrew (1979), culture is related to the less rational
and instrumental tasks in an organization as well as the more expressive social
tissue that give those tasks meaning, such as the meaning of having a safety
policy, if procedures should be followed only when safety inspectors are
present, or if efficiency is the real area of priority and the safety-first policy
only serves as a function for external stakeholders. It is argued that, in order
for people to function within any given setting, they must have a continuing
sense of what reality is all about in order to be acted upon. In this setting,
culture is the system of such public and collectively accepted meanings with
regard to safety, operating for a given group at the time (Pettigrew, 1979).
34
One of the most influential anthropologists of modern time, Clifford Geertz,

regards culture as “webs of significance” spun by man and in which man are
suspended. According to Geertz (1973), cultural scientists should try to
interpret those webs in search of meaning and explanation. Geertz (1973)
concluded that culture is most effectively treated as a symbolic system. By
isolating the elements of the symbolic system, specifying the internal
relationship, the whole system may be characterized in general. Symbols are
the surface expression of the underlying cultural structure. Pettigrew (1979)
also emphasized symbols, languages, ideologies, beliefs, rituals, and myths as
an important part in the codification of meaning and emergence of normative
patterns. For example, in accordance with requirements of the ISM Code,
shipping companies have safety policies. From a cultural perspective, policies
such as safety first are not important in themselves. The importance stems
from how such policies, as a symbol, make sense for the organizational
members. A safety-first policy may be perceived as a statement aimed to
attract customers or directed toward other external stakeholders. In the case of
the Bow Mariner, it is likely that the company safety policy would be given
meaning as a façade maintained toward external stakeholders, which can
explain the actual safety-degrading behavior on board. The vessel did hold a
valid ISM certificate and documentation, but during normal operations they
were not acted upon and complied with. This understanding of culture also
establishes a relation between culture and climate—namely, that cultural
beliefs and meaning given to organizational factors are reflected in actual
behavior. In this particular case, the symbolic system is explained by the gap
among safety policy, guidelines, and actual behavior.
One of the most widely used organizational culture frameworks is probably

that of Edgar Schein (1992, 2004), a framework that built upon Pettigrew’s
35
cultural theory. This framework explains culture at three levels: (1) artifacts—
visible organizational structures and processes that are difficult to measure but
are felt and heard by individuals who enter a new culture; (2) espoused
values—norms, standards, and moral principles usually measured through
questionnaire surveys; and (3) basic underlying assumptions—unconscious
taken-for-granted beliefs, perceptions, thoughts, and feelings, which may be
understood by ongoing observations and participation. Schein (1992, p. 12)
formally defined culture as:
a pattern of shared basic assumptions that the group learned

as it solved its problems of external adaption and internal
integration that has worked well enough to be considered
valid and, therefore, to be thought to new members as the
correct way to perceive, think, and feel in relation to those
problems.
Schein regards culture as shared assumptions expressing consistent, clear, and

organization-wide consensus. However, in looking for organization-wide
consensus, important areas of conflict will be disregarded and lost. What is
regarded and learned as valid patterns of shared assumptions depends upon
how the group is defined, which in turn opens up for the existence of various
groups and subcultures within an organization.
Regarding the Bow Mariner, it is natural to assume that the group of Greek
officers and the group of Filipino crew differ with regard to what is shared,
which also brings about the notion of subcultures. In general, it is assumed
that subcultures are found within the shipboard departments’ deck, engine,
and galley. Analyses of other levels in the organization may give different
results, as the group may be defined as the whole fleet or as all shipping
36
companies. A cultural trait identified at the industry level may be related to

manning policies and the extended use of manning agencies and contract
employment, which most regard as the only possible solution to manning—a
solution that is taken for granted and not questioned by insiders, thereby
determining national subcultures. When distinguishing climate from culture,
climate is often suggested to arise from individuals whereas culture is
suggested to arise from group or interpersonal processes (Dansereasu &
Alutto, 1990)
Safety culture is assumed to be influenced by and seen as an integrated part of

an organizational culture and a product of equal processes. Organizational
culture is a relatively stable, multidimensional, holistic construct shared by
groups of organizational members that supply a frame of reference. It gives
meaning to and/or is typically revealed in certain practices manifested as
organizational climate. In the same way, Mearns and Flin (1999) described
safety climate as employees’ perceptions, attitudes, and beliefs about risk and
safety whereas safety culture is a more complex and enduring trait reflecting
fundamental values, norms, assumptions, and expectations. These cultural
elements can be seen through safety management practices, which again are
reflected in the safety climate and in actual behavior (Mearns & Flin, 1999).
In conclusion, most of what is true for safety culture is also considered true
for safety climate (Guldenmund, 2000).
3.2 Organizations, management and cultural change
Safety management and safety culture are all about change—a change toward
enhanced operational safety. Both Pettigrew (1979) and Schein (2004)
regarded individuals, including entrepreneurs and leaders, as important in the
process of creating and managing an organizational culture. According to
37
Schein, one of the most decisive functions of leadership is the creation,

management, and sometimes even the destruction of culture. Although Schein
regarded leaders as important in these processes, leaders are not regarded as
the only determiner of culture. Schein also stated that culture is a result of
complex group of learning processes only partially influenced by leaders.
Groups (e.g., departments or vessels that operate within similar situations)
may behave very differently from one another as the group dynamics differ.
Schein (2004) referred to culture as those elements of a group or an
organization that are most stable and least malleable. However, according to
Schein (2004), the group itself needs a certain degree of stability. Any group
with a stable membership and a history of shared learning will have developed
some level of culture, but a group having either a great deal of member
turnover and/or a history without any challenging events may lack shared
assumptions. Zohar (1980) identified a stable workforce with less turnover
and older workers as an organizational characteristic when determining safety
climate. A stable workforce is then vital for both climate and culture. Within
the segments of shipping that have large turnover and less group stability, it
might be questioned if they have developed any shared basic assumptions
(culture) or working practices (climate).
According to Schein (2004), some barriers to the development of an

integrated shared culture exist, including the insufficient stability of group
membership and the insufficient shared history of practice or the presence of
many subgroups with different kinds of shared experience. This may lead to a
situation of ambiguity and conflict. Joanne Martin (Martin, 1992)
distinguished three perspectives: (1) integration, (2) differentiation, and (3)
fragmentation. Martin’s (1992) recommendation is that an organization be
viewed from all three perspectives, as one perspective’s strengths are
38
another’s weaknesses. As a result, a greater understanding of an

organization’s culture and how to approach cultural change may be obtained.
From each perspective, different aspects of culture and cultural change are
captured. This view was also supported by Alvesson (1993).
From the integration perspective, culture is described as patterns of

manifestations shared by all members of the organization, and the
organization’s manager/leader is regarded as the primary source of cultural
change. One of the most renowned representatives within the integration
perspective is probably Edgar Schein (Frost, 1991; Richter & Koch, 2004;
Schein 2004). According to Schein’s theory, the organization’s leader is
regarded as being able to create and manage the culture in pre-given
directions. The development of subcultures is regarded as undesirable side
effects appearing when the organizations grow and mature. In such a
situation, Schein argued that the leader’s effort should focus on integrating the
variety of subcultures.
From the differentiation perspective, cultural manifestations are described as

sometimes inconsistent when consensus occurs only within the boundaries of
subcultures, which often is regarded as being in conflict with each other. With
regard to influence and change, greater importance is assigned to the
environment, and teams of leaders are ascribed to have secondary influence to
cultural change. Nick Pidgeon (1998) considered the importance of
subcultures, questioning whether a unified culture may be designed within a
large organization and arguing that existing subcultures should be given
attention regarding how they differ in (safety-related) priorities, perceptions,
and interpretations of emerging safety problems. How these aspects interact
with each other, existing power relations, and the like is of equally great
importance (Pidgeon, 1998).
39
While both the integration and differentiation perspectives focus on what is

shared and accounts of planned and directed goals change, the fragmentation
perspective regards ambiguity as the essence of culture. In this perspective,
culture is something that is constantly fluctuating, and no stable organization-
wide or subcultural consensus is supposed to exist. Cultural change is seen as
being in a constant flux, where the power of change is usually seen as being
diffused broadly in the environment and among organizational members. The
ambiguity is supposed to emerge from the complexity and unpredictability in
the organization and society; the fragmentation perspective does not assume
that the organizational members have similar reactions to these ambiguities.
Karl Weick and the theories of high reliability organizations (HRO) are well
known within this field (Weick & Sutcliffe, 2007).
Within shipping, subcultures, conflicts, ambiguity, stress, and

misunderstandings are likely to be present due not only to the lack of stability
of membership, but also to an insufficient shared history of practice. This
relates to both the specific work situation, but also, amongst others, to
nationality. National cultures are known to differ in aspects such as power
distance and the degree of human inequality, uncertainty avoidance, and how
they adapt to unstructured situations as well as how they integrate into a group
with regard to individualism versus collectivism (Hofstede, 2001). In
addition, value of life, safety standards, and risk perception are known to
differ between nationalities. Geert Hofstede perceived culture as mental
programming—a pattern of thinking, feeling, potential acting, and unwritten
rules of the social game that distinguishes the members of one group of
people from others. A certain culture is learned through the lifetime. What is
acquired in early childhood and once established is difficult to change
(Hofstede & Hofstede, 2005). Hofstede argued that layers of culture (e.g.,
40
organizational culture) acquired later in life tend to be more changeable.

Organizational acquirements and practices—the visible part of culture—are
regarded as faster and easier to change.
Based on the previous discussion, it is concluded that organizational safety

culture and behavior are subject to change by organizational practices and
structures, which includes safety management systems. Furthermore,
individuals in the organization (e.g., their leaders) do have a mediating effect
upon the formation of an organizational safety culture.
3.3 Organizational culture and safety management
Safety management systems have two interrelated main functions: to avoid
accidents and improve safety. Theories of accident causation and safety
management have progressed over time (Borys, Dennis & Legget, 2009). H.
W. Heinrich is considered to be a pioneer within safety management and
accident causation (Heinrich, Roos & Petersen, 1980) from the first age of
safety—namely, the technical age. With reference to the shipping industry,
the first development of SOLAS belonged to a technical age; the introduction
of the ISM Code represented a transition to the age of management systems
and culture. The age of human factors occurred in between, in which the view
of human error considered the interaction of human and technical factors
when exploring the causes of errors and accidents (Borys et al., 2009).
As previously highlighted, traditional safety management systems (SMS)

described in the ISM Code fall along a linear causality where attempts are
made to predict and prevent future incidents by reflecting upon previous
experience related to empirical safety control. The traditional SMS contains
several subsystems. First, a system for reporting and collecting experience
data from the vessel itself is required, followed by a system of data processing
41
(i.e., summarization and analysis) in order to reveal causal factors and

perform trend analyses, which form the basis for the development of safety
measures. Both identification of causes and remedial actions are closely
intertwined with how the organization addresses technological,
organizational, and individual factors. Irrespective of this, one critical system
requirement is the reliability and accuracy of input data (i.e., near miss and
accidents reports). As long as the input is reliable, the overall system
presupposes the possibility of revealing a root cause and develops efficient
measures in order to control operational safety (Kjellén, 2000). Although the
fundamental rationale of safety management has changed little over the years,
the rationale for the understanding of human error and causal factors, root
causes, and adequate safety measures has changed. However, quite
surprisingly, recent research also points to insufficient scientific evidence on
the effectiveness of systematic safety management to make recommendations
either in favor of or against them (Robson et al., 2007). Thus, the obvious
question is why: Is the reason found within the theoretical rationale of safety
management or within organizational understanding and applications of the
systems or is the reason found elsewhere?
The application of traditional safety management is questioned with a

distinction between small-scale accidents and larger more infrequent accidents
(Rundmo, 1996). According to Rundmo (1996), the empirical safety control,
in which measures are developed through the analysis of past events, is only
applicable for frequent and small-scale accidents such as ordinary work
accidents. When it comes to medium-size and more infrequent accidents (e.g.,
groundings and collisions) and large-scale accidents with very low
probabilities (e.g., the capsize of the Herald of Free Enterprise and the fire at
the Scandinavian Star; see Table 1), traditional safety management is not
42
considered to be applicable. Such accidents are often too unique and complex
to grasp, and it will not be possible to single out some isolated underlying
causes or develop measures that cover all involved risk—a risk that in the first
place is considered too complex to understand. Others, such as Scott Sagan
(1993) and Charles Perrow (1999), questioned the possibility of foresight and
preventions of accidents through empirical safety management.
Within the age of human factors, human error is regarded as primary cause of
accidents. In more recent theories and subsequent ages of safety, human error
is seen as a consequence of faults deriving from other parts in the organization
or environment, complexity, interactions, and/or organizational culture. A big
difference between human error as a cause and human error as a consequence
is seen in the characteristics of remedial actions. With human error cited as
the cause of failure, the tendency for safety measures is to seek to control
human behavior with inter alia procedures and checklists. Then, when the real
cause is found to lie elsewhere in the organization, such measures may not
clearly be the answer to the underlying problem and incidence of failure, and
accidents will continue to occur. The following sections shed light on these
issues from the theoretical perspective of accident causation and prevention.
The Man Made Disaster model
Barry Turner is presumably among the first to regard latent conditions as a

primary cause in accident causation. With the development and introduction
of the Man Made Disaster model, accidents and disasters are proposed to
develop through a long change of events leading back to root causes like lack
of information and misperception among individuals (SINTEF, 2003). Turner
argued that this is a result of an organizational culture where information and
interpretations of hazard signals fail. Thus, a typical accident can be traced
43
back to initial beliefs and norms—culture and climate—that do not comply

with existing operational reality. From this perspective, accident development
is viewed as a process, often over years, developing from an interaction
between human and the organizational arrangements of the socio-technical
system (SINTEF, 2003). With reference to accidents such as the capsize of
Herald of Free Enterprise, a common understanding of the crew—given by
the management of the organization—was a general understanding that the
vessel should leave some minutes ahead of schedule, even if it involved
putting pressure on those who did not move fast enough. The inherent risks of
such a practice were not questioned. From the Man Made Disaster
perspective, systematic safety management should deal with these
breakdowns in the interpretation of information. For example, are some
danger signals or causes systematically disregarded or misunderstood? As
with the Bow Mariner, why did none of the involved stakeholders (e.g.,
charterers, vetters, inspectors, and flag state) manage to reveal the degrading
shipboard safety situation?
As such, existing cultural beliefs and norms are the essence of the Man Made
Disaster theory. According to Turner, such cultural beliefs and norms might
be formally laid down in rules and procedures or more tacitly taken for
granted and embedded within working practices. An accident is then assumed
to occur because of inaccuracy or inadequacy in the accepted norms and
beliefs and of a discrepancy between the way the world is thought to operate
and the way it really operates (Pidgeon & O'Leary, 2000). When
acknowledging the weaknesses in traditional safety management and failure
of foresight, these institutional barriers to effective learning should, according
to Pidgeon and O’Leary (2000), be addressed. The aim for efficient safety
management should be to let all organizational members develop a safety
44
imagination that breaks the pattern of becoming overly fixated with prescribed
patterns, simplification, and ignorance. Pidgeon and O’Leary (2000)
presented seven guidelines for fostering a safety imagination: (1) attempt to
fear the worst, (2) use good management techniques to elicit varied
viewpoints; (3) play the “what if” game with potential hazard; (4) allow no
worst case situation to go unmentioned; (5) suspend assumptions about how
the safety task was completed in the past; (6) approach the edge of a safety
issue with a tolerance of ambiguity, as newly emerging safety issues will
never be clear; and (7) force oneself to visualize “near miss” situations
developing into accidents. It is argued that such safety imagination is a critical
facet of organizational learning and an effective safety culture.
Normal Accident Theory
The Man Made Disaster theory also highlights how system vulnerability
arises from unintended and complex interactions between contributory
preconditions (Pidgeon, 2000), which may be linked to Charles Perrow’s
theory of normal accidents (NAT). However, in contrast to Barry Turner,
Perrow regards some systems to be to complex and interactive to avoid
organizational accidents completely. In such systems, safety management is
regarded as futile as accidents are doomed to happen due to the system
characteristics.
The development of NAT started with Perrow’s exploration of the 1979

accident at the Three Mile Island nuclear plant. During this investigation,
Perrow was struck by the fact that the present accident literature
overwhelmingly blamed the operators (Perrow, 1999). Perrow then looked
into accident reports from various industries, such as mining, aircraft, and
marine accidents, and evolved the alternative theory that risk is a result of two
45
dimensions—interactions and complexity—rather than human error. It is

argued that the operator is free from blame as the overall system, interactions,
and interdependencies of events are incomprehensible for a critical period of
time. Like Rundmo (1996), Perrow distinguishes between small-scale,
frequent personal accidents and other medium-sized accidents and larger-scale
accidents. Risk is, according to Perrow (1999), regarded as something
designed into organizations in the form of complex systems with tight
couplings. The paradox is, when barriers and other safety measures are built
into the system to increase safety, often it is the complexity that is increased at
the cost of safety. Perrow (1999) regards such error-prone organizations as
impossible to manage safely in the long run, thereby resulting in the notion of
normal accident. In the maritime setting, Perrow (1999, p. 230) described the
risk-inducing complexity as follows:
The ship itself, with its power plant explosive mixture,

steering apparatus, and draft in shallow channels is important,
but so are other ships, the insurance industry, the fragmented
shipping industry, attempts are regulation, rules of the road,
dangerous cargoes, national jealousies and interests, and, of
course the horrendous environmental problems of fog, ice,
and storms.
In his analysis, Perrow also pointed to conflicting organizational goals,

production pressure, and organizational pressure related to risk taking,
authoritarian structure on board, and inappropriate leadership. One of
Perrow’s (1999) main points was to regard all human constructions (e.g.,
vetting, class, regulatory bodies, flag state, insurance, manning companies) as
systems and not as collections of individuals or representatives of ideologies.
Dangerous accidents lie in these systems based on how the different parts fit
46
together and how they interact. From this point of view, safety management is
about reducing the system complexity and/or loosening the couplings to
reduce interactivity. From the perspective of NAT, Turner’s notion of “safety
imagination” is a dead end for organizational safety as long as complexity,
incomprehensible interdependencies, and tight couplings are present.
High Reliability Organizations Theory
The High Reliability Organizations Theory (HRO) developed as a result or a

continuance of Perrow’s rather pessimistic message—namely, that accidents
are inevitable in some systems or organizations due to their characteristics
(Roberts, 1990). The concept of safety culture constitutes a central difference
between NAT and HRO. Whereas NAT argues that in some systems accidents
are inevitable, HRO argues that even in the most vulnerable and error-prone
systems, safety culture has characteristics that can counteract the inherent
system risk.
In many ways, HRO is in line with the arguments given by Turner, Pidgeon,
and Perrow. HRO recognized that everything that may fail during operations
has not yet been experienced; therefore, a system based on experience
feedback (as the ISM code) is doomed to fail on its own premises. HRO also
recognizes that not all incidents may be deduced to detect all possible failure
or error modes. Recognizing that the world is complex, unstable, unknowable,
and unpredictable, the HRO perspective maintains reluctance to the
simplification inherent in traditional risk assessment. Both procedures and
checklists may represent simplification of measures when a complex,
unstable, unknowable, and unpredictable working situation is attempted to be
controlled by preplanned prescriptions. This view is shared by Perrow, and
47
both NAT and HRO are skeptical towards those who heavily rely on risk
assessment.
In HRO, more attention is given to the real work going on in frontline

operations. By empowering those doing the actual work, operators have the
possibility to solve the situations themselves, based on their own experience
and knowledge. This is a contrast to traditional safety management from the
era of human factors with a focus on the control of human behavior. Within
the HRO paradigm of HRO, frontline personnel operate by using pre-planned
descriptions, but it is accepted that in real situations deviations will occur.
One of the HRO’s key points is mindfulness, which is related to the concept
of safety culture and similar principles such as Turner’s safety imagination.
Mindfulness is understood as a combination of alertness, sensibility,
flexibility, and adaptability. This perspective argues that unexpected events
should be handled by creating a mindful infrastructure by following five main
principles: (1) continuously tracking small failures, (2) resisting
oversimplification, (3) being sensitive to operations, (4) maintaining
capability for resilience, and (5) monitoring the shifting locations of expertise.
The violation of these principles is regarded as a setback toward the more
traditional approach, where simple diagnoses are accepted, frontline expertise
is overridden by faith in risk analysis, and safety measures are developed
detached from operations. The HRO view implies that the operators gain more
responsibility, so other parts of the organization have to give them the
possibility to act. In other words, control is taken from the upper levels of the
organization in favor of lower levels.
Within HRO, safety culture is seen as essential in managing risk. All of

Johanne Martin’s three perspectives of safety are adopted: “Each form of
culture handles ambiguity differently: Integration denies it, differentiation
48
selectively clarifies it, and fragmentation accepts it. In a mindful culture, all
three forms of culture are present” (Weick & Sutcliffe, 2007, p. 112). HRO
does not reject the fact that organizational members have shared values and
beliefs. However, with regard to safety these shared patterns are not regarded
as vital for the outcome. The shared orientations are accommodated
differently in all situations, and the chain of events and patterns of interactions
between people fall under the influence of situational conditions as stress,
misunderstandings, interpretation, and others conditions specific to each chain
of event. From an HRO perspective, more weight is placed on fragmentation
than on differentiation and more on differentiation than on integration.
Managing risk and safety culture
James Reason (Reason, 2001) developed a widely used practical definition

and approach to safety culture. Barry Turner and the Man Made Disaster
model provided much of the conceptual foundation for Reason’s work
(Pidgeon & O'Leary, 2000). Reason’s approach to safety management and
safety culture is also to a large degree adopted by the HRO perspective
(Weick & Sutcliffe, 2007). Based on the organizational culture literature,
Reason (2001) differs between two theoretical stands: those who regard
culture as something an organization has and those who regard culture as
something the organization is. Reason favors the former approach and thus
regards culture as something changeable and manageable by organizational
practices.
Like Turner, NAT, and HRO, Reason is also concerned about the
organizational factors that trigger an accident. According to Reason (2001),
accidents by their nature are not directly controllable, as many of the causal
influencing factors lay outside organizational control and influence. As such,
49
rather than struggling vainly to exercise direct control over behavior,

accidents, and incidents, organizational managers should measure and
improve the processes of underlying factors, such as training, procedures,
planning, budgeting, goal conflicts, and others. Thus, an efficient SMS system
should help identify those conditions most needing correction and not be
limited to non-compliance of global rules. Attention should also be directed
toward the quality of these global rules (e.g., accuracy, relevance, availability,
and workability of procedures). The information reported into the system
should embrace organizational factors as well as local workplace factors and
unsafe acts. The cultural factor is linked to commitment, competence, and
cognizance within the organization as a whole.
Reason (2001) regards safety culture as a cornerstone in efficient safety

management in order to get the needed operational information. He identifies
safety culture using four critical subcomponents: (1) a reporting culture, (2) a
just culture, (3) a flexible culture, and (4) a learning culture. Together, these
interact to create an informed culture. According to Reason, an informed
culture is one that collects operational experience data that are characterized
by an organization’s climate in which members feel free to report without
experiencing negative, unfair, or in other ways meaningless consequences.
Attention is also given to how the interpretation of information and outcome
is influenced by the overall company policy. Reason is more preoccupied with
organizations’ internal organizing than how safety is influenced by other
organizations and macroeconomics than by national and international
conditions such as politics, laws and regulations.
50
The theory of Practical Drift – why organizations fails
Safety culture in shipping is often depicted side by side with compliance to

prevailing procedures and other safety measures. However, if compliance is to
be a valid key performance indicator, it is presupposed that these safety
measures are appropriate for the actual action. The development of
standardized measures fitting all real-life situations are a challenging if not an
impossible task considering the complexity and unpredictability of most
situations. This is also pointed out by NAT and HRO. The Practical Drift
Model (PDM) provides an explanation for how and why organizations
experience such gaps among standardized measures, real situations, and
actions, referred to as practical drift (Snook, 2000). PDM combines HRO and
NAT in two ways, emphasizing how different degrees of mindfulness will
depend on different situations and contexts. During their lifetime,
organizational systems develop both tight and loose couplings, which they
shift in between—tight to loose couplings and back again—as the various
sub-units within the operative part of the organization alternate between a low
and high degree of interdependence. The model also captures both contextual
and temporal factors when explaining why incidents and accidents occur,
along with practical drift from the global rules, such as standards, procedures,
and checklists.
Standardized rules are often designed according to organizational lifecycle,

where the governance structure is top-down oriented. Using organizational
safety management systems, organizational managers put large effort into
developing extensive routines and procedures in order to make the
organization robust and resilient against future unforeseen events. When
designed, tight couplings and rule-based action of logic is assumed to
51
characterize the organization. This, in contrast to an operational situation, is

assumed to be loosely coupled, and the real world does not act in accordance
with the organizational design. When organizations experience that
unforeseen events do not occur as expected, the attention toward the
limitations and inadequacy in routines and procedures become a part of
everyday life and practice and are therefore more relaxed. Others (e.g.,
Hollnagel, 2004) point to real-life work processes that are irregular and
unpredictable in contrast to work regulations—either explicitly by procedures
and instructions or implicitly by rules, standards, or good practice. Another
issue arises when these rules are developed detached from the operational
situations (e.g., in shipping by shore personnel); consequently, accuracy,
relevance, availability, and workability of the rules may be low. Moreover,
company policy may favor efficiency over safety.
As a result, the sharp end operators—when aware of this situation—will be

able to break the strict rules without fear of sanctions or punishment. On a
local basis, breaking strict rules may actually get the job done quicker and
more efficiently. The operator may be “rewarded” for achieving additional
goals in the organization. During such a process of de-coupling, the
organization will become increasingly free from global rationality.
Subcultures with their own logic of action based on experiences and tacit
knowledge develop, and the operators accordingly drift further away from a
rule-based system to a more task-based system.
According to PDM, accidents and incidents occur when the system suddenly
and stochastically becomes tightly coupled, as with the Bow Mariner,
Scandinavian Star, and Herald of Free Enterprise. In such situations, the
involved operators are forced to act on the assumption that all others act in
accordance with the original rules and procedures initially designed. The
52
operators become trapped in a game in which trusting their own logic of

actions is the only solution while they must simultaneously base their decision
on the assumption that others are following the general rules.
After an unwanted event, the outcome is often even more tightly designed
control criteria. James Reason describes such an approach to safety
management as a person-oriented approach, which may also be perceived as
belonging to a blame culture, implying crew shortcomings as the cause of
error. With human error cited as the cause of failure, the tendency is for safety
measures to seek to control human behavior by developing more procedures
and checklists. When the real cause is found to lie elsewhere in the
organization, such measures may clearly not be the answer to the underlying
problem and incidence of failure, and unwanted events will continue to occur.
Global rules may be perceived as less and less meaningful; thus, constantly
relying on them will further undermine the safety system as more procedures
are violated and the local units drift even further apart from global rationality.
Less reporting of experience data could also be a consequence in the longer
run.
3.4 Safety culture and measurable outcome variables
The IAEA has built its concept of safety culture upon Edgar H. Schein’s
three-level model and regards safety culture as both attitudinal as well as
structural relating to both the organizations and the individuals (IAEA, 1991).
Based on this conceptual definition, INSAG and IAEA regard systematic
safety management, which belongs to the first layer in Schein’s model (i.e.,
artifacts), as a tool for promoting a strong safety culture and achieving a good
safety performance. This may be measured in the second layer, espoused
53
values, using a questionnaire survey. The third layer, basic assumptions, may
be captured by observation.
Using previous research within the oil industry, Rundmo (1996) demonstrated
that risk perception and risk behavior are significantly correlated, but also
relatively independent from each other. The association between risk
perception and risk behavior is then caused by the fact that the same predictor
affects both variables. Crew risk perception and other subjective assessments
are suggested as good indicators of the safety level, but not as predictors for
risk behavior. It is further suggested that employees’ behavior to a great
extent is constrained by the conditions under which they work. When the
working conditions are not perceived to be satisfactory, employees know that
the occupational risk is higher; they feel more unsafe, which will affect their
risk perception (Rundmo, 1996). Risk perception as a measurement variable is
supported by Zohar (1980), who used climate research to assume that an
individual’s perception focuses on the organizational environment,
organizational control system, and safety management system.
Drawing from the Man Made Disaster theory, safety culture is the product of
cultural beliefs and behavioral norms regarding hazards, which are laid down
in the organizational control system and thus reflected in procedures and
guidelines. This is in accordance with Schein (1992, 2004), who regards
organizational control systems (e.g., safety management systems) as a
manifestation of basic underlying cultural assumptions. In this respect,
companies approach to safety management is a manifestation of the
underlying beliefs and thoughts on how safety management should be
performed. Pettigrew (1979) relates the cultural concept to how everyday
tasks and objectives in the organization are expressed as well as their
meaning, including how the organizational members comprehend the
54
importance of the organizational control system and derived safety measures

in relation to other directions given by the company. This approach aligns
with Geertz’s (1973) work referring to the social processes where meaning
and sense making arise. Social processes are influenced by crew composition
and human resource policies. Personal beliefs and values are also antecedents
to behavioral norms. Traditional cultural researchers also regard beliefs and
values as less changeable as they are acquired through a lifetime and therefore
deeply rooted within the individual. However, Hofstede and Hofstede (2005)
argue that beliefs and values learned through an organizational context are
more changeable as they are acquired at a later stage of life. The
organizational context and organizational control system are assumed to
directly impact behavioral norms.
When managing behavioral norms by means of an organizational control

system, Kjellen (2000) defined the system as one that provides the
information needed for safety and signaling related to health and safety
matters. In this regard, Reason (2001) regarded safety culture as the
cornerstone for ensuring the information flows as needed. Reason also
addressed the safety management system as a whole, with regard to how the
incoming experience data are analyzed and processed into safety measures.
From this, the state of the safety management system is regarded as a
measurement outcome variable.
55
Research methodology
4 Research methodology
This chapter describes the research methods applied in this thesis. Based on
the discussion in Chapter 3, climate will not be separated from culture, but in
theory and research methodology the recommendation from both perspectives
will be taken into consideration. This synthesis of qualitative and quantitative
methods is regarded as important in order to understand how culture is created
through social processes while quantitative methods simultaneously say
something about how widespread certain patterns of behavior and perceptions
are within the industry and statistical associations. The application of a multi-
method approach is also supported by others. According to Cooper (2000), the
triangulation of different methods allows the researcher to take a multifaceted
view of safety culture, so that the interrelationships among psychological,
behavioral, and situational factors can be examined with a view to establish
antecedents, behaviors, and outcomes within the specific contexts.
Triangulation allows the employment of each method’s strengths in order to
overcome the others’ weaknesses. Both Rousseau (1990) and Schein (1992,
2004) identified that different layers of culture are amenable to different
research methods. For example, the fundamental content of culture is assumed
to be unconscious and highly subjective. The organizational members’ basic
assumptions, values (what is important), and beliefs (how things work) as
well as culture’s social construction are difficult to assess without interactive
probing. Moreover, the organizational members’ fears and defenses are
difficult to elicit without interaction, which gives reason for an ethnographic
methodology. On the other hand, the organizational members’ patterns of
behavioral norms (how things are done) are far more accessible to observation
from outsiders and respond to structured instruments and quantitative
methods.
56
According to Geertz (1973), cultural analysis is (or should be) guessing at

meanings, assessing guesses, and drawing explanatory conclusions from
better guesses. However, the previously described examples illustrate how
such guesses might be wrong if not related to individuals’ own experiences
and national, situational, and/or historical context. Street Corner Society is a
method of participant observation where becoming native, without being too
attached, is a part of the research strategy (Whyte, 1991). In accordance with
Street Corner Society, the researcher lives with the community in order to
understand the nature of the field, learn to understand the group, and build
trust and credibility. Whyte also demonstrated how such qualitative studies
may be expressed and presented in a more quantitative format. Whyte has
become a major spokesman for the advantages of integrating research
methods, including those typically associated with quantitative research
(Bryman, 1991). As such, there is no adversative relationship between cultural
field studies from an anthropological perspective representing qualitative
methods and the quantitative methods typically from the psychological
perspective.
From the author’s own experience, the ethnographic approach has been
valuable not only to understand, but to correct misunderstandings. For
example, when conducting the first field studies, one Philippine mate
discussed her former Norwegian captain, who refused any crew members to
whistle on the bridge. Both the Philippine mate and I interpreted this as an
indication of authoritarian leadership. However, months later, after getting a
better understanding of seamanship in a national historical context, I found
that this interpretation was wrong and in fact related to superstition, which is
quite common among the older generation of seafarers. In the days of sails,
the seafarers needed wind, and they whistled to call for the wind. Nowadays,
57
with engine propulsion, wind is no longer wanted as wind causes waves.

Thus, whistling is not allowed as it calls for the wind. In other situations,
misinformation may be a result of deliberately withholding information as the
group being researched does not trust the researcher. In another field study, I
observed that all engine crew wore helmets in the engine room, as required.
However, after one week at sea, once I had become familiar with the group, I
learned that usually no one wore their helmets. They only applied the rules
when they had a third party on board—an outsider.
Apart from this, also favoring an integrated use of methods, cultural

interpretation, and theoretical development is suggested to follow its own
unplanned course in the search of grasping and analyzing “the web of
significance” and structure of symbols and meaning, plunging more deeply
into the same ideas. These theoretical formulations do not make much sense
or hold much interest apart from the context of interpretation. Indeed, a safety
culture study carried out within the anthropological tradition alone could have
brought descriptions, interpretations, and understanding of how safety is
interwoven with symbols and cultural elements on a single vessel or in one
department. However, the developed theories would not make any sense
outside that unit. Thus, within the organizational aim of enhancing safe
operations in general, this approach would not be very useful.
4.1 Quantitative research and questionnaire survey
A quantitative design incorporating a questionnaire is used to develop an
understanding of the manifestation of culture and to give direction for the
subsequent qualitative studies. In order to grasp the underlying dimension of
safety culture through the use of questionnaire, a high level of instrument
quality—namely, reliability and validity—is required.
58
The starting point of every questionnaire is item generation, concerning which

questions and themes should be included. This stage is also related to content
validity. Although Hinkin (1995) suggests a strong theoretical framework as a
starting point, followed by a sorting process allowing for the deletion of
conceptually inconsistent items, Guldenmund (2007) suggests two different
approaches for questionnaire development. First, a descriptive model of the
construct can be used as a starting point—namely, a normative or theoretical
approach. Second, theories and results of previous research can be used in
combination to construct a new questionnaire, which is a more pragmatic
approach.
However, some inherent difficulties exist in both of these frameworks. First,

if starting with a theoretical approach, a lack of theoretical consensus may
result in different themes, scales, and items depending upon the researchers’
theoretical stand. Second, as a result of the first point, previous research may
be difficult to use due to, inter alia, the variety of themes, scales, and items
used. Moreover, when established theory is derived from empirical research,
the theory itself may be misleading due to the vast amount of far-from-
validated measurements in use. Hale (2000) perceived part of the problem as
being induced by the tendency for each researcher to start from scratch by
developing his or her own instrument. Hardly any scales have been reused in
the same form in more than one study, and they can therefore not be
systematically refined and improved by combined research efforts across
several research groups. It is important to keep in mind that the cultural
disagreements and differences related to epistemology, ontology, and
methodology, to a great extent, may explain the many different measurements
in use.
59
Theory plays a key role in how measurement is conceptualized, and the lack
of theoretical consensus poses a clear challenge to researchers. Under such
circumstances, it is especially important to be mindful of measurement
procedures and development. In addition, it is considered impossible to reach
theoretical progress without adequate measurement (Hinkin, 1995). As
described here, the situation within the field is complicated, and some
important questions are raised regarding whether the methodological
disagreements derive from a lack of theoretical consensus or if theoretical
consensus is lacking due to methodological differences and the use of flawed
measures. In order to overcome some of these problems, a previously
developed instrument was used in this thesis
Questionnaire development
In order to examine safety culture, a questionnaire developed by Studio

Apertura, a constituent centre of The Norwegian University of Science and
Technology (NTNU), in collaboration with the Norwegian DNV and the
research institution SINTEF, was used. Their development was based on a
theoretical review and an evaluation of eight preexisting questionnaires—four
developed in Norway, two in Denmark, and two in the United Kingdom. The
evaluation was carried out according to five criteria: (1) foundation
(theoretical foundation, documentation, and premises for application), (2)
thematic width, (3) practical experience of use, (4) the ability to describe and
measure safety culture, and (5) the ability to be used at multiple levels
(individual, group, team, company). A more thorough description of the
development is available (Antonsen & Norges teknisk-naturvitenskapelige
universitet. Institutt for sosiologi og statsvitenskap, 2009; SINTEF, 2003;
Studio Apertura, 2004). The questionnaire was previously used to survey
60
safety culture on board supply vessels and found to be acceptable for use
within merchant shipping. The full questionnaire and letter of introduction are
included in Appendix 1.
Survey sample and respondents’ demographics
The research population is Norwegian-controlled dry cargo and liquid carriers

above 500 gross tons. A total of 150 target group vessels were randomly
selected from the 953 vessels within the Norwegian Shipowners’
Association’s list of members for 2005. The target group of 150 vessels
represented approximately 15% of the overall population, which was
considered to be large enough to be representative of the population as a
whole (Neuman, 2000). A sample of 10% is recommended, but some
withdrawals were expected; thus, a 5% margin was included in the original
sample. The sample was stratified with regard to status of the vessel’s flag
register (white, grey, or black listed flag) and type of vessel (general cargo,
bulk carrier, oil tanker, gas tanker, or chemical tanker).
Following the initial selection, telephone calls were made to each company to
ask for their participation. Thirty-one companies, with a total of 83, vessels
agreed to participate while 45 companies with 67 vessels in total declined.
Reasons for not participating included:
 Being unable to contact the company despite repeated efforts (23

vessels, 16 companies).
 The vessel was not owned by a Norwegian party and therefore was not
defined as Norwegian controlled (15 vessels, 8 companies).
61
 Ship management was outsourced to a non-Norwegian country and

therefore was not defined as Norwegian controlled (14 vessels, 8
companies).
 The company refused to participate (12 vessels, 10 companies).
 The remaining vessels were sold (3 vessels, 3 companies).
The population was later redefined, and vessels managed from a non-
Norwegian country were not considered to be Norwegian controlled.
In total, 1,574 questionnaires were distributed to 83 tankers and bulk/dry

cargo carriers; 76 vessels from 29 companies returned a total of 1,262 forms,
resulting in an individual response rate of 80.2%, a vessel response rate of
91.5%, and a company response rate of 93.5%. The questionnaires were
returned from 40 liquid bulk carriers (liquid tanker) and 36 dry bulk carriers
(dry cargo); 63% of the respondents were employed on a liquid tanker and
37% on a dry cargo vessel. Twenty-two nationalities were represented, with
the majority from the Philippines (65.5%), followed by Norway (9.2%),
Poland (8.1%) and Russia (5.5%). Unfortunately, no company with vessels
flying a black-listed flag was willing to participate.
The further validation process (results are included in the articles in part II) is
based on the following premises and methodological guidelines.
Validity and reliability through theoretical conceptualization

The starting point when developing a questionnaire and scales is conceptual
definition, which specifies the theoretical basis. A questionnaire is normally
comprised of several dimensions or constructs represented by several partly
overlapping items, called multidimensional scales. When generating the item
62
pool, each item making up a construct should reflect the latent variables
underlying the theme (i.e., the different features or dimensions of the safety
culture concept). In order to truly reflect the underlying feature, the items in
each dimension should ideally have a common cause (i.e., local management)
or consequence (i.e., work practices). Thus, an underlying assumption is that
the items reflecting one single construct are unidimensional. In other words,
within each measured dimension, items are strongly associated with each
other while simultaneously representing a single dimension of the concept.
Three reasons for using a multi-item measure instead of a single-item measure
are noted. First, an individual item is not reliable due to a considerable
random measurement error. Second, an individual item lacks precision and
can only categorize people into a relatively small number of groups. Third, an
individual item lacks scope, and it is very unlikely that a single item may
represent a complex theoretical concept (DeVillis, 2003; Gliem & Gliem,
2003; Hair, 1998; Shevlin, Miles & Bunting, 1997; Spector, 1992). Summing
up, single items are considered to be less valid, less accurate, and less reliable
than multi-item constructions. It is also suggested that a scale should consist
of a minimum of three items in order to be robust (Pett, Lackey & Sullivan,
2003). However, due to collinearity, the use of multiple items could represent
a problem in regression models and when independent variables are created
by summing items in a scale. Such an additional method represents a
procedure that does not control for the effect of measurement error.
Regression parameter estimates may be attenuated or increased (Shevlin et al.,
1997).
When ensuring the conceptual definition, the primary concern is content

validity, which is a requirement for construct validity (Hinkin, 1995). Content
validity is the degree to which elements of the measurement are relevant to
63
and representative of the underlying safety culture concept. Determining

whether the scale or item-set has good content validity can be done from a
number of sources of relevant theory, empirical literature, and expert
judgment. Construct validity concerns the degree to which inferences can
legitimately be made from the operationalized constructs in the questionnaire
to the theoretical concepts on which those operationalizations were based.
When using multidimensional scales, both the convergent validity of the
respective subscales (i.e., the degree to which the items within a particular
subscale measure the same unidimensional construct) and their discriminant
validity (i.e., the degree to which the items in different subscales measure
different rather than the same construct) need to be considered. Both content
and construct validity are concerned with how the measurement fits with the
theoretical foundation and power of generalization—namely, external
validity. When the objective of a study is to establish a causal relationship
(i.e., using regression analysis), internal validity is of particular consideration,
referring to the confidence placed on the assessed cause-effect relationship.
The internal validity of the conclusions reached depends on the reliability and
validity of the questionnaire or scales used (Neuman, 2000; Raubenheimer,
2004).
Other aspects that should be taken into consideration are whether the items are
measuring a perception or an attitude. Perceptions are considered more
volatile and mostly oriented toward the current workplace conditions, whereas
attitudes are considered to be less open to change, more durable, and
developed through experiences both inside and outside the workplace. Cooper
(2000) cautions against the use of measurements that include attitude scales
due to the risk of muddying the construct. Previous research has shown that
attitudinal questions have more positively skewed responses than the
64
perceptional questions and may therefore influence the analytical results.

Moreover, Cooper (2000) indicated that the mix of attitudinal-perceptual
questions is one explanatory reason that different factor structures emerge
across research groups.
Measurement errors threaten the validity of the conclusion about the

relationship between the constructs. Method bias has both a systematic and a
random component, with the systematic error in particular being considered a
major problem. One source of method bias is apparent when the same
measurement is used for all constructs, making it difficult to assess the
strength of the bias. The direction also varies, and the observed relationships
may be either inflated or deflated. Potential sources of common method biases
are produced by a common source or evaluator (e.g., social desirability,
consistency motive acquiescence, or positive and negative affectivity)
whereas method effects are caused by an item’s characteristics (i.e.,
complexity, ambiguity, scale format, and negatively worded items). Method
effects are caused by item context produced by the measurement context
(Podsakoff et al., 2003). Podsakoff et al. (2003) presented several approaches
to addressing this problem. However, others regard the common method
biases as an urban legend, claiming that the supposed effect on correlations is
overstated (Spector, 2006). Spector (2006) further pointed to the fact that, as
long as there is uncertainty related to the presence and size of a possible bias,
applying methods, inter alia statistical methods, in order to control the bias
effect might produce biases itself, as one might control for something that
does not exist. However, being aware of the possible problems makes it easier
to consider them during the process of developing a valid measurement.
65
Validity and reliability through factor analysis

When data are collected, factor analysis is a common method for validation of
the questionnaire’s conceptual definition. Factor analysis defines the
underlying structure of the interrelationship (correlation) between the
variables in the questionnaire data by defining a set of common underlying
dimensions known as factors. However, not only do the constructs share that
they are facets of the same concept, but correlations could also be—due to
similarities in measurement—common source and/or common method
(DeVillis, 2003; Podsakoff et al., 2003). If the data are not biased, the
extracted factors comprise internally consistent and correlated items that
externally differ from the other factors. Thus, the extracted factors are
assumed to have discriminant and convergent validity. The convergent
validity is further assessed by the means of scale analysis and inter-item
statistics.
For the method of factor analysis to be appropriate, a certain sample size is

required. Preferably, the sample size should be 100 or larger. As a general
rule, it is suggested to have at least five times as many observations as there
are variables to be analyzed. Some even propose 20 cases for each variable.
Small sample sizes or low variable-case ratio lead to higher chances of
“overfitted” data (i.e., deriving factors that are sample specific with little
generalizability) (Hair, 1998). Hair (1998) also pointed out that correlations in
small samples could be deemed significant and appear in the factor analysis
just by chance. In addition, if no items are substantially correlated, factor
analysis is not applicable. The method’s applicability is commonly tested by
Barlett’s test of sphericity, which should be significant, and Kaiser-Mayer-
Olkin (KMO), which should exceed 0.60 (Hair, 1998).
66
Factor analyses can be done from an exploratory or confirmatory perspective.

Exploratory techniques are often more useful early in the validation process,
while confirmatory techniques are far more common when the instrument in
question has been previously validated. The most common method for
extracting factors is Principal Component Analysis (PCA), which is
considered suitable when the research purpose is data reduction or
exploration, but should not be used in causal modeling. When the research
purpose is theory confirmation and causal modeling, Common Factor
Analysis (CFA) (e.g., Principal Axis Factoring [PAF]) is most suitable (Hair,
1998). An important tool for interpreting factors is factor rotation (Hair,
1998). Varimax rotation is the most common method for interpretation.
However, the varimax rotation method, which belongs to the group of
orthogonal rotation techniques, may be problematic to use. Orthogonal
techniques assume that the underlying factors are independent, but from the
theoretical perspective dimensions of, for example, safety culture, they are not
regarded as independent but as an integration of various sub-facets. To
validate a questionnaire, CFA is recommended with an oblique rotation
technique. Oblique techniques allow for correlation between factors and are
preferable when the researcher’s aim is to obtain several theoretically
meaningful factors or constructs (Field, 2005; Hair, 1998; Pett et al., 2003).
Moreover, confirmatory analysis is recommended when the final objective is
structural equation modeling (SEM) analysis (Hoyle, 1995).
Each item’s, or variable’s, “fit” with the underlying dimension is represented

by factor loading. Factor loadings range from 1 to -1; the closer to 1, the
better the representation of the underlying dimension. The definition of a
significant loading depends upon the sample size. A small sample size
requires higher loading than a large sample. In addition, variables’
67
communalities should be assessed. Communality refers to the total amount of

variance an original variable shares with all other variables included in the
factor analyses. Variables with low loadings and low communalities should be
considered for deletion (Hair, 1998). By one rule of thumb in CFA, loadings
should be 0.7 or higher to confirm that independent variables identified a
priori are represented by a particular factor. However, such high loadings (≥
0.70) are not typical, and real-life data may not meet this criterion. Thus,
some researchers—particularly for exploratory purposes—use a lower level
such as 0.4 for the central factor and 0.25 for other factors (Raubenheimer,
2004). On the other hand, factor loadings must be interpreted in light of
theory and their practical significance, not by arbitrary cutoff levels alone. In
addition, items with multiple significant loadings at various factors should be
deleted, as this is a sign of multidimensionality (Hair, 1998) and, thus, not
discriminant valid. In a multidimensional scale, it is recommended that a
minimum of three items load significantly in each factor. The more items
there are per factor, the more likely is it that the factor will replicate as
originally constructed (Pett et al., 2003; Raubenheimer, 2004). When using a
validated questionnaire, the extracted factors should be similar to the
theoretical construct as operationalized.
Factor correlation analysis is often done in order to check construct validity,

which is viewed as the extent to which an operational measure truly reflects
the underlying safety culture concept as well as whether it operates in a
consistent manner.
Validity and reliability through scale analysis

Various statistics can be selected in order to estimate the reliability of scale
and items, including alpha models, split-half models, Guttman models, and
68
parallel and strict parallel models. Cronbach’s alpha is extensively reported as

the most commonly accepted measure for internally consistency reliability
(Hinkin, 1995; Shevlin, Miles, Davies, & Walker, 2000). Internal consistency
(convergent validity) is statistically tested by Cronbach’s alpha coefficient
and inter-item statistics. Although no consensus exists with regard to the
Cronbach’s alpha coefficient, usually a value above 0.7 is considered
acceptable, although some advocate for a level of 0.8 or better, especially
when a new scale is being evaluated (Netemeyer, Sharma & Bearden, 2003;
Raubenheimer, 2004). Others suggest that when dealing with psychological
constructs, values below even 0.7 can, realistically, be expected because of
the diversity of constructs being measured (Field, 2005).
Although high reliability is generally cited as evidence of good psychometric

properties of a scale, it is noted that Cronbach’s alpha value on its own should
be used with caution (Shevlin et al., 2000). The value depends upon the
number of items in the scale and is a function of, inter alia, the inter-item
correlation and the item-total correlation. Thus, the inter-item statistics should
also be examined. Inter-item statistics, or convergent validity, are related to
the extent to which different scale items assumed to represent a construct
converge on the same construct. Convergent validity is the degree to which
multiple attempts to measure the same concepts agree, which may be tested
by the item-total correlations. Rules of thumb suggest that the item-total
correlation should exceed 0.5 (Hair, 1998) or 0.4 (Field, 2005) and the inter-
item correlation should exceed 0.3 (Hair, 1998), but not 0.8. An inter-item
correlation exceeding 0.8 suggests that items are duplicates of one another
(Pett et al., 2003). Moreover, Shevlin et al. (2000) argued that a high estimate
of Cronbach’s alpha may indicate the presence of systematic error, such as
scales deviating from unidimensionality. In such cases, extraneous variables
69
can make a substantial contribution to inflating the Cronbach’s alpha value

rather than the actual dimension being measured. Indeed, when the factor
loadings of the dimension being measured are low, the presence of systematic
errors can greatly inflate the estimate of Cronbach’s alpha, especially with
large sample sizes (Shevlin et al., 2000). These findings substantiate the
importance of reporting item statistics so that the presence of
unidimensionality can be evaluated, along with factor loadings and cross
loadings.
Causal relationship through structural equation modeling

Structural equation modeling (SEM) is applied to test the causal relationship
between the components deriving from factor analysis. Through SEM
analysis, it is possible to estimate multiple and interrelated dependence
relationships. SEM is focused on testing causal processes inherent in theory.
Moreover, this method has the ability to represent unobserved concepts, as
safety culture, in these relationships and account for measurement error in the
estimation process (Hair, 1998). The structural relations between tested
variables are specified with both a theoretical and empirical foundation.
SEM is an extension of both factor analysis and regression analysis. The

method serves purposes similar to multiple regressions, but in a way that takes
into account the modeling of interactions, nonlinearities, measurement error,
and correlated error terms. SEM also considers when the independent
variable(s) as well as the dependent variable are measured with multiple
indicators, such as when extracted using factor analysis. Hence, the
advantages of SEM compared to multiple regressions include more flexible
assumptions. SEM analysis also opens up the possibility to explore multiple
relationships simultaneously, where regression analysis only examines a
70
single relationship at the time, holding all other variables constant (Hair,
1998).
The final model is evaluated with goodness-of-fit criteria assessing the overall
model. Assessing the goodness-of-fit is not as straightforward as with other
multivariate dependence techniques (e.g., multiple regressions) as no single
test best describes the “strength” of the SEM model. Instead, a number of
goodness-of-fit measures have been developed; when used in combination,
the results are assessed from three perspectives: overall fit, comparative fit to
a base model, and model parsimony. However, there is no consensus of what
accurate levels of fit are, as none of the measures (except the chi-square
statistics) have an associated statistical test. According to Hair (1998), several
guidelines have been suggested, but ultimately each researcher must decide
whether the model fit is acceptable (Hair, 1998).
4.2 Qualitative research design
The term qualitative research refers to any kind of research that produces
findings not arrived at by means of statistical procedures or other means of
quantification. Although some of the qualitative data may be quantified, as
with census data, the analysis itself is a qualitative one (Strauss & Corbin,
1990). The most prominent qualitative research techniques employed in this
study are: (1) document study, (2) case studies, (3) interviews, (4)
participatory observation, and (5) participation in maritime forums.
Document study
Document study is an indirect method of data collection that does not require
participation of the subjects involved. Official maritime accident investigation
71
reports (CPEM, 1999; Danish Maritime Authority, 2009; Justis-og

politidepartementet 1991; National Transportation Safety Board, 1990; United
States Coast Guard, 2005) are studied in order to understand how safety
culture, safety management, and context interact and influence the course of
events.
In addition, administrative safety management documentation has been

studied in the four different companies selected for the case study (further
information about these cases follows). This includes safety meeting minutes,
reported events, root cause analyses, procedural manuals and checklists, and
other available relevant documentation. Analyses were performed with the
intention of understanding the companies’ approaches to safety management
and priority areas. Of particular interest were safety information data
analyses—namely, how the experience information was categorized and their
approach to identifying causes in cases of near-miss, incidents, and accidents.
Such analyses occurred with reference to the previous mentioned ages of
safety and to which degree technical, personal, and underlying organizational
causes were identified. The analyses were also seen in relation to safety
measures and changes done upon the processed information—namely, if
changes aimed at introducing more control in the form of procedures and
checklists or if changes were done in other levels of the organizational
structures and policies (e.g., manning policies). This part of the document
analysis was complemented with interviews.
The document studies have given valuable insights into understanding how
safety culture, safety management, and context interact and influence
accidents. These studies also provided a better understanding of shipping
companies’ approach to safety management. They have enabled the study of
72
past events and issues to identify changes over time; however, some
limitations exist. The documents are not representative; therefore, findings
cannot be generalized. Which documents were made available depended upon
each company; thus, a full comparative study between cases could not be
performed. In addition, all documentation should be considered as biased as it
represents the view of its authors (Sarantakos, 1998)
Case studies
Case study research involves studying individual cases—in this case shipping
companies—in their natural environment for a long(er) period of time,
employing a number of methods of data collections and analyses. Four case
studies were carried out in two tanker companies and two dry cargo
companies. The statistical results from survey data analyses were used as
criteria for selecting cases and focus area—namely, mixed crew nationality
and ship management. Thus, a better understanding of the structure,
processes, and complexities underlying the statistical results were achieved.
Administrative document studies (as previously described), interviews, and
participatory observation were performed along with seminar participation.
The information gathered is used to illustrate, explain, and expand the
quantitative findings. However, some of the drawbacks with case study as a
method are poor representativeness and poor replicability (Sarantakos, 1998).
Interviews
In contrast to document studies, interviewing requires direct interaction with

the respondents and heavily relies on their involvement, participation, and
contribution. Both formal and informal interviews were performed. Open
formal interviews were carried out with shore-side personnel working in
73
selected case companies within the department of safety management and/or

manning. When available, top-level management was also interviewed. The
interview process had to be adapted to the subjects’ availability; thus,
individual interviews were conducted in some cases and group interviews in
others. The scope of the interviews also changed over time as a better
understanding of the industry was acquired. Therefore, the first interviews are
more superficial in character then the last. All interviews were recorded.
When interviewing the sailing personnel, a more informal and ethnographic

approach was selected. Safety issues, behavioral norms, violation of
standards, and the like are sensitive issues. Most crews are not Norwegian and
do not enjoy fixed employment; thus, many fear losing their job from being
open about the situation. Most interviews conducted in the field studies were
done as a part of daily conversations after a trusting relationship had been
established. None of these interviews were recorded.
Participatory observation
Participatory observation at sea was also carried out. During the study, vessels
from various companies sailed for different periods of time, ranging from one
to two weeks.4 During the field studies, I participated in daily work activities
such as ballast tank cleaning, mooring operation, loading, and acting as watch
keeper at night. However, in a 24-hour society, leisure time is also of great
importance. During my spare time, I participated in leisure activities and
games. I experienced that spending time with the crew was essential to
4
The shipping industry is known to be a highly transparent industry. Thus, to ensure that all participants
remained unidentifiable, the number of vessels sailed is not included.
74
gaining their trust, and the first days no questions of a more sensible character
were asked. When interacting with the crew, informal interviews were carried
out as part of the daily conversation. Both interviews and observations aimed
to understand how the ship management influenced the work on board and to
understand team and group processes from within the group. Although the
crew knew about my research area in more general terms, research questions
were never explained in detail to avoid biases.
Participatory observation has the advantage that group processes and

management practices may be observed in their natural environment. In
addition, it is possible to retrieve firsthand information that respondents are
unable or unwilling to offer during formal interviews (e.g., deliberate
violations of safety standards). However, one limitation is that participatory
observation can only be employed with smaller groups; thus, findings cannot
be generalized. Although behavior is directly observed, the method does not
offer frequency of behavior. The method is also exposed to observer bias,
selective perception, and memory and offers no control measure regarding
bias, attitude, and opinion of the observer. In particular, the latter were
experienced during the field study, especially when working with lower-paid
crew from the third world originating from poor conditions. I had much
empathy for this group. Being aware of my own bias gave me some control
over the situation in order to remain objective.
Participation in maritime forums

Participation in maritime forums and seminars as a method was used for two
reasons: (1) to gain an understanding of the maritime context and
interrelationships and (2) to ensure quality assurance of results. Inspired by
William Foote Whyte and the Street Corner Society, an ethnographic
75
approach was adopted to gain an understanding of the research area and build
up credibility and trustworthiness. As a part of this, the maritime aspects
became integrated into my own social life as well as my research. Early in the
study, I reallocated myself from my institute’s economical and administrative
department, where I was originally employed, to the department of nautical
sciences. Along the road, I also became chairman of the Nautical Institute
Norway Branch, an international organization working to improve the safety
and efficiency of shipping. I then became an official member of the
Norwegian delegation participating in IMO meetings. Thus, by carrying out
research at the maritime industry, I became a part of the industry itself.
Several presentations of my results have been given at national and

international conferences, both for governmental and non-governmental
industrial stakeholders and company conferences. Feedback from the
audience has been used to interpret statistical results and define critical areas
for further investigation. As with participatory observation, going native
was—and still is—an area of concern.
4.3 Applied methods and statistics in articles
An overview of methods for data collection and statistics applied in the

different papers is presented in Table 2. Although not explicitly stated in all
articles, all quantitative results are interpreted in a qualitative framework.
76
Table 2.
Overview of Methods for Data Collection and Statistics
Article Id.
Data collection 1 2 3 4 5 6
Maritime document study x x x x x x
Field study at sea --- --- x x --- x
Shore interviews --- --- x x --- x
Maritime/safety related conference attendance --- --- --- x --- x
IMO attendance --- --- --- --- --- x
Survey comments x --- x --- --- x
Survey quantitative data x x x x x x
Analysis
Exploratory factor analysis (EFA) x x x x x x
Confirmatory factor analysis (CFA) x --- --- --- --- ---
Cronbach’s alpha x x --- --- x ---
Correlation (Pearson’s r) x x --- --- x ---
Inter-item, item-total statistics x x --- --- x ---
Analysis of variance (ANOVA) --- x --- --- --- ---
Linear regression analysis --- x --- --- --- ---
Logistic ordered regression analysis --- --- --- --- x ---
Development of factors used in analysis
Summarized items from extracted factor x x --- --- --- ---
structure
Transformed standardized factor scores --- --- --- --- x ---
Summarized items based on theoretical --- --- x x --- x
relationship
77
Research results
5 Research results
This section summarizes each article, including the objective, applied method,
main results, conclusions, and interrelationships. All articles are related to the
three research questions developed for the purpose of this thesis:
 What characterizes safety culture and safety management within the

shipping industry?
 What is the relationship between safety culture and safety performance

 What characterizes shipping companies’ application of the safety

management concept?
Finally, a structural model testing the causal relationship among the latent
dimensions of safety culture is presented.
5.1 Summary and results of article 1
Francis Group.
The objective of this article was threefold: (1) explore and analyze the
shipboard characteristics of safety culture; (2) elaborate upon which factors
affect the shipboard safety culture; and (3) use the results to set the direction
for future studies. The first two objectives are stated in the article.
78
Research results
The article first states the theoretical approach to safety culture, along with the
methodological framework. Safety culture is perceived as a reciprocal
interrelated fusion of three main elements. The first element is internal
psychological factors, including each individual’s attitude and perception
toward safety, work situation, and organization. This element was measured
using a questionnaire. In addition, formal and informal interviews were
performed to get a more comprehensive understanding of what is happening
inside people’s heads. The second element is observable safety-related
behavior, or what individuals are actually doing on board. Safety-related
behavior, which is perceived to be partly determined by the first element (i.e.,
psychological factors), was also measured by a questionnaire, in which
individuals reported their own behavior in working situations. In addition,
participatory field studies were conducted to observe actual behaviors.
Behavioral patterns are regarded as a manifestation of shipboard culture and
of a culture existing on a higher level in the organization (i.e., organizational
factors). Finally, the third element is organizational factors (e.g., employment
policy, safety management policy, approach towards efficiency versus safety),
which were measured by case studies and interviews. However, the
questionnaire also comprised questions related to organizational factors. In
this approach, the contextual influence is also important. Distinctive
characteristics of the shipping context are, inter alia, life and work on board as
a total institution and a 24-hour society.
A mixed method approach was applied, where retrieved data were integrated
into the interpretation, although this was not explicitly stated in the article.
Consequently, psychological measures traditionally referred to as safety
climate and safety culture traditionally explored by qualitative methods were
not distinguished. In this article (and the others), statistical data were all
79
Research results
interpreted in a qualitative framework. It was recommended that the

questionnaire not be used as a single method; thus, it was referred to as a
safety culture questionnaire.
The statistical approach was based on a combination of exploratory and

confirmatory factor analyses, using the method of principal component. The
exploratory factor analyses were carried out with orthogonal varimax rotation.
In the confirmatory factor analysis a one-factor solution on each construct was
performed. Based on a comparison of the exploratory and confirmatory factor
analyses, five factors were found to be valid: (1) crew interaction, (2)
reporting practices, (3) competence, (4) local management, and (5) working
situation (proactive work practices).
The qualitative data indicated that the shipping companies’ crewing strategy,
which includes employment terms, rotation systems, and policy toward
shipboard management, are interrelated with how safety culture is manifested
on board. Considering seafaring as a 24-hour society and the geographical
distance between the on-shore organization and the vessel may affect both the
quality of those systems and plans developed on shore and their
implementation on the vessels. Thus, ship management was identified as a
key factor to a sound safety culture along with the on-shore crewing strategy.
Oltedal, H., & Wadsworth, E. (2010). Risk perception in the Norwegian
Management, 37(6), 601-623.
Based on the findings from the first article, the second paper aimed to assess
the relationship among shipboard safety, safety culture, and shore-based
80
Research results
organizational factors, using risk perception as a proxy variable for the

general safety level on board. The objective of this article was to assess the
relationship between risk perception and the dimensions of safety culture.
With regard to assessing the relationship between risk perception and safety
culture, the article introduced risk perception as a dependent variable, thereby
expanding the trinity methodological framework presented in the first article
(i.e., person, behavior, and situation) with a measurable dependent variable.
Risk perception has been found to be adequate as a dependent variable, as
previous research indicates that risk perception and risk behavior are strongly
correlated. It is suggested that risk perception is a good indicator for safety
level in general, which is constrained by the situation and context.
Explorative principal component analysis with varimax rotation was applied

in order to explore the latent dimensions of safety culture. Eight factors were
identified, providing a good representation of the concept of shipboard safety
culture: (1) competence, (2) interpersonal relationship (crew interaction), (3)
shore orientation, (4) ship management (local management), (5) proactive
work practices, (6) feedback, (7) demand for efficiency, and (8) reporting
practices. Furthermore, a one-way analysis of variance (ANOVA) was carried
out to explore any associations between the demographic (nationality, age,
department, and vessel type) and organizational (work description) variables
as well as both dependent variables (i.e., risk perception and the independent
safety culture dimensions). A linear regression analysis (OLS) was
subsequently conducted to assess the associations between risk perception and
the dimensions of safety culture, controlling for any potentially influential
demographic and organizational factors.
81
Research results
Based on the results, the overall safety in respondents’ working situations was
perceived to be very high, which may indicate a relatively good safety
standard. The ANOVA analysis showed significant differences between age
groups in the dimensions of interpersonal relationship and shore orientation.
The type of vessel indicated significant differences on the dimensions of
competence, local management, and feedback. Those working on dry cargo
vessels perceived the feedback on reported experience data to be better than
those working on liquid tankers. However, those working on liquid tankers
had a better perception of their own level of competence and their local
management than those working on dry cargo vessels. Finally, work
description (i.e., teamwork versus individual) showed significant differences
on all dimensions of safety culture. Crews working on a team perceived the
dimensions of competence, interpersonal relationship, local management,
feedback, and reporting practices to be better than those working on an
individual basis. Crews working on a team also felt less demand for efficiency
and perceived the shore side of the company to be more safety orientated.
The regression analysis indicated that local management, working practices,

and reporting practices have a positive association with risk perception while
demand for efficiency has a negative association with risk perception. The
working situation has a positive association when work is performed on a
team in contrast to when work is performed on an individual basis. None of
the demographic data were significantly associated with risk perception.
For future research, it was suggested to further examine the characteristics of

teamwork along with the concept of group identity. It would also be of
interest for future research to examine risk perception both in general and in
82
Research results
relation to potential differences among nationalities, along with differences

among white, grey and black listed flags of registration.
Given that the overall research topic in the current thesis concerns safety
management on dry cargo vessels and liquid tankers, the remaining articles
address safety management within dry cargo vessels and liquid tankers in
greater detail, starting with liquid tankers. The third article also suggests
differences with regard to safety management within the two sectors.

The aim of this article is twofold: (1) describe safety management within the
liquid tanker sector and (2) identify factors that influence safety management
performance.
The theoretical rationale of traditional safety management systems was

introduced for the first time in this article. Traditional safety management was
presented as a system containing four sub-systems: (1) reporting and
collection of experience data from the vessel; (2) data processing,
summarizing, and analysis; (3) development of safety measures; and (4)
implementation. The topic of focus was presented in a situational context
where safety concerns need balance to ensure profits and economical
concerns. Safety management as such is related to the International Safety
Management (ISM) code. This article was the first to explicitly state that it
83
Research results
adopted a multi-method approach combining surveys, case studies, field

studies, interviews and other qualitative information, although all the
concerned articles are based on a multi-method approach.
With regard to the statistics and factor analysis, the survey items were
grouped as they relate to the information flow in a safety management
systems—more precisely: (1) crew’s reporting practices, (2) analysis and
follow up by shore side, (3) procedures and checklists, and (4) perceived
balance between commercial pressure and safety concerns. Explorative
principal component analysis was carried out in each group in order to
examine the items’ interrelationships. Shore-side development of safety
measures was also analyzed, although most findings stemmed from the
qualitative data.
The results indicated a situation with substantial underreporting of experience

data from the vessels. Such underreporting may be explained by the crew’s
fear of negative consequences, a complicated reporting system, and a lack of
understanding of the overall safety management system. The development of
measures tends to focus on controlling human actions, often in the form of
excess use of procedures and checklists. This situation was traced back to a
person-oriented approach in safety management. Moreover, procedures and
checklists are often perceived as being problematic to use in daily shipboard
operations. In order to turn such a situation around, it was suggested that the
seafarers’ experience be taken seriously, with regard to both reasons for
underreporting and their experience with new measures. On board, the ship
management was identified as a factor strongly influencing the shipboard
situation, and it is suggested that the shore side pay more attention to that
84
Research results
element. Other organizational factors that were suggested to influence the

situation included employment conditions and crew stability.
As the overall research topic in the current thesis concerns safety management
on both dry cargo vessels and liquid tankers, this article focused on safety
management within the dry cargo sector in order to compare the findings with
the liquid tank sector. The aim of this article was to (1) describe safety
management within the dry cargo industry, (2) identify factors that may
influence safety management performance, and (3) compare the current
situation within the dry cargo to the liquid tanker industry. This was the first
article to provide a description of the two sectors (i.e., liquid and dry cargo)
with the purpose of conducting comparative exploration and analysis. The two
sectors were introduced in order to make each sector’s major safety
challenges visible.
This article served as a follow-up to the third article and, thus, followed a
similar structure related to the information flow of a traditional safety
management system. Although a multi-method approach was applied, only
areas in which statistical data were available for both sectors were presented
and analyzed—namely, (1) crew’s reporting practices, (2) procedures and
checklists, and (3) perceived balance between commercial pressure and safety
85
Research results
concerns. An explorative principal component analysis was carried out at each

group in order to examine the items’ interrelationships.
This article pointed to a central custom-related difference between the two

sectors. Although customers of dry cargo shipping have fewer safety-related
requirements, the tanker sector is extensively embedded in the Norwegian oil
industry, which explains the extended focus on safety in general, safety
management, attitudes, etc., within the tanker industry compared to dry cargo.
Underreporting of minor incidents and near misses are more present within
the dry cargo industry. In both sectors, reporting frequency is correlated with
feedback given upon reported events. Although both sectors have substantial
underreporting, our data indicated that the safety campaigns, which are typical
for the oil industry, have some positive effects on reporting practices.
However, within the dry cargo industry, such feedback is perceived as better
than within the tanker sector. The analysis suggested that the shore tanker
organization is not prepared to manage the growing workload that increased
reporting brings about. The organization fails to obtain feedback’s motivating
effect, which again may counteract the effect of safety campaigns. However,
it was suggested that the dry cargo industry could benefit from such
campaigns when it comes to increasing crews’ awareness and recognition of a
near miss, along with a better understanding of reporting’s importance in
safety-management systems. It was recommended that the shore side provide
resources for the potential increase in number of reports placed and proper
follow-up. When following up the reports, the development of new measures
should also be considered carefully, and alternatives to the development of
new procedures should be developed. The experiences from the tanker
industry suggested that the development of a constantly increasing, detailed,
and extensive procedural system may undermine safety. Moreover, the
86
Research results
analysis suggested that safety measures be initiated by internal and industrial

need, not external demands from the customer. When externally initiated,
safety management may be less integrated into the operational part of the
organization. It was also suggested that this external demand is related to the
existence of a poor procedural system.
Both the third and the fourth articles pointed to a situation with substantial
underreporting of experience data. Reporting is regarded as a critical
cornerstone in formal safety management. Some influencing factors were
suggested, and the fifth article further assessed the relations among reporting
practices, safety culture, and organizational factors.
Oltedal, H., & McArthur, D. (2010). Reporting practices in merchant
shipping, and the identification of influencing factors. Safety Science, 49(2),
331-338.
This article pursued three aims: (1) assess the relationship between reporting
practices and safety culture, (2) explore the influence of familiarity with local
managers, and (3) further explore differences between vessel type (i.e., liquid
and dry cargo).
Reporting practices are regarded as a cornerstone when working with

systematic safety management, and this article was the first to statistically
explore factors affecting reporting practices. A review of safety research
within the maritime sector indicated that—although underreporting of
experience data is regarded as a major problem within the industry—little
research has been done. With a foundation in both theory and results derived
87
Research results
from research within other high-risk sectors, barriers to experience data

reporting were summarized in the following categories: (1) fear of
disciplinary action or of other people’s reactions; (2) risk acceptance, where
incidents are regarded as a part of the job or unpreventable; (3) useless, as
reporting does not lead to any changes; and (4) practical reasons like time
pressure or a complicated reporting system.
In the statistical analysis, explorative principal component analysis with

varimax rotation was carried out, followed by a scale reliability analysis.
Seven factors were extracted and found to be valid, reflecting crews’
perceptions of (1) their own competence, (2) interpersonal relationships
among the crew, (3) shipboard management, (4) work practices, (5) feedback
on reported safety information, (6) shore orientation to safety, and (7)
perceived demand for efficiency. An ordered logistic regression was then
carried out in order to explore the relationships between extracted factors and
reporting practices. In the analysis, the dependent variable reporting practices
were measured with four possible outcomes: (1) never/seldom, (2) sometimes,
(3) often, and (4) always.
The results indicated that high competence, a good and open interpersonal
relationship among the crew, safety-oriented management, execution of
proactive work practices, and feedback on reported events all increase the
odds of being in a higher category of the reporting frequency measure. On the
other hand, shore orientation downgrading safety and prioritizing efficiency
increase the odds of being in a lower category of the reporting frequency
measure. With regard to feedback, vessel type, and management, the effect of
these variables is dependent on the category of the dependant variable. The
effect of both vessel and feedback is larger when moving between the higher
88
Research results
categories. Crews who have been working with their closest manager for more
than one year tend to report more often. However, the effect of being familiar
with one’s superior is larger when moving between the lower categories.
None of the identified factors should be addressed in isolation from each

other. As followed from the discussion, they are all important and mutually
dependent. Thus, the internal relationship between the identified dimensions
of safety culture should be further explored using, for example, structural
equation modeling and/or path-analysis, as performed in this thesis.
Oltedal, H., & Engen, O. A. (2011). Safety Management in Shipping—

Making sense of limited success. Safety Science Monitor, submitted
This article sought to (1) explore the gaps between the safety ambitions (in the
form of, e.g., rules and procedures) and operational practice on board and (2)
identify possible pitfalls when relying on safety through a system perspective
(as described in articles three through five), without focusing on its human
interrelationships.
A multi-method approach combining surveys and case studies, including field

studies and interviews, was applied. Previous articles (i.e., three through five:
Oltedal, 2010; Oltedal & Engen, 2010; Oltedal & McArthur, 2010) all pointed
to substantial weaknesses in current safety management; thus, this article
explored the theoretical rationale behind traditional safety management.
Normal Accident Theory (NAT) and High Reliability Organization Theory
(HRO) were outlined, with an emphasis on how they explain and make sense
of safety, risks, and accidents. The conceptual framework of the Practical
89
Research results
Drift Model (PDM), which seeks to combine NAT and HRO and explain why
seemingly well-regulated organizations (e.g., within shipping), develop traits
that may evolve into big accidents and disasters, was applied. The PDM
model allows us to look for possible new explanations of the success and
failure of safety strategies. In an effort to question what the scientific safety
literature has offered to make sense of the gaps between safety ambitions and
the practical outcome, some examples of gaps from our own data were
examined more closely.
The structure of the article followed the four stages of the PDM model, with
each stage being supported by qualitative and quantitative data. (1) Design
refers the stage in which organizational managers or designers develop
extensive routines and procedures in order to make the organization robust
and resilient against attacks and unforeseen events. (2) Engineered refers to an
operational situation in which the routines and procedures are first applied and
experienced to not always match the real situations. (3) Applied refers to
situations in which the designed control measures are substituted with a logic
of action based on individuals’ experiences and tacit knowledge. (4) Fails
refers to a situation of change (e.g., major unforeseen event such as ship
collisions), when individual units are forced to act on the assumption that all
others are acting in accordance with the original rules and procedures as they
were initially designed. The actors are then trapped in a game where trusting
their own logic of actions is the only solution while they must simultaneously
base their decision on the assumption that others are following the general
rules.
The article concluded that the industry could gain from abandoning the
person-oriented approach, where control measures are designed to control
human actions, often in the form of the excess use of procedures and
90
Research results
checklists. These measures developed through a traditional safety

management system are standardized to fit all—whether in a fleet of 5 vessels
or 100 vessels. This creates a paradox when confronting actual work
situations, where operations are never the same. The vessels are different, the
people, constellations of people, power figurations, weather and so on. A
standardized measurement will therefore never align with reality. Yet human
actions and deviations are compared to the standard and found to be
erroneous. In an attempt to gain control, new and even more detailed
measurements may be developed, thereby creating a vicious cycle resulting
from the anxiety of not being in control. Organizations should abandon such
person-oriented approaches in their search for causal and influencing factors.
5.7 Causal relationships between components of

safety culture
In order to test the multidimensionality of the theoretical safety culture
construct, a first-order confirmatory factor analysis (CFA) was carried out,
resulting in six dimensions that were found to be a reliable reflection of the
safety culture concept: (1) company orientation, (2) local management, (3)
crew interaction, (4) competence, (5) proactive working practices, and (6) risk
perception. All dimensions and their interrelationships were further discussed
in articles number 1 (Oltedal & Engen, 2009), number 2 (Oltedal &
Wadsworth, 2010) and number 5 (Oltedal & McArthur, 2010) and are briefly
presented below.
Company orientation reflects the crew’s perception of the shore organization.

A high score indicates that the company has a reactive approach and statistics
are a major concern, whereupon the crew perceives the safety work to be a
façade. A low score indicates that the company is proactive and cares about
91
Research results
the human consequences of hazardous situations, whereupon the crew

perceives the company’s safety work to be a real priority area.
Competence reflects the crew’s perception of their own training and ability to
work safely and handle critical and hazardous situations. A high score
indicates that crews see themselves as having a high level of competence in
these areas; a low score indicates a low competence level.
Local management reflects crew’s perception of their closest manager as a

role model as well as the manager’s engagement and interest in ensuring
safety in work operations. A high score indicates good safety management; a
low score indicates poor safety management.
Proactive work practices reflect performance of proactive activities, such as

safe job analysis and hazard identification, as well as how safety is prioritized
in daily operations. A high score indicates that proactive work practices are
performed on a regular basis, whereas a low score indicates the infrequent
performance of proactive work practices.
Group interaction reflects the relationship amongst the crew, including their
problem-solving abilities, form of communication, and sharing of safety
information. A high score indicates the presence of a good interpersonal
relationship amongst the crew, whereas a low score indicates that the on-
board group interaction is poor.
Risk perception is an indicator related to the onboard safety and the crew’s
own assessment of the probability that they or any other crewmembers will
have an accident on board the vessel during the next 12 months. A high score
indicates a low probability of an accident to occur; a low score indicates a
high probability that an accident will occur.
92
Research results
In the SEM structural model (see Figure 4) the latent dimensions of safety
culture are tested for causal relationships. Postulated causal relationships are
grounded in both theory and empirical results. The hypothesis is that the
shipping company’s orientations toward safety at the shore side of the
organization influence the safety culture on board the vessel; thus, company
orientation towards safety is set as an exogenous variable. The outcome
measurement variable is represented by the crew’s overall risk perception.
Each path was analyzed and evaluated; paths with no significant effect were
removed from the model.
Figure 4: Structural model testing for validity of causal structure
All paths in Figure 4 are reasonable and consistent with the theoretical
construct. The model shows a good fit (RMR=0.044, CFI=0.957, and
RMSEA=0.052). The root mean square residual (RMR) represents the
average residual value derived from the fitting of the variance-covariance
matrix for the tested hypothesized model. In a well-fitting model, this value
will be small (0.5 or less). The Comparative Fit Index (CFI) is a measure of
93
Research results
complete covariation in the data. CFI ranges from zero to 1.00, with values
close to 0.95 being advised. The root mean square error of approximation
(RMSEA) takes the error of approximation in the population into account.
RMSEA values of less than 0.05 are indicative of a good fit between the
hypothesized model and the observed data. RMSEA values as high as 0.08
represent reasonable errors of approximation (Byrne, 2010). The standardized
regressions weighted (Stand. Reg.) together with significance value (P) are
presented in Table 3.
Table 3.
Standardized Regression Weights and Significance Value of Structural Model
Dimension of safety culture Stand. Reg. P

Local management. <--- Company orientation -.390 ***
Crew interaction. <--- Local management. .494 ***
Competence <--- Local management .169 ***
Competence. <--- Crew interaction. .575 ***
Proactive work <--- Local management .300 ***

Risk perception <--- Company orientation -.256 ***
Risk perception <--- Crew interaction. .131 ***
Risk perception <--- Local management .122 .002
Proactive work <--- Competence .233 ***

i
Proactive work <--- Crew interaction. .194 ***
Standardized total effect of each latent variable is further presented in Table 4.

Standardized total effect is the sum of each latent dimension’s direct and
mediated effect.
94
Research results
Table 4.
Standardized Total Effects
Local management
Squared Multiple
Crew interaction
Company orient.
Risk perception
Proactive work
Correlations
Competence
practices
Local management -.390 .000 .000 .000 .000 .000 .152
Crew interaction -.193 .494 .000 .000 .000 .000 .244
Competence -.177 .453 .575 .000 .000 .000 .456
Risk perception -.329 .187 .131 .000 .000 .000 .151
Proactive work
-.195 .501 .328 .233 .000 .000 .362
practices
The results from the SEM analyses indicate that company orientation has a
direct effect upon the performance of local management on board the vessel
(-.390) as well as a direct effect on the general risk perception (-.256). When
safety work in the company is not perceived as a genuine effort, local
management is perceived as being less safety oriented while the risk level on
the vessel is generally perceived to be higher. Management/leadership style,
commitment, and visibility are also among the most commonly measured
95
Research results
dimensions in safety research in general (Flin, Mearns, O'Connor, & Bryden,

2000). Within the maritime sector, research initiated by the Maritime and
Coastguard Agency in the United Kingdom identified various core leadership
qualities necessary for effective safety leadership (Maritime and Coastguard
Agency, 2004). On board, these qualities were primarily geared toward the
captain as a key leader of safety, but also toward lower ranks with leadership
responsibilities.
Local management demonstrated a direct effect upon the competence

dimension (.169), proactive work practices (.300), risk perception (.122), and
crew interaction (.494), suggesting that local management plays a major role
in the development of safety culture on board. Proactive work practices
increase the chances of revealing potentially dangerous situations; thus,
preventive measures should be introduced in advance of operations. Error
detection and correction are also assumed to be stimulated by teamwork
(Kontogiannis & Malakis, 2009). The relationship between local management
and company orientation (-.390) indicates that the company is the driving
force for how management is performed. The relationship between company
orientation and local management is one directional, indicating that
experiences from shipboard management do not influence the overall
company policy. Likewise, management-relationships are one directional,
indicating that the management is not adapted to the shipboard situation—an
individual is either a good manager or he/she is not. Traditionally, no
requirements concerning formal management competence or training have
been established within shipping. Such requirements were recently adopted by
amendments of the STCW convention, which will put into force on January 1,
2012.
96
Research results
Furthermore, competence has a direct effect on proactive work practices

(.233), indicating that well-informed crews trained to handle risk are better in
performing proactive work practices. However, quite surprisingly, proactive
work practices have no direct effect on the general level of risk perception.
Competence includes both training acquired on board the vessel and training
acquired on shore. Also included are the orientation new crew members
receive when joining the vessel, which is often referred to as familiarization
and safety-related drills carried out on board. Within safety research,
competence is among the top five most commonly measured themes (Flin et
al., 2000). Minimum training requirements are covered by international
conventions and regulations developed by IMO, with parts are required to be
performed onboard. For example, every crew member must participate in at
least one abandon ship and one fire drill every month. These drills should, as
far as practicable, be conducted as if there were an actual emergency
(International Convention for the Safety of Life as Sea, 1974, International
Maritime Organization, 2009). On board, the captain and the ship
management are ultimately responsible for how such drills—as well as other
onboard training arrangements—are carried out.
Finally, crew interaction has a direct effect upon the competence dimension
(.575), proactive work practices (.194), and risk perception (.131). This
relationship is also one-directional; thus, crew interaction is suggested to be a
driving force when it comes to how onboard competence-increasing activities
are performed as well as proactive work practices. The relationship between
local management and crew interaction is one directional, suggesting that
management plays a major role in how crewmembers interact, but
management does not adjust its style to the crew’s characteristics and
dynamics. A trusting relationship is regarded as a key factor in forming a
97
Research results
safety culture (Reason, 2001). The importance of group interaction and

teamwork has also been suggested as influencing accidents to a greater extent
than individual unsafe acts (Barnett, Gatfield & Habberley, 2010; Mitropoulos
& Cupido, 2009).
5.8 Summing up and presentations of main

conclusions
The main conclusions drawn from the research indicate that the shore side of
the company is the driving force for onboard safety culture. Thus, undesirable
onboard working practices may be traced back to shore-side organizational
decisions. The onboard conditions are influenced by the shore side in three
ways: (1) the organization’s manning policy (e.g., contract arrangements, mix
of nationalities, and rotation systems), which establish the premises for the
cultural development; (2) the approach to safety management, which lay the
foundation for the reporting culture and system efficiency; the industry to a
large degree is person oriented, with excessive use of standardized measures;
and (3) shipboard management has a major mediating influence as the shore-
side organization’s contribution as a support function sets the stage for
shipboard management performance. All three areas are further discussed in
the following chapter.
98
Discussion
6 Discussion
This chapter presents a discussion of the overall results from all articles
included in this thesis. The discussion is related to the research questions
given in the introduction:
1. What characterizes safety culture and safety management within the

shipping industry?
2. What is the relationship between safety culture and safety performance
3. What characterizes shipping companies’ application of the safety
management concept?
At the international level, safety management within shipping is carried out

through the delegation of authority from the UN to IMO, delegating
responsibility for developing and maintaining workable safety regulations.
Each coastal state is responsible for the enforcement of maritime regulations,
while each flag state is responsible for ensuring compliance with international
regulations. The results of this study indicate that the structure of the
administrative authorities has some deficiencies as some maritime
administrations may be involved in shipping activities without fulfilling their
safety responsibilities. In many instances, they are unable to detect real safety
threats on board when performing inspections. However, the shipping
companies have the primary responsibility for the safe operations of ships and
the welfare of the crew. Safety management is regulated through the ISM
Code, which was developed by the IMO. One of the intentions behind the
ISM Code has been to develop a safety culture within shipping. Study results
indicate several deficiencies related to both the industry’s application of safety
99
Discussion
management and the underlying theoretical rationale. It is important to

emphasize that there are variations within companies.
Articles 3 and 4 indicated that the application of safety management in

relation to culture is associated with two areas. First, safety culture is
associated with a reporting culture. A vessel with a high reporting frequency
is regarded as having a good safety culture. Reporting emphasizes what has or
could have gone wrong, while experience data and suggestions for
improvement are less emphasized. This restricts the crew’s ability to learn and
improve. The qualitative results point to several situations where less serious
episodes (e.g., a misallocated knife) has been reported while more serious
episodes (e.g., gas on deck) have not. Reasons for not reporting include fear
of being blamed by the ship management or shore organization. The review of
investigation reports shows that human error is still identified as the major
cause of maritime accidents, which impedes the possibility of revealing
underlying organizational causes. An inherent intention of the ISM Code was
to move the shipping industry application of safety management from the age
of human factors to the age of organizational factors. This study indicates that
the application of the code still operates within the era of human errors,
resulting in safety measures aiming to control human performance through
procedures and checklists. Second, a safety culture is associated with
compliance with the developed safety measures, but without questioning
whether or not the given safety measures are adequate for the situation.
Depending upon the type of operation, the crew is expected to relate to safety
directives given not only by their own company, but also by the operator,
charterer, and customer. In many instances, these directives are not applicable,
are incompatible, or do not reflect the onboard operation. Better integration of
operating personnel’s experience and expertise could improve this situation as
100
Discussion
safety measures in some cases are developed by people with no seagoing

experience. The overall study results indicate that such weaknesses are related
to the shore side of the organization. The SEM analyses support the
relationship between shore-side and shipboard practices. In his thesis, Soma
(2005) found that safety is a quality of the ship owner rather than the vessel.
Within the industry—especially within the liquid tanker sector—reporting and

compliance are attempted by safety campaigns aimed at altering the crew’s
attitudes. However, according to the theoretical discussion, personal attitudes
are less amendable by such means. Despite the campaigns, the results indicate
that a substantial underreporting of experience data occurs. It could be time to
focus attention on shore-side personnel. Launching safety campaigns aimed at
shore-side personnel would enable them to gain a better understanding of their
own roles as a support function and how their practice affects operational
practice—not only with regard to safety management per se, but also
concerning the influence of other parts of the organization (e.g., commercial
pressure and crew resource management).
Another drawback is standardization. Measures are generally standardized to

fit all vessels within a company fleet. The shore personnel fail to understand
the diversity created by situational circumstances in operations, crew
constellations, the vessel’s technical condition, and so on. Accordingly, the
crew bypasses the problem with standardization and develops their own
deviating working practice. In particular, article 6 dealt with this subject.
These processes may be seen in relation to Snook’s theory of practical drift
(Snook, 2000). Formal instructions given by the company are complied with
on paper, but onboard the vessels working practices are adapted to the
situation. However, as the crew at this point uses their experience and tactical
101
Discussion
knowledge to ensure safe operations, deviations from standardized procedures

should not uncritically be put on par with unsafe operations.
The standardization of safety measures is necessary in relation to crew

organization and lack of crew stability. Life and work on board are known to
be a highly formalized, hierarchical, and authoritarian organization, with a
clear chain of command, clear communication lines, levels of authority, and
clearly defined tasks and activities that are more or less the same on all vessel.
Such organization makes it possible to handle crew as changeable
components. Indeed, within the industry, such an organizational structure is
necessary in order to handle the lack of crew stability. However, the lack of
both crew stability and standardization may be problematic for other safety
reasons as well. A lack of crew stability may be a barrier to change, and safety
management is all about change—changes of work practices and attitudes and
others; thus, crew may not see the benefit of their efforts. By the time
necessary changes are highlighted, crew have most likely signed onto another
vessel, where they might experience the same problems. Shipboard
management may be problematic as the ship’s management is not familiar
with each crewmember’s capabilities and limitations; by the time they finally
get to know them, the crewmember changes vessel. Finally, crewmembers
lack familiarity with the specific ship, the rest of the crew, and the
management. The influence and importance of crew stability is discussed in
articles 1, 3, 5, and 6. Crew stability also relates to team dynamics; article 2
indicates that the onboard safety level is perceived as better when work is
performed as a team.
The overall study results further identified the onboard management as a

driving force. Management and leadership as mediating factors have support
in climate research from Zohar (1980), who found management’s attitude to
102
Discussion
be of major importance in fostering a safe work environment. This is also

supported by cultural research done by Pettigrew (1979) and Schein (1992,
2004), who both regard individuals (entrepreneurs and leaders, respectively)
as important in the processes of creating and managing an organizational
culture. In articles 2 and 5, ship management was statistically associated with
risk perception and reporting practices. The shipboard management is
responsible for implementing prevailing regulations and company policies.
The SEM analyses demonstrated that local management has a direct effect
upon all shipboard dimensions of safety culture and that the quality of the
shipboard management is influenced by the shore side of the company.
Evidence indicates that the ship officers are getting dual instructions regarding
safety performance as well as demand for efficiency from other parts of the
organizations. Thus, safety and efficiency are balanced in daily operations,
where efficiency and corner-cutting activities are rewarded in absence of
accidents, but are simultaneously identified as causal factors when accidents
occur. Ambiguous instructions from shore should be counteracted by a strong
shipboard management. However, according to the results, few ship managers
have formal management and leadership education. In addition, shipboard
management is not evaluated by crewmembers on a regular basis. By
recognizing the importance of ship management, the company should provide
support in order to ensure adequate onboard management and leadership—for
both the crew and the shore side.
Shipping companies’ understanding of safety management has also resulted in

more administrative work, which makes less time available for attention to
operations. The constant development of new procedures increases the
system’s complexity as defined by Perrow (1999), resulting in the industry
becoming increasingly prone to accidents. The work situation is less easy to
103
Discussion
understand when the procedural framework does not fit the situation and
when each procedure is not internally coordinated with the overall framework.
With reference to Perrow (1999), fewer procedures may simplify the system
characteristics and increase resilience. To this end, an understanding of when
standardized measures and procedures are appropriate should be reached. This
is not to suggest that standardized measures and procedures should be
completely abandoned. However, the crew should be able to trust their
relevance and applicability, which relates to the casual rationality inherent in
safety management systems along with limitations of foresight.
Safety management as described by the ISM code is based upon causal

rationality, where attempts are made to prevent future events by reflecting
upon previous experiences. However, insufficient scientific evidence exists on
the effectiveness of systematic safety management to make recommendations
either in favor of or against them (Robson et al., 2007), which may be related
to their application. As pointed out by Rundmo (1996), the application of
traditional safety management is questioned with a distinction between small-
scale accidents and larger and more infrequent accidents (Rundmo, 1996).
Measures developed through the analysis of past events are in theory only
applicable for frequent and small-scale accidents, such as ordinary work
accidents. When it comes to medium-size and larger and more infrequent
accidents, traditional safety management is not considered to be applicable.
Such accidents are often too unique and complex to grasp or to single out
some isolated underlying causes. When the course of event is unclear, it is
difficult to develop measures that cover all involved risk as the risk in the first
place is considered too complex to be fully understood. This highlights the
second point: the limitation of the foresight of future events. The study results
104
Discussion
indicated that the shipping industry does not differ sufficiently between these
types of events and, thus, applies the same logic independent of type of event.
The results from the study point to a situation in which serious accidents are
on the rise, despite the introduction of the ISM Code and systematic safety
management. This can also be seen in relation to the different ages of safety
and the recent development of an adaptive age. The adaptive age of safety is
characterized by a shift from reliance on systems supported by safety culture
to operations. The focus on operations is also in accordance with the theories
of HRO and mindfulness. When recognizing the limitations of safety
management systems and safety rules, which attempt to control human
behavior, it is proposed that adaptive cultures should be embraced (Borys et
al., 2009). Consequently, resilience engineering requires a change in
perspective from human variability as a liability and in need of control to
human variability as an asset in a situation getting out of control, thereby
making it important for safe operations. Embracing variability as an asset
challenges the comfort of management and, thus, may meet resistance from
the industry. However, with reference to the standardization and development
of global rules, the industry could ask whether they are made for comfort and
to simplify the work for shore personnel or to support safety in daily
operations. A consequence of safety management is to shift the focus to how
crew is coping in daily operations under constantly shifting circumstances as
well as learn from their adaption processes. This should also be seen in
relation to the fact that most of the time, when crewmembers adapt, the
operations are still performed in accordance with the super-eminent objectives
(Hollnagel, 2009). In order to follow such an approach, the crew should be
able to trust in the applicability of procedures and the safety measures that
ought to be in place. Second, crewmembers should have the possibility to
105
Discussion
develop skills, competences, and tactical knowledge in order to handle any

unexpected, infrequent situations that cannot be prevented through traditional
safety management and standardized measures. Third, when things do go
wrong, organizations should remove themselves from a person-orientated
approach in which operators are blamed.
Safety culture encompasses not only what is done at the operational level, but
also at all levels in the company. In order to improve safety, companies
should look for influencing factors derived from organizational structures and
policies. The shipping industry is known to employ crewmembers from
various nationalities through manning agents located in each country, without
giving the crewmembers a fixed company employment, thereby resulting in
instability. With this in mind, one might question if some of the vessels do
have a (safety) culture at all as they lack stability in the group. Without
structural stability, they have not been given the possibility to develop the
deeper and less conscious levels of cultural patterns and sense making; nor do
they have the possibility to develop the experience and tactical skills
necessary to handle new and emerging unexpected situations as their efforts
are orientated toward making sense of what is unfamiliar—namely, new crew,
new power relations, new social constructions, and the like. The study results
indicate that loose conditions of employment affect the overall safety
management system (e.g., in the form of more underreporting of experience
data due to fear of negative consequences and a lack of trust). By changing
such a manning strategy, the company could have better possibilities to create
a positive safety culture and build competence so that the crewmembers have
better premises for handling unsafe situations.
106
Concluding remarks
7 Concluding remarks
This chapter includes a discussion of the major limitations for the thesis with
regard to methods and measurements. Suggestions for future research and
implications are also addressed.
7.1 Methodological limitations
The major methodological limitations concern the validity of the
questionnaire and sample characteristics. The questionnaire did not show the
ability to cover all aspects of the safety culture concept, which is a limitation
affecting the overall validity of the study. Due to low reliability, several items
and constructs were excluded from the further statistical analysis. One
explanation is poor representative reliability across subpopulations or groups
of people (Hair, 1998; Neuman, 2000). It may be fair to assume that some
groups (e.g., Norwegian employees or senior officers) are better informed
about their company’s strategic and tactical management and operations and,
therefore, are better placed to answer some of the questions related to the
company. A second issue concerns the constructs itself. DeVillis (2003)
focused on the fact that the items constituting a construct or dimension should
share a common cause or consequence. Some of the questionnaire constructs
did not meet this latter requirement.
Biases could also be produced by national differences, languages, and

response style. In a cross-national study, Harzing (2005) found that English
language survey versions tended to be more homogenized, potentially
obscuring cross-national differences. McCrae (2001), who studied
Norwegians and Filipinos, did not find such differences. Given that
questionnaire language could potentially bias the results, the questionnaire
was also made available in Norwegian, Polish, and Tagalog. Taking this into
107
Concluding remarks
consideration, cross-cultural comparisons of results are not performed in the

current study. In addition, indicated differences at the organizational level
(e.g., between type of vessel and employment terms) should not be
overestimated.
The survey data are representative of vessels flying a white and grey flag
only, as those registered under a black listed flag did not want to participate.
As participation was voluntary on behalf of the company, it is assumed that
those participating do, in general, emphasize safety in their operations; thus,
the results are biased in a positive direction. Moreover, the survey data are
only representative of members of the Norwegian Shipowners’ Association.
The possibility exists that the results are subject to the common method bias
(Podsakoff et al., 2003) due to the data deriving from a common source (e.g.,
a common scale for different questions). Potential statistical remedies have
been suggested. Spector (2006) is skeptical of the merits of such approaches.
He argued that—given that it is not possible to know the existence or extent
of any possible bias—treating it could in fact introduce more bias than what
existed in the first place. He recommended using a multi-method strategy so
that results do not rely exclusively on the results of one questionnaire. In the
current research, case studies, interviews, participatory, and field studies were
used to validate the data.
As with the survey data, a question of validity arises to the qualitative data. In
order to be certain that the elements of culture identified by qualitative
methods, Hopkins (2006) recommended consulting the members of the
culture. If the members of the culture fail to recognize the description of their
culture, the description must be called into question. Accordingly, all results
in this thesis were also presented to several people working within the
108
Concluding remarks
industry; they expressed that they believe the results to be giving an accurate
representation of the situation.
7.2 Theoretical limitations
The principal objective of this thesis has been to examine the role of safety
culture for safety management and vice versa. Limitations also follow from
the theoretical stand and research perspective. By focusing on cultural
influences on safety management, other areas of equal importance give way.
Research with other perspectives (e.g., professional culture, national culture,
or a sociotechnical approach) would bring about different results. For
example, technological changes have unquestionably left their mark on both
operational safety and the organizational structure of the industry. The
shipping industry has, since the early 1960s, steadily adopted the automation
and integration of new technology (Alderton, 2004). Yet despite the
introduction of new technology partly intended to increase safety by, for
example, reducing human error, new technology may also be the cause of new
and emerging risk (Schager, 2008). This could be a mismatch between
ergonomic aspects and the human information processing system,
overreliance in technology that may fail, loss of operational skills and
experience necessary for handle critical and unexpected situations, or changes
in the social and organizational system.
7.3 Future research
With estimation that 75% to 96% of marine casualties are caused by some
form of human error (Anderson, 2003; Rothblum, 2000; Wagenaar &
Groeneweg, 1987), human error is possibly overemphasized as a causal
explanation for accidents at sea. Research on organizational and structural
factors in shipping accidents indicates that the human element is identified as
109
Concluding remarks
a causal factor without addressing the relationship to underlying

organizational and structural factors. Still, the need exists to trace the human
factors to conditions resulting from decisions taken at higher organizational
levels (IMO, 2010b). A new investigation into accident reports can possibly
identify other organizational and structural factors related to shipping
accidents. If reviewed and analyzed according to, for example, the accident
model developed by James Reason (2001), latent organizational influences,
local workplace factors, preconditions for unsafe acts, and unsafe acts can be
identified. With new findings discussed in light of theories developed by
Snook (2000) and Hollnagel (2009), among others, alternative explanations
can be put forward, such as how frontline personnel make sense of
organizational safety communications and adapt their work practices through
social relations and psychological mechanisms, thereby moving safety
management and research into the new era of the adaptive age (Borys et al.,
2009; Hollnagel, 2009; Snook, 2000). The adaptive age embraces adaptive
cultures and resilience engineering and requires a change in perspective from
human variability as a liability and in need of control to human variability as
an asset and important for safety. Efforts could also be made to better identify
and measure the social processes among workers, along with further
exploration of the relationship among management, leadership, and safety-
related matters Thus, better insights into how behavioral norms interact with
and are formed by the social life on board could be achieved.
7.4 Final remarks
This thesis has explored the safety culture and safety management within
shipping in relation to current theories of safety management and safety
culture. The major limitations of the research along with implications for
110
Concluding remarks
safety practitioners and researchers—previously addressed in this thesis—can

be summarized as follows.
Survey: Parts of the applied questionnaire showed several deficiencies, and

results may be biased due to common method, psychometric properties,
language, and characteristic with the sample, which may affect the validity of
the conclusions. Future research should strive to develop an instrument in
order to reduce such biases.
Research model: The strengths and limitations of both qualitative and

quantitative research should be acknowledged, and future research should be
open to a multi-method approach.
Safety researcher: As the theories of safety management are developing over

time, safety researchers should strive to develop a better understanding of the
limitations of current safety management systems and be open to research
within the prevailing adaptive age.
Safety practitioners: In practical applications of safety management, one

should rely less on safety through standardized measures and experience data.
This includes understanding the difference between events where such
measures are applicable and unexpected events where it is adequate to support
competence-promoting activities so that the operators have the ability to adapt
their behavior to new situations. The human inferential capacity in handling
unexpected situations should not be underestimated in relation to technology.
111
References
8 References
Alderton, T. (2004). The Global seafarer: Living and working conditions in a

globalized industry. Geneva: International Labour Office.
Alvesson, M. (1993). Cultural perspectives on organizations. New York:

Cambridge University Press.
Anderson, P. (2003). Cracking the code: The relevance of the ISM code and
its impact on shipping practices. London: Nautical Institute.
Antonsen, S. & Norges teknisk-naturvitenskapelige universitet. Institutt for

sosiologi og statsvitenskap. (2009). Safety culture theory, method and
improvement. 2009, 47.
Barnett, M., Gatfield, D., & Habberley, J. (2010). Shipboard Crisis

Management: A Case Study [Online]. Retrieved from:
http://disruption.solent.ac.uk/mhfr/resources/RINA%202002.pdf
Borys, D., Dennis, E., & Legget, S. (2009). The fifth age of safety: The
adaptive age. Journal of Health & Safety Research & Practice, 1(1),
19-27.
Bryman, A. (1991). Street Corner Society as a Model for Research into

Organizational Culture. In P. J. Frost, L. F. Moore, M. R. Louis, C. C.
Lundberg, & J. Martin (Eds.), Reframing organizational culture (p.
205). Newbury Park, CA: Sage.
Byrne, B. M. (2010). Structural equation modeling with AMOS: Basic

concepts, applications, and programming. New York: Taylor &
Francis group.
Christophersen, J. G. (2009). Sikkerhetsstyring i skipsfarten 1998-2008:

Bakgrunnsfaktorer for reguleringsmessig etterlevelse og overtredelse
av ISM-koden. Det juridiske fakultet: Universitetet i Oslo. 2009, 22
Cooper, M. D. (2000). Towards a model of safety culture. Safety Science,

36(2), 111-136.
112
References
Corbett, A. (2009). Thousands of ships slip through inspection net.

Tradewinds. 9 January 2009.
CPEM. (1999). Report of the enquiry into the sinking of ERIKA of the coasts
of Brittany on 12 December 1999.
Danish Maritime Authority. (2009). The grounding of MIRABELLE on 16

December 2008. Denmark: Division for Investigation of Maritime
Accidents.
Dansereasu, F. J., & Alutto, J. A. (1990). Level-of-analysis Issues. In B.

Schneider (Ed.), Organizational climate and culture (pp. 193-236),.
San Francisco, CA: Jossey-Bass.
DeVillis, R. F. (2003). Scale Development, Theory and Applications.London:

Sage Publications.
Ek, Å. (2006). Safety culture at Sea and Aviation Transport (Doctoral Thesis,
Lund University).
Field, A. (2005). Discovering statistics using SPSS (and sex, drugs and rock
'n 'roll) (2nd ed.). London: Sage.
Flin, R., Mearns, K., O'Connor, P., & Bryden, R. (2000). Measuring safety
climate: Identifying the common features. Safety Science, 34(1-3),
177-192.
Frost, P. J. (1991). Reframing organizational culture. Newbury Park, CA:

Sage.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Glendon, A. I., & Stanton, N. A. (2000). Perspectives on Safety Culture.

London: Brunel University Research Archive.
Gliem, J. A., & Gliem, R. R. (2003). Calculating, Interpreting, and Reporting

Cronbach's Alpha Reliability Coefficient for Likert-Type Scales.
Midwest Research to Practice Conference in Adult, Continuing, and
Community Education. Columbus, OH.
113
References
Gregory, D., & Shanahan, P. (2010). The Human Element: A Guide to Human
Behaviour in the Shipping Industry, The Stationery Office (TSO).
Guldenmund, F. W. (2000). The nature of safety culture: A review of theory

and research. Safety Science, 34(1-3), 215-257.
Hair, J. F. (1998). Multivariate data analysis (5th ed.). Upper Saddle River,
NJ: Prentice Hall.
Hale, A. R. (2000). Culture's confusions. Safety Science, 34(1-3), 1-14.
Harzing, A. W. (2005). Does the use of English-language questionnaires in

crossnational research obscure national differences? International
Journal of Cross Cultural Management, 5(2), 213-224.
Håvold, J. I., & Nesset, E. (2009). From safety culture to safety orientation:
Validation and simplification of a safety orientation scale using a
sample of seafarers working for Norwegian ship owners. Safety
Science, 47(3), 305-326.
Håvold, J. I., & Norges teknisk-naturvitenskapelige universitet. Institutt for

industriell økonomi og teknologiledelse. (2007). From safety culture
to safety orientation: Developing a tool to measure safety in shipping.
Norges teknisk-naturvitenskapelige universitet. Institutt for industriell
økonomi og teknologiledelse. 2007, 180.
Heinrich, H. W., Roos, N., & Petersen, D. (1980). Industrial accident

prevention: A safety management approach (5th ed.). New York:
McGraw-Hill.
Hinkin, T. R. (1995). A review of scale development practices in the study of

organizations. Journal of Management, 21(5), 967-988.
Hofstede, G. (2001). Culture's consequences: Comparing values, behaviors,

institutions, and organizations across nations (2nd ed.). Thousand
Oaks, CA: Sage.
Hofstede, G., & Hofstede, G. J. (2005). Cultures and organizations: Software

of the mind (2nd rev.). New York: McGraw-Hill.
114
References
Hollnagel, E. (2009). The ETTO Principle: Efficiency-Thoroughness Trade-

Off Why Things That Go Right Sometimes Go Wrong. Aldershot:
Ashgate.
Hollnagel, E. (2004). Barriers and accident prevention. Aldershot: Ashgate.
Hopkins, A. (2006). Studying organizational cultures and their effects on

safety. Safety Science, 44, 875-889.
Hoyle, R. H. (1995). Structural equation modeling: Concepts, issues, and

applications. Thousand Oaks, CA: Sage.
International Atomic Energy Agency (IAEA). (1991). Safety Culture. Vienna:

IAEA.
IMO STW 41/4, 3. (2009, July). Unlawful Practices Associated with

Certificate Competency. Sub-Committee on Standards of Training
and Watchkeeping 41st Session.
International Convention for the Safety of Life at Sea (1974), & International
Maritime Organization. (2010). SOLAS : Amendments 2008 and
2009. (pp. 78). London: International Maritime Organization.
International Maritime Organization (IMO). (2010a). ISM code: International

safety management code and guidelines on implementation of the ISM
code (3rd ed.). London: International Maritime Organization.
International Maritime Organization (IMO). (2010b). Study on human and

organizational factors by WMU. London: Subcommittee on flag state
implementation.
115
References
International Maritime Organization (IMO). (2007a). Comprehensive Review

of the STCW Convention and the STCW Code—Communication and
leadership skills (STW/39/3). London: Sub-Committee on Standards
of Training and Watchkeeping 39st Session.
International Maritime Organization (IMO). (2007b). Comprehensive Review

of the STCW Convention and the STCW Code—Leadership and
managerial skills (STW 41/7/13). London: Sub-Committee on
Standards of Training and Watchkeeping 41st Session.
International Maritime Organization (IMO). (2007c). Role of the human

element: Near miss information. London: International Maritime
Organization.
International Maritime Organization (IMO). (2005). Role of the Human

Element—Assessment of the impact and effectiveness of
implementation of the ISM Code (MSC 81/17/1). London:
International Maritime Organization.
Justis-og politidepartementet. (1991). Scandinavian Star-ulykken, 7 april

1990. Oslo: Norges Offentlige Utredning.
Kjellen, U. (2000). Prevention of Accidents through Experience Feedback.

London: Taylor & Francis.
Kontogiannis, T., & Malakis, S. (2009). A proactive approach to human error

detection and identification in aviation and air traffic control. Safety
Science, 47(5), 693-706.
Lamvik, G. M. (2002). The Filipino seafarer: A life between sacrifice and

shopping. Trondheim: Dept. of Social Anthropology, Norwegian
University of Science and Technology.
Lappalainen, J. (2008). Transforming Maritime Safety Culture: Evaluation of

the impacts of the ISM Code on maritime safety culture in Finland.
Finland: Centre for Maritime Studies, University of Turku.
Mansell, J. N. K., & SpringerLink. (2009). Flag State Responsibility. Berlin,

Heidelberg: Springer-Verlag Berlin Heidelberg.
116
References
MARISEC. (2008). Shipping Industry Flag State Performance Table.

London: Maritime International Secretariat Services Ltd (Marisec)
MARISEC. (2006). Shipping Industry Guidelines on Flag State Performance.

London: Maritime International Secretariat Services Ltd (Marisec).
Maritime and Coastguard Agency. (2004). Driving Safety Culture:

Identification of Leadership Qualities for Effective Safety
Management. Aberdeen: Maritime and Coastguard Agency.
Martin, J. (1992). Cultures in organizations: Three perspectives. New York:

Oxford University Press.
McCrae, R. R. (2001). Cross-cultural research on the five-factor model of

personality. In W. J. Lonner, D. L. Dinnel, S. A. Hayes, & D. N.
Satler (Eds.), Online Readings in Psychology and Culture (Unit 6,
Chapter 1). Washington: Center for Cross-Cultural Research, Western
Washington. Retrieved from http://www.wwu.edu/~culture
Mearns, K. J., & Flin, R. (1999). Assessing the State of Organizational

Safety—Culture or Climate? Current Psychology, 18(1), 5.
Mitropoulos, P., & Cupido, G. (2009). The role of production and teamwork
practices in construction safety: A cognitive model and an empirical
case study. Journal of Safety Research, 40(4), 265-275.
Mitroussi, K. (2003). The evolution of the safety culture of IMO: A case of

organisational culture change. Disaster Prevention and Management,
12(1), 16-33.
National Transportation Safety Board. (1990). Grounding of U.S. Tankship

EXXON VALDEZ on Bligh Reef, Prince William Sound Near Valdez,
AK March 24, 1989.
Netemeyer, R.G., Sharma, S. & Bearden, W.O. (2003). Scaling procedures :

issues and applications, Sage Publications, Thousand Oaks, Calif.
Neuman, W. L. (2000). Social research methods: Qualitative and quantitative

approaches (4th ed.). Boston: Allyn and Bacon.
117
References
Olsen, E. (2009). Safety climate and safety culture in health care and the
petroleum industry: Psychometric quality, longitudinal change, and
structural models. Stavanger: University of Stavanger, Faculty of
Social Sciences.
Oltedal, H. (2010). The use of safety management systems within the

Norwegian tanker industry—do they really improve safety? In R.
Bris, C. Guedes Soares, & S. Martorell (Eds.), Reliability, Risk and
Safety: Theory and Applications (pp. 2355-2362). London: Taylor &
Francis Group.
Oltedal, H., & McArthur, D. (2010). Reporting practices in merchant

shipping, and the identification of influencing factors. Safety Science,
49(2), 331-338
Oltedal, H., & Wadsworth, E. (2010). Risk perception in the Norwegian

shipping industry and identification of influencing factors. Maritime
Policy & Management, 37(6), 601-623.
safety management systems within Norwegian dry cargo shipping. In
J.M. Ale, I.A. Papazoglou, & E. Zio (Eds.), Reliability, Risk and
Safety (pp. 2118-2125). London: Taylor & Francis Group
Østreng, D. (2007). I samme båt-Forholdet mellom sjøfolk på multinasjonale

skip. Oslo: Universitetet i Oslo, Institutt for sosiologi og
samfunnsgeografi.
Paris MOU. (2010). Port State Control—In the offing. Annual Report 2009.
ParisMou.
Paris MOU. (2007). On course for safer shipping–25 years. ParisMou.
Perrow, C. (1999). Normal accidents: Living with high-risk technologies.

Princeton, NJ: Princeton University Press.
Peterson, M. F., Ashkanasy, N. M., & Wilderom, C. P. M. (2000). Handbook

of organizational culture & climate. Thousand Oaks, CA: Sage.
118
References
Pett, M. A., Lackey, N. R., & Sullivan, J. J. (2003). Making sense of factor
analysis: The use of factor analysis for instrument development in
health care research. Thousand Oaks, CA: Sage.
Pettigrew, A. M. (1979). On Studying Organizational Cultures. Administrative

Science Quarterly, 24(4), 570-581.
Pidgeon, N. (1998). Safety Culture: Key theoretical issues. Work & Stress,
12(3), 202-216.
Pidgeon, N., & O'Leary, M. (2000). Man-made disasters: Why technology

and organizations (sometimes) fail. Safety Science, 34(1-3), 15-30.
Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003).

Common Method Biases in Behavioral Research: A Critical Review
of the Literature and Recommended Remedies. Journal of Applied
Psychology, 88(5), 879.
Raubenheimer, J. E. (2004). An item selection procedure to maximize scale

reliability and validity. South African Journal of Industrial
Psychology, 30(4), 59-64.
Reason, J. (2001). Managing the risks of organizational accidents. Aldershot:

Ashgate.
Richter, A., & Koch, C. (2004). Integration, differentiation and ambiguity in

safety cultures. Safety Science, 42(8), 703-722.
Riksrevisjonen. (2010). Riksrevisjonens undersøkelse av Sjøfartsdirektoratets

saksbehandling ved dokumentkontroll av fartøy og sjøfolk. Oslo:
Nærings-og handelsdepartementet.
Roberts, K. H. (1990). Some Characteristics of one Type of High Reliability

Organization. Organization Science, 1(2), 160-176.
Robson, L. S., Clarke, J. A., Cullen, K., Bielecky, A., Severin, C., Bigelow, P.
L., Irvin, E., Culyer, A., & Mahood, Q. (2007). The effectiveness of
occupational health and safety management system interventions: A
systematic review. Safety Science, 45(3), 329-353.
119
References
Rousseau, D. M. (1990). Assesing organizational culture: The case for

multiple methods. In B. Schneider (Ed.), Organizational climate and
culture (pp. 153). San Francisco, Calif.: Jossey-Bass.
Rothblum, A. R. (2000). Human Error and Marine Safety. Paper presented at

the National Safety Council Congress and Expo, Orlando, October
13-20.
Rundmo, T. (1996). Associations between risk perception and safety. Safety

Science, 24(3), 197-209.
Sagan, S. D. (1993). The limits of safety: Organizations, accidents, and

nuclear weapons. Princeton, NJ: Princeton University Press.
Sarantakos, S. (1998). Social research (2nd ed.). Houndmills: Macmillan.
Schein, E. H. (1992). Organizational culture and leadership (2nd ed.). San

Francisco: Jossey-Bass.
Schein, E. H. (2004)., Organizational culture and leadership (3rd ed.). San

Francisco: Jossey-Bass.
Schager, B. (2008). Human error in the maritime industry: How to

understand, detect and cope. Sweden: Marine Profile Sweden AB.
Shea, I. P. (2005). The organizational culture of a ship: A description and

some possible effects it has on accidents and lessons for seafaring
leadership (PhD dissertation, University of Tasmania).
Shevlin, M., Miles, J. N. V., & Bunting, B. P. (1997). Summated rating

scales. A Monte Carlo investigation of the effects of reliability and
collinearity in regression models. Personality and Individual
Differences, 23(4), 665-676.
Shevlin, M., Miles, J. N. V., Davies, M. N. O., & Walker, S. (2000).

Coefficient alpha: A useful indicator of reliability? Personality and
Individual Differences, 28(2), 229-237.
SINTEF. (2003). Sikkerhetskulturer i transport: En kunnskapsoversikt.
120
References
Snook, S. A. (2000). Friendly fire: The accidental shootdown of U.S. Black

Hawks over Northern Iraq. Princeton, NJ: Princeton University Press.
Soma, T. (2010). It takes management to reduce accidents [Online]. Retrieved

from
http://www.dnv.com/industry/maritime/publicationsanddownloads/pu
blications/dnvtankerupdate/2008/no12008/ittakesmanagementcommit
menttoreduceaccidents.asp
Soma, T. (2005). Blue-chip or sub-standard - a data interrogation approach

to identify safety characteristics of shipping organisations (DrIng
thesis, NTNU Trondheim).
Sorensen, J. N. (2002). Safety culture: A survey of the state-of-the-art.

Reliability Engineering & System Safety, 76(2), 189-204.
Spector, P. E. (2006). Method Variance in Organizational Research: Truth or

Urban Legend? Organizational Research Methods, 9(2), 221.
Spector, P. E. (1992). Summated rating scale construction: An introduction.

Newbury Park, CA: Sage Publications.
Stopford, M. (2009). Maritime economics (3rd ed.). London: Routledge.
Strauss, A. L., & Corbin, J. M. (1990). Basics of qualitative research:

Grounded theory procedures and techniques. Newbury Park, CA:
Sage.
Studio Apertura. (2004). Sikkerhetskulturer i transportsektoren. Metoder for

kartlegging av sikkerhetskultur: Evaluering av noen eksisterende
verktøy. Trondheim: Studio Apertura.
Tharaldsen, J. (2011). "In safety we trust": Safety, risk and trust in the
offshore petroleum industry. (DrPhil thesis, UiS, Stavanger)
The United Nations Scientific Committee on the Effects of Atomic Radiation.

(1988). UNSCEAR 1988 Report Annex D Exposures from the
Chernobyl accident.Vienna: The United Nations Scientific
Committee on the Effects of Atomic Radiation.
121
References
Tradewinds. (2007). "Evangelia" slips through the cracks.Tradewinds.
Turner, B. A. (1978). Man-made disasters. no. 53, New York : Crane, Russak,
Wykeham science series
United States Coast Guard. (2005). Investigation into the explosion and
sinking of the chemical tanker Bow Mariner in the Atlantic Ocean on
February 28, 2004, with loss of life and pollution. Washington, DC:
United States Coast Guard.
Wagenaar, W. A., & Groeneweg, J. (1987). Accidents at sea: Multiple causes

and impossible consequences. International Journal of Man-Machine
Studies, 27(5-6), 587-598.
Weick, K. E., & Sutcliffe, K. M. (2007). Managing the unexpected (2nd ed.).
San Francisco, CA: Jossey-Bass.
Whyte, W. F. (1991). Street Corner Society: Excerpts from the Appendix to

the 1955 Edition. In P. J. Frost, L. F. Moore, M. R. Louis, C. C.
Lundberg, & J. Martin (Eds.), Reframing organizational culture (p.
173). Newbury Park, CA: Sage.
Wiegmann, A. D., Zang, H., von Thaden, T., Sharma, G., & Mitchell, A.
(2002a). Safety Culture: A Review. Illinois: Aviation Research Lab
Institute of Aviation, University of Illinois.
(2002b). A Synthesis of Safety Culture and Safety Climate Research.
Illinois: Aviation Research Lab Institute of Aviation, University of
Illinois.
(2002c). Safety Culture: A concept in chaos? To appear in the
Proceedings of the 46th Annual Meeting of the Human Factors and
Ergonomics Society. Santa Monica, Human Factors and Ergonomics
Society.
Winchester, N., Alderton, T., & Seafarers International Research Centre.

(2003). Flag state audit 2003: Introduction & appendices. Cardiff:
Seafarers International Research Centre.
122
References
Zhang, H., Wiegmann, D. A., von Thaden, T. L., Sharma, G., & Mitchell, A.
A. (2002). Safety culture: a concept in chaos? Conference
Proceedings.
Zohar, D. (1980). Safety Climate in Industrial-Organizations—Theoretical

and Applied Implications. Journal of Applied Psychology, 65(1), 96-
102.
Zohar, D. (2008). Safety climate and beyond: A multi-level multi-climate

framework. Safety Science, 46(3), 376-387.
Zohar, D. (2010). Thirty years of safety climate research: Reflections and

future directions. Accident Analysis & Prevention, 42(5), 1517-1522.
123
Part II
Article 1
Francis Group.
Article 2
Management, 37(6), 601-623.
Article 3

Article 4
124
Article 5
Oltedal, H. & McArthur, D. (2010). Reporting practices in merchant shipping,

and the identification of influencing factors. Safety Science, 49(2), 331-338.
Article 6
Oltedal, H. A., & Engen, O. A. (2010). Safety Management in Shipping—

Making Sense of limited Success. Safety Science Monitor, submitted.
125
Article 1
Oltedal, H. A., & Engen, O. A. (2009). Local management and its
impact on safety culture and safety within Norwegian shipping. In S.
Martorell, C. Guedes Soares & J. Barnett (Eds.), Safety, Reliability and
Risk Analysis: Theory, Methods and Applications (pp. 1423-1430).
London: Taylor & Francis Group.

Safety, Reliability and Risk Analysis: Theory, Methods and Applications – Martorell et al. (eds)
© 2009 Taylor & Francis Group, London, ISBN 978-0-415-48513-5
Local management and its impact on safety culture and safety

within Norwegian shipping
H.A. Oltedal
University College Stord/Haugesund, Norway
O.A. Engen
University of Stavanger, Norway
ABSTRACT: This paper addresses safety culture on tankers and bulk carriers and which factors affect the
safety culture onboard vessels. The empirical setting for the study is the Norwegian shipping industry. Safety
management is a challenging issue within shipping for several reasons. First of all, life and work onboard a vessel
is a 24 hour activity and the crew has few possibilities of interacting with the surrounding society. Secondly
the geographical distance between the on-shore organization and the vessel may affect both the quality of those
systems and plans developed on shore and their implementation on the vessels. The ship management is thus
identified as a key factor to a sound safety culture along with the on shore crewing strategy.
1 INTRODUCTION along with the need for reorganization to improve effi-

ciency. This resulted in a cut in the crewing level. Later
In this paper we will discuss the safety culture within in the 80s a global recession caused further structural
the Norwegian shipping industry with tankers and bulk changes; flagging-out, use of external crewing agen-
carriers, and identify which organizational factors may cies and signing on crew from developing countries
affect this particular safety culture. and lower wages (Bakka, Sjøfartsdirektoratet 2004).
In Norway, shipping has for several centuries However, the shipping industry is today facing new
been the principal trade, and Norway as a maritime manning related challenges as there is a global short-
nation has roots way back in the Viking age. Today age of manpower, this is due to three main challenges:
Norway is one of the five largest shipping nations in First, it is less attractive nowadays to work in the ship-
the world, after Greece, Japan, Germany and China. In ping industry. Second, the recruitment for ship crews
the third quarter of 2007 the Norwegian foreign-going has been slow. This has resulted in the third situation
fleet comprised 1,795 ships, the highest number ever where the liquefied natural gas (LNG) shipping sec-
in Norwegian history, of which about 49 percent are tor is drawing crew from the tanker industry, and the
flying under the Norwegian flag (Nærings- og handels- tanker industry in turn is drawing people from the dry
departementet 2007). The remaining 51 percent may bulk sector.
register in any of the world’s more than 150 flag states. In 1894 the British Board of trade carried out a
Norwegian shipping companies employ some 57,000 study which showed that seafaring was one of the
seamen from more than 60 different nationalities and world’s most dangerous occupations, and it still is
of which about 30 percent are Norwegian Nationals (Li, Shiping 2002). Regulations in order to reduce
(Norwegian Shipowners’ Association). The crew may the risk at sea were introduced about 150 years ago.
be recruited and managed by the shipping company These regulations initially encompassed measures to
itself, or by one of the world’s many professional crew rescue shipwrecked sailors, and further requirements
hiring companies. Within the Norwegian fleet, most for life-saving equipment, seaworthiness and human
sailors are contract-employees working on different working conditions. Traditionally the safety work has
vessels during each enrolment, which results in con- focused on technical regulations and solutions even
tinually shifting working groups. The situation today though experience and accident statistics indicate that
is a result of a structural change dating back to the most of the accidents at sea somehow were related
60s and 70s when technical development allowed for to human performance (Bakka, Sjøfartsdirektoratet
bigger vessels with more automation and monitoring, 2004). However, a few very serious accidents at seathat
1423
occurred in the late 80’s resulted in a change towards Internal
how safety was organised, and more focus was given Person
Safety Psychological
to the human barriers and how the seafarers’ work- Factors
ing conditions were affected by organisational and Climate
managerial factors—both on shore and at sea. Along
with this the term safety culture started to gain a CONTEXT
foothold also within shipping. The idea of safety cul- External Observable Factors
ture within shipping was officially introduced on the
4th November 1993 by the adoption of a new reso- Situation Behavior
lution, the present SOLAS Convention 1974 Chapter Organisational Safety
IX, entitled ‘‘Management for the Safe Operation of l Factors Behavior
Ships and for Pollution Prevention’’, also known as the
International Safety Management Code (ISM Code)
(Le Meur 2003). Figure 1. Reciprocal safety culture model (adopted from
Hence, the main purpose of this paper is to elaborate Cooper, 2000).
the following questions:
– What characterises safety culture on tankers and adoption and internal integration, that has worked
bulk carriers? well enough to be considered valid and, there-
– Which factors affect the safety culture on board fore, to be taught to new members as the correct
vessels? way to perceive, think and feel in relation to those
With reference to shipping, this article will more problems’’(Schein 2004).
concretely analyse crewing strategies such as out- Further we have decided to use a methodologi-
sourcing of crewing management and the extended use cal framework presented by Cooper (2000), and the
of contract employment instead of permanent employ- application of this framework will be discussed below.
ment. Our hypothesis is that these conditions may con- Cooper (2000) introduces a reciprocal model of safety
tribute to an unfavourable and error-inducing working culture that allows the multi-faceted and holistic nature
environment, i.e. poor communication between shore of the concept to be fully examined by using a trian-
management and the ship management and the remain- gular methodology approach, depicted in Figure 1.
ing crew, unworkable procedures, lack of loyalty to Cooper’s (2000) model contains three elements:
the organisation, dysfunctional interaction, fear of
reprisals, which again counteract the development of 1. The subjective internal psychological factors i.e.
a safety culture. attitude and perceptions.
2. Observable on-going safety related behaviour.
3. The organisational situational features.
2 APPROACH TOWARDS SAFETY CULTURE
According to Cooper (2000) these elements reflect
There seems to be no clear consensus concerning those accident causation relationships found by a
the ontological, epistemological, and methodological number of researchers, such as John Adams, Herbert
questions related to the topic of safety culture. The William Heinrich and James Reason. The investi-
main differences seem to be gation of several serious shipping accidents such
as Herold Free Enterprize (Department of Transport
1. Definition of the scope of safety culture and the 1987), Exxon Valdes (National Transportation Safety
relationship between culture and climate. Board 1990) and Scandinavian Star (Justis- og poli-
2. Which methods are regarded as most suitable for tidepartementet 1991) is also congruent with their
measurement. findings. The Herold Free Enterprize accident was
3. The relationship to other organisational (safety- partly caused by members of the crew not following
related) aspects (Cooper 2000, Guldenmund 2000, best practice, but was also due to managerial pres-
Neal, Griffin & Hart 2000, Peterson, Ashkanasy & sure from the organization’s upper level to sail as early
Wilderom 2000, Sorensen 2002, Yule 2003). as possible, along with other mistakes made by the
However, it is not the scope of this paper to prob- on-shore management. Two years later when the US
lematise the concept of safety culture. As a point tanker ‘‘Exxon Valdes’’ grounded, the accident inves-
of departure we will apply Schein’s definition of tigation determined several probable causes linked to
organisational culture: human errors induced by managerial faults in the upper
levels of the organisation. At the vessel, the third mate
‘‘A pattern of shared basic assumptions that the failed to properly manoeuvre the vessel, possibly due
group learned as it solved its problems of external to fatigue and excessive workload. The master failed
1424
to provide a proper navigation watch, possibly due to of their actions (Dekker & Dekker 2006). An alterna-
impairment from alcohol. At the onshore part of the tive view is to recognise human error not as a cause
organisation, the shipping company fails to supervise of accidents, but as a consequence or symptom of
the master and provide a rested and sufficient crew organisational trouble deeper within the organisation,
for the ‘‘Exxon Valdez’’. In addition to this effective arising from strategic or other top level decisions.
pilot and escort services were lacking. The follow- This includes resource allocation, crewing strategy
ing year, in 1990, there was a fire on the passenger and contracting (Dekker, Dekker 2006, Reason 2001,
liner ‘‘Scandinavian Star’’. In the aftermath of this Reason, Hobbs 2003). An organisation is a complex
accident the investigation brought into focus organ- system balancing different, and often also conflicting,
isational and managerial faults with regard to a lack goals towards safety and production in an aggressive
of competence and training, but also weaknesses in and competitive environment (Rasmussen 1997), a
the wider social-technical system. These weaknesses situation that to a large extent is current within ship-
consisted of ownership separated from management, ping. The BBS approach towards safety often implies
unsatisfactory control routines by the flag state and, that more automation and tighter procedures should be
in general, an overall maritime system with a lack of added in order to control the human actions. However,
transparency. Further, Cooper’s (2000) three elements the result may be that more complexity is added to the
will be outlined more in detail, starting with safety system. This in combination with the organisation’s
related behaviour. struggle to survive in the competitive environment,
leads to the system becoming even more prone to acci-
dents (Perrow 1999) (Dekker & Dekker 2006, Reason
2.1 The importance of safety related behaviour
2001, Reason & Hobbs 2003). However, the concept
Herbert William Heinrich work (published in 1931 of focusing on the human side of safety is not wrong.
Industrial Accident Prevention) is the basis for the the- After all, the technology and production systems are
ory of Behaviour-Based Safety (BBS), which holds operated, maintained and managed by humans, and
that as many as 80–90 percent of all workplace acci- as the final barrier towards accidents and incidents
dents are caused by human error and unsafe acts they are most of the time directly involved. The pro-
(Tinmannsvik, 2008). Schein’s (2004) definition of ponents of the BBS approach argue that behaviour
culture does not clearly address observable behaviour control and modification may bring a shift in an organ-
patterns, but behaviour is regarded to be partly isation’s safety culture, also at the upper level, but this
determined by a person’s perceptions, feelings and is most likely if the focus is not exclusively addressing
thoughts. However, Schein (2004) regards behavioural observed deficiencies at the organisation’s lower lev-
patterns as a manifestation of a culture existing at a els (DeJoy 2005). DeJoy (2005) calls attention to three
higher level in the organisation, and not as culture apparent weaknesses related to the BBS approach:
itself. When it comes to BBS, the current theories
1. By focusing on human error it can lead to victim-
posit that safety culture, and a reduction of accidents
blaming.
may be achieved through continuous attention to three
2. It minimises the effect of the organisational
domains:
environment in which a person acts.
1. Environmental factors such as equipment, tools, 3. Focusing on immediate causes hinders unveil-
physical layout procedures, standards, and temper- ing the basic causes, which often reside in the
ature. organisational environment.
2. Person factors such as people’s attitudes, beliefs,
Due to this, we will also include the organisational
and personalities.
environment in the safety culture concept, as proposed
3. Behaviour factors, such as safe and at-risk work
by Cooper (2000).
practices, referred to as the Safety Triad (Geller
2001).
2.2 The relation to organisational factors
When adopting this approach humans are seen
as a cause of accidents, whereupon interventions When human error is not seen only as a cause of acci-
to enhance safety are aimed at changing attitude dents, but as a symptom and consequence of problems
or behavior (i.e. poster campaigns, training, proce- deeper inside the organisation, or what Reason (2001,
dures and so on, or changing the technology they 2003) refers to as latent organisational factors, empha-
operate). This orientation towards risk and safety man- sis is placed on weaknesses in strategic decisions made
agement has traditionally been and still is adopted from at the top level in the organisation. These strategic
the majority of the shipping companies. The BBS- decisions may reflect an underlying assumption about
approach has been criticised for placing too much the best way to adapt to external factors and to achieve
responsibility on the people operating the systems, internal integration, and if they are common for most
assuming that they are responsible for the outcome shipping companies an organisational culture may also
1425
be revealed (Schein 2004). Schein (2004) also stresses Table 1. Questionnaire constructs and number of items.
the importance of leadership. The top management
influences the culture as only they have the possibil- Construct Number of items
ity of creating groups and organisations through their
Top management’s safety priorities 3
strategic decisions. And when a group is formed, they
Local management 7
set the criteria for leadership and how the organisation Procedures & guidelines 7
will support and follow up their leaders. The leaders, at Interaction 18
all levels on shore and at the vessel, are also key figures Work situation 8
in the development of a safety culture. It is their task Competence 5
to detect the functional and dysfunctional elements of Responsibility & sanctions 7
the existing culture, and to channel this to the upper Working environment 9
levels of the organisation. In return, the upper levels of Reporting practices 10
the organisation should give their leaders the support
necessary in order to develop the desired culture.
response rate of 80% and a vessel response rate of
91.5%. The survey was carried out in 2006.
3 METHODOLOGICAL IMPLICATIONS
3.3 Statistical analysis
Cooper’s (2000) framework put forward the impor-
tance of methodological triangulation in order to grasp The Statistical Package for the Social Sciences (SPSS)
all facets of the cultural concept. The internal psy- v.15.0 was used to perform all of the analysis, which
chological factors are most often assessed via safety included descriptive statistics, exploratory factor anal-
climate questionnaires. Our approach is to start with ysis (EFA), confirmatory factor analysis (CFA) and
such a survey in order to gain insight into the seafarer’s bivariate correlation analysis.
perceptions and attitudes related to safety, along with With regard to the EFA, the principal component
self-reported work behaviour related to risk taking, analysis with Varimax rotation was carried out. The
rule violation and accident reporting. The survey also factors were extracted based on the three following
includes questions related to crewing strategy, which analytical criteria: (1) Pairwise deletion, (2) Eigen
opens up the possibility of assessing the relation- value more than 1, and (3) factor loading more than
ship between the organisational situation and actual 0.50. Of the extracted factors, all factors with 2 or
behaviour. The survey results are used to determine fewer items were removed, based on the notion that a
which organisational factors are most likely affect the factor should be comprised of at least three items to be
safety culture, and to define research areas for a further robust (Pett, Lackey & Sullivan 2003). A confirmatory
qualitative study. factor analysis (CFA), using a one-factor solution for
each construct, has also been performed. The advan-
tage of the CFA is that it allows for more precision in
3.1 Development of questionnaire items
evaluating the measurement model (Hinkin 1995), and
The survey instrument was developed by Studio the results were compared with the EFA for providing
Apertura in collaboration with DNV and SINTEF. The validity evidence based on the hypothesis that a valid
development was based on an evaluation of seven instrument should produce similar results.
already existing questionnaires in comparison with Each factor was then evaluated using the Kaiser-
various theoretical views of the safety culture con- Meyerto-Olkin (KMO) parameter, and only factors
cept (Studio Apertura 2004). Minor adjustments were with KMO value at 0.60 or above were included in
made after a pilot for use within the tanker and bulk the further analysis (Hair 1998).
carrier sector. This resulted in a questionnaire with This was followed by a scale-reliability test. For that
constructs and accompanying number of items as pre- purpose, the Cronbach’s Alpha coefficient of internal
sented in table 1. All items were measured on a 5 point consistency was calculated, and evaluated along with
likert scale ranging from strongly disagree to strongly inter-item statistics. Cronbach’s Alpha is a measure
agree, or very seldom/never to very often/always. of scale reliability concerned with the proportion of a
scale’s total variance that is attributable to a common
source, presumed to be the true score of the latent
3.2 Questionnaire sample
construct being measured. In our case that will be
A total of 1574 questionnaires were distributed to 83 the safety culture. Usually a value above 0.7 is con-
randomly selected Norwegian controlled tankers and sidered acceptable, although some advocate an alpha
bulk carriers. All vessels were flying a flag on the Paris level of 0.8 or better (Netemeyer, Sharma & Bearden
MOU white or grey list. 76 vessels returned a total 2003). As the alpha value is a function of, inter
of 1262 completed forms, which gives an individual alia, the average inter-item correlation; the inter-item
1426
correlation and item-total statistics have also been reporting practices, (3) competence, (4) local manage-
evaluated. Rules of thumb suggest that the item-total ment, and (5) work situation. With regard to the ‘‘local
correlation should exceed .50, and the inter-item cor- management’’, ‘‘competence’’ and ‘‘work situation’’
relation should exceed .30 (Hair 1998), but it should factor both EFA and CFA result in final solutions con-
not exceed .80. An inter-item correlation exceeding sisting of the same items, but with minor differences
.80 suggests that items are duplicates of one another in factor loading. The CFA included three more items
(Pett, Lackey & Sullivan 2003). Then the remaining in the ‘‘interaction’’ factor than the EFA, and the final
items went through a last CFA, a five-factor solution, factor, ‘‘reporting practices’’ resulted from only the
in order to provide each factor’s explained variance. CFA.
Finally, correlation analysis has been carried out Four of the constructs did not pass the reliability
in order to evaluate the construct validity, which is tests. The first, ‘‘top management’s safety priorities’’,
viewed as the extent to which an operational measure was excluded due to low representative reliability
truly reflects the underlying safety culture concept, across subpopulations. This construct also consisted of
and if they operate in a consistent manner. too few items. The remaining three constructs, ‘‘proce-
Based on this analytical process, five factors (1) dures and guidelines’’, ‘‘responsibility and sanctions’’
interaction, (2) reporting practices, (3) competence, and ‘‘working environment’’ were excluded due to
(4) local management, and (5) work situation were low validity, mostly resulting from poor theoretical
found to be reliable and valid. The aforementioned relationship within the items of each construct.
factors are presented further in detail in the next For the further analysis the results from the CFA are
section. used. The 5 factors in question are presented in Table 2
along with number of items and explained variance.
Each factors Cronbach’s alpha value and inter item
4 RESULTS statistics is presented in table 3.
The alpha values range from .808 to .878, and the
4.1 Results from descriptive analysis internal item statistics are all within the recommended
Regarding demographics, 21 different nationalities are levels. The five factors are therefore considered to be
represented. The Filipino contingent forms the largest a reliable and valid reflection of the underlying safety
group constituting 63% of the sample, followed by culture concept.
the Norwegian group with almost 11%, and the Polish Further, table 4 presents the correlation coefficients
which represents 9%. The last major group was the between the factors, or safety culture dimensions. All
Russians with 6%. The other remaining 17 nationali- correlations are significant at the 0.01 level (2-tailed)
ties were represented in a range from 3% to 1%.
There is also great variation with regard to employ-
ment conditions. All in all, 12% of the sample Table 2. Final factors, number of items and explained
consists of permanent employees, of whom 80% are variance.
Norwegian and 16% from the European Union. 91% of
the Norwegians are permanent employee. The remain- Number Explained
ing 9% are apprentices, substitutes or newly employed Factor of items variance
on probation. Only 3% of the non Norwegian sailors
Interaction 8 35.63%
are permanent employees. With regard to the Filipino
Reporting practices 5 9.77%
seafarers, the largest nationality, 99.6% are contract Competence 4 7.12%
employees, most on 9 month contracts (62%), fol- Local management 3 5.96%
lowed by 6 month contracts (27%). The extended use Work situation 3 5.08%
of contract employment is reflected in their experience.
All in all, 85% had three years or more experience
within shipping in general. However, 69% of the sam-
ple had worked on the current vessel for only 1 year Table 3. Final factors, Cronbach’s alpha and inter-item
or less. statistics.
The employment terms were in general different for
the captains. The captains normally do not have sailing Inter-item Item-total
periods that exceed 6 months. The most typical sailing Factor Alpha range range
period for the captains is 3 months or less.
Interaction .878 .360–.606 .520–.724
Reporting practices .808 .335–.761 .491–.668
4.2 Results from factor analyses Competence .839 .497–682 .628–.712
Local Management .866 .692–.716 .724–774
From the 9 theoretical safety culture constructs, a Work situation .817 .512–.749 .554–.739
five factor solution was derived, (1) interaction, (2)
1427
Table 4. Factor correlation matrix. Pearson’s r. dates for signing on and signing off. Schein (2004)
points out that lack of stability may be a threat to the
F1 F2 F3 F4 F5 possibility of developing a culture: ‘‘( . . . ) there must
be a history of shared experience, which in turn implies
F1: Interaction 1
some stability of membership in the group.’’ Even if
F2: Reporting
practices .352 1 the crew as a group is in constant change, they all have
F3: Competence .639 .323 1 common history as seafarers. So even if lack of stabil-
F4: Local ity within the group indicates that a common culture
management .474 .362 .367 1 should not develop on the ship, a common culture of
F5: Work how to act and interact may have developed amongst
situation .494 .322 .441 .444 1 the seafarers, and when a new crewmember is signed
on a new vessel, he knows what is expected from him.
However, the question is if such a culture is a safe
culture? Reason (2001, 2003) emphasize that to reach
The five safety culture dimensions correlate in a safe culture, the organisation should strive for an
a positive direction, which is consistent with the informed culture where those who manage and operate
theoretical concept, and they are therefore considered the system, both on board and on shore, have current
to be a valid reflection of the underlying safety culture knowledge about the factors that determine the safety
construct. of the system as a whole, which again depends on that
the crew on board are prepared to report their errors
and near misses, and the reporting practice is one of
5 DISCUSSION the dimensions deriving from the analyses, explaining
9.77% of the variance. This dimension also includes
All three constructs have a good alpha level, and as feedback on reported events. In order to attain good
the alpha levels are concerned with the variance that reporting practices, the organisation should strive to
is common among the items, these constructs also create an atmosphere of openness, trust and loyalty.
reflect the areas where it is possible to speak about Integrating into the group is also a survival mecha-
safety culture. With reference to Cooper’s framework nism, and every crewmember will most likely make an
towards safety culture, we will further discuss how effort to integrate. If not, he would most likely have a
the organisation’s factors such as crewing strategy, hard time during his contract period with no possibility
witch includes employment terms, rotations system to leave the vessel and the other crewmembers. How-
and policy towards the on board shipping management, ever, to compromise oneself and be open about one’s
may affect the on board safety culture and climate own mistakes is not always an easy task, especially not
represented by the identified dimensions. The organ- in an unknown working environment. Something that
isation’s structural factors are all to be found within may reinforce the crewmembers’ fear of reporting their
Cooper’s element of situation, while the identified own mistakes is the ongoing practice that each crew
safety culture dimensions are to be found within the member is evaluated by their senior officer/captain,
elements of person and behaviour. and based on this report get recommended or not
Interaction is the dimension accounting for the recommended for re-hire. Interviews have revealed
largest proportion of the total explained variance, with that this evaluation practice differs. Some practise
35.63%, meaning that with regard to safety culture an open evaluation where all parts are involved, with
most of the variance in the original data is explained focus on how to improve the evaluated crew’s short-
by this dimension. When taking into account how comings, and where the shore organisation seeks to
distinctive a ship is as a work place, this is no sur- ensure that the evaluation is conducted in as objec-
prise. A ship may be characterised as a total institution tive a way as possible. At other vessels, the evaluation
since both work and leisure time happen at the same is closed for insight by the evaluated and may also
place and with few possibilities to interact with the be highly subjective. Some of the respondents have
surrounding world (Goffman 1968). In such a set- expressed that by reporting, their next contract may be
ting the crew members are socialised into a common at stake, or they may meet with other negative conse-
culture and rules of interaction. Schein (2004) refers quences. So, lack of stability and constantly changing
to this as internal integration. The interaction climate working groups may sacrifice a trusting and open envi-
is characterised by lack of stability within the crew ronment, and thus also the sailors’ commitment to
due to different terms of employment. First of all, safety.
permanent employment seems to be reserved for the A crew committed to safety is essential, but not
Norwegian sailors. Sailors of other nationalities are enough. Lack of competence may cause a situation
almost all contract employees. In addition, the length where the crew do not identify potential dangerous sit-
of contract varies and all crew members have different uations, or create them. Competence, which accounts
1428
for 7.12% of the total variance, is in this setting com- who are always very insulting towards jr. officers and
prised of activities performed on board the vessel, and rating.’’ Schein (2004) regards the leader as an impor-
is all under the control of the captain, training, drills tant key figure in the cultural development. At sea
and familiarisation when signing on. Also, the com- the captain holds a key role. The captain is the one
petence dimension does correlate strongly with the in command at every vessel, and according to Schein
interaction dimension with a correlation coefficient (2004) the captain’s orientation will affect the work-
at .639. This indicates that a situation when the sailors ing climate, which precedes the existence of a culture.
are feeling confident with the nature of their task also So, in a situation where lack of crew stability impedes
results in a better interaction climate where conflicts the development of a safety culture, the role of the
are more likely to be absent. As with the interaction captain is even more vital. Also, it is important to
climate, competence will also be affected by the crew take into account that the leadership style that is prac-
stability. A crew member that is constantly signing on tised on board not only affect the sphere of work, but
new vessels and that has to interact with new crew also time off. However the Captains themselves may
members and leaders, uses more effort adapting to not be aware of their own importance, or how they
the new situation, working out how things are done affect safety. Most Captains, or other department lead-
at that specific vessel, the informal structure onboard ers for that matter, do not have managerial training
and so on. When more stability is provided, more effort or education. When adopting a cultural view towards
may be placed on upgrading their competence, and the safety, as in this research, in as opposed to a behaviour
competence will be kept within the vessel and organi- based view, more emphasis is placed on organisational
sation. Both the training activities and crewing strategy factors. Decisions regarding crewing strategy, employ-
may be controlled by the ship management, and thus ment terms and managerial development programmes
these safety culture dimensions are also, to a certain are all strategic decisions made on shore. With ref-
degree, controllable. erence to Schein’s culture definition, we will argue
The dimension of work situation consists of pro- that the safety culture originates within the organ-
active activities as Safe Job Analysis (SJA), safety isation on shore. Based on Scheins’ definitions of
evaluations and the possibility they have to prioritize culture there ought to exist a pattern of shared basic
safety in their daily work. So how may the organisa- assumptions that may solve the problems the shipping
tion affect this? For one, they may supply sufficient industry is facing. Our case however has revealed an
crew. Today many vessels are sailing with a smaller offshore practise characterised by extended use of con-
crew at the same time as new standards and resolutions tract employment, lack of stable working conditions
like the ISM-code increase the amount of paperwork on board the vessels, and little or no use of managerial
to be done. Both own observations and interviews training and development. This practice does not pro-
reveal that inter alia check lists and SJA are done in a mote a good safety culture and is considered to has a
mechanical manner. This may originate from various negative effect on the overall safety level.
reasons such as an overload of work, no understanding
of the importance of those activities, lack of feedback
or improper planning by the local management. 6 CONCLUSIONS
The local management dimension, accounts for
5.96% of the explained variance, and the direct effect The aim of this paper was to analyse the character-
of local management is relatively small. However, istics of the safety culture on Norwegian tankers and
local management is considered to have an indirect bulk carriers, and identify what organisational factors
effect on the safety climate through the managers, may affect the safety culture on board vessels. Statis-
or senior officers, affect on the interaction climate, tical analysis identified five safety related dimensions
competence and training activities, reporting practices on board the vessels: interaction climate, reporting
and the work situation. Again we wish to focus on practices, competence, local management and work
the importance of stability within the work group. situation. Within shipping the interaction climate is
Most captains have a sailing period of 3 months or characterised by unstable working conditions. Under
less, while most of the non Norwegian ratings have such conditions it is difficult to achieve and maintain
a sailing period of 9 months. Most senior officers a stable crew, and proper management becomes even
also have a shorter sailing period then an ordinary more important. Also the Captain has a vital role, as
rating. Then a rating possibly has to deal with sev- he has the possibility to directly affect all the other
eral different leaders during his stay. And each captain safety related aspects through his own leadership style.
and department manager’s leadership style may vary, The Captains, officers and ratings normally have dif-
and are sometimes even destructive, as shown by fol- ferent employment terms and shift terms. This may
lowing comment from a Pilipino engineer. ‘‘The only jeopardise the development of a sound safety culture
problem on board is the treatment of senior officers to as the crew has a poor possibility of developing com-
the lowest rank. (. . . ) There are some senior officers mon behaviour practices and a mutual understanding
1429
of how to do things right. As neither the Captains Le Meur, C. 2003, Maritime Safety Culture.
nor the officers normally have any managerial train- Li, K.X. & Shiping, Z. 2002, ‘‘Maritime professional safety:
ing, their leadership styles often affect the safety in prevention and legislation on personal injuries on board
a negative direction. The on board situation is to a ships’’.
large extend considered to be created by the on-shore Nærings- og handelsdepartementet 2007, Stø kurs: reg-
jeringens strategi for miljøvennlig vekst i de maritime
crewing strategy and management policy. næringer, Nærings- og handelsdepartementet, Oslo.
In order to develop a sound safety culture on-board, National Transportation Safety Board 1990, Grounding of
the shipping companies should go in new directions U.S. Tankship EXXON VALDEZ on Bligh Reef, Prince
and pursue a crewing strategy which offers more William Sound Near Valdez, AK March 24, 1989.
favourable employment terms and fixed shifts for all Neal, A., Griffin, M.A. & Hart, P.M. 2000, ‘‘The impact
nationalities, and strive for a more stable workforce. of organizational climate on safety climate and individual
Another measure would be to accept the Captain’s behavior’’, Safety Science, vol. 34, no. 1–3, pp. 99–109.
and department managers’ roles as leaders, and offer Netemeyer, R.G., Sharma, S. & Bearden, W.O. 2003, Scaling
managerial development. A final measure will be procedures: issues and applications, Sage Publications,
Thousand Oaks, Calif.
to develop a policy and system that ensure proper Norwegian Shipowners’ Association, www.rederi.no.
onboard management. Perrow, C. 1999, Normal accidents: living with high-risk
technologies, Princeton University Press, Princeton, N.J.
Peterson, M.F., Ashkanasy, N.M. & Wilderom, C.P.M. 2000,
REFERENCES ‘‘Handbook of organizational culture & climate’’ in Sage,
Thousand Oaks, Calif., pp. 193.
Bakka, D. & Sjøfartsdirektoratet 2004, Hundre år for sikker- Pett, M.A., Lackey, N.R. & Sullivan, J.J. 2003, Making sense
het til sjøs: Sjøfartsdirektoratet 1903–2003, Direktoratet, of factor analysis: the use of factor analysis for instrument
Oslo. development in health care research, Sage, Thousand
Cooper, M.D. 2000, ‘‘Towards a model of safety culture’’, Oaks, Calif.
Safety Science, vol. 36, no. 2, pp. 111–136. Rasmussen, J. 1997, ‘‘Risk management in a dynamic soci-
DeJoy, D.M. 2005, ‘‘Behavior change versus culture change: ety: a modelling problem’’, Safety Science, vol. 27, no.
Divergent approaches to managing workplace safety’’, 2–3, pp. 183–213.
Safety Science, vol. 43, no. 2, pp. 105–129. Reason, J. 2001, Managing the risks of organizational
Dekker, S. & Dekker, S. 2006, The Field guide to understand- accidents, Ashgate, Aldershot.
ing human error, Ashgate, Aldershot. Reason, J. & Hobbs, A. 2003, Managing maintenance error:
Department of Transport 1987, mv Herald of Free Enterprise. a practical guide, Ashgate, Aldershot.
Geller, E.S. 2001, The psychology of safety handbook, Lewis Schein, E.H. 2004, Organizational culture and leadership,
Publishers, Boca Raton, FL. 3rd edn, Jossey-Bass, San Francisco.
Goffman, E. 1968, Asylums: essays on the social situ- Sorensen, J.N. 2002, ‘‘Safety culture: a survey of the state-
ation of mental patients and other inmates, Penguin, of-the-art’’, Reliability Engineering & System Safety,
Harmondsworth. vol. 76, no. 2, pp. 189–204.
Guldenmund, F.W. 2000, ‘‘The nature of safety culture: a Studio Apertura 2004, Sikkerhetskulturer i transportsek-
review of theory and research’’, Safety Science, vol. 34, toren. Metoder for kartlegging av sikkerhetskultur:
no. 1–3, pp. 215–257. Evaluering av noen eksisterende verktøy.
Hair, J.F. 1998, Multivariate data analysis, 5th edn, Prentice Tinmannsvik, R.K. 2008 Robust arbeidspraksis: hvorfor
Hall, Upper Saddle River, N.J. skjer det ikke flere ulykker på sokkelen? Tapir akademisk
Hinkin, T.R. 1995, ‘‘A review of scale development practices forl., Trondheim.
in the study of organizations’’, Journal of Management, Yule, S.J. 2003, ‘‘Senior manager’s transformational lead-
vol. 21, no. 5, pp. 967–988. ership behaviours for safety’’. Australian Journal of
Justis- og politidepartementet 1991, Scandinavian Star— Psychology, vol. 55, no. Supplement.
ulykken, 7 april 1990, Justis- og politidepartementet,
Oslo.
1430
Article 2
shipping industry and identification of influencing factors. Maritime
Policy & Management, 37(6), 601-623.

MARIT. POL. MGMT., NOVEMBER 2010,
VOL. 37, NO. 6, 601–623
Risk perception in the Norwegian shipping industry

and identification of influencing factors
HELLE OLTEDAL*y and EMMA WADSWORTHz

yFaculty of Technology, Business and Maritime Education,
Stord/Haugesund University College (HSH), Bjørnsonsgate 45,
Haugesund 5528, Norway
zSeafarers International Research Centre (SIRC), Cardiff University,
52 Park Place, Cardiff CF10 3AT, UK
In the recent years, safety culture in relation to shipboard safety and organiza-
Downloaded By: [Swets Content Distribution] At: 11:23 1 November 2010
tional factors has received increasing interest within the shipping industry.
This study examines this relationship, where risk perception is an indicator for
shipboard safety. Data was derived from a survey carried out in 2006, where 1262
questionnaires were collected from 76 vessels. Explorative factor analyses were
used to extract factors of safety culture. Analysis of variance was used to assess
the associations between the safety cultural factors and demographic and
organizational variables. Finally, linear regression analysis was carried out to
assess the association between risk perception and safety culture, controlling for
the influence of demographic and organizational variables; age, department,
vessel type and nature of work. The results suggest that safety-oriented shipboard
management style, performance of proactive working practices and good
reporting practices all contribute to a better perception of shipboard safety,
while a high demand for efficiency contributes to a more negative perception of
the safety level. Further, safety is perceived as better when work is performed as a
team. To gain a better understanding of risk perception and safety at sea, it would
be helpful to further examine the characteristics and influences of teamwork.
1. Introduction
The term ‘safety culture’ first appeared after the Chernobyl accident in 1986 [1, 2].
Since then there has been considerable interest in the concept and how it relates to
other safety related organizational, sociological and individual conditions, with some
researchers suggesting that it may be the main recent issue in organizational
safety [3]. However, there is still considerable debate about safety culture with no
clear consensus yet on a number of fundamental aspects, including: its definition and
scope; the theoretical issues underpinning it; its relationship to safety climate and to
other organizational factors and characteristics; and how best to measure it [4–9].
Nevertheless, safety culture has been implicated by a number of recent industrial
accident investigations (see, e.g. [1, 10–12]) and it is widely accepted that
organizations with a strong safety culture are more effective at preventing workplace
accidents and injuries [13]. Broadly, it is assumed that an organization’s strategic
approach towards business, operational management and safety is reflected in the
local working situation in the form of patterns of staff safety related behaviours,
*To whom correspondence should be addressed. E-mail: helle.oltedal@hsh.no
Maritime Policy & Management ISSN 0308–8839 print/ISSN 1464–5254 online ß 2010 Taylor & Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/03088839.2010.514954
602 H. Oltedal and E. Wadsworth
perceptions and attitudes, and this is what is commonly referred to as the

organization’s ‘safety culture’ [6, 7, 9, 14, 15].
In this article, risk perception is used as an indicator of the actual shipboard safety
level, as previous research indicates that risk perception and risk behaviour are
strongly correlated [16]. Also, recent research within the maritime industry has
suggested risk perception to be a good indicator for safety level in general [17]. An
important conclusion of Rundmo’s study [16] was the indication that behaviour is
largely constrained by working conditions. When these conditions are not judged to
be at a satisfactory level of safety, employees know that the risk of suffering an
accident is enhanced and consequently they feel unsafe. Perceived risk, therefore,
may be directly affected by an organization’s safety management, demands for
efficiency and other local workplace and organizational factors. Reason [5] has
identified three organizational driving forces for safety; commitment, competence
and cognizance. Commitment refers to motivation and resources. Organizational
leaders, such as shipboard management, are important motivational driving forces.
Management at local level often reflects the organization’s strategic orientation,

whether to pursue commercial fortunes or good safety records and practice. A highly
committed organization also makes available sufficient recourses, such as time,
equipment and people. Competence is related to the organization’s safety manage-
ment systems; whether they get in the right experience data, the quality of data, how
the organization acts upon it in the form of feedback, training and development of
other safety measures. However, neither commitment nor competence is enough if
the organization does not have cognizance, or awareness, of the dangers in their
operation. Such awareness may be nurtured by active discussions within working
groups, proactive activities such as safe job analysis and other activities which give
safety and risk continuous attention. These three Cs, commitment, competence and
cognizance, may be referred to as cornerstones in safety culture [5].
Seafaring is widely acknowledged to be, comparatively, a very high risk
occupation [18–21]. An examination of the statistics for loss of life and personal
injuries has ranked seafaring as the second most dangerous occupation in the
world [20]. Fatal accident rates among British seafarers, for example, are over 26
times higher than those of other workers [19]. Within the maritime setting, it is
traditionally estimated that about 80% of accidents (i.e. individual injury and
navigational accidents such as groundings and collisions) are attributable to human
error [22, 23]. Such a ‘person oriented’ focus is not unique to the maritime industry
[24, 25]. As a result of such a person-oriented focus, procedures and standards are in
place to control the behaviour of seafarers, although with limited success, as
accidents involving human error seem to be on the rise [26]. However, there has been
a recently increasing interest in other potentially influential factors which are found
in the organizational and situational context, inducing human error through
unfavourable working conditions. This is a move away from blaming individuals
towards a more holistic understanding of accidents and their causes [15, 22]. The
influence of organizational issues may appear in the form of productivity pressure,
poor training or inadequate manning, which may result in unfavourable working
conditions, such as increased demand for efficiency and work stress and lack of both
resources and competence [5, 23]. Hetherington et al. [23] reviewed the current status
of safety in the maritime industry, and the contribution of human factors in the
causal chain of events in shipping accidents. She indicates that a particular
combination of low manning level, fatigue, stress, work pressure, communication,
Risk perception in the Norwegian shipping industry 603
environmental factors, and long periods of time away from home are workplace
dangers. When moving away from a person-oriented focus, blaming the individuals,
it is such organizational factors that ought to be brought to light and reformed.
Within the maritime setting, the notion of safety culture and its relation to safety
was first introduced by the International Safety Management (ISM) code [15, 27].
Until recently, however, safety culture has been given relatively little attention within
the shipping industry [23, 28]. The aim of this study, therefore, is to explore further
the concept of safety culture and its relationship to safety by using risk perception as
a proxy variable. Our contribution to the domain of maritime research is to bring
forward an estimation of the functional relationship between risk perception and
safety culture by using regression analysis.
2. Method
2.1. The Norwegian shipping industry
On 1 January 2010, the Norwegian-controlled fleet totalled 40.4 million dwt. Since 1
January 2009 the fleet has decreased by 3%. The number of ships in the fleet stands
at 1836. The total fleet has decreased by 40 ships since January last year. The number
of ships flying the Norwegian flag has decreased by 48 during the same period, and
46% of the ships now fly the Norwegian flag (NOR, The Norwegian Ordinary Ship
Register or NIS, The Norwegian International Ship Register) [29]. The other 54%
may be registered in any of the world’s 183 ship registers [30]. On an international
basis, the various flag states performances are assessed on standards of safety,
environmental and social performance and are maintained and enforced by flag
states, in full compliance with international maritime regulations. Based on their
performance, the flag states are then categorized into three subgroupings; the white
list (good performance), the grey list (mediocre performance) and the black list (poor
performance) [31].
The Norwegian-controlled foreign-going fleet and distribution of type of vessel
and flag are given in Table 1 (29).
On 1 January 2010, seafarers of the nationalities given in Table 2 were employed
on Norwegian controlled ships (29).
Table 1. Norwegian-controlled foreign-going fleet, 1 January 2010: type of vessel and

registration.
Type of vessel Number of NOR Number of NIS Foreign flag Total
Passenger vessels 6 9 11 26
Gas tankers – 54 68 122
Chemical tankers 1 139 167 307
Shuttle/storage tankers 6 8 37 51
Other oil tankers 2 50 32 84
Combined carriers – 12 – 12
Bulk carriers – 58 93 151
Other dry cargo vessels 41 154 398 593
Offshore service 204 100 186 490
Total 260 584 992 1836
Table 2. Crew by country of domicile: employed on ships.
Country of domicile Number of people Country of domicile Number of people
Philippines 21 570 Sweden 490

Norway 10 370 UK 455
India 6135 Bangladesh 380
Russia 4020 Estonia 320
Poland 2505 Brazil 285
Ukraine 1140 Canada 190
Latvia 960 Portugal 160
Romania 920 Spain 150
China 660 Other countries 1960
Croatia 580
Total 53 250
Statistics from the Norwegian Maritime Directorate, up to 2008, indicate a

reduction in most types of accidents except for groundings which has been increasing
since 2005, with a marginal decrease in 2008—80, compared to 90 groundings
in 2007. In addition, 23 foreign vessels grounded in Norwegian waters in 2008.
The Norwegian Maritime Directorate points to causes related to navigation errors,
deficient watch keeping, inattention and use of non-authorized electronic equipment.
A frequent common underlying cause of groundings is fatigue. However, work
related and individual accidents/injuries have shown a pronounced decrease, with a
total of 451 registered personal accidents/injuries on Norwegian vessels in 2008, a
decrease from more than 1000 in 2003 [32].
2.2. Questionnaire development

In order to examine safety culture, a questionnaire developed by Studio Apertura, a
constituent centre of The Norwegian University of Science and Technology
(NTNU), in collaboration with the Norwegian Det Norske Veritas (DNV) and the
research institution SINTEF (Selskapet for INdustriell og TEknisk Forskning) was
used. The development was based on a theoretical review and an evaluation of the
eight pre-existing questionnaires, four developed in Norway, two in Denmark and
two in the UK [33, 34]. The evaluation was carried out according to the following
five criteria: (1) foundation (theoretical foundation, documentation and premises for
application), (2) thematic width, (3) practical experience of use, (4) the ability to
describe and measure safety culture, and (5) the ability to be used at multiple levels
(individual, group, team, company). The questionnaire was made up of 12 sections;
background, demographic information and risk perception, with the remaining
10 sections representing dimensions of safety culture; top management’s safety
priorities, local management, procedures and guidelines, interaction, work situation,
competence, responsibility and sanctions, working environment, learning from
incidents and description of the organization. Except for the dependent variable risk
perception, all constructs were measured on five-point Likert scale ranging from
‘strongly disagree’ to ‘strongly agree’, or from ‘very seldom/never’ to ‘very often/
always’. Risk perception was measured on a scale ranging from one to ten, with
1 indicating ‘very bad’ and 10 indicating ‘very good’.
Table 3. Original population, redefined population and number of participating vessels.
Number of Approximate Number of Total response, %

Original vessels in number of Norwegian- responding of Norwegian
population NSA memberlist controlled vessels vessels controlled
Bulk carrier 172 127 20 16

Chemical tanker 257 171 17 10
Combined carrier 20 13 2 15
Gas carrier 121 102 6 6
General cargo 255 215 19 9
Oil tanker 79 30 5 16
Shuttle tanker 49 43 7 16
Total 953 702 76 11
White flag 799 589 66 11
Grey flag 91 67 10 15
Black flag 63 46 0 –
Total 953 702 76 11

Norwegian flag 492 362 39 11
Other flags 461 340 37 11
Total 953 702 76 11
2.3. Survey sample

A total of 150 vessels were randomly selected from the 953 vessels within the
Norwegian Shipowners’ Association’s (NSA) list of members for 2005 for the target
group of liquid tankers, general cargo and bulk carriers. This number represented
approximately 15% of the overall population, which was considered to be large
enough to be representative of the population as a whole [35]. A sample of 10% is
recommended, but some withdrawals were expected, and thus a 5% margin was
included in the original sample. The sample was stratified with regard to status of the
vessel’s flag register (white, grey or black listed flag) [31, 36] and type of vessel
(general cargo, bulk carrier, oil tanker, gas tanker and chemical tanker). Following
the initial selection, telephone calls were made to each company to ask for their
participation. Thirty-one companies with 83 vessels agreed to participate, and
45 companies with 67 vessels declined. Reasons for not participating included the
following:
. Being unable to contact the company despite repeated efforts (23 vessels,
16 companies).
. The vessel was not owned by a Norwegian party, and therefore not defined as
Norwegian controlled (15 vessels, 8 companies).
. Ship management was outsourced to a non-Norwegian country, and therefore
not defined as Norwegian controlled (14 vessels, 8 companies).
. The company refused to participate (12 vessels, 10 companies).
. The remaining vessels were sold (3 vessels, 3 companies).
The population was redefined, and vessels managed from a non-Norwegian
country were not considered Norwegian controlled. The original and redefined
populations (Norwegian controlled) are given in Table 3, along with the number and
percentage of responding vessels in each category. In total, 1574 questionnaires were
distributed to 83 tankers and bulk carriers. Seventy-six vessels from 29 companies
returned a total of 1262 forms, which gives an individual response rate of 80.2%,
a vessel response rate of 91.5% and a company response rate of 93.5%.
Unfortunately, no company with vessels flying a black listed flag was willing to
participate.
2.4. Statistical analyses

The Statistical Package for the Social Sciences (SPSS) v.16.0 was used to perform all
the analyses, which included descriptive statistics, exploratory factor analysis (EFA),
correlation analysis, analysis of variance (ANOVA) and linear regression analysis
(LRA).
With regard to the EFA, principal component analysis with Varimax rotation was
carried out, in order to explore the latent underlying dimensions reflecting the
concept of safety culture on board. Factors were extracted based on the three
following analytical criteria: (1) pairwise deletion, (2) eigenvalue more than 1.0 and
(3) factor loading more than 0.50. This was followed by a scale reliability test
evaluated upon the following three criteria: (1) Cronbach’s alpha coefficient 40.70,
(2) item-total correlation40.40 and (3) inter-item correlation40.30 and50.80 [37, 38].
Correlation analysis was performed to evaluate the construct validity. To be a true
reflection of the underlying safety culture dimension, factors should correlate in a
consistent manner with the theoretical concept.
One-way ANOVA was carried out to explore any associations between the
demographic (nationality, age, department and vessel type) and organizational
(work description) variables and both the dependent variable risk perception and the
independent safety culture dimensions.
Finally, LRA was carried out in order to assess the associations between risk
perception and the dimensions of safety culture, controlling for any potentially
influential demographic and organizational factors. Forward stepwise regression
analysis was carried out in one block (entry p50.05).
3. Results
3.1. Demographics
Of the 1262 returned questionnaires, the following three subgroups were excluded
from the analyses: the captains, whose closest managers are shore-based and not
shipboard personnel (n ¼ 66); catering personnel (n ¼ 126), who have fewer maritime
education requirements than deck and engineering personnel; and a group of
questionnaires forwarded by a Philippine crewing agency (n ¼ 81). The latter group
were excluded as there was no way to control whether the employing company was
within the target group (i.e. Norwegian controlled). The remaining 989 responses
(78%) were used in the analyses presented below. See Table 4 for demographical
distribution.
Only one of the respondents was female, and the rest male. Filipino seafarers
dominate the sample, followed by Polish and Norwegian.
3.2. Dependent variable

The dependent variable, risk perception, was measured on a 10-point scale (from 1
‘very bad’ to 10 ‘very good’) with the question ‘All in all, how would you assess the
safety in your working situation?’ Overall safety level was perceived as relatively
good, with a mean at 8 (standard deviation (SD) ¼ 2).
Table 4. Demographic profiles of respondents.
N N
Age Gender
531 years 30.7% 304 Male 99.9% 973
31–40 years 28.2% 279 Female 0.0% 1
41–50 years 28.8% 285 Missing 15
51–60 years 10.4% 103 Total 989
460 years 0.7% 7
Missing 11
Total 989
Vessel type Nationalities
General cargo/bulk 43.1% 426 Filipino 63.0% 623
Liquid tanker 56.9% 563 Polish 8.4% 83
Missing – Norwegian 8.2% 81
Total 989 Russian 5.9% 58
Indian 3.4% 34
Romanian 2.8% 28
Chinese 2.7% 27
Others (13 nations) 5.6% 34
Missing 21
Total 989
Table 5. ANOVA analysis between risk perception and groups.
N Mean SD SE df F Sig.
Age 823 8.09 1.714 0.060 4 1.796 0.128

531 years 263 8.06 1.669 0.103
31–40 years 229 7.96 1.778 0.117
41–50 years 234 8.12 1.765 0.115
51–60 years 91 8.47 1.537 0.161
460 years 6 7.33 0.816 0.333
Vessel type 831 8.09 1.710 0.063 1 0.221 0.638
General cargo/bulk 360 8.06 1.772 0.093
Liquid tanker 471 8.11 1.662 0.077
Work description 736 8.07 1.713 0.063 1 50.799 0.000
Individual 158 7.24 2.210 0.176
Team 578 8.30 1.710 0.063
Note: SE, standard error; df, degrees of freedom.
The results from ANOVAs assessing differences in risk perception by the

demographic (age and vessel type) and organizational (team or individual work)
factors are presented in Table 5.
The results show no significant differences between age groups or vessel types
However, there were significant differences between those who work on an individual
basis and those who work in a team. When work is on an individual basis, risk
perception is higher (indicated by a lower mean score corresponding to poorer
perception of safety (i.e. higher perception of risk)).
Table 6. Factor reliability test.
Factor Cronbach’s alpha Inter-item range Item-total range
Competence 0.907 0.549–0.785 0.646–0.820

Interpersonal relation 0.856 0.369–0.718 0.543–0.710
Shore orientation 0.745 0.313–0.422 0.470–0.530
Ship management 0.868 0.654–0.730 0.721–0.777
Work practices 0.788 0.345–0.735 0.450–0.702
Feedback 0.780 0.416–0.761 0.449–0.722
Demand for efficiency 0.754 0.368–0.723 0.417–0.692
Reporting practices 0.842 0.736–0.736 0.736–0.736
3.3. Factor analysis

Explorative factor analysis of the items in the questionnaire resulted in an eight
factor solution, explaining 67.10% of the total variance. The Kaiser–Meyer–Olkin
(KMO) statistic was very high (0.888) indicating the existence of a strong
relationship between the items. The factor solution is presented in the Appendix
along with each item’s factor loading. The resulting factors, or dimensions of safety
culture, were as follows:
(1) Competence level—training and education.
(2) Interpersonal relationship among the crew.
(3) Shore orientation towards risk and safety management.
(4) The ship management’s safety engagement.
(5) Proactive work practices.
(6) Feedback from reported conditions.
(7) Demand for efficiency.
(8) Reporting practices of near miss and incidents.
All factors returned a scale reliability estimate (Cronbach’s alpha) above 0.70. All
inter-item statistics are also between the recommended levels of 0.30 and 0.80, and
item-total above 0.40. As given in Table 6, the overall evaluation of the scales
demonstrates a good internal consistency, and therefore also a good representation
of their underlying safety culture dimensions.
As given in Table 7, all the eight safety culture dimensions correlate in consistence
with the theoretical concept, both internally and with the dependent variable of risk
perception (DV). They are therefore considered to be a valid reflection of the
underlying safety culture construct. Table 7 presents the correlation coefficients
between the factors (corr.), along with significance level (sig.) and number of
respondents (N). All correlations are significant at the 0.001 level (two-tailed).
3.4. Univariate analyses

The results from the analyses of variance testing for differences on the safety culture
dimensions by demographic (age and vessel type) and organizational (team or
individual work) factors are presented in Table 8. When significant differences are
found, number of respondents, mean and SD for each subcategory are also
presented.
Summarized results from the ANOVA are given in Table 9.
The detailed results in Table 8, which are summarized in Table 9, show significant
differences between age groups on the dimensions of interpersonal relationship and
Table 7. Factor correlation matrix.
F1 F2 F3 F4 F5 F6 F7 F8
F1 Competence Corr. 1
Sig.
N
F2 Interpersonal relations Corr. 0.569** 1
Sig. 0.000
N 884
F3 Shore orientation Corr. 0.283** 0.412** 1
Sig. 0.000 0.000
N 800 764
F4 Local management Corr. 0.331** 0.407** 0.354** 1
Sig. 0.000 0.000 0.000
N 935 880 791
F5 Working practices Corr. 0.432** 0.490** 0.279** 0.473** 1
Sig. 0.000 0.000 0.000 0.000
N 933 874 788 922
F6 Feedback Corr. 0.238** 0.243 0.251** 0.280** 0.273** 1
Sig. 0.000 0.000 0.000 0.000 0.000
N 933 752 714 788 790
F7 Demand for efficiency Corr. 0.251** 0.351** 0.507** 0.344** 0.309** 0.245** 1
Sig. 0.000 0.000 0.000 0.000 0.000 0.000
N 907 853 771 898 901 775
F8 Reporting practices Corr. 0.298** 0.321** 0.329 0.346** 0.372** 0.469** 0.318**
Sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000

N 831 782 782 819 817 742 801
DV Risk perception Corr. 0.277** 0.321** 0.338** 0.406** 0.374** 0.207** 0.358** 0.316**
Sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
N 815 759 691 804 800 692 784 724
Note: **denotes significance at 0.001 level.

609
Table 8. ANOVA analysis between factors and groups.
Dimension (in bold) N Mean SD SE df F Sig.
Competence
Age 952 21.39 0.75 0.08899 4 4.384 0.002
531 years 297 21.28 2.74
31–40 years 273 20.97 2.91
41–50 years 276 21.65 2.54
51–60 years 99 22.15 2.50
460 years 7 20.71 4.39
Vessel type 963 21.41 1.75 0.08853 1 4.835 0.028
Liquid tanker 551 21.58 2.71 0.11527
Dry cargo 412 21.18 2.79 0.13743
Individual 156 20.67 3.07 0.24561
Team 577 21.56 2.60 0.10074
Interpersonal relations
Age 888 25.03 3.18 0.10684 4 6.072 0.000
531 years 279 24.81 3.36 0.20089
31–40 years 257 24.46 3.20 0.19955
41–50 years 257 25.46 2.96 0.18475
51–60 years 90 26.07 2.83 0.29848
460 years 5 25.40 3.21 1.43527
Vessel type 898 25.14 3.18 0.10616 1 0.995 0.322
Individual 145 23.52 3.61 0.29992
Team 532 25.28 2.94 0.12725
Shore orientation
Age 797 12.57 3.67 0.13004 4 3.583 0.007
531 years 256 12.85 3.73 0.23336
31–40 years 229 13.00 3.67 0.24241
41–50 years 228 11.82 3.43 0.22703
51–60 years 78 12.60 3.95 0.44712
460 years 6 12.33 3.01 1.22927
Vessel type 805 12.56 3.70 0.13076 1 1.022 0.312
Individual 126 13.71 3.31 0.29451
Team 496 12.17 3.72 0.16692
Local management
Age 942 12.70 2.26 0.07364 4 2.141 0.074
Vessel type 953 12.70 2.27 0.07369 1 5.999 0.014
Liquid tanker 546 12.86 2.11 0.09047
Dry cargo 407 12.49 2.46 0.12201
Individual 153 11.87 2.70 0.21802
Team 569 12.98 2.16 0.09055
Work practices
Age 937 17.39 0.28 0.07433 4 1.848 0.118
Vessel type 948 17.39 2.27 0.07378 1 0.004 0.950
Individual 154 16.36 2.44 0.19689
Team 565 17.65 2.16 0.09082
(continued )
Table 8. Continued.
Dimension (in bold) N Mean SD SE df F Sig.
Feedback
Age 800 11.03 2.93 0.10347 4 1.398 0.233
Vessel type 808 11.03 2.93 0.10285 1 7.429 0.007
Liquid tanker 465 10.78 2.90 0.13470
Dry cargo 343 11.35 2.92 0.15777
Individual 126 10.24 3.04 0.27055
Team 488 11.58 2.80 0.11696
Demand for efficiency
Age 912 7.14 2.64 0.08726 4 1.208 0.306
Vessel type 922 7.13 2.64 0.08692 1 0.243 0.622
Individual 144 7.76 2.61 0.21709
Team 561 6.85 2.63 0.09899

Reporting practices
Age 912 7.14 2.64 0.08726 4 1.208 0.306
Vessel type 922 7.13 2.64 0.08692 1 0.243 0.622
Individual 136 6.76 2.07 0.17781
Team 509 7.42 2.03 0.09017
Note: SE, standard error; df, degrees of freedom.
shore orientation. Those of age 41 and above perceive interpersonal relationships to

be better than those of age 40 and below. Those between 31 and 40 years perceive the
shore organization to be less safety orientated than the other groups. The group
between 41 and 50 years had the best perception of the shore organization’s safety
orientation.
Type of vessel showed significant differences on the dimensions of competence,
local management and feedback. Those working on dry cargo vessels perceive the
feedback on reported experience data to be better than those working on liquid
tankers. However, those working on liquid tankers had better perception towards
level of competence and their local management than those working at dry cargo
vessels.
Work description, teamwork versus individual, showed significant differences on
all dimensions of safety culture. Those working in a team perceived the dimensions
of competence, interpersonal relationship, local management, feedback and report-
ing practices to be better than those working on an individual basis. Those working
in a team also felt less demand for efficiency and perceived the shoreside company to
be more safety orientated.
3.5. Results from regression analysis

Table 10 presents the results from the regression analysis.
The regression results indicate that local management, working practices and
reporting practices have a positive association with risk perception, and that demand
for efficiency has a negative association with risk perception. Moreover, working
situation has a positive association when work is performed as a team in contrast to
612
Table 9. Summarized results: ANOVA analysis between groups and safety culture dimensions.
Interpersonal Shore Local Working Demand for Reporting

Competence relations orientation management practices Feedback efficiency practices
Group
Age 0.002 0.000 0.007 0.074 0.118 0.233 0.306 0.306
Vessel type 0.028 0.322 0.312 0.014 0.950 0.007 0.622 0.622
Work description 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
H. Oltedal and E. Wadsworth
Table 10. Results from regression analysis.
R2 change B SE B
Constant 3.126 0.687

Local management 0.184 0.145 0.033 0.200*
Demand for efficiency 0.086 0.146 0.027 0.236*
Working practices 0.051 0.131 0.033 0.173*
Reporting practices 0.031 0.139 0.035 0.169*
Work description 0.016 0.491 0.163 0.119*
Note: R2 ¼ 0.35, *p50.001.
when work is performed on an individual basis. None of the demographic data were
significantly associated with risk perception.
Risk perception and significant safety culture dimensions will be further discussed
in the following section.
4. Discussion
4.1. Risk perception
Overall safety in the respondents’ working situation is perceived as very high, which
may indicate a relatively good safety standard. This result is surprising as seafaring is
considered to be a high-risk occupation [19, 20, 39]. One explanation for the good
risk perception is that the sample is representative only for vessels flying a white or
grey flag, which is presumed to perform better on safety. A second explanation is
based upon voluntary participation, that the companies that do not prioritize safety
may be those who did not participate. A third explanation for the high score is
negligence of danger as a coping strategy in a high-risk environment. Similar high
scores on safety are also seen among fishermen and employees of offshore service
vessels, even though both are considered high-risk industries [40]. A fifth explanation
is that risk perception is not a good indicator of ‘objective’ risk levels after all when it
comes to the maritime offshore industry. A final explanation may be related to the
implementation of many policies focusing on safety, such as the ISM Code and
International Ship and Port Facility Security (ISPS) Code, which may have
improved the overall standard of safety on board.
However, there is also another side to the picture of feeling safe. Perception is
assumed to influence behaviour in such a way that more risky actions are taken by
those who feel at greater risk, while fewer are taken by those who feel safer [41].
Hence, the perception of safety per se, may induce accidents. Similar situations are
also suggested at the organizational level, in that a low accident rate gives an
impression of safety, which may or may not reflect the reality, resulting in
safety being down prioritized, and thus the organization becoming more prone to
accidents [5].
The regression analysis suggests that variance in risk perception is associated
with local management, working practices, demand for efficiency, reporting practices
and work description (i.e. if work is normally done on an individual basis or in
a team).
4.2. The importance of working in a team

Whether or not a person works in team or on an individual basis is an organizational
decision. The results suggest that when working in a team, the perception of safety is
better than when working on an individual basis. It is of particular interest that team
versus individual work resulted in significant differences on all the eight dimensions
of safety culture, as well as the risk perception variable, suggesting that teamwork is
associated with better safety culture overall. When working alone, one has to rely on
one’s own experience and competence alone, both during normal work duties and if
something unexpected should occur. In teamwork, however, each individual
contributes with unique experience and competence, which may contribute to
enhanced situational awareness and handling of unexpected situations. Competence
is related to training and education, which along with experience are impor-
tant elements to shape skill-, rule- and knowledge-based activities and/or decisions
[5, 14, 42]. Crew members with high levels of experience, education and training are
expected to be better able to recognize, handle and control hazardous shipboard
situations [43]. Inadequate competence related conditions, with consequent lack of

risk awareness, have also been found to be an important influencing factor in
collisions and groundings [22], along with inadequate teamwork [43].
The results also suggest a strong relationship between the interpersonal
relationship dimension and competence. When competence relates to skill-, rule-
and knowledge-based activities and/or decisions, safety in practice relates to people
questioning the established mind set, work plans and the like. Such behaviour is
assumed to not only increase situational awareness, but also lead to increased
competence. The interpersonal relationships amongst crew are assumed to be a
determinant factor for creating an open and trusting relationship facilitating such
behaviour [5]. Also working in a team and on-the-job training have been found to be
important within navigation, as not all skills can be learned through formal
education and training or simulation [44]. Research initiated in 1990 by the US
Office of Naval Research, with scenario-based training being a core concept, also
points to the importance of such training for, inter alia, adequate measurement of
team performance, possibilities for diagnosing team performance and knowledge
about how to remediate poor performance (both process and outcome measures)
[45]. This research [45] has also identified four dimensions underlying good
teamwork: (1) information exchange, (2) communication, (3) supporting behaviour
and (4) team leadership.
The importance of teamwork has received attention from others. For example,
Mitropoulos and Cupido [46] found that accidents are more influenced by social
relations in teamwork rather than unsafe acts.
4.3. The importance of management

Perception of management, i.e. management commitment, management style,
management visibility, etc. is the most commonly measured dimension in safety
research in general [42] and also the most replicated across studies [47]. Our results
also suggest an association between management and risk perception. By sincere
safety engagement, involvement and follow-up, the management style at the vessels is
assumed to convey to the crew the importance of safety in their work at the vessel,
and for the organization as a whole. However, at sea, most department leaders, or
the captain for that matter, do not have managerial training or education. Onboard
management style is therefore left to each individual, and may vary substantially
from vessel to vessel or from one sailing period to another. The management
orientation will affect the shipboard safety by creating a shipboard atmosphere, i.e.
whether it is possible to report all kinds of experience data without getting sanctions
[15]. Across other studies, two critical functions of management have been found; to
help the team to compete a task, and to keep members maintained and function as,
inter alia, developing a positive culture [48]. From the construction industry, also
considered a highrisk occupation, it has been found that successful managers use the
following management practice: conduct new workers orientation, watch out for
vulnerable crew members, analyse productivity problems with the crew, respond to
good work, and create a calm and friendly job atmosphere [46]. It is also suggested
that effective management should be adjusted to the situation. Some studies indicate
that when situations are highly standardized or when the team members are
competent, directive leadership becomes unnecessary [48]. When leading and
working in a team, the management’s shifting role may be challenging. However,
team leaders can be trained to shift their role from: being a supervisor; to being a
team member engaging in command-and-control activities; or to being a facilitator

who helps teams diagnose and discuss their own performance in an open, trusting
and constructive manner [45]. Also, how to manage depends on the situation, be it
normal operations or an emergency or crisis [43]. The ship management’s importance
for safety within the shipping industry is also suggested in other research [15, 23, 49].
4.4. The importance of proactive work practices

Proactive work practices are activities where safety is discussed with others, or by
means of, inter alia, safe job analysis and safety evaluation. The intention is to
identify weaknesses and to take remedial actions before something fails or becomes
critical, and such activities may also be performed by single individuals [5, 50]. Our
results suggest that when proactive activities are performed, the working environ-
ment is perceived as safer. Our results also indicate that proactive activities are more
frequently performed when working in a team. These activities are performed most
efficiently when the working team are actively in dialogue, exploring all possibilities
of modes of failure with no restrictions. As discussed above, when a team is
composed of members with a variety of experience and competence, the possibility
for error detection increases, as long as the interpersonal relationship among the
team members allows it. For example, other research points to a situation in a
harbour, where each of the team members committed at least one error. However,
the error was detected and corrected by another team member [44] and thus possible
failure avoided. Seifert and Hutchins [44] also suggest that group identity appeared
to provide motivation for individuals to participate in minimizing error. Each person
in the team feels responsible not only for his own job, but for all parts and processes.
Hence proactive work practices are not limited to formal activities done before each
specific work operation, such as safe job analysis and the use of pre-printed error
detection forms and tools, but happen continuously during each work operation.
One pitfall is to become overly fixed within prescribed patterns of thinking [51].
Other more recent research also points to the problem of error detection in planning,
as opposed to on-the-job error detection, as this requires people to strive to see things
in a new way each time [52]. The same research [52] also favours on-the-job error
detection and correction stimulated by teamwork. When it comes to the possibility of
foresight, another distinction may be made between emergency and crisis [43].
An emergency is defined as a situation outside normal operating parameters, and the
necessary corrective actions are more often based on documented procedures.

Crises, on the other hand, differ as they might be more difficult to successfully handle
based on pre-planned procedures. These situations are often more complex
and unpredictable than emergencies, and thus require a creative and flexible
response [43].
It has been observed that within shipping, the more formal proactive approaches
sometimes are performed in a mechanical manner, and not as intended [15].
Thus, these may also be less efficient than the error detection and prevention that
happens during operations on an informal basis.
4.5. The importance of demand for efficiency

Work pressure is also noted to be one of the most common themes in safety surveys
[42]. As shipping is an international and global activity, increased competitiveness,
cost reduction and demand for efficiency will always, to a certain degree, be present.
When, and if, the company communicates this through the organization, the demand
for efficiency may result in people crossing the boundaries for safe behaviour [53].
The origin of demand for efficiency may also be traced back to organizational
design. For example, error diagnosis and safety may improve by including more crew
members in the form of expertise in redundant units within the system. However, this
also increases overall manning costs due to frequently underutilized expertise and
training [44]. If a specific area or department is constantly double manned in order to
handle unplanned and critical events better, the extra personnel may be redundant
most of the time, i.e. during ‘normal’ operations. The result is that organizations may
choose to have a manning level designed for normal operations, presupposing that
everything occurs as planned. Thus, the race for cost-effectiveness directly affects
thoroughness when situations are deviating from the pre-planned, or in more hectic
operations. The resulting high workload will then be followed by work stress,
perceived demand for efficiency and trespassing of safety boundaries.
We suggest that there are parallels between demand for efficiency due to economic
concerns and responding to emergencies and crises. The common factor is decision
making under stress. When a situation is deviating from normal operations, i.e. an
escalating marine emergency, it is suggested that two fundamental and interrelated
skills are required: (1) situation assessment (what is the problem) and (2) decision
making (what shall I do). Moreover, it is noted that experienced people, as opposed
to novices, are generally those making the decisions. Also, if decisions need to be
made quickly, time pressure becomes an additional stressor [43].
From other research, stressors have been divided into the following three
categories: (1) characteristics of the environment (e.g. noise, heat, cold, vibration,
altitude) that could have a direct physical impact on operators; (2) task complexity
or individual-level stressors (e.g. time pressure, workload, information uncertainty
and negative feedback) that occur as a result of the functions required in the process
of individual performance; and (3) teamwork factors (e.g. lack of cohesion,
communication delays, intra-group conflict and role ambiguity) that could act
either as ‘stressors’ or as ‘stress moderators’. Team characteristics affect the
operator’s perception of resources and may either exacerbate or alleviate the effects
of individual-level stressors [54].
Our results suggest that when working in a team, the crew felt less demand for
efficiency, and teamwork may thus be a counteracting buffer.
Nevertheless, both on the individual and organizational levels, the trade-off

should be balanced so that efficiency or thoroughness does not happen at the cost of
acceptable safety. Other research also points to demand for efficiency as one of the
most frequent reasons for violations of procedures and checklists [55], and an
important influencing factor in collisions and groundings [22] as within the offshore
oil industry [44]. Research from Wright (1986), cited by Mitropoulos and Cupido
(2009 [40]), highlights production pressure and the consequent development of
shortcuts, suggesting that while this did not present a problem for regular employees
who were familiar with work conditions and processes, the shortcuts presented a
more hazardous problem for contract employees less familiar with the production
system [46].
4.6. The importance of reporting practices

A near miss and incidents are situations that could have serious outcomes under
slightly different circumstances. They are more frequent than accidents, and regarded
as useful information for preventing accidents. A good reporting culture where
people are willing and motivated to report their own near misses and mistakes is
regarded as a cornerstone in order to achieve a safe culture. Interpersonal
relationships, i.e. openness, communication, problem solving and appreciation, lay
the foundations for the possibility of such reporting without the risk of negative
consequences or reactions such as blame [5]. This is supported by our results, which
indicate that interpersonal relationship is associated with reporting frequency. Our
results also suggest that when working in a team, the crew members are more likely
to place a report than when working on individual basis. One reason could be
positive group dynamics emphasizing safety and the importance of reporting.
Another reason could be that the near miss or incidents have been observed, and
therefore are more difficult to cover up. Reporting practices were also negatively
associated with demand for pressure, which suggests that when having high work
load, the crew find less time to place a report.
A good reporting culture is the basis for an organization to be informed about the
risks in their working situation. When managers are informed, necessary precautions
may be taken, which is assumed to bring increased safety. Previous research from the
shipping industry suggests that reporting practices are influenced and characterized
by the following: (1) lack of feedback given upon the reports—with feedback
perceived to be inadequate, too general and delayed, (2) lack of understanding of the
importance of the reports as preventive measures, (3) difficulties with defining a near
miss, which result in fewer reports, (4) fear of being blamed and negative
consequences; and finally (5) a complicated reporting system which is difficult to
relate to and time consuming [55].
5. Conclusion, limitations and suggestion for future research

The objective of this article was to assess the relationship between risk perception
and safety culture in the Norwegian-controlled shipping industry, and to identify
which factors affect risk perception. Risk perception has been used as a proxy
measure of the overall safety level on board vessels.
The results suggest that a safety-oriented shipboard management style, perfor-
mance of proactive working practices and good reporting practices, all contribute to
a better perception of the on-board safety level, while a high demand for efficiency
contributes to a poorer perception of the safety level.
Description of work—team versus individual work situation—had significant
association with risk perception. The results suggest that safety is perceived as better
when work is performed as part of a team, than on an individual basis. Also, when
working in a team, management was perceived to be more safety oriented, proactive
working practices were carried out more frequently and near misses and incident
reports were reported to be placed more frequently. Moreover, when working in
groups, the perception of efficiency pressure was less. However, none of these factors
should be addressed in isolation from the others. As followed from the discussion,
they are all important and mutually dependent.
The sample is representative of vessels flying a white and grey flag only, as those
registered under a black listed flag did not want to participate, As participation was
voluntary on behalf of the company, we assume that those participating do
emphasize safety, in general, and the development of a sound safety culture. Thus, it
is not possible to generalize the results to the overall population.

To understand risk perception and safety at sea, it would be helpful to further
examine the characteristics of teamwork along with the concept of group identity.
Shipping is distinctive among industry sectors in this respect in part because its teams
are often international and less stable due to different employment terms, and in
addition because the vessel is a 24 h society influencing all aspects of life on board,
not simply working processes. It would also be of interest to examine risk perception
both in general and per factor among the different nationalities of the respondents.
We also suggest that future research should address differences between white, grey
and black listed flags of registration.
Acknowledgements
The authors thank Neil Ellis at Seafarers International Research Centre (SIRC) for
the useful comments and suggestions.
References
1. INTERNATIONAL ATOMIC ENERGY AGENCY, 1991, Safety Culture (Vienna: IAEA).
2. INTERNATIONAL ATOMIC ENERGY AGENCY, 1992, The Chernobyl Accident: Updating of
INSAG-1: INSAG-7 (Vienna: IAEA).
3. TURNER, B. A. and PIDGEON, N. F., 1997, Man-Made Disasters, 2nd ed. (Oxford, UK:
Butterworth–Heinemann).
4. CHOUDHRY, R. M., FANG, D. and MOHAMED, S., 2007, The nature of safety culture:
A survey of the state-of-the-art. Safety Science, 45(10), 993–1012.
5. REASON, J., 2001, Managing the Risks of Organizational Accidents (Aldershot, England;
Burlington, VT: Ashgate).
6. COOPER, M. D., 2000, Towards a model of safety culture. Safety Science, 36(2), 111–136.
7. GULDENMUND, F. W., 2000, The nature of safety culture: A review of theory and
research. Safety Science, 34(1–3), 215–257.
8. NEAL, A., GRIFFIN, M. A. and HART, P. M., 2000, The impact of organizational climate
on safety climate and individual behavior. Safety Science, 34(1–3), 99–109.
9. PETERSON, M. F., ASHKANASY, N. M. and WILDEROM, C. P. M., 2000, Handbook of
Organizational Culture & Climate (Thousand Oaks, CA: Sage).
10. CAIB, 2003, Columbia Accident Investigation Board Report. http://caib.nasa.gov/news/
report/volume1/default.html, retrieved 10 March 2010.
11. CULLEN, W. D., 2001, The Ladbroke Grove Rail Inquiry. Part 2 Report. http://
www.railwaysarchive.co.uk/docSummary.php?docID=48, retrieved 10 March 2010.
12. HIDDEN, A., 1989, Investigation into the Clapham Junction Railway Accident. http://
www.railwaysarchive.co.uk/docSummary.php?docID=36, retrieved 10 March 2010.
13. BARAM, M. and SCHOEBEL, M., 2007, Safety culture and behavioral change at the
workplace. Safety Science, 45(6), 631–636.
14. GULDENMUND, F. W., 2007, The use of questionnaires in safety culture research – an
evaluation. Safety Science, 45(6), 723–743.
15. OLTEDAL, H. A. and ENGEN, O. A., 2008, Local management and its impact on safety
culture and safety within Norwegian shipping. In: Safety, Reliability and Risk Analysis
Theory, Methods and Applications, edited by M. Artorel, C. Guedes and J. Barnett
(Valencia, Spain: European Safety and Reliability Association), pp. 1423–1430.
16. RUNDMO, T., 1996, Associations between risk perception and safety. Safety Science,
24(3), 197–209.
17. GRABOWSKI, M., AYYALASOMAYAJULA, P., MERRICK, J. and MCCAFFERTY, D., 2007,
Accident precursors and safety nets: Leading indicators of tanker operations safety.
Maritime Policy and Management, 34(5), 405–425.
18. HÅVOLD, J. I., 2005, Safety-culture in a Norwegian shipping company. Journal of Safety
Research, 36(5), 441–458.
19. ROBERTS, S. E. and MARLOW, P. B., 2005, Traumatic work related mortality among
seafarers employed in British merchant shipping 1976–2002. Occupational and
Environmental Medicine, 62(3), 172–180.
20. ANDERSON, P., 2003, Cracking the Code: The Relevance of the ISM Code and its Impact on
Shipping Practices (London: Nautical Institute).
21. HANSEN, H. L. and PEDERSEN, G., 1996, Influence of occupational accidents and deaths
related to lifestyle on mortality among merchant seafarers. International Journal of
Epidemiology, 25(6), 1237–1243.
22. MACRAE, C., 2009, Human factors at sea: Common patterns of error in groundings and
collision. Maritime Policy and Management, 36(1), 21–38.
23. HETHERINGTON, C., FLIN, R. and MEARNS, K., 2006, Safety in shipping: The human
element. Journal of Safety Research, 37(4), 401–411.
24. CRICHTON, M., 2005, Attitudes to teamwork, leadership, and stress in oil industry drilling
teams. Safety Science, 43(9), 679–696.
25. HEINRICH, H. W., ROOS, N. and PETERSEN, D., 1980, Industrial Accident Prevention:
A Safety Management Approach, 5th ed. (New York: McGraw-Hill).
26. FRANK, J. and OSLER, D., 1998, Accidents double on fleet boom, officer shortage,
Lloyd’s List, 21 February.
27. LE MEUR, C., 2003, Maritime Safety Culture. http://www.ismcode.net/research_projects/
chrystelle_safety_paper.pdf, retrieved 10 March 2010.
28. HÅVOLD, J. I., 2000, Culture in maritime safety. Maritime Policy and Management, 27(1),
79–88.
29. NORWEGIAN SHIPOWNERS’ ASSOCIATION, 2010, Quarterly Information – Shipping and
Offshore Activities, No. 1 2010. http://www.rederi.no/default.asp?V_ITEM_ID=
502&AID=6874&TEMPORARY_TEMPLATE=79, retrieved 10 March 2010.
30. LLOYD’S REGISTER FAIRPLAY, 2008, World Fleet Statistics, retrieved from International
Maritime Organization upon request.
31. MARISEC, 2008, Shipping Industry Flag State Performance Table. http://www.marisec.
org/flag-performance/FlagStatePerformanceTable09.pdf, retrieved 10 March 2010.
32. NORWEGIAN MARITIME DIRECTORATE, 2009, Annual Report 2008. http://www.sdir.no/
upload/Om%20sj%C3%B8fartsdirektoratet/%C3%85rsmeldinger/%C3%A5rsmelding
%202008%20til%20nett.pdf, retrieved 10 March 2010.
33. SINTEF, 2003, Sikkerhetskulturer i transport: En kunnskapsoversikt. Report No. STF22
A03300, ISBN: 82-14-02830-2.
34. STUDIO APERTURA, 2004, Sikkerhetskulturer i transportsektoren. Metoder for
kartlegging av sikkerhetskultur: Evaluering av noen eksisterende verktøy.
Arbeidsrapport 1b/2004.
35. NEUMAN, W. L., 2000, Social Research Methods: Qualitative and Quantitative
Approaches, 4th ed. (Boston: Allyn and Bacon).
36. MARISEC, 2006, Shipping Industry Guidelines on Flag State Performance. http://
www.marisec.org/flag-performance/, retrieved 10 March 2010.
37. HAIR, J. F., 1998, Multivariate Data Analysis, 5th ed. (Upper Saddle River, NJ: Prentice
Hall).
38. FIELD, A., 2005, Discovering Statistics Using SPSS: (and sex, drugs and Rock ’n’ Roll),
2nd ed. (London: Sage).
39. HÅVOLD, J. I., 2005, Safety-culture in a Norwegian shipping company. Journal of Safety
Research, 36(5), 441–458.
40. BYE, R. and LAMVIK, G., 2007, Professional culture and risk perception: Coping with
danger on board small fishing boats and offshore service vessels. Reliability Engineering
and System Safety, 92(12), 1756–1763.
41. MEARNS, K. and FLIN, R., 1995, Risk perception and attitudes to safety by personnel in
the offshore oil and gas industry: A review. Journal of Loss Prevention in the Process
Industries, 8(5), 299–305.
42. FLIN, R., MEARNS, K., O’CONNOR, P. and BRYDEN, R., 2000, Measuring safety climate:
Identifying the common features. Safety Science, 34(1–3), 177–192.
43. BARNETT, M., GATFIELD, D. and HABBERLEY, J., 2010, Shipboard Crisis Management:
A Case Study. http://disruption.solent.ac.uk/mhfr/resources/RINA%202002.pdf,

retrieved 10 March 2010.
44. SEIFERT, C. M. and HUTCHINS, E. L., 1992, Error as opportunity: Learning in a
cooperative task. Human-Computer Interaction, 7(4), 409–435.
45. SALAS, E. and CANNON-BOWERS, J. A., 1998, Making Decisions Under Stress: Implications
for Individual and Team Training (Washington, DC: American Psychological
Association).
46. MITROPOULOS, P. and CUPIDO, G., 2009, The role of production and teamwork practices
in construction safety: A cognitive model and an empirical case study. Journal of Safety
Research, 40(4), 265–275.
47. MEARNS, K., WHITAKER, S. M. and FLIN, R., 2003, Safety climate, safety management
practice and safety performance in offshore environments. Safety Science, 41, 640–680.
48. KÜNZLE, B., KOLBE, M. and GROTE, G., 2010, Ensuring patient safety through effective
leadership behaviour: A literature review. Safety Science, 48(1), 1–17.
49. HÅVOLD, J. I., 2010, Safety culture and safety management aboard tankers. Reliability
Engineering and System Safety, 95(5), 511–519.
50. KJELLÉN, U., 2000, Prevention of Accidents Through Experience Feedback (London:
Taylor & Francis).
51. PIDGEON, N. and O’LEARY, M., 2000, Man-made disasters: Why technology and
organizations (sometimes) fail. Safety Science, 34(1–3), 15–30.
52. KONTOGIANNIS, T. and MALAKIS, S., 2009, A proactive approach to human error
detection and identification in aviation and air traffic control. Safety Science, 47(5),
693–706.
53. RASMUSSEN, J., 1997, Risk management in a dynamic society: A modelling problem.
Safety Science, 27(2–3), 183–213.
54. KONTOGIANNIS, T. and KOSSIAVELOU, Z., 1999, Stress and team performance: Principles
and challenges for intelligent decision aids. Safety Science, 33(3), 103–128.
55. OLTEDAL, H., 2009, The use of safety management systems within the Norwegian tanker
industry – do they really improve safety? In: Reliability, Risk, and Safety: Theory and
Applications, edited by R. Bris, C. Guedes Soares and S. Martorell (London: Taylor &
Francis), pp. 61–64.
Appendix
Eight factor rotated solution
Factors

Item description Competence relations orientation management practices Feedback efficiency practices
I have received the education that is 0.861 0.141 0.057 0.092 0.161 0.080 0.101 0.047
necessary in order to handle critical
or hazardous situations
I have received the education that is 0.846 0.246 0.067 0.069 0.095 0.072 0.022 0.054
necessary in order to work safely
I have received the training that is 0.837 0.198 0.069 0.079 0.143 0.067 0.044 0.059
necessary in order to handle critical
or hazardous situations
I have received the training that is 0.797 0.274 0.112 0.093 0.079 0.083 0.023 0.053
necessary in order to work safely
New crew members get a thorough 0.650 0.239 0.095 0.124 0.197 0.072 0.092 0.093
introduction to safety related issues
I feel appreciated by my co-workers 0.152 0.788 0.095 0.072 0.164 0.065 0.042 0.121
I feel appreciated by the ship 0.152 0.753 0.127 0.114 0.094 0.073 0.150 0.140
management
My co-workers do their jobs in a way 0.272 0.721 0.049 0.176 0.120 0.090 0.066 0.160
that make me feel safe
We solve problems and conflicts in a 0.268 0.670 0.115 0.152 0.203 0.012 0.134 0.061
good manner
My co-workers can communicate 0.321 0.591 0.234 0.024 0.135 0.012 0.042 0.064
effectively in English
The working environment is character- 0.207 0.563 0.159 0.162 0.126 0.093 0.181 0.104
ized by openness and dialogue
I feel that the work we do onboard is 0.043 0.231 0.704 0.061 0.104 0.007 0.045 0.132
too little appreciated by the
company
(continued )
621
622
Appendix. Continued.
Factors

Item description Competence relations orientation management practices Feedback efficiency practices
What we learn in courses is not relevant 0.073 0.049 0.689 0.022 0.092 0.104 0.087 0.079
in practice
Reporting itself takes too much time 0.094 0.093 0.676 0.067 0.095 0.006 0.077 0.091
In my company, they are more preoc- 0.076 0.024 0.640 0.162 0.022 0.163 0.272 0.025
cupied with the statistics than the
human consequences of an incident
When an undesirable incident has 0.069 0.283 0.535 0.207 0.032 0.012 0.237 0.134
occurred, people are more preoccu-
pied with placing blame than finding
the cause of the incident
Does your closest superior follow up to 0.140 0.085 0.140 0.826 0.237 0.105 0.106 0.090
H. Oltedal and E. Wadsworth
ensure that all work on board is done

in a safe manner?
Is your closest superior clear in his 0.131 0.163 0.099 0.823 0.178 0.071 0.100 0.097
engagement to ensuring his co-w-
orkers safety?
Is your closest superior a good role 0.100 0.220 0.114 0.802 0.106 0.086 0.100 0.088
model when it comes to attending to
his own and others safety?
Do you carry out a ‘safe job analysis’/ 0.139 0.114 0.058 0.0.73 0.845 0.015 0.100 0.083
‘risk analysis’ before high-risk
operations?
Do you carry out safety evaluation 0.195 0.147 0.133 0.132 0.819 0.074 0.090 0.094
before new working methods, tools
or routines are introduced?
Do you have the possibility to prioritize 0.179 0.290 0.081 0.222 0.619 0.073 0.108 0.037
safety first in your daily work?
Do you discuss safety issues with your 0.162 0.232 0.039 0.232 0.501 0.213 0.040 0.042
co-workers?
Do you receive constructive feedback 0.093 0.038 0.136 0.098 0.050 0.874 0.075 0.091
from the captain on the conditions
you report?
Do you receive constructive feedback 0.073 0.060 0.006 0.173 0.060 0.858 0.108 0.106
from the captain on the conditions
you report?
Do you get information from incidents/ 0.102 0.101 0.067 0.031 0.123 0.635 0.019 0.217
accidents on other vessels?
Due to captain’s demand for efficiency 0.122 0.109 0.166 0.123 0.012 0.069 0.858 0.062
we sometimes have to violate
procedures
Due to the company’s demand for 0.065 0.061 0.244 0.017 0.122 0.024 0.846 0.086
efficiency we sometimes have to
violate procedures
Do you ever feel forced to continue 0.013 0.235 0.153 0.160 0.124 0.135 0.521 0.102
your work even if safety may be
threatened?
Do close calls get reported in writing? 0.109 0.120 0.138 0.112 0.069 0.284 0.114 0.833
Do minor incidents get reported in 0.138 0.123 0.092 0.168 0.161 0.212 0.139 0.819
writing?
Eigenvalue 9.257 2.645 2.101 1.803 1.444 1.324 1.197 1.028
Explained variance 29.861 8.533 6.776 5.815 4.656 4.241 3.862 3.318
Notes: KMO ¼ 0.888. Loadings in bold indicate the factor in which the item was placed.
623

Article 3
Oltedal, H. A. (2010). The use of safety management systems within

the Norwegian tanker industry—Do they really improve safety? In R.
Bris, C. Guedes Soares, & S. Martorell (Eds.), Reliability, Risk and
Safety: Theory and Applications (pp. 2355-2362). London: Taylor &
Francis Group.

Article 4
Oltedal, H. A., & Engen, O. A. (2010). Tanker versus dry cargo—The
use of safety management systems within Norwegian dry cargo
shipping. In J.M. Ale, I.A. Papazoglou, & E. Zio (Eds.), Reliability,
Risk and Safety (pp. 2118-2125). London: Taylor & Francis Group.

Reliability, Risk and Safety – Ale, Papazoglou & Zio (eds)
© 2010 Taylor & Francis Group, London, ISBN 978-0-415-60427-7
Tanker versus dry cargo—The use of safety management systems

within Norwegian dry cargo shipping
H.A. Oltedal
Stord/Haugesund University College, Haugesund, Norway
O.A. Engen
University of Stavanger, Stavanger, Norway
ABSTRACT: In this article, we explore how the Norwegian controlled dry cargo shipping industry
compares to the tanker industry regarding reporting practices, procedures and checklists, and how the
industry balances commercial pressure and safety concerns. The article presents several differences
between the two sectors, i.e. underreporting of minor incidents and close calls are present to a larger degree
within the dry cargo sector than within tank. Also within the dry cargo industry, the procedural system
is experienced as being more useful in daily operations. The statistical data derive from a survey carried
out in 2006, supported by qualitative information from two case studies. The analysis suggests that the
differences originate partly from each sector’s relationship with their respective customers. Customers of
dry cargo shipping demand fewer requirements with regard to safety management and liquid tankers are
to a larger degree embedded in the oil industry, with a more mature safety management system. However,
it is suggested that this relationship between the liquid tanker industry and their customers, brings about
changes based on external demands, in contrast to internal needs. This may explain both better reporting
practices and the development of an extensive procedural system within liquid tankers.
1 INTRODUCTION it was formally adopted by IMO November 1993.

For oil tankers, chemical tankers, gas carriers, bulk
During the second half of the 1980s and early carriers and cargo high speed craft of 5000 gross
1990s, there seemed to be a significant increase of tonnage and upwards, implementation became man-
maritime accidents. Just to mention a few examples: datory no later than 1th of July 1998, and extended
In 1980 the ore/bulk/oil carrier MV Derbyshire was to remaining categories of ships in 2002.
lost at sea with all 42 crew and two supernumer- Seaborne world trade may be divided into two main
ary wives on board during a typhoon off Japan. In groups such as tank and dry cargo. Tank, or trans-
1989 the oil tanker Exxon Valdes ran aground off portation by tankers, refers to cargo in liquid form, as
the coast of Alaska causing extensive environmen- oil products, chemicals, vegetable oil and wine, trans-
tal damage and in 1992 the ore carrier Aegean Sea ported in tanks. Dry cargo refers to all cargo which is
broke in two at La Coruna. transported in dry form, traditionally split into two
From 1990 to mid-May 1997 alone, a total of groups; general cargo and bulk. General cargo refers
99 bulk carriers were lost at sea, with the death of to cargo that is susceptible for loading in general,
654 people. It became apparent that the international nonspecialized stowage areas or standard shipping
shipping industry was perhaps no longer capable of containers; e.g., boxes, barrels, bales, crates. The bulk
regulating itself and action was needed to reverse carrier carries dry cargoes which are shipped in large
the downward spiral of maritime calamity. Studies quantities and do not need to be carried in packaged
collected and reviewed by the International Mari- form. The principal bulk cargoes are grains, such as
time Organization (IMO), showed that maritime wheat, coal, iron ore, bauxite, phosphate and nitrate.
accidents involved people making mistakes, mistakes (IMO 1999, Skipsfartens bok 2008).
which may be traced back to on-shore decisions The academic literature dealing with human
(IMO 1999). This triggered the development of factors, safety and safety management systems within
the International Safety Management (ISM) code. merchant shipping is found to be limited, which
The ISM codes’ intention is to make the shore is also noted by other studies, see Hetherington,
side management responsible for effective safety Flin & Mearns (2006), Christophersen (2009),
management and the on board safety situation, and Håvold (2005). It has been even more difficult to
2118
retrieve academic literature dealing with differences The customers/charterers of liquid tankers are
of safety management between liquid and dry cargo known to emphasize safety, both environmental and
shipping, However, differences in reporting practices individual, when shipping their cargo. One of the
between dry cargo and tanker vessels are suggested most significant safety initiatives was introduced by
by Ellis et al. (2010), Oltedal et al. (2010). Safety the Oil Companies’ International Maritime Forum
management through the ISM-Code empathize (OCIMF) in 1993, when launching a Ship Inspection
organizational learning and improved safety through Report Programme (SIRE). This programme was
collection of experience data, e.g., by means of self originally launched to specifically address concerns
reported accidents, incidents and near-misses. The about sub-standard shipping, and is today a unique
information is processed and used for development tanker risk assessment tool of value to charterers,
of remedial actions to prevent future individual ship operators, terminal operators and government
incidents and accidents, and total loss of vessel and bodies concerned with ship safety (SIRE 2010).
cargo (Kjellén 2000, Anderson 2003). However, up Despite of this, recent research into the
to present time, far more effort is given to techni- Norwegian controlled tanker industry’s safety
cal safety than human factors, management and management system indicates a situation with
organizational factors. This misalignment in safety substantial under-reporting of experience data,
work may explain why maritime accidents are still such as self experiences, near misses and incidents.
on the rise (Christophersen 2009). Underpinning the Moreover, reports that were received by shore side
ISM-Code and the safety management system is the appeared to have been “massaged”, to present
idea of a safety culture being developed in the indus- a better state of affairs than was in fact the case.
try, and incorporation of the human elements. The situation is suggested induced by a blame
Research carried out in the aftermath of ISM- culture, where the seafarers fear for the conse-
Code implementation suggests industrial opposi- quences of reporting. Other factors counteracting
tion towards implementation, increased workload reporting are suggested to be a complicated and
of paperwork, and the potential situation of self paper producing reporting system and inadequate
incrimination by producing documentation which in feedback upon the reports. Also, although safety is
turn may be used as evidence (Christophersen 2009, the oil companies’ official priority, the sailors often
Anderson 2003, Oltedal 2010). experienced commercial pressure and efficiency to
This article explores three main research be the shore sides’ priority (Oltedal 2010).
questions:
1. What is the present safety management situation 2.2 Description of the dry cargo industry
within the dry cargo industry?
In volume, the bulk industry accounts for 50%
2. What are the main factors explaining the safety
of world seaborne trade (Skipsfartens bok 2008).
situations on dry cargo?
Customers/charterers of dry cargo shipping are
3. Finally, what is the current safety situation
as varied as their cargo, which includes everything
within dry cargo compared to the liquid cargo
from bananas to clothing and timber. According
industry?
to our sources (see methodology) the typical dry
cargo customers/charterers do not place the same
2 SHIPPING DRY VERSUS LIQUID emphasis on safety, more importance is placed on
CARGO price and time.
Dry cargo has been especially exposed to acci-
Although both dry and liquid cargo face similar dents, and after the increase in accident rates in the
challenges, some are more characteristic for each 90s, IMO worked with new standards concerning
group. For dry cargo it is of fundamental impor- technical and structural issues. However, it was
tance that the cargo is properly distributed through- noted that a number of accidents involving bulk
out the ship so that the structure is not overstressed carriers have occurred as a result of inadequate
and the ship has an adequate standard of stability loading and unloading and that safe practices could
(Skipsfartens bok 2008). A challenge typical for liq- prevent such accidents in future. On some occa-
uid tankers is emission of cargo and environmental sions inadequate loading could be traced back to a
pollutions. A shipping accident involving spill and situation where terminals, which were often owned
environmental pollution is apparently more often by the cargo owners or charterers of the ship, were
subject to public opinion and criticism. putting pressure upon the ships to amend their
loading plans or to load cargo to suit them, with
little consideration about the overall safety of the
2.1 Description of the tanker industry
ship, Thus safety was threatened by commercial
In volume, the tanker industry accounts for 35%–40% concerns (IMO 1999). Own field data indicates that
of world seaborne trade (Skipsfartens bok 2008). this description also depicts today’s situation.
2119
3 METHODOLOGY checklists, and (3) perceived balance between
commercial pressures and safety concerns. For
A multi-method approach combining surveys and facilitating comparison, statistical results from the
case studies, including interviews, is adopted. The tanker industry are presented in brackets, however
survey aims to indicate the interrelated patterns of the only when relevant to discussion and conclusion.
seafarers’ perceptions of and attitudes towards safety Each section is brought to a close with the
and issues related to the formal safety management discussion of the presented data and results. The
system. The qualitative results aim to give a more qualitative information is used as a framework for
thorough understanding of the overall situation, interpretation of the quantitative results.
underlying processes and the seafarers’ experiences
concerning own practice and safety management.
4.1 Reporting practices on board
3.1 Survey Survey questions related to reporting practices
are presented in Table 1, together with question
A total of 1574 questionnaires were distributed to
reference (ref) and number of respondents (n).
83 randomly selected Norwegian controlled liquid
The descriptive statistics, together with mean and
tankers and dry cargo carriers. 76 vessels returned
standard deviation, are presented in Table 2.
a total of 1262 completed forms. 455 of these
Explanation for interrelationship between survey
respondents worked on 35 dry cargo vessels.
questions (Oltedal 2010): Reporting practices (R1,
The survey forms part of a major safety culture
R2 & R5) are assumed to improve if the crew gets
survey carried out in 2006, performed with a
feedback from the report follow up (R3), and when
validated instrument (Oltedal, Engen 2009). The
the importance of reporting and its role in safety
results from liquid tanker industry are previously
management is comprehended (R4). A blame pol-
published in Oltedal (2010).
icy (R6) is supposed to reduce reporting (R1 & R2)
and increase the probability of a cover up of own
3.2 Case studies
mistakes (R5) (Reason 2001). A blame policy is
Two dry cargo companies were studied in 2008 and
2009. In order to ensure the companies’ anonym- Table 1. Reporting practices—survey questions.
ity, company specific information is retained. In
both companies data are retrieved from participa- Ref n
tion in captains’ conferences where safety was an
issue, and both formal and informal interviews R1 Do minor incidents get reported in writing? 406
with seafaring captains and shore-based personnel R2 Do close calls get reported in writing? 405
involved in safety related matters. R3 Do you receive constructive feedback from 391
the company on the conditions you report?
3.3 Statistical analysis R4 Reporting is important to prevent 451
the recurrence of accidents and incidents.
All statistical analysis is performed using SPSS R5 Are reports of undesirable incidents ever 383
version 16.0. Principal Component Analysis (PCA) “fixed up” to cover mistakes?
with Varimax rotation and Pairwise deletion is car- R6 When an undesirable incident has occurred, 444
ried out in order to examine the survey items’ interre- people are more preoccupied with placing
lationship, and in terms of their common underlying blame than finding the cause of the incident.
dimension. The extracted rotated component matrix
with factor structure and loadings is presented.
The loading represents the correlation between the Table 2. Reporting practices—descriptive statistic.
variable and the extracted factor(s), with estimates 1 2 3 4 5
ranging from 0 to 1.00. Items that load strongest Ref (%) (%) (%) (%) (%) Mean Std
on a given factor are considered to be most like the
factor, and thus the underlying latent dimension R1 7.1 10.6 22.9 30.8 28.6 3.63 1.20
(Hair 1998, Pett, Lackey & Sullivan 2003). R2 6.2 12.3 23.0 37.0 21.5 3.55 1.14
Descriptive statistics for each item, includ- R3 10.0 8.2 18.7 31.7 31.5 3.66 1.72
ing percentage frequency distribution, mean and R4 0.2 0.2 2.0 59.0 38.6 4.35 0.56
standard deviation are presented. R5 42.0 22.2 14.9 13.3 7.6 2.22 1.32
R6 14.6 49.1 12.4 20.9 2.9 2.48 1.07
4 PRESENTATION OF RESULTS R1 through R3 and R5: 1 very seldom/never, 2 seldom,
3 sometime, 4 often and 5 very often/always.
The results are presented in sections as follows; R4 and R6: 1 strongly disagree, 2 disagree, 3 not
(1) crews’ reporting practices, (2) procedures and sure, 4 agree and 5 strongly agree.
2120
when people are blamed for an unwanted situation Table 4. Reporting practices—factor structure and
rather than the situational circumstances. loadings.
The descriptive data in table 2 point to a situation
of substantial underreporting. A total of 40.6% Ref Fac. 1 Fac. 2
of the respondents state that they never or only R1 Do minor incidents get 0.854 .226
sometimes report minor incidents (tank: 35.4%) and reported in writing?
41.5% that they never or only sometimes report close R2 Do close calls get reported in 0.845 .209
calls (tank: 36.3%). In addition, 36.9% state that writing?
they never or only sometimes get constructive feed- R3 Do you receive constructive 0.782 0.230
back upon reported events (tank: 48.1%). However, feedback from the company
as few as 0.4% of the respondents disagree with the on the conditions you report?
statement that reporting is important to prevent the R4 Reporting is important to 0.433 .141
recurrence of accidents and incidents. A total of prevent the recurrence of
35.8% of the respondents admit to sometimes, often accidents or incidents.
or always fixing up the reports to cover up own mis- R5 Are reports of undesirable 0.054 0 .900
incidents ever “fixed up” to
takes (tank: 36.5%), and 23.8% perceive that a blame
cover mistakes?
culture is present in the organization (tank: 25.7%). R6 When an undesirable incident 0.600
.375
The respondents were also asked to list their rea- has occurred, people are more
sons for not reporting incidents. Each respondent preoccupied with placing
could mark up to three of eleven pre-specified blame than finding the cause
alternatives, or write any other reason if not listed. of the incident.
The three major reasons are listed in Table 3.
In order to report an incident or miss, the situ- *KMO 0.672.
ation has to be recognized as reportable. In both *Underlined loadings indicate the factor on which the
companies, both seafaring and shore personnel were item was placed.
presented to various near miss, incidents and acci-
dents scenarios (all similar but escalating situations)
and asked to define each situation as reportable or in prevention of future incidents and accidents is also
not, and categorize its accident potential. The results indicated. Within the liquid tanker sector, the latter
indicate that both shore personnel and to a larger item, reporting’s importance, was negatively associ-
degree seafarers have problems recognizing a near ated with undesired reporting practices. The descrip-
miss. This was reflected in both companies’ safety tive results imply that underreporting of close calls
management system with few reports. and minor incidents are present in a larger degree
The factor structure matrix and loadings are within the dry cargo sector than within the tank.
presented in Table 4. Both sectors have equal ranking of reasons
Table 4 indicates two underlying reporting for not reporting an incident (table 3), however
dimensions. One reflecting desired reporting within dry cargo 5% more of the respondents
practice (factor 1) and one reflecting undesired have listed “the incidents did not have any
reporting practices (factor 2). serious consequence” as the primary reason for
not reporting an incident. Both second and third
4.2 Section discussion reason listed for not reporting incidents are related
to fear for negative reactions and blame.
The factor structure (factor 1) indicates relationship The factor structure (factor 2) indicates
between the tendency to report and feedback given by relationship between an organizational blame
the company on reported events. An interrelationship culture and the tendency to alter the relation of
with the crews’ recognition of reporting’s importance how things happened in the reports, in order to
cover up own or others’ mistakes.
Table 3. Reasons for not reporting an incident—from
The results from factor 1 suggest that reporting
survey. frequency may increase if feedback upon reported
events is improved. Moreover, if the crewmem-
Response bers understand the importance of reporting
(%) to prevent future events, the results suggest that
reporting frequency may increase. However, the
The incident did not have any 35.7 respondents report less than the respondents within
serious consequence the tank sector, but are at the same time more sat-
There could be negative reactions 29.1 isfied with feedback given on the reported condi-
from my co-workers
tions. We suggest two possible explanations for
I am afraid that the information 20.6
will be used against me.
this relationship and sector differences. (1) Within
dry cargo fewer reports are placed, which gives less
2121
workload on shore personnel handling incoming the safety role they play (P2). They are also assumed
reports and thus better feedback may be provided. to be more workable if they reflect the reality of the
If this hypothesis should be true it also suggests working process. That is ensured through involving
that the liquid tanker industry has not provided the crew in the development of procedures, and by
sufficient resources/personnel at shore in order to paying attention to crews’ experience (P3). If these
handle incoming reports. (2) The factor structure factors (P2 & P3) are not present, the procedural sys-
for dry cargo and liquid cargo differed at one item, tem may be perceived as confusing and difficult to
which reflects the respondents’ understanding of relate to (P4 & P5) (Reason 2001). Factor analysis
reporting’s role in preventing future incidents and also included item R5 “When an undesirable incident
accidents. This suggests that within dry cargo, has occurred, people are more preoccupied with plac-
increased reporting may be achieved by measures ing blame than finding the cause of the incident” Our
aiming at improving the crews’ understanding of the hypothesis is that organizations inclined to focus on
interrelationship between experience information the human as a cause, as in a person model (Reason
and prevention. Within the tanker sector, safety 2001), are also more inclined to react by adding more
campaigns towards sailing personnel are quite procedures, which may result in a procedural system
common, and if the latter suggestion should be true, difficult to relate to.
it also suggests that these campaigns should have The descriptive data in Table 6 show that 94%
some effect upon increased reporting practices. of the respondents regard the procedures as help-
ful in their work (tank: 93%), and 91.1% states
4.3 Procedures and checklists to have received good training in the companies’
procedures (tank: 90.3%). 68.9% state that they
Survey questions related to the perception of have got the opportunity to influence and form the
procedures are presented in Table 5, with question procedures (tank: 67%). A total of 19.1% feel it is
reference (ref) and number of respondents (n). difficult to know which procedures are applicable
The descriptive statistics, together with mean and (tank: 22.1%), and 10% that procedures are diffi-
standard deviation, are presented in Table 6. cult to understand or poorly written (tank:17.3%).
Explanation for interrelationship between survey The respondents were also asked to list their rea-
questions (Oltedal 2010): Good procedures are sons for not following procedures. Each respond-
supposed to be helpful in the work (P1). It is ent could mark up to three of seven pre specified
assumed that procedures are more easily put into options, or write any other reason if not listed. The
operational practice if the user has received training three most frequent reasons are listed in Table 7.
in how to understand and apply the procedures, and Results from the qualitative studies indicate that
procedures and checklists are used in a less extent
Table 5. Procedures—survey questions. within dry cargo than within the tanker sector. The
interviewees’ explanation is that their customers
Ref n (dry cargo) do not make demands for specific use
and development of such. According to the inter-
P1 The procedures are helpful in my work. 451 viewees, this is resulting in a procedural system more
P2 I have received good training in the 450 understandable, simplified and easy to follow.
company’s procedures. Factor structure matrix and loadings are
P3 We have the opportunity to influence 431 presented in Table 8.
and form the procedures. Table 8 indicates two underlying dimensions
P4 I feel that it is difficult to know 439 of procedures. One reflecting helpful procedures
which procedures are applicable.
(factor 1) and one reflecting poor procedures
P5 The procedures are difficult to 447
understand or are poorly written. (factor 2).
4.4 Section discussion
Table 6. Procedures—descriptive statistics. The factor structure (factor 1) indicates that
procedures are perceived as helpful when sufficient
1 2 3 4 5
Ref (%) (%) (%) (%) (%) Mean Std
Table 7. Reasons for not following procedures.
P1 0.00 1.6 4.4 61.6 32.4 4.25 0.61
Response
P2 0.7 2.4 5.8 63.3 27.8 4.15 0.69 (%)
P3 3.2 13.0 14.8 57.3 11.6 3.61 0.963
P4 13.0 49.9 18.0 18.2 0.9 2.44 0.96 The work will be done faster. 47.6
P5 14.8 62.2 13.0 8.7 1.3 2.20 0.84 The procedures do not work as intended. 39.7
I feel pressured because I am overloaded 35.6
P1 through P5: 1 strongly disagree, 2 disagree, 3 not with work.
sure, 4 agree and 5 strongly agree.
2122
Table 8. Procedures—factor structure and loadings. and driven by internal need. Our hypothesis is
that when the procedural systems are formed by
Ref Fac. 1 Fac. 2 external influence, in contrast to internal need, it is
manifested in a procedural system perceived as less
P1 The procedures are helpful 0.835 .185
in my work. suitable, adequate and helpful in daily operations.
P2 I have received good training 0.772 216
in the company’s procedures. 4.5 Balance between commercial pressures
P3 We have the opportunity to 0.720 0.019 and safety concerns
influence and form the
procedures. Survey questions related to crew’s perception of
P4 I feel that it is difficult to 0.001 0.853 their priority of safety versus efficiency are pre-
know which procedures sented in Table 9, together with question reference
are applicable. (ref) and number of respondents (n). The descrip-
P5 The procedures are difficult .100 0.807 tive statistics, together with mean and standard
to understand or are
poorly written.
deviation, are presented in Table 10.
R5 When an undesirable .300 0.574 Explanation for interrelationship between survey
incident has occurred, questions (Oltedal 2010): The priority of safety ver-
people are more preoccupied sus efficiency is supposed to be communicated from
with placing blame than the organization’s top management (SE2) down to
finding the cause of the the lower levels of the organization (SE1), which are,
incident.
*KMO 0.634. Table 9. Balance between commercial pressures and

*Underlined loadings indicate the factor on which the safety concerns—survey questions.
item was placed.
Ref n
training in using the procedures are provided, and SE1 I am confident that my company always 448
when the users get involved in the development prioritizes the crew’s safety.
process. On the other hand, the procedural system SE2 The on-shore top management in my 450
is perceived as less helpful in companies that are company prioritizes safety before economy.
perceived as blame oriented (factor 2). The item SE3 Do you have the possibility to prioritize 459
relationships in factor 2 suggest that poor proce- safety first in your daily work?
dures and company blame policy do have a com- SE4 Do you ever feel forced to continue your 454
mon underlying organizational factor, which is work even if safety may be threatened?
reflected in their safety management orientation. SE5 The number of crewmembers is not sufficient 447
In the factor structure representing dry cargo, to ensure safety on board.
SE6 The management doesn’t care how we do 445
most variance was explained by factor 1, helpful
our work as long as the work gets done.
procedures. Within liquid cargo, most variance was
SE7 I experience that safety is more a façade 442
explained by poor procedures. The first extracted than a real priority area.
factor always explains the largest amount variance,
and this difference suggests that the dry cargo sector
to a greater degree are represented by helpful proce- Table 10. Balance between commercial pressures and
dures, and the liquid cargo sector to a larger degree safety concerns—descriptive results.
are represented by a poor procedural system.
It is also noted a difference among stated top 1 2 3 4 5
three reasons for not following procedures. In Ref (%) (%) (%) (%) (%) Mean Std
dry cargo, pressure due to work overload were
stated among the top three, and within liquid SE1 1.6 2.5 10.5 46.2 39.3 4.19 0.84
tanker poor procedures were stated among the SE2 2.4 3.2 10.9 55.1 28.4 4.04 0.86
top three. The other reasons were ranked equally. SE3 0.7 1.5 4.8 38.1 54.9 4.45 0.72
SE4 44.7 20.0 28.2 6.8 0.2 1.98 1.01
This difference also suggests that poor procedures
SE5 17.0 37.8 13.6 26.2 5.4 2.65 1.19
are more representative for the liquid tanker
SE6 20.7 45.6 16.4 13.7 3.6 2.34 1.06
sector. Although both sectors do have to relate to
SE7 17.2 39.4 10.0 26.2 7.2 2.67 1.24
minimum standards given by IMO and flag state
administration, within the liquid tanker sector, the SE1, SE2 & SE5 through SE7: 1 strongly disagree,
procedural system is to a larger degree formed and 2 disagree, 3 not sure, 4 agree and 5 strongly
driven by customer demands. Within dry cargo agree. SE3 & SE4:1 very seldom/never, 2 seldom,
the procedural system is to a larger degree formed 3 sometime, 4 often and 5 very often/always.
2123
ideally, aligned (Reason 2001). However, deviation Table 11. Balance between commercial pressures and
from official policies and goals may be experienced safety concerns—factor structure and loadings.
(SE6, SE7).The level of stress experienced is interre-
lated with crewing level. Crewing constitutes a major Ref Fac. 1 Fac. 2
part of operational expenditures, and how safety is pri- SE1 I am confident that my 0.819 .223
oritized may be reflected in the on board crewing level company always prioritizes
(SE5), which again influences the crew’s opportunity the crew’s safety.
to prioritize safety in their daily work (SE3, SE4). SE2 The on-shore top management 0.767 .106
The descriptive statistics show that 85.5% of the in my company prioritizes
respondents are confident that their company does safety before economy.
prioritize the crews’ safety (tank: 90.6%). A little less, SE3 Do you have the possibility 0.710 .113
83.5% do feel that the top management prioritizes to prioritize safety first
safety before economy (tank: 83.8). 93% states to often in your daily work?
or always having the possibility to prioritize safety SE4 Do you ever feel forced .512 0.455
to continue your work even
first (tank: 94.6%). However, 35.2% do sometimes
if safety may be threatened?
or often feel forced to continue work even if safety
SE5 The number of crewmembers .128 0.745
should be threatened (tank: 32.5%). Moreover, 17.3% is not sufficient to ensure
agree with the statement that the management does safety on board.
not care how the work is done, as long as the work SE6 The management doesn’t .363 0.721
is done (tank: 13%). 33.4% percent experience safety care how we do our
to be a façade more than a real priority area (tank: work as long as the work
35.6%), and 31.6% that the number of crewmembers gets done.
is not sufficient to ensure safety (tank: 27.5%). SE7 I experience that safety .050 0.685
The results from case studies indicate that is more a façade than
the sailing officers experience an efficiency a real priority area.
pressure due to commercial concerns. However, in
*KMO 784.
one of the companies, the perceived pressure was *Underlined loadings indicate the factor on which the
apparently more prominent than real pressure. From item was placed.
discussions between officers and commercial agents
it was apparent that perceived pressure originated the crew experience management to disregard how
from misunderstanding in communication and lack work is done together with lack of sufficient crew
of understanding of the other parts’ work tasks and in order to ensure safety (factor 2).
responsibilities. In the other company, demand for The factor structure differs from the tanker
more efficient operations and turnover were clearly industry in one aspect; within dry cargo most of
expressed by the top management. However, in the the variance is explained by safety perceived as
latter company the case study was carried out in the priority (factor 1): Within the tanker industry most
period of current financial crises, with low freight of the variance was explained by safety being a
rates and earnings, and thus demand for efficiency façade and with efficiency being the real priority.
could be regarded necessary in order to survive as a The results from the case studies suggest that
company. It was also found that safety and attitudes demand for efficiency is a reality also within dry
campaigns were implemented in a lesser degree than cargo. However, within dry cargo the crew is to
within the tanker sector. Factor structure matrix a lesser degree exposed towards official safety
and loadings are presented in Table 11. programs, initiated by external partners. It seems
Table 11 indicates two underlying reporting that external safety initiatives are potentially
dimensions. One reflecting when safety is perceived less integrated in the organization, compared to
to be a priority (factor 1) and one reflecting when when initiated by internal needs. However, the
demand for efficiency is perceived to be a priority safety initiative within the tanker sector may also
(factor 2). have given the crew higher expectations regarding
safety priority. Thus the discrepancy may also be
4.6 Section discussion explained by differences in expectations, as well as
differences in work practices
The factor structure (factor 1) indicates that when
top management and the company is perceived to
prioritize the crew’s safety, the crew also perceive 5 CONCLUSION
to have the possibility to prioritize safety in their
daily work. No business company exists with the The tanker sector is extensively embedded in the
primary objective to be safe. It is also indicated that Norwegian oil industry. The Norwegian oil industry
when the safety priority is perceived as a façade, is dominated by a limited number of big enterprises
2124
with “machine bureaucratic structures” well suited Christophersen, J.G. 2009, “Sikkerhetsstyring i skipsfarten
for co-operation with legal authorities and their rules 1998–2008: bakgrunnsfaktorer for reguleringsmessig
and regulations (Lindøe et al. 2010). This explains etterlevelse og overtredelse av ISM-koden”, vol. no. 22,
the extended focus on safety, safety management, pp. 699.
Ellis, N. & Bloor, Michael, Sampson, Helen 2010,
attitudes etc. within the tanker industry compared to “Patterns of seafarer injuries”, Maritime Policy &
dry cargo. Although both sectors do have substan- Management, vol. 37, no. 2, pp. 121–128.
tial underreporting, our data indicate that the safety Hair, J.F. 1998, Multivariate data analysis, 5th edn, Pren-
campaigns, as typical for the oil industry, have some tice Hall, Upper Saddle River, N.J.
positive effects upon reporting practices. However, Håvold, J.I. 2005, “Safety-culture in a Norwegian ship-
within the dry cargo industry the feedback upon ping company”, Journal of Safety Research, vol. 36,
reported events are perceived as better than within no. 5, pp. 441–458.
the tanker sector. One hypothesis is that the shore Hetherington, C., Flin, R. & Mearns, K. 2006, “Safety
tanker organization is not prepared to manage the in shipping: The human element”, Journal of Safety
Research, vol. 37, no. 4, pp. 401–411.
increased work load that increased reporting brings IMO and the safety of bulk carriers, International Mari-
along. Feedback is assumed to motivate reporting, time Organization, 1999, Retrieved 14 March 2010
and when the shore organization is not prepared to at http://www.imo.org/includes/blastDataOnly.asp/
handle increased reporting, this may counteract with data_id%3D7987/BULK99.FIN.pdf.
the effect of safety campaigns. However, the dry Kjellén, U. 2000, Prevention of accidents through experi-
cargo industry could have a positive effect from such ence feedback, Taylor & Francis, London.
campaigns, in increasing the crews’ awareness and Lindøe, P.H., Engen, O.A. & Olsen, O.E. 2010, “Responses
recognition of a near miss, along with better under- to accidents in different industrial sectors.” Safety
standing of reporting’s importance in safety manage- Science. Forthcoming.
Oltedal, H.A. & Engen, O.A. 2009, “Local management
ment systems. The shore side is thus recommended and its impact on safety culture and safety within
to have resources available for potential increase in Norwegian Shipping” in Safety, Reliability and Risk
number of reports placed, and proper follow up. Analysis Theory, Methods and Applications, eds. M.
When following up the reports, the development of Artorel, C. Guedes & J. Barnett, 2008th edn, European
new measures should also be considered carefully, Safety and Reliability Association.
and alternatives to development of new procedures Oltedal, H. 2010, “The use of safety management systems
developed. The experiences from the tanker within the Norwegian tanker industry—do they really
industry suggest that development of a constantly improve safety?” in Reliability, Risk, and Safety:
increasing, detailed and extensive procedural system Theory and Applications, ed. Bris, Guedes Soares &
Martorell, Taylor & Francis Group.
might undermine safety. Moreover, the analysis Oltedal, H. & McArthur, D. “Reporting practices in
suggests that safety measures should be initiated merchant shipping, and identification of influencing
by internal and industrial need, not external factors”. Submittet in Safety Science, April 2010.
demands from customer. When externally initiated, Pett, M.A., Lackey, N.R. & Sullivan, J.J. 2003, Making
safety management might be less integrated in sense of factor analysis: the use of factor analysis for
the operational part of the organization. It is also instrument development in health care research, Sage,
suggested that this external demand is related to the Thousand Oaks, Calif.
existence of a poor procedural system. Reason, J. 2001, Managing the risks of organizational
accidents, Ashgate, Aldershot.
SIRE, 2010, Retrieved 15 March 2010 at http://www.
REFERENCES ocimf.com/pages.cfm?action sire_introduction2.
Skipsfartens bok: Norges maritime næringer, 2008,
Anderson, P. 2003, Cracking the code: the relevance of Breakwater Publishing.
the ISM code and its impact on shipping practices,
Nautical Institute, London.
2125
Article 5
Oltedal, H. & McArthur, D. (2010). Reporting practices in merchant

shipping, and the identification of influencing factors. Safety Science,
49(2), 331-338.

Author's personal copy
Safety Science 49 (2011) 331–338
Contents lists available at ScienceDirect
Safety Science
journal homepage: www.elsevier.com/locate/ssci
Reporting practices in merchant shipping, and the identification

of influencing factors
H.A. Oltedal ⇑, D.P. McArthur
Faculty of Technology, Business and Maritime Education, Stord/Haugesund University College, N-5528 Haugesund, Norway
a r t i c l e i n f o a b s t r a c t
Article history: The objective of this paper is to identify the factors determining the reporting frequency of experience
Received 29 April 2010 data e.g. incidents and accidents. The empirical setting is the Norwegian controlled merchant fleet. Data
Received in revised form 16 August 2010 were collected from a survey carried out in 2006, where 1262 questionnaires were gathered from 76 ves-
Accepted 14 September 2010
sels. The data were subjected to explorative factor analysis, method of principal component and varimax
rotation. Seven factors, representing latent dimensions of safety culture, were extracted. Internal consis-
tency (Cronbach Alpha) and scale reliability were found to be acceptable. The factor scores were used in
Keywords:
an ordered logistic regression to examine the factors’ relationships to reporting frequency. The results
Reporting practices
Safety management
show that enhanced safety related training, a trusting and open relationship among the crew, safety ori-
Safety culture ented ship management, performance of pro-active risk identification activities and feedback on reported
Ordered logistic regression events all are significantly related to higher reporting frequency. On the other hand, demand for effi-
Shipping ciency and lack of attention to safety from shore personnel, are significantly related to lower reporting
Seafaring frequency. The results also show a significantly lower reporting frequency among those who have worked
with their local manager less than 1 year. Bulk and dry cargo vessels also show significantly lower report-
ing frequency than those working on liquid bulk carriers.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction situation in working operations. Thus, under-reporting of safety

related events constitutes a major threat to the efficiency and
Information about incidents, near-misses, operational failures utility of a safety management system.
and successes are crucial when taking a pro-active approach to There is an extensive literature considering reporting practices
safety. Such information is often collected and processed into basic and the factors which influence these. Particular interest has been
data for remedial actions through the use of a formal safety man- shown in high-risk sectors such as civil aviation and nuclear power
agement system, which constitutes a cornerstone of organizational plants. Attention has also been given to road and rail transporta-
learning (Kjellén, 2000; Reason, 2001). Within the maritime indus- tion, health care and oil & gas related activities. Although the
try, at the international administrative level, guidelines for formal importance of reporting practices in determining the outcome of
safety assessment systems have already been developed. In these, a safety management system has been acknowledged in the liter-
identification of potential hazards is the first step (International ature, little attention has been given to the topic within merchant
Maritime Organization, 2007a). At company level, safety manage- shipping. This is somewhat surprising given that seafaring is
ment is regulated by the International Safety Management (ISM) regarded as a high-risk occupation (Anderson, 2003; Hansen and
code, which requires shipping companies to have a system of Pedersen, 1996; Håvold, 2005; Roberts and Marlow, 2005).
reporting and collecting experience data (International Maritime A few studies which examine merchant shipping have been
Organization, 2002). Although various systems for safety manage- retrieved by searching through the Science Direct database for peer
ment do exist, safety management per se is not only a system prop- reviewed publications. The findings indicate a culture of under-
erty. It is not sufficient simply to have a system for the collection of reporting of safety information, which represents a shortcoming
safety related data. The system efficiency is determined by its in the their safety management (Psarros et al., 2010; Oltedal,
human interrelationships (crew, shore personnel, analysts and 2010; International Maritime Organization, 2007b; Ellis et al.,
others). A fundamental pillar of safety management is that infor- 2010). An analysis of accident data from the Lloyd’s Register
mation reported into the system is reliable and reflects the actual FairPlay (LRFP) and the Norwegian Maritime Directorate (NMD)
for vessels registered in Norway, suggests that only 30% of acci-
⇑ Corresponding author. Tel.: +47 20 70 26 44; fax: +47 938 26 187. dents experienced were reported (Psarros et al., 2010). Own re-
E-mail address: helle.oltedal@hsh.no (H.A. Oltedal). search indicates that 36% of shipboard tanker crew never or only
0925-7535/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssci.2010.09.011
332 H.A. Oltedal, D.P. McArthur / Safety Science 49 (2011) 331–338
sometimes report near-misses or minor incidents (Oltedal, 2010). 2009). The reporting of an incident or accident is often a time-con-
Within the dry cargo sector, around 40% state that they never or suming activity, and it is therefore reasonable to expect that de-
only sometime report a minor incident or near-miss (Oltedal and mands for efficiency will lead to less reporting, as well as an
Engen, 2010). Differences in reporting practices between dry cargo increase in situations that should have been reported in the first
and tanker vessels are also suggested by others (Ellis et al., 2010). place.
The International Maritime Organization (IMO) expresses an Another factor influencing people’s propensity to report is how
awareness of the under-reporting of safety information, and high- the organization handles blame and punishment (Reason, 2001).
lights that this must be improved (International Maritime Reason (2001) emphasises the importance of organizations devel-
Organization, 2007b). oping a just culture. In a just culture, there is an atmosphere of
trust in which people are encouraged or even rewarded for provid-
1.1. Theoretical foundation ing essential safety-related information, but in which there is also a
clear line between acceptable and unacceptable behaviour. For
Although various safety management models exist (see Kjellén, example, if efficiency by unsafe acts is knowingly accepted during
2000), of fundamental importance in all of them is the collection of normal operations, the operator should not be blamed and pun-
safety information from the operational productions system, used ished when that very same action is a causal factor in an unwanted
with the intention of preventing future accidents and other un- event.
wanted episodes. A model of a safety management system is illus- Feedback of the results to the reporter is considered a motivat-
trated in Fig. 1. ing factor, as the individuals understand the relevance of reporting
Fig. 1 shows the flow of information in a safety management and see that is taken seriously. However, feedback on reported
system. The shaded areas represent the safety information system, events in the form of constantly developing new procedures may
that is, the part of the safety management system which provides also be perceived as blame, as it is signalling that the operator
information needed for decisions and signalling related to safety did something wrong. Such a person oriented approach also has
matters. Reporting and the collection of data on, for instance, oper- another drawback. When searching for human error, the organiza-
ational accident risk by means of self reported accidents, incidents tion often does not look behind the operators when looking for
and near-misses is regarded as a critical function of the overall sys- causal factors, which could be found elsewhere in the organization
tem. Information processing and the development of remedial ac- such as, for instance, a pressure for efficiency.
tions all depend on the reliability and accuracy of the reported Reporting practices, how the organization handles blame, pun-
information (Kjellén, 2000). ishment and feedback, and the general functionality of the organi-
When it comes to the under-reporting of safety related events, zation’s safety management, are regarded as key elements in an
Kjellén (2000) uses behavioural theory to explain different propen- organization’s safety culture (Reason, 2001). Management charac-
sities to report. Behavioural theory is primarily used to explain teristics and leadership, at all levels in the organization both shore-
why people deliberately violate safety rules. It focuses on the side and shipboard, are regarded as major enablers and barriers
consequences of behaviour and how these consequences affect with respect to the development of an efficient safety culture
people’s judgments in relation to recurring situations, and their in- (Maritime and Coastguard Agency, 2004), and thus also an ade-
grained action patterns in these situations. For instance, when per- quate reporting culture.
forming an operation, people may have two action alternatives to
choose between: one which is considered safe and one considered 1.2. Previous research
unsafe. Experience shows that people are inclined to choose the
unsafe alternative when this alternative provides a consequence The under-reporting of safety-related information has been
regarded as positive, such as saving time or other resources. When identified as a problem in several industries. For example, it is sug-
people in turn get positive feedback for their efficiency, deliber- gested that up to 68% of all workplace accidents and injuries are
ately violating safety rules becomes a valuable skill, and over time not captured in the national injury surveillance systems, which
a part of the organizational culture. This safety-efficiency trade-off were set up by the Occupational Safety and Health Administration
has also been noted by other researchers, for example (Hollnagel, (OSHA) and the Bureau of Labor Statistics (BLS) in the United States
(US) (Rosenman et al., 2006). Others suggest that 81% of injuries
experienced remain unreported (Probst et al., 2008). Within the
health industry, barriers against reporting include fear of reprisals,
lack of confidentiality, time constraints and lack of post-reporting
feedback (Espin et al., 2007).
Although Psarros et al. (2010) concludes that under-reporting
constitutes a major problem within the maritime industry, the rea-
sons for non-reporting are not addressed (Psarros et al., 2010). Ellis
et al. (2010) suggest that under-reporting is more frequent on gen-
eral cargo vessels compared to others. However, these conclusions
are drawn based on differences in aggregate data in administra-
tions, and not reporting practices directly. Barriers against report-
ing are not reliably identified, however differences in national/
cultural risk perception have been suggested.
A recent cross-industrial review indicates that a fear of blame
and punishment (legal, organizational or from co-workers) is the
most commonly cited barrier to reporting (van der Schaaf and
Kanse, 2004). The International Maritime Organization also
addresses fear of being thought blameworthy or of being disci-
plined, embarrassment, fear of legal liability, and so on, as the main
barriers towards reporting (International Maritime Organization,
Fig. 1. Safety management system (adapted from Kjellén, 2000). 2007b). van der Schaaf (2004) has also identified other common
H.A. Oltedal, D.P. McArthur / Safety Science 49 (2011) 331–338 333
barriers such as: a lack of follow-up, managerial issues such as no Being unable to contact the company despite repeated efforts
commitment or distrust, time demands, a paper producing report- (23 vessels, 16 companies).
ing system and a lack of understanding of what constitutes a near- The vessel was not owned by a Norwegian party, and therefore
miss and incident. Own research indicates that all of these factors not defined as Norwegian controlled (15 vessels, 8 companies).
are present within the shipping industry (Oltedal, 2010). Moreover, Ship management was outsourced to a non-Norwegian country,
van der Schaaf (2004) reports that a ‘macho’ environment may and therefore not defined as Norwegian controlled (14 vessels,
discourage reporting, and that injury in such environments is 8 companies).
regarded as a part of the work. Based on his review, van der Schaaf The company refused to participate (12 vessels, 10 companies).
(2004) identifies four main categories of reporting barriers: (1) fear The remaining vessels were sold (3 vessels, 3 companies).
of disciplinary action (as a result of a ‘‘blame culture’’ where those
who commit an error are punished) or of other people’s reactions The population was redefined, and vessels managed from a non-
(embarrassment); (2) risk acceptance (incidents are part of the Norwegian country were not considered Norwegian controlled.
job, cannot be prevented, the ‘macho’ perspective); (3) useless On an international basis, the various flag states’ performances
(perceived attitudes of management taking no notice, not likely are assessed on standards of safety, environment and social perfor-
to do anything about it) and (4) practical reasons (too time-con- mance and are maintained and enforced by flag states, in full com-
suming, too difficult). pliance with international maritime regulations. Based on their
performance, the flag states are then categorized into three sub
1.3. The objective of the paper groupings; the white list (good performance), the grey list (medio-
cre performance) and the black list (poor performance) (MARISEC,
It is evident that the under-reporting of safety information and 2008). It turned out that companies operating vessels flying a black
risk, is a problem within the merchant shipping industry, yet we listed flag declined to participate in the study. Thus, the sample is
have not been able to find any research which empirically explores representative only for vessels flying a white and grey listed flag.
which factors affect the on-board reporting practices. As stated The results of the study were based on self-completed question-
earlier, a considerable amount of research has been undertaken naire data. In total, 1574 questionnaires were distributed to 83
looking at reporting practices and the factors influencing it within tankers and bulk carriers. 76 vessels from 29 companies returned
other industries. However, working within merchant shipping is in a total of 1262 forms, which gives an individual response rate of
many ways different from shore-based activities. Life and work on- 80.2%, a vessel response rate of 91.5% and a company response rate
board is a 24-h-a-day activity, where the crew has little opportu- of 93.5%. Each vessel received a package with individual question-
nity to interact with the surrounding society. Most of the sailors naires and a sealable return envelope. On each vessel, the safety
are contract employees, comprised of multiple nationalities, with delegate received instructions regarding administration, purpose
a typical sailing period of 9 months at a time. Also, due to its global and anonymity. Vessels not returning any questionnaires were
nature, the industry is highly exposed to competition. (Oltedal and reminded up to four times. The survey was administrated during
Engen, 2009). We have therefore identified a need for more re- the spring/summer of 2006.
search on the factors influencing reporting practices in the setting
of merchant shipping. 2.2. Questionnaire development
The objectives of the study are:
The questionnaire was developed by Studio Apertura (2004) (a
(1) To determine the structure of safety culture in Norwegian constituent centre of The Norwegian University of Science and
controlled merchant shipping by liquid tankers and bulk car- Technology (NTNU), in collaboration with the Norwegian DNV
riers using exploratory factor analysis. and the research institution SINTEF (2003)) as a part of a pro-
(2) Explore the relationship between on-board reporting prac- gramme for research in risk and safety in transport (RISIT) founded
tices and safety culture factors. by The Research Council of Norway. The main part of the question-
(3) Check for differences in reporting practices and familiarity naire was made up of 10 sections representing the following
with their local manager. dimensions of safety culture: top management’s safety priorities,
(4) Check for differences in reporting practices between type of local management, procedures and guidelines, interaction, work
vessel; liquid tanker versus bulk / dry cargo carrier. situation, competence, responsibility and sanctions, working envi-
ronment, learning from incidents, and description of the organiza-
2. Method tion. All these constructs were measured on five point Likert scales
ranging from ‘strongly disagree’ to ‘strongly agree’, or from ‘very
2.1. Survey samples and administration seldom/never’ to ‘very often/always’.
Håvold and Nesset (2009) discuss the issue of language and re-
The survey forms part of a PhD-research project which explores sponse style. They proceed using only an English and Norwegian
the relationship between shipboard safety culture and manage- version of their questionnaire. McCrae (2001) studied Norwegians
ment on Norwegian controlled liquid- and drybulk cargo vessels, and Filipinos who completed questionnaires both in their own lan-
with a second goal of looking for differences between the two ves- guage (Norwegian/Tagalog) as well as in English. He found no
sel classes. significant differences in the mean responses. However, in a
The sample was randomly selected from the Norwegian Ship- cross-national study, Harzing (2005) found that English language
owners’ Association’s list of members for 2005, and the sample versions tended to be more homogenized, potentially obscuring
was stratified with regard to type of vessel (general cargo, bulk car- cross-national differences.
rier, oil tanker, gas tanker and chemical tanker). Initially, 150 ves- Given that forcing the respondents to answer in the official
sels were selected, which constitutes about 15% of the population. working language of English could potentially bias the results,
Following the initial selection, telephone calls were made to each the questionnaire was also made available in Norwegian, Polish
company to ask for their participation. Thirty-one companies and Tagalog. This covered the main languages in use on-board. In
agreed to participate with 83 vessels. Forty-five companies with addition, during the fieldwork it became obvious that there was
67 vessels declined to participate. Reasons for not participating significant variation in the English-language abilities of the crew.
included: This provides further support for the approach adopted here. All
participants were issued with questionnaires in their own national Table 1

language and English. They were free to choose which version to Factor scale-reliability test, number of items and explained variance.
return. Factor N Explained Cronbach’s Inter-item Item-total

ID items variance Alpha range range
2.3. Demographics F1 4 30.621 .914 .651–.806 .781–.821

F2 4 10.102 .825 .453–.680 .604–.673
F3 3 7.917 .867 .647–.710 .731–.779
Questionnaires were returned from 40 liquid bulk carriers F4 3 7.296 .819 .527–.741 .564–.733
(liquid tanker) and 36 dry bulk carriers (dry cargo). 63% of the F5 3 5.612 .790 .440–.767 .470–.732
respondents were employed on a liquid tanker and 37% at a dry F6 4 5.391 .669 .262–.405 .410–.527
cargo vessel. The sample was male dominated (92.5% of the F7 2 4.366 .852 .743–.743 .743–.743
respondents). 22 nationalities were represented. The majority from
the Philippines (65.5%), followed by Norway (9.2%), Poland (8.1%)
and Russia (5.5%). Just over 56% of respondents were under the Factor 2; interpersonal, reflecting the relationship amongst the
age of 40. crew, problem solving abilities, form of communication and
sharing of safety information.
3. Results Factor 3; management, reflecting the crews’ perception of their
closest manager as a role model, engagement and interest in
The Statistical Package for the Social Sciences (SPSS) v.16.0 and ensuring safety in work operations.
STATA v. 10.1 were used for the data analysis. Factor 4; work practices, reflecting performance of pro-active
activities as safe job analysis and hazard identification, and
3.1. Dependent variable: reporting frequency how safety is prioritised in daily operations.
Factor 5; feedback, reflecting feedback given to crew on reported
The dependent variable, reporting frequency, was measured by safety information, and experience reports from other vessels.
the item: ‘‘Do minor incidents get reported in writing”. A total of 13% Factor 6; shore orientation, reflecting the shore-side part of
stated ‘never or seldom report’ (N = 164), 20.7% stated that they organisation’s attitude towards and prioritising of safety.
sometimes did report (N = 261), 31.8% that they often reported Factor 7; efficiency, reflecting the relationship between per-
N = 401) and 23.5% that they always reported minor incidents. ceived demand for efficiency and safe working practices.
The remaining 11.1% were missing (N = 140).
3.3. Ordered logistic regression
3.2. Factor analysis
The dependent variable ‘‘reporting frequency” has an ordinal
All 1262 responses were submitted to explorative principal nature, with four possible outcomes (1 = never/seldom, 2 = some-
component factor analysis with Varimax rotation, in order to iden- times, 3 = often and, 3 = always). It is possible to handle such data
tify the latent underlying dimensions of safety culture. The data in a number of ways. Ordinary least squares is seldom appropriate
were deemed appropriate for analysis, according to the Kaiser– for such data since it requires a continuous dependant variable. To
Meyer-Olkin measure of sampling value of .858, and significant convert our dependant variable to a continuous variable would re-
Barlett’s test. Factors were extracted based on the following three quire an assumption about the distance between each of the cate-
analytical criteria: (1) Pairwise deletion, (2) Eigenvalue more than gories. Rather than impose such an assumption, we choose an
1.0 and (3) factor loading more than 0.50. Items that failed to attain alternative technique: the ordered logistic regression model.
minimum loading of 0.5, or which loaded significantly on more The ordered logit regression model can be seen as an extension
than one factor, were omitted. This resulted in the extraction of of standard logistic regression. Standard logistic regression utilises
7 factors, explaining 71.305% of the total variance. a dichotomous dependant variable. For a polytomous variable, this
This was followed by a scale-reliability test. Each factor was can be extended to a multinomial logit model. However, this ne-
evaluated based on the following three criteria: (1) Cronbach’s glects the fact that the variables are ordinal, and assumes that they
Alpha coefficient > 0.70 (2) item-total correlation > 0.40, (3) in- are simply nominal. The ordinal logistic regression model utilises
ter-item correlation > 0.30, <0.80. However, these cut-off points this important information. In the model, m 1 equation are
are rules of thumb, and no clear consensus with regard to where simultaneously estimated, where m is the number of categories
the cut-off points exists (Hair, 1998; Field, 2005; Pett et al., 2003). of the dependant variable. The equations are formed by pooling
Each item’s theoretical significance was also taken into account. the data i.e. category 1 versus categories 2, 3 and 4, categories 1
Factor correlations analysis was carried out to evaluate the con- and 2 versus categories 3 and 4 and then categories 1, 2 and 3 ver-
struct validity, which concerns the theoretical relationship be- sus category 4.
tween the factors. (Hair, 1998; Field, 2005). All extracted factors The assumption made is that the effect of the independent vari-
were found to be a valid and reliable representation of the under- ables on the dependant variable is independent of the category.
lying safety culture construct. This assumption may be referred to as the proportional-odds
Each factor’s explained variance, Cronbarch’s Alpha value, inter- assumption or the parallel regressions assumption. Like any
item range and item-total range are presented in Table 1. Extracted assumption, it is not always met in practice. It can be tested by
factor structure, item description and loadings are presented in estimating a generalised ordered logit regression (which does not
Table 2. Loadings in bold indicate the factor onto which the item impose coefficient equality across the equations) and comparing
was placed. it to the ordered logit regression using a likelihood ration test.
The seven extracted factors in Table 2, representing the on- We performed this test and obtained a p-value of 0.036. This indi-
board safety culture, were labelled as followed (label in italic): cates that at the 5% level of significance, the assumption of parallel
regressions is not met.
Factor 1; competence, reflecting the crews’ perception of their There are a number of options available at this stage. We can as-
own training and education, in order to work safely and handle sume that the failure of the assumption is due to sampling variabil-
critical and hazardous situations. ity and proceed with the standard logistic regression. We could
Table 2
Seven factor rotated solution with factor loadings, and explained variance.
F1 F2 F3 F4 F5 F6 F7
I have received the training that is necessary in order to handle critical or hazardous situations .856 .165 .111 .163 .053 .015 .055
I have received the education that is necessary in order to handle critical or hazardous situations .855 .152 .117 .168 .078 .034 .070
I have received the education that is necessary in order to work safely .849 .246 .063 .103 .089 .096 .031
I have received the training that is necessary in order to work safely .813 .270 .097 .087 .082 .102 .037
We receive sufficient safety-related information when we sign on/start a new sailing period .191 .749 .137 .153 .042 .138 .039
We receive sufficient safety-related information when we start a new watch .239 .746 .151 .139 .106 .097 .016
We solve problems and conflicts in a good manner .232 .745 .143 .173 .016 .080 .098
The working environment on-board is characterized by openness and dialog .157 .730 .124 .087 .091 .120 .127
Does your closest superior follow-up to ensure that all work on-board is done in a safe manner .115 .142 .848 .193 .105 .120 .080
Is your closest superior a good role model when it comes to attending to his own and others’ safety .093 .212 .821 .104 .079 .139 .059
Is your closest superior clear in his engagement to ensuring his co-workers’ safety .141 .163 .817 .182 .072 .140 .076
Do you carry out a ‘‘Safe Job Analysis”/”Risk Analysis” before high-risk operations .130 .131 .136 .871 .048 .063 .075
Do you carry out a safety evaluation before new working methods, tools, or routines are introduced .185 .151 .169 .840 .115 .108 .036
Do you have the possibility to prioritize safety first in your daily work .190 .288 .202 .622 .112 .051 .101
Do you receive constructive feedback from the company on the conditions you report .087 .063 .103 .029 .883 .089 .041
Do you receive constructive feedback from the captain on the conditions you report? .078 .041 .152 .056 .875 .022 .077
Do you get information from incidents/accidents on other vessels? .066 .089 .020 .120 .684 .101 .022
In my company, they are more preoccupied with the statistics than the human consequences of an .089 .021 .098 .030 .154 .784 .091
incident
Reporting in itself take too much time .115 .078 .059 .177 .031 .709 .029
I experience that safety work is more a facade than a real priority area .070 .104 .063 .024 .110 .606 .222
When an undesirable incident has occurred, people are more preoccupied with placing blame than finding .085 .249 .204 .027 .016 .560 .193
the cause of the incident
Due to the captain’s demand for efficiency we sometimes have to violate procedures .093 .134 .130 .039 .091 .228 .878
Due to the company’s demand for efficiency we sometimes have to violate procedures .066 .093 .059 .146 .052 .253 .865
also move to a multinomial logit and disregard the ordinal nature Table 3
of the data. However, more insight can be gained by proceeding Results from the constrained generalised ordered logistic regression analysis with
factor scores (FAC). Odds ratios are presented.
with the generalised logistic regression. This model allows the
parameter estimates to vary by category. However, it may be the Beta Odds ratio Std. err. z P > |z|
case that the parallel regression assumption is violated only with FAC1 competence 1.4949 0.1387 4.33 0.0000
some of the included regressors. It is possible to test for this by FAC2 interpersonal 1.1971 0.1095 1.97 0.0490
imposing equality restrictions across equations on parameter esti- FAC3 management 1.5573 0.1489 4.63 0.0000
mates and testing the validity of these constraints using a Wald FAC4 work practices 1.3319 0.1262 3.02 0.0020
FAC5 feedback 1.8090 0.2943 3.64 0.0000
test. If there are good reasons, a priori, to select such constraints FAC6 shore orientation 0.6094 0.0594 5.08 0.0000
then these can be tested. Otherwise, a stepwise regression FAC7 efficiency 0.7581 0.0942 2.94 0.0030
approach can be used to test which variables violate the assump- Vessel (tank = 1) 0.5328 0.3104 2.03 0.0430
tion. In our case, there was no reason to suspect that some vari- Management (exp. > 1 year = 1) 2.9817 0.3215 3.40 0.0010
ables were more likely than others to violate the assumption. For Deviations from Proportionality 2
this reason, we chose the stepwise approach. The resulting model FAC5 Feedback 1.4424 0.1578 2.32 0.0200
Vessel (Tank = 1) 0.8117 0.2967 0.70 0.4820
is far more parsimonious than the unconstrained generalised logis-
Management (exp. > 1 year = 1) 0.9719 0.3092 0.09 0.9260
tic regression. The results are presented in Table 3.
Deviations from Proportionality 3
The results are presented as a set of ‘base’ coefficients and then
FAC5 Feedback 1.6840 0.2202 2.37 0.0180
deviations from proportionality. These deviations are calculated by Vessel (Tank = 1) 1.6483 0.3701 1.35 0.1770
taking the ratio of coefficients between equations (since the model Management (exp. > 1 year = 1) 0.4595 0.3915 1.99 0.0470
is presented using odds ratios). For example, the parameters given Constants
under the heading ‘‘Deviations from Proportionality 2” are ob- CONS1 1.6840 0.2792 6.05 0.0000
tained by dividing equation 1 by equation 2. So, for example, the CONS2 1.6483 0.2153 0.89 0.3750
odds ratio for category 3 for the variable ‘Vessel’ can be calculated CONS3 0.4595 0.2634 6.54 0.0000
n = 473
as 1.6483 0.5328 = 0.8782. Parameters which are constrained to ‘ ¼ 526:055
be equal across equations are not shown. Pseudo R2 = 0.1561
The results from the model are in line with expectations. The Wald v2 (15) = 154.99 p-value = 0.0000
odds ratios are in the expected order of magnitude and are jointly
significant. The factors competence, interpersonal, management,
work practices and feedback all increase the odds of being in a
higher category of the reporting frequency measure. The two 3.4. Validity and limitations
factors reflecting shore orientation and efficiency significantly
decrease the odds of being in a higher category of the reporting fre- There is a possibility that some of the relationships reported in
quency measure. Crews on tanker vessels are more likely to have a this section are subject to the common method bias (Campell and
lower reporting frequency than those on bulk vessels. Crews who Fiske, 1959; Podsakoff et al., 2003). This results from the fact that
have been working with their closest manager for more than one the data come from a common source. For example, a common
year tend to report more often. scale to the different questions. Potential statistical remedies have
The three variables which did not meet the parallel regressions been suggested. Spector (2006) is sceptical of the merits of such
assumption were FAC5 Feedback, Vessel and Management. This approaches. He argues that given that it is not possible to know
means that the effect of these variables is dependent on the cate- the existence or extent of any possible bias, treating it could in
gory of the dependant variable. fact introduce more bias than existed in the first instance. He
recommends using a multi-method strategy so that results do not port their own mistakes and experiences. Interpersonal relation-
only rely on the results of one questionnaire. In this research, case ships amongst crew, in practice, also reflects the degree to which
studies, interviews, participatory and field studies have been used the crew shares safety-related information when changing shifts,
to validate the data. The results have also been presented to people and more informal processes of sharing safety-related information
working within the industry who have expressed that they believe during operations. For both subordinate and superior and manage-
the results to give an accurate representation of the true situation. rial positions, additional challenges may arise in relation to multi-
national-crew and unstable crewing with low stability within
teams. When signing on a new ship, new crew will be unfamiliar
4. Discussion with the ship management’s and closest superior’s management
style as well as fellow crewmembers and the on-board working cli-
The results suggest that there is a positive association between mate. The seafarers require time to familiarize and adjust to the
the respondent’s perception of their local manager’s leadership new situation. For instance, if the ship management on the seafar-
skills and reporting frequency (P > 0.001, odds ratio = 1.5573). A ers’ previous vessel were blame oriented, this seafarer will most
leadership style where the manager is perceived as a good role likely join the new vessel with this latest experience in mind,
model ensures and follows up that all work on-board is done in a and be cautious about reporting their own mistakes for fear of
safe manner, were positively related to increased reporting being blamed or sanctioned. In time, the seafarer will learn how
frequency. This relationship is also supported by other research. the management is oriented on that particular vessel. The problem
Perception of management, such as management/leadership style, is even more pronounced when the seafarer is constantly changing
commitment and visibility, is the most commonly measured vessel, with new management each sailing period, and thus has to
dimension in safety research in general (Flin et al., 2000). Within go through this familiarization process each time. The manage-
the maritime sector, research initiated by the Maritime and Coast- ment style is known to vary within the sector, and poor shipboard
guard Agency in the United Kingdom identified various core lead- management and leadership are identified in other research
ership qualities as being necessary for effective safety leadership (Oltedal and Engen, 2010; Knudsen, 2003).
(Maritime and Coastguard Agency, 2004). Shipboard, it was found The results indicate a positive relationship between compe-
that these qualities were primarily geared towards the captain as a tence and reporting frequency (odds ratio: 1.4949, P > 0.0000).
key leader for safety, however also for lower ranks with leadership Competence is among the top five most commonly measured
responsibilities. However, perceived gaps between desirable lead- themes within safety research (Flin et al. 2000), and refers to the
ership qualities and what is currently being exhibited were also perception of own skills and ability to handle critical and hazard-
identified. With reference to the explosion and sinking of the ous situations, and their ability to perform their work in a safe
chemical tanker ‘‘Bow Mariner” (United States Coast Guard, manner. In our analysis, competence is comprised of two sub fac-
2005), poor leadership skills were on the agenda of the IMO (Inter- ets, formal education and training. Minimum training require-
national Maritime Organization 31 October, 2007). ments are covered by international conventions and regulations
The results also suggest increased reporting frequency for those developed by the IMO, where parts are required to be performed
who have worked with their closest superior/manager for one year on-board. For example, it is required that every crew member par-
or more. This ‘‘leadership-familiarity” variable did not meet the ticipates in at least one abandon ship and one fire drill every
parallel assumptions, and thus it is indicated that the effect is month. Also that these drills, as far as practicable, be conducted
dependent on the category. Apparently the effect of being familiar as if there were an actual emergency (International Convention
with your superior is larger when moving from the category ‘‘never for the Safety of Life at Sea (1974), International Maritime Organi-
or seldom” to ‘‘sometimes” report (P > 0.001, odds ratio = 2.9817), zation 2009). On-board, the captain and ship management are ulti-
and the effect decrease slightly when moving to the higher catego- mately responsible for how such drills, and other on-board training
ries of ‘‘sometimes” to ‘‘often”, and ‘‘often” to ‘‘always”. Leadership arrangements, are carried out. Experience from the field shows
and management do have various facets, and may be seen as being that in this area there are large variations. On some vessels, if
both social and cognitive in nature. Social skills include things such performed, the drills are arranged as mustering, while others are
as team building, consideration of others, conflict resolution etc. arranged for realism. Variation in on-board training efficiency
We would suggest that such social skills are of particular impor- may be a result of various situations such as a lack of time due
tance within this maritime setting. Work at sea may be character- to demand for efficiency. The results indicate a negative relation-
ized as a total institution, as defined by Goffman (1968), where ship between demand for efficiency and reporting practices (odds
both work and leisure time happen at the same time, with few ratio: 0.7581, P > 0.0030). Demand for efficiency, caused by inter
and limited possibilities to interact with the surrounding world. alia commercial pressure, has been one of the most frequent rea-
In such settings, leadership and management style influences work sons for violations of procedures and checklists (Oltedal, 2010),
and social life in a more all-embracing manner, including the inter- as well as an important influencing factor for collisions and
personal relationship among crew members. Although the research groundings (MacRae, 2009).
regarding managements importance for safety within the off-shore However, with regard to competence, we suggest that one sig-
shipping industry is scarce, research from other industries has nificant cause could be that the minimum requirements for leader-
established a relationship between management, leadership and ship and managerial skills stated in the international conventions
safety related matters (Geldart et al., 2010; Wu et al., 2008; and regulations are inadequate (also noted by IMO in (Interna-
Vredenburgh, 2002; Zohar, 1980), thus management does not only tional Maritime Organization 31 October, 2007)). It is therefore
have a direct effect upon safety and reporting practices, but also an up to each maritime educational establishment to decide to what
indirect effect by influencing the other factors in the model, further degree managerial and leadership should be covered, or to each
discussed below. shipping company with regard to the provision of further educa-
The results also indicate that the interpersonal relationship tion. Insufficient managerial and leadership skills could also result
among the crew influences reporting practices (odds ratio: in time pressure and demand for efficiency through inadequate
1.1971, P > 0.0490). With regard to reporting practices, interper- planning and resource management.
sonal relationships relate to, inter alia, the degree of trust and open We also suggest that the same relationships exist with regard to
communication amongst the crewmembers. Reason (2001) regards the factor concerning pro-active work practices (leadership, mana-
a trusting relationship as a key factor in getting individuals to re- gerial skills, time available and competence interrelationship).
Work practices refers to the degree to which the crew perform pro- and when efficiency is given importance decreases reporting fre-
active activities like safe job analyses and safety analysis before quency. The three variables which did not meet the parallel regres-
risk activities, and to which degree they have the opportunity to sions assumption were Feedback, Vessel and Management. This
prioritize safety in their daily work. Such activities increase the means that the effect of these variables is dependent on the cate-
chances of revealing potentially dangerous situations, and thus in- gory of the dependant variable. Crews who have been working
creases the amount of reportable safety information (Kjellén, with their closest manager for more than one year tend to report
2000). more often.
The results indicate that pro-active work practices have a posi- However, none of the identified factors should be addressed in
tive relationship with reporting frequency (odds ratio: 1.3319, isolation from each other. As the discussion made clear, they are all
P > 0.0020). important and mutually dependent. It would therefore be of fur-
Feedback on reported events is held to stimulate organizational ther interest to explore the internal relationships between the
learning, and thus better premises for safety improvements, by, in- identified dimensions of safety culture, for example with structural
ter alia, sharing of experience of near-misses and incidents, as equation modelling and/or path-analysis.
guidelines for corrective actions as well as being a motivator for in- The data are representative of vessels flying a white and grey flag
creased reporting (Reason, 2001). A positive relationship between only, as those registered under a black listed flag did not want to
reporting frequency and feedback is also shown in our data (odds participate. As participation was voluntary on the behalf of the com-
ratio: 1.8090, P > 0.0000). However, the relationship did deviate pany, we assume that those participating do, in general, emphasise
from proportionality meaning that the effect of feedback is depen- safety in their operations, and the development of a sound reporting
dent on the category of the dependant variable reporting fre- culture. However, it would be of further interest to address potential
quency. Thus, the effect from feedback is larger when moving differences between flag of registration and safety.
between the higher categories ‘‘sometimes” to ‘‘often”, and ‘‘often” It would also be of interest to further explore the difference
to ‘‘always”, than when mowing between the lower categories with regard to type of vessel. The differences in reporting fre-
‘‘never or seldom” to ‘‘sometimes” Treatment of non-conformance quency could be a result of other variables related to type of vessel,
and development or remedial actions is normally done by shore as for example customer specific.
personnel. To what degree seafaring personnel are involved in
these processes varies within companies. Shipboard feedback is gi- References
ven by the captain and/or shipboard management, who in turn re-
ceive the information from the shore-side of the company. Thus Anderson, P., 2003. Cracking the Code: The Relevance of the ISM Code and its Impact
on Shipping Practices. Nautical Institute, London.
the quality of feedback given is not only dependent upon the ship
Campell, D.R., Fiske, D.W., 1959. Convergent and discriminant validation by
management, but also the shore-side’s orientation towards safety. multitrait-multimethod matrix. Psychological Bulletin 56 (2), 81–105.
This relationship between shipboard reporting practices and Ellis, N., Bloor, Michael, Sampson, Helen, 2010. Patterns of seafarer injuries.
shore-side safety orientation is also indicated by the results (odds Maritime Policy & Management 37 (2), 121–128.
Espin, S., Regehr, G., Levinson, W., Baker, G.R., Biancucci, C., Lingard, L., 2007. Factors
ratio: 0.6094, P > 0.0000). When the shore-side’s safety effort is influencing perioperative nurses’ error reporting preferences. AORN 85 (3),
perceived as a facade and person orientated, it is reflected in lower 527–528 (pp. 530–532, 534–536, 539–543).
reporting frequency. Field, A., 2005. Discovering statistics using SPSS: (and sex, drugs and rock ‘n ‘roll),
second ed. Sage, London.
Finally, it is indicated that the reporting frequency is lower on Flin, R., Mearns, K., O’Connor, P., Bryden, R., 2000. Measuring safety climate:
dry cargo vessels than on tanker vessels. The relationship deviates identifying the common features. Safety Science 34 (1–3), 177–192.
from proportionality, meaning that the effect of type of vessel is Geldart, S., Smith, C.A., Shannon, H.S., Lohfeld, L., 2010. Organizational practices and
workplace health and safety: a cross-sectional study in manufacturing
larger when moving between the higher categories ‘‘often” to companies. Safety Science 48 (5), 562–569.
‘‘always”, than when mowing between the lower categories ‘‘never Goffman, E., 1968. Asylums: essays on the social situation of mental patients and
or seldom” to ‘‘sometimes”. Differences between the liquid and dry other inmates. Penguin, Harmondsworth.
Hair, J.F., 1998. Multivariate data analysis, fifth ed. springer, berlin.
cargo sector are further discussed in Oltedal and Engen (2009). Hansen, H.L., Pedersen, G., 1996. Influence of occupational accidents and deaths
related to lifestyle on mortality among merchant seafarers. International
5. Conclusion, limitations and suggestions for future research Journal of Epidemiology 25 (6), 1237–1243.
Harzing, A.W., 2005. Does the use of english-language questionnaires in cross-
national research obscure national differences? International Journal of Cross
This article has studied the factors which influence the fre- Cultural Management 5 (2), 213–224.
quency of reporting of experience data such as data on accidents/ Håvold, J.I., 2005. Safety-culture in a Norwegian shipping company. Journal of
incidents. The previous research outlined in the paper has shown Safety Research 36 (5), 441–458.
Håvold, J.I., Nesset, E., 2009. From safety culture to safety orientation: validation
that under-reporting is a significant problem within the merchant and simplification of a safety orientation scale using a sample of seafarers
shipping industry. Under-reporting undermines the foundations working for Norwegian ship owners. Safety Science 47 (3), 305–326.
on which any safety management system is constructed. If acci- Hollnagel, E. 2009. The ETTO Principle: Efficiency-Thoroughness Trade-Off Why
Things That Go Right Sometimes Go Wrong, Ashgate.
dents/incidents are not reported then past mistakes cannot be International Convention for the Safety of Life at Sea (1974) & International
learned from and the probability of future accidents/incidents can- Maritime Organization 2009, SOLAS on CD-ROM + demo, IMO Publ., London.
not be reduced. In particular, the analysis presented in this paper is International Maritime Organization 31 October 2007, Comprehensive Review of
the STCW Convention and the STCW Code – Communication and leadership
important since it not only identifies the significance of potentially skills (STW/39/3), International Maritime Organization, London.
influential factors, but also quantifies the relative strength of these International Maritime Organization, 2007a. Formal Safety Assessment:
factors. This allows a better targeting of budgetary resources to im- Consilidatated text of the Guidelines for Formal Safety Assessment (FSA) for
use in the IMP rule-making process (MSC/Circ.1023-MEPC/Corc.392).
prove safety. International Maritime Organization, London.
In particular, the objective of this article was to assess the rela- International Maritime Organization, 2007b. Role of the human element: Near miss
tionship between reporting practices and dimensions of safety cul- information. International Maritime Organization, London.
International Maritime Organization, 2002. International safety management code:
ture, management and vessel in the Norwegian controlled shipping
ISM code: and revised guidelines on implementation of the ISM code by
industry. The results indicated that high competence, a good and administrations, second ed. International Maritime Organization, London.
open interpersonal relationship among the crew, a safety oriented Kjellén, U., 2000. Prevention of accidents through experience feedback. Taylor &
management, execution of pro-active work practices and feedback Francis, London.
Knudsen, F., 2003. If you are a good leader I am a good follower.
upon reported events all increase reporting frequency. The two MacRae, C., 2009. Human factors at sea: common patterns of error in groundings
dimensions reflecting when shore orientation downgrade safety and collision. Maritime Policy & Management 36 (1), 21–38.
MARISEC, 2008. Shipping Industry Flag State Performance Table. Maritime Psarros, G., Skjong, R., Eide, M.S., 2010. Under-reporting of maritime accidents.
International Secretariat Services Ltd (Marisec), London. Accident Analysis & Prevention 42 (2), 619–625.
Maritime and Coastguard Agency, 2004. Driving Safety Culture: Identification of Reason, J., 2001. Managing the risks of organizational accidents. Ashgate, Aldershot.
Leadership Qualities for Effective Safety Management. Roberts, S.E., Marlow, P.B., 2005. Traumatic work related mortality among seafarers
McCrae, R.R., 2001. Cross-cultural research on the five-factor model of personality. emplyed in British merchant shipping 1976–2002. Occupational and
In: Lonner, W.J., Dinnel, D.L., Hayes, S.A., Satler, D.N., (Eds.), Online Readings in Environmental Medicine 62, 172–180.
Psychology and Culture (Unit 6, Chapter 1), Center for Cross-Cultural Research, Rosenman, K.D., Kalush, A., Reilly, M.J., Gardiner, J.C., Reeves, M., Luo, Z., 2006. How
Western Washington. <http://www.wwu.edu/~culture>. much work-related injury and illness is missed by the current national
Oltedal, H.A., Engen, O.A., 2009. Local management and its impact on safety culture surveillance system? [Article]. Journal of Occupational & Environmental
and safety within Norwegian shipping. In: Artorel, M., Guedes, C., Barnett Medicine 48 (4), 357–365.
Safety, J. (Eds.), Reliability and Risk Analysis Theory, Methods and Applications, SINTEF, 2003. Sikkerhetskulturer i transport: En kunnskapsoversikt. (STF
2008th ed. European Safety and Reliability Association. 22A03300).
Oltedal, H., 2010. The use of safety management systems within the Norwegian Spector, P.E., 2006. Method variance in organizational research: truth or urban
tanker industry – do they really improve safety? In: Bris, R., Guedes Soares, C., legend? Organizational Research Methods 9 (2), 221.
Martorell, S. (Eds.), Reliability Risk and Safety: Theory and Applications. Taylor Studio Apertura. 2004. Sikkerhetskulturer i transportsektoren. Metoder for
& Francis Group. kartlegging av sikkerhetskultur: Evaluering av noen eksisterende verktøy.
Oltedal, H.A., Engen, O.A., 2010. Tanker versus dry cargo – The use of safety United States Coast Guard, 2005. Investigation into the explotion and sinking of the
management systems within Norwegian dry cargo shipping. In: Ale, B.J.M., chemical tanker Bow Mariner in the Atlantic Ocean in February 28 2004 with
Papazoglou, J.A., Zio, E. (Eds.), Reliability, Risk and Safety. Taylor & Francis loss of life and pollution. United States Coast Guard, Washingthon.
Group, London. van der Schaaf, T., Kanse, L., 2004. Biases in incident reporting databases: an
Pett, M.A., Lackey, N.R., Sullivan, J.J., 2003. Making sense of factor analysis: the use empirical study in the chemical process industry. Safety Science 42 (1), 57–67.
of factor analysis for instrument development in health care research. Sage, Vredenburgh, A.G., 2002. Organizational safety: Which management practices are
Thousand Oaks, Calif. most effective in reducing employee injury rates? Journal of Safety Research 33
Podsakoff, P.M., MacKenzie, S.B., Lee, J., Podsakoff, N.P., 2003. Common method (2), 259–276.
biases in behavioral research: a critical review of the literature and Wu, T., Chen, C., Li, C., 2008. A correlation among safety leadership, safety climate
recommended remedies. Journal of Applied Psychology 88 (5), 879. and safety performance. Journal of Loss Prevention in the Process Industries 21
Probst, T.M., Brubaker, T.L., Barsotti, A., 2008. Organizational injury rate (3), 307–318.
underreporting: the moderating effect of organizational safety climate. Zohar, D., 1980. Safety climate in industrial-organizations – theoretical and applied
Journal of Applied Psychology 93 (5), 1147–1154. implications. Journal of Applied Psychology 65 (1), 96–102.
Article 6
Oltedal, H. A., & Engen, O. A. (2010). Safety Management in

Shipping—Making Sense of limited Success. Safety Science Monitor,
submitted.

Safety Management in Shipping: Making sense of limited
success
H.A. OLTEDAL*, O.A. ENGENh
*Faculty of Technology, Business and Maritime Education, Stord/Haugesund University

College (HSH), Haugesund, Norway.
h
SEROS-Centre for Risk Management and Societal Safety, University of Stavanger,
Stavanger, Norway.
Abstract
This article addresses the gaps between regulative safety ambitions and operational
results and practice within the maritime sector. In particular, the article focuses on the gaps
within safety ambitions in the form of, for example, rules and procedures and practise,
exploring possible pitfalls when relying on safety from a system perspective without paying
attention to its human interrelationship. This examination applies the Practical Drift Model
(PDM) conceptual framework developed by Scott A. Snook (2000). PDM combines the
Normal Accident Theory (NAT) and High Reliability Organization Theory (HRO) to explain
why seemingly well-controlled and sophisticated organisational systems may in reality
increase their risk potential. In fact, existing scientific safety literature has not offered much
space in which to make sense of the gaps between theory and practice. Hence, we more
closely examine certain examples of such gaps from our own data collected from the
Norwegian controlled shipping industry. Applying a multi-method approach combining
surveys and case studies, including field studies and interviews, we conclude that maritime
accidents to a large degree derive from a mismatch between the local demands of the situation
and global rules designed by planners.
1 Introduction
Shipping is a global industry as most vessels are affected by legislation in many countries
and abide international rules; in addition, owners can choose flag states and labour markets in
which to recruit crews. Such interdependencies also make shipping a highly complex
business. In the safety area, many stakeholders (e.g., crew, shipping companies, unions,
industrial bodies, national and international regulators) constantly act and react to internal and
external changes. Although safety is to a large extent regulated and procedures and guidelines
for best practices are in place, severe accidents and incidents still frequently happen. 1
International safety regulations stem from the International Maritime Organization (IMO)
in the form of conventions, protocols, and resolutions. At other levels, safety is addressed by
regional, flag state, and port state regulations (Kuo, 2007), which are transformed into
shipping companies’ safety management systems in the form of procedures and standards.
Companies’ safety activities are regulated by the International Safety Management (ISM)
1
Two of particular note are the Norwegian ship Langeland, which sank in a storm off the coast of Sweden (all
six crewmen were presumed dead0, and Full City, which went aground and polluted the southern Norwegian
coastline. Both accidents happened during the same storm on the night of 31 July 2009.
1
Code, which was fully implemented on 1 July 2002 (719 International Maritime Organization,
2010). Although procedures and standards are in place to control risk and seafarers’
behaviour, accidents involving human error also seem to be on the rise (Soma, 2008).
Statistics from Lloyds’ Fairplay (2010), shown in Figure 1, indicate the navigational accident
frequency (collisions, contacts, and wrecked/stranded vessels) from 1993 to 2009 in relation
to the world’s fleet size.
Figure 1: Navigational accident frequency 1993-2009 (Source: Lloyds' Fairplay, 2010)
Navigationalaccidentfrequency
0,0200
0,0150
0,0100
0,0050
0,0000
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Navigationalaccidentfrequency
Fleet size in number of: Crude oil tankers over 100,000 dwt, Chemical tankers over 10,000 dwt,
Containers over 20,000 dwt, RoRo cargo over 10,000 dwt, Bulk over 50,000 dwt.
As Figure 1 indicates, the frequency of serious accidents—especially navigational accidents—

has been increasing since 2002, despite the introduction of the ISM Code.
The European Maritime Safety Agency (EMSA) has suggested links among accident
numbers, loss of life, and economic activity, especially when accidents occur because ships
and seafarers are being worked harder (European Maritime Safety Agency, 2010).
Surprisingly, the trend indicates that the number of accidents is higher during economic
upturns. According to EMSA (2010), during economic downturns, supply overcapacity, high
levels of ship scrapping, lower operating speeds, and generally less pressure to meet tight
deadlines are seen to be the main reasons for reductions in accidents; the opposite is the case
during economic upturns. In addition, our own research demonstrates how well-intended
pressures to induce safer practices can at times be either ineffective or directly counter-
productive. Furthermore, EMSA (2010) noted that the great majority of accidents include a
human error component; seafarers often make mistakes under difficult circumstances, such as
bad weather, fatigue, task overload, and training shortcomings. Thus, we would stress the
importance of including and paying attention to this human interrelationship when working
with safety.
1.1 Objective of the Article

This article addresses the gaps between safety ambitions in the form of establishing, for
example, rules and procedures and practise, where we also explore possible pitfalls when
relying on safety through a system perspective without paying attention to its human
interrelationship, especially those in the sharp end—namely, the seafarers themselves. In
order to accomplish such objectives, we will consider how the Normal Accident Theory
(NAT) and High Reliability Organization Theory (HRO) explain and make sense of safety,
2
risks, and accidents. We will also show how these theories do not sufficiently explain why
seemingly well-controlled and sophisticated organisational systems in certain contexts and
situations tend to increase their risk potential.
In order to make sense of the gaps between safety ambitions and the practical outcome, we
will examine data collected from the Norwegian controlled shipping industry. By
reconsidering the applicability of HRO and NAT, we apply the Practical Drift Model (PDM)
conceptual framework developed by Scott A. Snook (2000). PDM aims to combine NAT and
HRO and presumably better explain why seemingly well-regulated organisations develop
traits that may evolve into significant accidents and disasters. As such, the PDM model
enables us to look for possible new explanations of the success and failure of safety strategies.
The article is organised as follows. In the next section, we briefly present our theoretical
framework by introducing NAT and HRO. Thereafter, we introduce PDM and explain the
relevance of the model. The following section presents relevant data from the survey of
Norwegian-controlled shipping industry as well as the chosen methodological framework.
The findings and discussion follow the logic of the PDM, which organises the presentation of
the qualitative data. Finally, we summarise existing theoretical explanations and suggestions,
supplementing them with some of our own new findings to formulate hypotheses for future
theoretical research.
2 Theoretical Approaches to Safety Management
According to Scott and Davies (2007), NAT belongs to the open system paradigm and
thus includes organisational members’ intentions, power, loyalty, and identity along with the
influence of the environmental context outside the organisation itself. Systems (organisations
and their surroundings) are characterised as error avoiding, error neutral, and error inducing,
based upon a number of organisational and contextual features. NAT brings forward the
notions of interactions and couplings, whereupon systems are placed in a matrix depending
upon the characteristics of their interactions and couplings—namely, linear or complex
interactions and tight or loose couplings. Accidents are explained by the complexity in the
interactions and tight couplings. Therefore, error-inducing systems should strive to reduce the
complexity and/or loosen up the system couplings.
In order to handle the risk inherent in such systems, some managerial advice has been
given. According to Perrow (1999), a system with linear interactions, loose couplings, and
few complex interactions may handle both centralised and decentralised (component) failure.
A linear and tightly coupled system is expected to be best managed in a centralised manner, as
(component) interaction is visible and expected. Such a linear system is also fit for
standardised measures and procedures, as there is little variation in the production system.
Meanwhile, in a complex and loosely coupled system, decentralisation is desirable in order to
cope with unexpected complex interactions that enable people to develop substitutions and
alternative paths. In such a system, standardised measures are less fit, as the complexity
brings along more variation in operations. Risk in these three mentioned combinations of
systems is supposed to be manageable; however, Perrow (1999) pointed out a problem with
managing systems that are both tightly coupled and complex. In such systems, centralisation
is desirable for coping with tight couplings while decentralisation is desired for coping with
the unplanned interactions of failure. According to Perrow (1999), an organisation cannot be
simultaneously centralised and decentralised. The essence of NAT is that, no matter how hard
we try, these kinds of systems are ultimately bound to fail.
Although NAT is built strongly upon a system perspective, the premise is not traditional
system theory (see Bertalanffy, 1971, for traditional definitions). In NAT’s understanding of
3
system, the aim is to understand specific events, such as the failure of an operator to close a
valve and examine it in the context of the failure (Perrow, 2004). This specific context could
be the operators’ mental models, which again are the result of training, company ideology,
experience, intentions, and power figurations. Each element—industry structure,
organisational structure, group structure, and the like—should again be examined in its own
context.
Although NAT apprehends accidents as inevitable in interactively complex and tightly
coupled systems, HRO applies a contrary strategy in order to develop nearly error-free
organisations (Weick and Sutcliffe, 2007). HRO advocates pay more attention to internal
organisational processes than system characteristics, yet they also assume that the research
objects all are interactively complex and tightly coupled. Recognising that the world they face
is complex, unstable, unknowable, and unpredictable, the HRO perspective is reluctant to
accept the simplification and standardisation inherent in traditional risk assessment. This view
is shared by Perrow (1999). Both NAT and HRO are sceptical towards those who rely highly
on risk assessment when developing preventive safety measures. In HRO, more attention is
given to the real work going on in frontline operations. By empowering the persons doing the
actual work, operators are enabled to solve the situations themselves, based upon their own
experience and knowledge. They operate using pre-planned descriptions, but it is accepted
that real situations will include deviations from standardised measures as procedures.
One of the HRO’s key points is “mindfulness”. We understand mindfulness as a
combination of alertness, sensibility, flexibility, and adaptability. The perspective argues that
unexpected events should be handled by creating a mindful infrastructure by following five
main principles: (1) continually track small failures, (2) resist oversimplification, (3) be
sensitive to operations, (4) maintain capability for resilience, and (5) monitor shifting
locations of expertise. The violation of these principles is regarded as a step backwards
towards the more traditional approach, where simple diagnoses are accepted and frontline
expertise is overridden by faith in risk analysis and developed measures detached from the
operations. The HRO view implies that operators gain more responsibility, resulting in the
need for other parts of the organisation to give them the possibility to act. In other words,
control is taken from the upper levels of the organisation in favour for lower levels.
However, the definitions of both complexity and interaction are relative. A new system
may be perceived as more complex than a system that has existed for some time. Interactions
may be unknown and not comprehensible in the early dawn of new technologies; through
experience, the interactions become more understandable and visible. A system may be
perceived as complex for a new operator, but as simple for an experienced one. A system may
also be perceived as simple in certain situations, but turn complex if changes occur in its
context. Therefore, we would like to incorporate a new dimension, acquaintance, representing
the knowledge and experience one has with the system and context. Acquaintance
complements the concept of mindfulness, particularly when incorporating time as an
important factor in the analysis.
PDM (Snook, 2000) combines HRO and NAT in two ways. First, it emphasises how
different degrees of mindfulness depend on different situations and contexts and how
organisational systems during their lifetime develop both tight and loose couplings. The main
objective of the model is to capture both contextual and temporal factors when explaining
why incidents and accidents occur. More precisely, PDM consists of three dimensions: (1)
situational couplings, (2) different logic of actions, and (3) time. The first dimension,
situational coupling, refers to Karl Weick’s statement that it is not the existence or non-
existence of loose couplings that is of a crucial determinant of organisational functioning over
time, but rather the patterning of loose and tight couplings (Weick, 1976). This statement
underlines how organisations shift from tight to loose couplings and back again as the various
4
sub-units within the taskforce alternate between low and high degrees of interdependence.
The second dimension, logic of action, refers to what “new institutionalism” defines as
contextually dependent mindsets or frames that influence peoples’ behaviour (DiMaggio,
1994). Here, mindsets and frames refer to norms, scripts, routines, or habits possessed by the
actors within an organisational setting. Snook’s (2000) premise in PDM is that organisational
members shift back and forth between rule- and task-based logics of action depending on
shifting contextual factors. The third dimension, time, refers to the lifetime of an organisation,
which is characterised by shifting periods of tight and loose couplings and different kinds of
logic of actions. The dynamics of PDM are driven by the twin muscles of behaviourally
anchored logic of action and stochastically determined situational coupling (Snook, 2000).
This duality implies that periods of stabilisation with ruled-based logic of action will be
replaced by periods of instability characterised by a more task-based logic of action. The latter
invites a higher risk potential and greater possibilities for accidents.
1)Design 4)Failed
Tightlycoupled Tightlycoupled
Ruledbasedlogic Taskbasedlogicof
ofaction action
2)Engineered 3)Applied
Looselycoupled Loosely coupled
Rulebasedlogic Taskbasedlogic of
ofaction action
Unstable Stable
Figure 2: The Practical Drift Model (Based on Snook, 2000)
The different boxes in Figure 2 refer to various phases of the organisational development
and different characteristics of situational couplings and actions of logic. Box 1 in the model
is denoted as “design” and refers to the stage of an organisational lifecycle in which the
governance structure is top-down oriented. The organisational managers or designers have put
significant efforts into developing extensive routines and procedures in order to make the
organisation robust and resilient against attacks and unforeseen events. This implies a
technocratic character of the organisation, which also assumes tight couplings and a rule-
based action of logic.
Box 2 is denoted as “engineered” and refers to an operational situation that is more
loosely coupled. During operations, the rules of logic, which are designed for tight couplings,
do not always match the situations. When organisations based on strict, high reliability
principles experience that the unforeseen attacks and events do not occur as expected, the
attention focused on the routines and procedures become part of everyday life and practice
5
and, consequently, more relaxed. The logic of action based on the rules of “continuously red
alert” tends to normalise, which results in a tendency for de-coupling organisational systems.
The real world does not act in accordance with organisational design; as actors become
aware of this, they will be able to break the strict rules without any fear of sanctions or
punishment. On a local basis, breaking the strict rules may actually get the job done more
quickly and efficiently. The accumulated experiences from the real world of everyday practice
will thus push the organisation away from the originally designed strict rules and create
unstable situations. Such instability generates further pressure for change, as evident in the
shift towards box 3 in the model, which is denoted as “applied”.
The shift from “engineered” to “applied” involves a change of mentality among the
actors and a transition from tight to loosely coupled organisational systems. This process is
what Snook (2000) calls “practical drift”. During such a process of de-coupling, the
organisation becomes increasingly free from the global rationality that characterises “design”
and develops sub-units with their own defined rationality and internal governance structures.
Such sub-units are autonomous entities whose logic of action is based on experiences and
tacit knowledge. Accordingly, the actors drift further away from a rule-based to a more task-
based system. The locally independent sub-units solve everyday problems in an applied task-
based manner. The problem-solving processes become pragmatic responses to what the single
units (e.g., work groups or vessels) justify as their own understanding of “the rules”.
Box 4 “fails” indicates a situation in which the system suddenly and stochastically
becomes tightly coupled. The circumstances may include an incident in which single units
confront one another (e.g., two ships on a collision course). In such situations, the actors in
the individual units are forced to act on the assumption that all others act in accordance with
the original rules and initially designed procedures. This may create extreme instability and,
in the worst case, a catastrophe. The actors are actually trapped in a game where trusting their
own logic of actions is the only solution, yet they must simultaneously base their decisions on
the assumption that others are following the general rules. From “failed”, the organisation
once again enters the phase of “design”. In “redesign”, the organisation tries to control for the
experienced unwanted outcome, often in the form of even tighter designed control criteria.
Thus, the model indicates a recovery phase in which actors learn from mistakes and re-
introduce a top-down-based governance structure with tight couplings and logic of action
based on generally accepted rules. Real-life accidents often induce such processes. Therefore,
the model indicates certain determinism—namely, “peaceful” organisations without repeated
redesign processes inevitably drift towards ruin. On the one hand, this is the old story of “the
unrocked boat” (Reason, 2001). On the other hand, the model indicates how structural
constraints or enforcements also become influenced by external contexts and either fortify the
lack of mindfulness or increase mindfulness based on locally based logic of actions. Both
ways of practical drift may be equally dangerous.
The practical drift model thus emphasises the organisational contradiction between
decentralisation and control, where decentralised decision making is needed to permit flexible
responses to surprises while increasing the degree of complexity requires centralisation and
discipline. Bridging the gap between safety goals and actual performance aims to reduce this
kind of contradiction and create organisational designs that may simultaneously inhibit the
logic of hierarchical control and decentralised decision making.
3 Methodological Considerations
This article adopts a multi-method approach combining surveys and case studies,
including field studies and interviews. Safety management and its interrelationship with
6
human behaviour are, in accordance with current accident and safety management theories
(e.g., Cooper, 2000; Reason, 2001; Turner and Pidgeon, 1997; Weick and Sutcliffe, 2007),
understood as a multi-layer construction comprising organisational members’ fundamental
safety-related assumptions and values (what is important), beliefs (how things work), and
patterns of behavioural norms (how things are done). The fundamental assumptions that
individual group members have towards safety management and the explanation of actual
behaviours at operational level are difficult to understand and explain without interactive
probing and qualitative methods. Fundamental assumptions are assumed to be reflected in
behavioural norms and perceptions, which are far more accessible through quantitative
methods. Once integrated, these factors are commonly referred to as safety culture (Reason
and James, 2001; Reichers and Arnon, 1990)
The current study was carried out in two phases. The first phase was a self-completion
questionnaire survey of crewmen working on Norwegian controlled liquid tankers and dry
cargo vessels. A random sample was drawn from the Norwegian Shipowners’ Association’s
list of membership. The second phase involved a sub-sample and case studies selected from
the survey sample group. This second phase employed semi-structured interviews with
crewmen and shore personnel along with participatory observations. Thus, the quantitative
results are representative of the selected population and are applicable for generalisation.
Although the narratives presented in the articles are not applicable for generalisation, they
provide examples and understanding of how the gap between safety goals and actual
performance is created through social processes. In addition, some characteristics of the
interaction between the various organisational members and the existing safety control
systems are brought forward with the qualitative results.
Furthermore, a multi-method approach is applied so that the interrelationships between
the elements may be examined in order to establish antecedents, performance, and outcomes
(Bergman, 2008; Brannen, 2005; Cooper, 2000). More precisely, the survey aims to indicate
seafarers’ perceptions of and attitudes towards safety management and risk. The qualitative
results aim to provide a more thorough understanding of the overall situation, underlying
processes, and seafarers’ own situational experiences.
A theoretical and practical framework for the assessment of safety culture open to a
multi-method approach is provided by Cooper (2000). This frameworks includes (1)
subjective internal psychological factors (i.e., attitude and perceptions); (2) observable
ongoing safety-related behaviour; and (3) organisational situational features. According to
Cooper (2000), these elements reflect those accident-causation relationships previously found
by a number of researchers, such as John Adams (1995), Herbert William Heinrich (1980)
and James Reason (2001). The investigation of several serious shipping accidents, such as the
Herald Free Enterprise (Department of Transport, 1987), the Exxon Valdez (National
Transportation Safety Board, 1990), and the Scandinavian Star (Justis-og politidepartementet,
1991), is also congruent with their findings.
3.1 Questionnaire Development

The survey instrument was developed and validated by Studio Apertura in collaboration
with the Norwegian DNV and the SINTEF research institution (Studio Apertura, 2004). The
main part of the questionnaire comprises 10 sections representing specific safety-related
dimensions: (1) top management’s safety priorities, (2) local management, (3) procedures and
guidelines, (4) interaction, (5) work situation, (6) competence, (7) responsibility and
sanctions, (8) working environment, (9) learning from incidents, and (10) description of the
organisation. This article analyses the results from the third section on procedures and
guidelines. All items were measured on five-point Likert scales ranging from “strongly
7
disagree” to “strongly agree”. The respondents were also given the possibility to comment in
a specific situation.
3.2 Survey Samples and Administration

In total, 1,574 questionnaires were distributed to 83 tankers and bulk carriers. Each vessel
received a package with individual questionnaires and a sealable return envelope. On each
vessel, the safety delegate received instructions related to administration, purpose, and
anonymity. Vessels not returning any questionnaires were reminded up to four times. The
survey was administrated during the spring/summer of 2006.
A total of 76 vessels from 29 companies returned 1,262 forms, providing an individual
response rate of 80.2%, a vessel response rate of 91.5%, and a company response rate of
93.5%. In addition, 297 respondents provided written comments. For further information
regarding survey sample and administration, see Oltedal and Wadsworth (2010).
3.3 Survey Demographics

Questionnaires were returned from 40 liquid bulk carriers (liquid tanker) and 36 dry bulk
carriers (dry cargo); 63% of the respondents were employed on a liquid tanker and 37% on a
dry cargo vessel. The sample was male dominated (92.5% of the respondents), and 22
nationalities were represented. The majority of respondents were from the Philippines
(65.5%), followed by Norway (9.2%), Poland (8.1%), and Russia (5.5%). Just over 56% of
the respondents were under the age of 40. Only a few (11.5%) of the respondents (mostly
Norwegian nationals) had fixed employment within the shipping company. The remaining
88,5% were contract employees—43.6% with 9-month contracts, 21,5% with 6-month
contracts, and 4,5% with 3-month contracts. The remaining 18.9% had a different length of
contract, none of which was more than 1 year. Finally, when signing on a new vessel, 15.2%
of the respondents stated seldom or never signing on to the same vessel, 39.9% stated staying
on the same vessel sometimes, and the remaining 44.8% often or always stay on the same
vessel. On seven of the vessels, all respondents who returned the questionnaire stated that
they often or always sailed on the same vessel.
3.4 Case Studies

Four shipping companies—two tankers and two dry cargo companies—were approached
about conducting case studies from 2007 to 2009. In order to ensure the companies’
anonymity, company-specific information was retained. In all four companies, the HSQ
manager and Crewing Manager were interviewed. In one company, the safety management
system (SMS) data system was examined. In the other companies, all available statistics,
experience feedback from reported cases, and safety bulletins were examined. Information
was also retrieved through participatory observations at sea from two field studies,
participation in captains’ conferences where safety was an issue, and both formal and
informal interviews with seafaring captains and shore-based personnel involved in safety-
related matters.
The qualitative information aims to give a more thorough understanding of the processes
underlying the presented statistics. The narratives serve two important complementary
purposes. First, they contextualise the data in a broader and deeper sense than the quantitative
studies are able to offer. Second, they personalise the relationships between regulatory
regimes and their subordinates, illustrating what kind of meaning the agents put into their
action when deciding to follow or not to follow the prevailing regulatory rules. This article
emphasises seafarers’ view.
8
3.5 Statistical Analysis and Presentation of Results
All statistical analyses were performed using PASW (by SPSS) statistics version 18.0.
Descriptive statistics for each item, including the percentage of frequency distribution, mean,
and standard deviation, are presented. Principal Component Analysis (PCA) with Varimax
rotation and Pairwise deletion was carried out in order to examine survey items’
interrelationships and their common underlying dimensions. The extracted rotated component
matrix and loadings were presented. The loading represents the correlation between the
variable and the extracted factor(s), with estimates ranging from 0 to 1,00. Items that load
strongest on a given component are considered the most like the underlying latent dimension
(Hair, 1998; Pett, Lackey and Sullivan, 2003).
4 Rationale of Procedures in Maritime Safety Management and Research Findings
This section presents data related to operational procedures. The use of procedures in
error management involves two components: error reduction and error containment (Reason,
2001). According to Reason (2001), some problems are associated with the existing form of
error management, including the fact that error management is generally not informed by
current knowledge related to error and accident causation. Moreover, error management tends
to focus on personal and active failures rather than latent conditions and the situational
contributions to making errors. With reference to this article, latent organisational and
situational conditions are defined as (1) incompatible goals that include (a) group goal
conflicts when the informal norms of a work group are incompatible with the safety goals of
the organisation and (b) conflicts at the organisational level in which there is incompatibility
between safety and productivity goals; (2) procedures in relation to their quality, accuracy,
relevance, availability, and workability; and (3) training and problems that include the failure
to understand training requirements, the downgrading of training relative to operations, poor
task analysis, and the inadequate definition of competence requirements (Reason. 2001). In
the following discussion, our results related to these areas—namely, incompatible goals,
procedures, and training—will be further addressed in relation to the industries’ approach to
safety management.
In order to overcome human error as a cause in maritime accidents, the International
Safety Management (ISM) code (Resolution A.741 (18)) was adapted by the International
Maritime Organization (IMO) in 1994, with part of the objectives ensuring safe practices on
board ships as well as the establishment of safeguards against all identified risks (Anderson,
2003). The code was initiated after the Herald of Free Enterprise accident, when underlying
contributory causes to the accident were found to be poor safety management and failure on
the part of the shore side to give the proper and clear directions (Department of Transport,
1987). By introducing the code, operational safety is promoted through a formal SMS. A
SMS contains several sub-systems. First, a system for reporting and collecting experience
data from the vessel itself is required. This is followed by a system of data processing—
namely, summarisation and analysis in order to reveal causal factors and perform trend
analysis, which forms the basis of the development of safety measures and subsequently often
on forms related to areas such as standardised procedures and checklists (Oltedal, 2010).
Standardised procedures and checklists are also particularly aimed at controlling human error,
which is associated with the vast majority of accidents. An estimated 75-96% of marine
casualties are caused—at least in part—by some form of human error (Rothblum, 2000).
Although systems for safety management and development of measurements are in place,
accidents caused by human error still occur, procedures are violated, and best practices are not
followed. According to Standard P&I Club, aspects of the ISM/SMS system are not
9
effectively implemented (e.g., no coherent system to ensure that lessons are learnt from
mistakes), which impacts the effectiveness of safety measures (Standard P&I Club, 2010).
When it comes to the ineffectiveness of ISM/SMS systems, one particular issue will be put
forward herein: Although various systems for safety management do exist, safety
management per se is not only a system property. It is not sufficient to have a system for the
collection of safety-related data. The system’s efficiency is determined by its human
interrelationships (crew, shore personnel, analysts, and others). Thus, we should stress the
importance of including and paying attention to this human interrelationship when working
with safety.
Survey questions related to the perception of procedures are presented in Table 1, with
descriptive statistics included in percentages, together with mean and standard deviation.
Scale information and coding are 1=strongly disagree, 2=disagree, 3=not sure, 4= agree, and
5=strongly agree.
Table 1.
Descriptive Statistics, Procedures, and Checklists
Survey item description 1 2 3 4 5 Mean Std
Due to the company’s demand for 14.9 41.4 12.8 26.6 4.4 2.64 1.151
efficiency, we sometimes have to violate
procedures
Due to the captain’s demand for efficiency, 18.4 42.1 14.3 21.9 3.2 2.49 1.117
we sometimes have to violate procedures
I have received good training in the 0.9 2.6 6.1 58.9 31.5 4.18 0.728
company’s procedures
I feel that it is difficult to know which 14.4 47.0 17.3 19.0 2.3 2.48 1.028
procedures are applicable
The procedures are helpful in my work 0.6 1.4 4.8 55.6 37.6 4.28 0.678
The procedures are difficult to understand 15.3 57.3 12.6 12.8 2.0 2.29 0.941
or are poorly written
We have the opportunity to influence and 3.4 13.1 15.5 55.3 12.8 3.61 0.978
form the procedures
The descriptive statistics summarised in Table 1 show that 31% of the respondents
violate procedures due to the company’s demand for efficiency while approximately 26% do
so due to the captain’s demand for efficiency. In addition, about 90% of the respondents are
satisfied with the training received in their company’s procedures. Moreover, approximately
93% perceive procedures as helpful in their work, indicating that the respondent’s attitudes
towards procedures per se are good. Thus, when procedures are breached, it is more likely due
to reasons other than poor attitude. About 21% feel that it is difficult to know which
procedures are applicable, which indicates a procedural system not adequate for the situation
and, as a result, weaknesses throughout the overall safety management system. In addition,
approximately 15% find the procedures difficult to understand and poorly written. Finally,
about 68% stated that they have the opportunity to influence and form the procedures.
The interrelationship among the items is presented in Table 2, along with items’
component loadings. Based on the analysis, two components (i.e., factors) are extracted. The
matrix contains factor loadings for each variable (i.e., item) at each factor. Factor loading is
the means of interpreting the role each variable plays in defining each factor and represents
the correlation of each variable and factor. Higher loadings (e.g., > 0.3 or higher) indicate the
variable representative of the factor and underlying dimensions. Within each factor,
interrelated variables are clustered at the same component and are the key to understanding
10
the nature of each particular factor (Hair, 1998). The interrelationships are subject to the
following theoretical discussion in order to address their meaningfulness.
Table 2
Items’ Interrelationships
Loadings
Comp. Comp.
1 2
I1 Due to the captain’s demand for efficiency, we sometimes have to 0.845 -0.027
violate procedures
I2 Due to the company’s demand for efficiency, we sometimes have 0.815 -0.077
to violate procedures
I3 I feel that it is difficult to know which procedures are applicable 0.697 -0.093
I4 The procedures are difficult to understand or are poorly written 0.639 -0.212
I5 The procedures are helpful in my work -0.158 0.793
I6 I have received good training in the company’s procedures -0.186 0.760
I7 We have the opportunity to influence and form the procedures 0.014 0.686
*Comp. = component
The extracted rotated component matrix presented in Table 2 indicates two underlying
dimensions: component 1 and component 2. Component 1 reflects an inadequate procedural
system and approach (I1 through I4). Component 2 indicates an adequate procedural system
and approach (I5 through I7).
The items interrelated in component 1 indicate that shipping companies that exert
commercial pressure (I1 and I2) are also recognised by an overly complex procedural system
(I3) along with procedures that are poorly formulated (I4). High factor loadings on all items
included in component 1 denote that these features are likely to be present simultaneously in
these organisations. In addition, features denoted in items belonging to component 2 (I5
through I7) are likely to be absent as they have low and negative loadings at component 1.
The items interrelated in component 2, indicating companies where the procedural
system is perceived to be useful in daily operations (I5), also relate to good training in these
procedures (I6) and allow the crew to influence and form the procedures (I7). Companies with
these features are also less likely to serve as pressure to ensure efficiency as these items (I1
and I2) have low and negative loadings at component 2. For the same reason, these companies
are denoted by a procedural system perceived to be easier to relate to in daily operations.
Overall, 297 respondents provided written comments on the questionnaire; 62 made
written comments related to procedures and checklists. The comments were categorised into
the following three sub–groupings: (1) procedure quality and usability (n=33), (2) violation of
procedures due to commercial pressure (n=11), and (3) others (n=18). Comments were
provided on a volunteer basis and, thus, not representative of the population. However, they
provide valuable insights into characteristics perceived to be problematic. The results are
summarised in Table 3.
Table 3
Survey Comments Related to Procedures and Checklist
33 respondents Procedures and checklists are not applicable and do not reflect the situation
on board: too detailed, too many, and look like they have been developed
by people with no sea-going experience.
11 respondents Procedures are being breached due to commercial pressure.
18 respondents Other, such as the relevance of training and work-specific situations.
11
Respondents were also asked to list their reasons for not following procedures. Each
respondent could mark up to three of seven pre-specified options or write any other reason, if
not listed. The results are listed in Table 4, with the number of respondents ticking each
reason (N), percentage (N%) and valid percentage (N Valid %). When calculating valid
percentages, missing cases are omitted.
Table 4
Reasons for Violating Procedures Listed by Frequency
Reason for not following procedures N N% N Valid %
1 The work will be done faster 538 42.6 47.2
2 The procedures do not work as intended 488 38.7 42.8
3 There are too many procedures 394 31.2 34.6
4 I feel pressured because I am overloaded with work 365 28.9 32.0
5 It improved the quality of my work 312 24.7 27.4
6 I am not familiar with the applicable procedures 227 18.0 19.9
7 The rest of the crew does it 203 16.1 17.8
8 Others 35 2.8 3.1
Both Table 3 and Table 4 indicate that procedures are violated and checklists are not
followed due to high workloads and commercial pressures. This is supported by the other
survey comments, which are presented in Table 5.
Of the 297 respondents who provided written comments, 108 comments were related to
workload and commercial pressure. The comments were categorised into the following four
sub–groupings: (1) high demand for efficiency and time pressure (n=35), (2) low crewing
level related to work (n=33), (3) violation of rest hours due to high workload (n=30), and (4)
others (n=10). Survey comments related to workload and commercial pressure are presented
in Table 5.
Table 5
Survey Comments Related to Workload and Commercial Pressure
35 respondents High demand for efficiency and time pressure, especially when arriving in
and leaving port.
33 respondents The number of crewmembers is too low compared to work tasks, which are
constantly increasing in quantity—especially administrative
30 respondents Rest hours are not followed, mostly due to low crewing level and high
workload.
10 respondents Not possible to categorise, comments as “sorry, I am tired” and “I fell
asleep at watch”.
The qualitative results from case studies and interviews indicate that the extensive use of
procedures and checklists, including those provided by the shipping company itself as well as
by third parties as charterers and customers. Other shipping-related research also identifies an
increasing volume of regulations, controls, and administrative work as the main factor
negatively affecting on-board safety (Knudsen, 2009). Based on Dreyfus and Dreyfus’ expert
model, Knudsen (2009) also pointed out that rules and procedures are justified for
inexperienced people. However, research also indicates that use of such standardised
measures is more widespread within the tanker sector (Oltedal, 2010) than within bulk and
dry cargo (Oltedal and Engen, 2010). These differences are assumed to originate partly from
each sector’s relationship with their respective customers. Customers of dry cargo shipping
demand fewer requirements with regard to safety management while liquid tankers are to a
larger degree embedded in the oil industry, with a more mature safety management system.
12
However, it is suggested that this relationship between the liquid tanker industry and their
customers will bring about changes based on external demands in contrast to internal needs
for safety reasons. Although some sector differences are evident, the crew customarily deals
with procedures and checklists from own their company, charterer, customers, and oil
installation—of which all are different but at the same time standardised to fit all.
Case studies also indicate that bridge officers regard procedures and checklists as valuable
for safety reasons, but within certain limits. Procedures and checklists are also seen as
problematic as there are too many of them and they are too detailed and too standardised. The
crew experienced less standardisation and determined that the possibility to accommodate
procedures and checklists in accordance with the ship-specific situation would improve safety
more. Problems with completing checklists and following procedures mostly occur during
hectic operations, such as visiting and leaving ports.
5 Discussion and Narratives
The discussion and narratives are presented as they relate to Snook’s (2000) four phases of
(1) design, (2) engineered, (3) applied, and (4) failed, which includes moving from the first
phase (design) through engineered, applied, and into redesign. This section’s discussion
supports the statistical results with interview narratives, illustrating the situation in depth.
Although the statistical data reveal discrepancies between formal procedures and actual
performance, the narratives illuminate why such discrepancies come into existence and how
they are characterised. From the methodological point of view, they complement the survey
and give us a better supplementary understanding of what is actually taking place in the sharp
end of the maritime industry.
5.1 Design
Although progress has been made on international cooperation, the late 1980s and early
1990s experienced a series of maritime disasters (Anderson, 2003; International Maritime
Organization, 2010), such as the Herald of Free Enterprise in 1987, the Exxon Valdez in
1989, the Scandinavian Star in 1990, and the Estonia in 1994. International conventions alone
did not seem to produce the intended levels of safety. Thus, international regulatory bodies
began looking for ways to revise and improve safety regimes. Within the industry, SMS came
to play an important role in achieving and maintaining high levels of safety and reducing
losses resulting from accidents and incidents. Faith in the value of such systems became
widespread, and they were made mandatory through the International Safety Management
(ISM) Code (International Maritime Organization, 2005). The ISM code requires all
companies to develop, implement, and maintain an SMS, which includes the functional
requirements of a safety policy as well as instructions and procedures to ensure the safe
operation of ships in compliance with relevant international conventions and flag state
legislation. Within these requirements, each shipping company is free to find a functional
solution best fitted to its own organisation and operations. However, previous research has
shown that actual work and practice often do not reflect what is stated in the shipping
companies’ overall safety management policy and functional requirements (Oltedal, 2010;
Oltedal and Engen, 2010).
Statistical results related to the design phase (presented in Table 1 and Table 2) indicate
that the procedural system is perceived to be more helpful in daily operations when the crew
has the opportunity to influence the procedures during the design phase. The following
narrative (Narrative 1), given by a Norwegian captain, illustrated a (common) impression of a
13
safety management system designed in such a way that it is detached from daily operations
and does not pay attention to its human interrelationship.
Narrative 1: Norwegian captain about safety management system and procedures.

“(...) my impression is that the aim is to belch forth as much paperwork as possible. And
some of the things, procedures and checklists are not possible to relate to. They are useless
crap developed by someone who has never set foot on a vessel. Just look at this (referring to
pre-port arrival checklist) ‘Electronic positioning instruments checked–positions verified’. Of
course we verify our position. We do that all the time when manoeuvring. We do not need a
checklist to remind us to check where we are. And this ‘Pilot and Port Control have been
given proper ETA (Estimated Time of Arrival) notice’—and what do they mean by proper
ETA. You can be sure that the agent has been at the phone for the last days asking for ETA,
and you can never be 100% sure of your arrival. Things happen, and we are not the only
vessel entering the port. And all the other points are more or less the same (researcher’s
comment: more than 20 items to be checked). But what these checklists really do is draw our
attention away from what we are supposed to do: manoeuvre the vessel. We feel like we are
being treated like children; don’t they think that we know how to do our job? We even have
checklists for the checklists. We spend our days more or less filling in checklists, and for what
purpose? To have someone to blame if anything goes wrong?”
This story was shared with several other informants, who concurred that it depicted
reality on board their vessels. Narrative 1, supported by survey comments presented in Table
3, also illustrates two points brought forward by Snook (2000). First, planners—not
operators—must design a system in which they would never have to work. Second, planners
are writing for future work situations and, as such, have limited information to draw upon. In
this, it is important to keep in mind how procedures are (normally) dealt with within the
framework of a traditional SMS as well as its relationship to risk assessment. An SMS
consists of several sub-systems. First, a system of reporting and collecting experience data
from the vessel itself is required. This is followed by a system of data processing—namely,
the summarisation and analysis in order to reveal causal factors and perform trend analysis,
which forms the basis for the development of safety measures. One critical system
requirement is the reliability and accuracy of input data (i.e., near miss and accidents reports).
As long as the input is reliable, the overall system presupposes the possibility of developing
efficient measures in order to control operational safety (Kjellén, 2000).
Given the arguments related to human rationality, Perrow (1999) has been reluctant to
support the usability of risk analysis. In the absence of absolute rationality, during the SMS
processes, some risks are minimised and other maximised while information is categorised
and simplified in order to facilitate processing. Perrow (1999) suggested that availability
heuristics are used when examining all existing cases of a phenomenon, then basing their
judgment on all this experience as people tend to judge a situation in terms of the most readily
available case—namely, the one most easily remembered.
Other issues hampering the applicability of an SMS include planners’ lack of knowledge
of the operating system, planners taking risk but not facing the consequences of their own
decisions in running operational risk, a lack of education, and a lack of training in
probabilities and statistics (Perrow, 1999). Thus, in these processes, crucial information may
be lost. According to Perrow (1999), heuristics appear to work because in reality our world is
loosely coupled and, thus, has a lot of slack and buffers in it that allow for approximations
rather than complete accuracy. Another drawback in the applied SMS, further contributing to
uncertainty, is the underreporting of experience data. Other research results indicate that
approximately 35% of seafarers working on Norwegian controlled tanker vessels state that
14
they never or only sometimes report minor incidents. About 36% state that they sometimes or
always alter the reports submitted in order to cover up mistakes (Oltedal, 2010). Within the
dry cargo sector, approximately 40% of the respondents state that they never or only
sometimes report minor incidents while about 36% indicate that they sometimes or always
alter the reports submitted (Oltedal and Engen, 2010).
Under such circumstances, where experience data are missing or incorrect, it is even
more challenging to reveal underlying causes and influencing factors. Moreover, when
developing measures for a future situation, there is uncertainty with regard to what that
situation will display. The complexity of influencing factors and patterns of actions makes it
difficult to categorise events into pre-fixed schemes, such as tools for root cause analysis and
safe job analysis. In addition, new measures themselves may have unintended consequences.
Moreover, the analysis of real and potential accidents often takes place within a blame-
oriented environment, with a tendency to point to causes of human error; this may hinder
revealing other underlying causes (Oltedal, 2010). Such a situation is illustrated in Narrative
2.
Narrative 2: Norwegian mate concerning a not-applicable procedure.

“(…) in one situation, one of the able bodies cut himself with a knife. It was a small finger
cut, and nothing serious. The incident was reported, as we are required to, whereupon we
were instructed that it was no longer permitted to carry knives. In situations where knives
were required, a safe job analysis and risk assessment were to be performed first. However,
what they [people ashore working with SMS] do not understand is that not wearing a knife
may involve greater risk. What if someone gets entangled in a hawser, with a risk of being
dragged overboard, a situation where a knife would be quickly required to cut loose, what
shall we do? Run to the bridge and carry out a risk assessment first? However, we bypassed
this new regulation and did a general ‘safe job analysis’, which concluded that all able
bodies could wear knives on a general basis.”
When it comes to near misses and experience data, HRO argues that mindfulness and
preoccupations with small failures may improve the reliability of the operational system. An
HRO approach implies, inter alia, that the organisation is sensitive to operations and first-line
experiences, encourages alternative frames of reference, and creates an error-friendly learning
culture in which people seek feedback, share information, ask for help, and talk about error
and experiment (Weick and Sutcliffe, 2007). However, if mindfulness is to be possible, it
requires support from the top management and throughout the entire organisation.
Mindfulness, as a proactive activity, is resource demanding. Mindfulness implies that every
operator thinks differently about success and the possibility for failure along with the use of
creativity in order to imagine what can go wrong and how. Mindfulness implies that every
operator—if in any doubt at all—may stop operations and, if so, with support from the
organisation and co-workers, even if the situation ultimately turned out to be safe every time.
We do question if any organisations exposed to competition and constantly facing decisions
of safety versus efficiency are capable of having all organisational members in a mindful
mind mode for a longer period of time before efficiency comes to dominate safety. After all,
no organisation exposed to competition exists with the primary goal of being safe; they need
to become competitive in order to avoid bankruptcy.
Both Narratives 1 and 2 depict situations in which measures have been developed
detached from the operational system, with the consequence that the measures are
experienced as difficult to relate to in operations. Narrative 2 also effectively depicts where
first-line operators (seafarers) bypass the shore sides’ efforts in the design phase in adjusting
to an applied mode. Planning has a symbolic as well as functional aspect (Clarke, 1999).
15
According to Clarke (1999), organisations and experts use plans as a form of rhetoric—tools
designed to convince audiences that they ought to believe what the organisation says.
In particular, some plans have so little instrumental utility in them that they warrant the
label “fantasy document”. The usual presumption in social science is that the first step in an
adequate planning process is to assess fairly completely what the problem is; the second step
is to write a plan that addresses the problem, and the final step is to implement the plan.
However, just like Perrow (1999), both Snook (2000) and Clarke (1999) are sceptical towards
the utility of planning, especially in relation to future complex situations with a high degree of
uncertainty. The uncertainty planners face when working with SMS may very well result in
the development of overly burdensome rules or rules that are too detailed in their efforts to
reduce uncertainty by controlling (most) human actions. Such a reality exists within shipping,
as supported by the majority of survey comments presented in
Table 3. Moreover, the results in Table 4 indicate that about 43% of the survey
respondents breach procedures because they do not work as intended, about 35% because
there are too many procedures, and about 27% because deviation from procedures will
improve the quality of their work.
Similar to Perrow (1999), Snook (2000) identified part of the problem as inherent in the
rationale of the design phase in a SMS. More precisely, the assumption is that the operational
system is tightly coupled, yet most of the time the organisational sub-units are a loosely
coupled system. When a global rule is violated, operations usually go on as normal. On the
contrary, breaking global rules may be perceived as being more efficient and even rewarded
by the organisational management. Thus, a mismatch exists between the rule-based logic of
action and the presumed tight situational coupling, which brings us into the next phase:
“engineered”.
5.2 Engineered
“Engineered” is defined by the interaction of the rule-based logic of actions and loose
couplings. Although measures are originally designed for a tightly coupled reality, crews do
experience a loosely coupled reality as nothing happens when the rules are broken. When
such behaviours are not followed by some kind of reprimand, punishment, or any dangerous
situation, their substantive risk-reducing reason may be questioned by operators (Snook
2000). Moreover, the presence of real-world constraints of actions may be overlooked by
planners, including commercial pressures from the shore. The results in Table 4 show that the
top reason for violating procedures is the need to work faster, which was ticked by about 47%
of the respondents. Perceived pressure due to the overload of work is among the top four
reasons for violating procedures, with approximately 32% of survey respondents concurring.
Commercial pressure is also perceived to be present on board, as supported by survey
comments in Table 5, case studies, and previous research carried out within both the liquid
tanker sector (Oltedal, 2010) and the dry cargo sector (Oltedal and Engen, 2010). For
example, it is well known among seafarers that hours of work are regulated primarily in
theory only. Comments in Table 5 substantiate that rest hours are not followed due to high
workloads, among other issues. Numerous anecdotal accounts were presented of not only
some seafarers’ excessive hours, but also the manipulation of the hours for work/rest records
in an attempt to conceal the truth and suggest compliance. One narrative depicting such a
situation follows (Narrative 3).
Narrative 3: Norwegian captain about rest-hour regulations

“Rest-hour regulations? There is no such thing as rest-hour regulations! The regulations
say that we are supposed to have 8 hours of rest during the day, but when calling in port after
16
passing through the English Channel, it is not unusual to be at the bridge for up to 18 to 20
hours. When you sail from Ushant to Rotterdam in about 40 hours! Nowadays with crew
shortages, the bridge officers are getting promoted faster and are more inexperienced, and it
may be difficult to leave them alone on the bridge with all the responsibility. As the master, I
am in command of the vessel and responsible for safe navigation. Even though it is not
explicitly stated in IMO regulations that I should be at the bridge at all times when passing
through the channel, most companies have a standing order that the captain has to be on the
bridge in congested waters, such as the Malacca Strait, Singapore Strait, and the English
Channel. I know who they would blame if something happened—if the vessel grounded or
something—and I was not present. In addition, I am expected to send reports to the company
and charterers and so on and thereby have to leave the wheelhouse, contrary to standing
orders. If it is for the immediate safety of the vessel, it is okay to depart from the regulations.
And that is what we do. When we are short of manpower to do anything like maintenance,
painting, or anything, I make the remark that it is for safety reasons, even if it is not, and then
it is sometimes okay. Anyway, after passing through the channel, preparations for port arrival
have to be done. When the pilot is boarding, I have to be present at all times. When calling in
to the port, I may not have slept at all for 30 hours! The first people to meet me upon arrival
is the port state control, flag state control, classification, QA control, and vetting inspectors
from customers, charterers, or our own company. Sometimes they queue up and everyone
expects to be first in line. Each vetting lasts for about 6-8 hours and requires a lot of the crew
to be available—mostly officers; the most important people are kept awake for the longest
periods. The first thing I am asked is if I have had 8 hours of rest. I have to say ‘yes’; if not, I
am in trouble. The inspector knows that I am lying, and I know that the inspector knows that I
am lying, but they do not care as long as they can tick of the right box as okay. Apart from
this, I have to handle provisions, crew changes, ballasting, loading, and discharging in the
shortest time possible. And the return keeps you awake for another few hours. The situation is
impossible, and everybody knows that the regulations and ‘safety first’ slogans are all a
charade, but nobody cares as long as the paperwork is okay and they have someone to blame
if something should happen”.
The demand to maximise profit may induce management to promote efficient and unsafe
behaviour, which may result in a reduced error margin or the overstepping of the boundary for
functionally acceptable performance. Management pressure for efficiency may also result in
reduced crew level. The majority of the total operation expenses may be broken down into
crewing, insurance, repairs and maintenance, stores and lubrications, and management and
operations. Crewing costs typically amount to 30 to 65% of the total, depending on crew
nationality, and are the easiest-to-reduce expenses in order to avoid failure or increase profit.
Especially in light of the current economic recession, with the accompanying decline in the
freight rates, the search for cost-cutting initiatives is a motivator that ensures profits and
continued operations. Crew reductions have historically been the main instrument for cutting
costs (e.g., the shipping crisis in the late 1970s and early 1980s) and still are a likely area for
cost reductions (MacDonald, 2006).
The appropriate crew level also depends upon other areas, such as operations, trading
areas, and frequency of calling in ports. Too few crew members are more notable when
calling in and leaving port—operations that have a higher demand for efficiency (Oltedal,
2010; Oltedal and Engen, 2009). When arriving, staying in, or leaving port, many things
happen in a short period of time that are not specified in the principles for safe manning
(International Maritime Organization [IMO], 1999), but affect the workload. This could be the
need for piloting, discharging, loading, crew change, provisioning, bunkering, inspections,
and the like. Fatigue is also considered to be influenced by the demand for efficiency and fast
17
turnover rate (Smith, 2007). Under such fluctuating work pressure, it is even less likely that a
standardised global measure is appropriate for the actual situation. Our data also suggest that
the industry is well aware of the situation. The following comments on rest-hour regulations
were made at a conference for operators, ship management, charterers, and inspectors.
“(…) I uphold that everybody is well aware that rest-hour regulations are not complied with.
Even when doing the planning, we know it is impossible to comply with the regulations. But
there is not that much we can do. If we plan for regulation compliance, we might lose
contracts”.
This comment was followed by a question:
“But if we agree that the violation of rest-hour regulations is a near miss (researcher’s
comment: everyone agreed). Should not such violations be reported as near misses? And if
you did, what would happen?”
The following answer was provided:
“If we openly admit that the rest-hour regulations are violated frequently, we will not get the
next contract”.
Undoubtedly, these IMO conventions (e.g., rest-hour regulations) have improved safety
in many areas. Nevertheless, companies and crew find ways to bypass the regulations if
needed, and the control systems intercept only some of the concerning actors. However, what
is striking in this situation is organisations’ awareness of the problem and that they—with full
awareness—design plans that are not in accordance with reality. This disconnection between
the real world of everyday practice and design is the driving force of the “practical drift”
(Snook, 2000). Snook (2000) defines “practical drift” as a phenomenon resulting from the
mismatch between the local demands of the situation and those of global design rules. As
such, the pervasive demands of day-to-day practice inevitably shift the logic of action from
one based primarily on formal rules to one driven more tightly by the task, loading schedules,
port arrivals, and efficiency. Subsequently, over time, pragmatic practice loosens the grip of
even the most rational and well-designed formal procedure, which brings us to the next phase:
“applied”.
5.3 Applied
When moving into the “applied” quadrant, rule-based logic of action yields to task-based
logic. The net effect of this practical drift is a general loosening of globally defined
rationality, in which task-based logic of action now matches the loosely coupled situation. In
this phase, the crew is no longer operating according to globally designed rationality. Rather,
it is replaced with the rationality of each local unit (e.g., a vessel, department, or team).
However, although the workforce now operates according to multiple, incrementally
emergent sets of procedures, each borne out of unique sub-unit logics grounded in the day-to-
day pragmatics or loosely coupled words, the operational system still remains resilient as long
as the system remains loosely coupled. As a result, the working environment may be very
different within each local unit (e.g., a vessel, department, or team). Crew and team instability
constitute additional challenges in this setting. As presented in section 3.3, survey
demographics, most non-Norwegian crew members are contract employees hired through
crewing agencies and do not work within a stable team. The demographics further show that
18
44.8% often or always sign in the same vessel each sailing period. However, about 45% of
these are spread out over various vessels, and most vessels experience various degrees of
instability within teams. Previous research also suggests that working within a well-
established team does have a positive effect upon safety culture in general as well as
interpersonal relationships (Snook, 2000; Oltedal and Wadsworth, 2010). Interpersonal
relationships relate to, inter alia, the degree of trust and open communication amongst
crewmembers. Reason (2001) regards a trusting relationship as a cornerstone for getting
individuals to report their own mistakes and experiences, which is fundamental for the basic
rationality of an SMS. For both subordinate and superior/managerial positions, additional
challenges may arise in relation to the multinational crew and unstable crewing with low
stability within teams.
When signing on a new ship, new crew is unfamiliar with the ship’s on-board
management style as well as fellow crewmembers and the on-board working climate. The
seafarers do need some time to familiarise themselves with and adjust to the new situation.
For instance, if the ship management at the seafarers’ previous vessel was blame oriented, the
seafarer will most likely sign on the new vessel with the latest experience in mind, thereby
being cautious about reporting his own mistakes for fear of being blamed or sanctioned. Over
time, each seafarer learns how management is oriented at the particular vessel. In other words,
seafarers become assimilated into the applied logic of their current unit. The problem is even
more pronounced when the seafarer is constantly changing vessels, working with new
management during each sailing period and experiencing this familiarisation process each
time. The management style is known to vary within the sector. Poor shipboard management
and leadership have also been identified in previous research (Oltedal and Engen, 2009;
Knudsen, 2004). Narrative 4, from the comments on a survey question, illustrates one
example of what happens when moving into the “applied” quadrant.
Narrative 4: Filipino mate concerning rest-hour registration

“(…) I once experienced being ordered to adjust my work and rest-period registration in a
way so as to comply with the regulations. To be open with this or to react in disagreement
may somehow jeopardise your next employment”.
Interpersonal relationships amongst crew, in practice, also reflect the degree to which the
crew shares safety-related information when changing shifts as well as more informal
processes of sharing safety-related information during operations; as such, these relationships
serve as a premise for the HRO principle of mindfulness. However, another issue not taken
into consideration by HRO is relations of power. When entering a relationship with others, we
constrain and are constrained by others as well as enabling and being enabled by others
(Stacey, 2007).
As Narrative 4 indicates, the power balance favours the individual giving the order.
However, such power balances are also constantly changing, shifting in favour of some while
going against others depending on the relative need they have for each other (Stacey, 2005).
The Filipino mate who shared the narrative is in need of employment; as a contract worker,
the power balance is against him. In this particular situation, regulations, restrictions, and
control systems are of little use. The HRO principles of mindfulness are also of little use as
the premises for mindfulness are not present. In other words, no premises exist for a trusting
and open relationship. However, by changing the employment terms, the power balance could
change, and the situation could become more favourable for compliance with external and
internal requirements.
Both NAT and HRO consider interaction to be a prime mover constituting networks or
systems at different levels that further interact with each other as the system influences the
19
individual’s behaviour within the organisation and in safety-related issues. Where NAT
emphasises system characteristics and power constellations as the main factors, HRO is
preoccupied with the internal structure, coordination, and goal specification. However,
according to Stacey (2005), these so-called forces (established premises) are taken as
powerful and stable conditions, arising outside of those who relate to them in daily operations,
in their own direct experience and influence. The individuals are supposed to comply with
global rules; however, they may feel alienated when relating with them and thus make a local
interpretation and adjustments of the standardised rules.
Making a departure from Stacey’s (2005) theories, international and national conventions
and legislations, such as work-hour regulations, should not be regarded as forces to which all
must comply, but rather a result of complex responsive processes of interaction. The point of
departure is how the various groupings of local interaction respond to the work-hour
regulation and how the regulation is expressed differently in many local situations. From our
own research, we find that the responses are expressed in numerous ways. Some have already
been mentioned as false registrations of work hours (see Table 5), erroneous justification of a
situation by an invented safety reason, ignorance, and so on. From the system perspective,
deviations in actual behaviour are controlled by feedback through control systems, such as
SMS. However, these are also responses that are adopted differently in each local interaction
(i.e., a practical drift in various directions).
Developing new conventions for vetting is the way IMO normally responds to risk. Each
of these changes is happening in small groups of local interactions among (small) groups of
individuals. Each group incorporates different responses, logic, interpretations, and situational
sense-making. Hence, we cannot talk about different systems, entities, or levels that are
detached from one another; rather, we should focus on changes that occur through continually
responsive processes of interaction. The result is that each operational unit over time drifts
further away from the globally established standards, and local responsive processes result in
the development of new local rationality and rules of engagement. However, although the
workforce operates in conflict with global rules, it is still assumed that people in other units
behave in accordance with the original set of established rules. Problems then occur when
either party miscodes the situation or fails to act according to the shared global rules or when
the situation again becomes tightly coupled.
5.4 Practical Drift from Applied to Failed

When operating in applied mode, each subgroup follows its own unique path of practical
drift. Each uneventful day that passes reinforces a steadily growing false sense of confidence
that everything is all right. However, in one rare stochastic fit, the system then becomes
tightly coupled, and applied, as the “normal” praxis becomes the cause of the accident
(Snook, 2000), as illustrated in the follows example. In December 2008, the Mirabelle, a
Maltese flag Norwegian operated general cargo ship, grounded shortly after departure from
port. The captain was alone at the bridge when grounding, which was identified as the main
cause of grounding. As he was alone, he handled the manoeuvring, the navigation, the look-
out, and the handling of the searchlight, thereby trying to do four jobs at once (Danish
Maritime Authority, 2009). One month later, on January 16, the Mirabelle grounded once
again, this time along the coast of Norway in the early morning. The navigator was alone at
the bridge and had fallen asleep on watch (O'Cinneide, 2009).
In this case, both accidents were in some way related to crewing level, resulting in the
first situation in work stress and in the second situation in fatigue. With reference to the
Mirabelle, an important factor is time. These factors with low crewing level had been present
for some time, although not intercepted in the ISM/SMS system for remedies. The longer the
20
passage of time between design and failure, the less the possibility of detecting the “real”
underlying causes, as crew may be undermanned and fatigued for a long time without any
mishaps. Groundings, along with ship collisions, are known to be the most typical type of
accidents at sea. Between 1996 and 2003, 652 groundings and collisions involving vessels
over 500 gross tonnes were reported to the Marine Accident Investigation Branch (MAIB)
under the United Kingdom’s Merchant Shipping Regulations alone (MAIB, 2004). The MAIB
investigations showed that a small number of causal factors are common to nearly all bridge
watch-keeping accidents and highlighted the following three principal areas of concern: (1)
fatigue and groundings: a third of all groundings involved a fatigued officer alone on the
bridge at night; (2) lookout and collision: two thirds of all vessels involved in collisions were
not keeping proper lookout; and (3) safe manning and role of the master: a third of all the
accidents that occurred at night involved a sole watch keeper on the bridge. Similarly, another
study undertaken by the Swedish Maritime Administration (SMA) found that, in 84% of 32
reported collisions and groundings occurring between 1997 and 2002, the accident occurred
between 2300 and 0800 hours. The study concluded that fatigue-related problems affecting
bridge watch keepers are, with very great probability, occurring more frequently than initially
believed (IMO sub-committee on standards of training and watch keeping).
Although fatigue, rest-hour regulation, and safe manning are recognised as causal factors,
accidents still happen due to these factors. An important mechanism to ensure compliance
with IMO requirements is ship inspections. However, the flag states are not alone in
conducting such inspections. Port states, classification societies, customers, charterers, cargo
owners, operators, and others also conduct inspections. Each year, flag states are evaluated
and ranked based on their effective enforcement of international rules, whereupon they are
placed on a “black list”, “grey list”, or “white list” (MARISEC, 2008). Some flags do not
manage or intend to uphold their responsibilities. The drawbacks with flag state control can in
theory be balanced by port state control (PSC). However, the PSC also faces problems. In
2007, the European PSC, Paris Memorandum of Understanding (ParisMOU), performed
22,877 inspections on 14,182 individual ships registered with 113 flags, which resulted in
1,250 detentions (ParisMOU, 2010).
The ParisMOU is an open register, and results from inspections, target areas and so on
are public information as well as important parameters towards customers. The ParisMOU
strategy is blaming and shaming. However, the latest review from Paris-based Equasis, which
provides PSC data to the shipping industry, reveals that in the same period (2007) no record
of inspection existed for some 17,910 ships. These ships avoid inspections and slip through
the PSC net. The most common types of vessels slipping through are old general cargo ships,
many of which have no classification record and are registered with flag states with a poor
safety record (Corbett, 2009). These ships may very well be the ones most in need of an
inspection.
We also suspect that vessels that are inspected and found to be operating according to
regulations are not necessarily doing this. Data from our own research indicate that some
companies or vessels deliberately try to avoid the regulations (e.g., comments in Table 5).
Vessels have work, overtime, and rest-hour registrations that do not reflect reality. Inspectors
may on some occasions reveal the problem by comparing different registrations (i.e., if the
registration indicates that the vessels are both bunkering and simultaneously have an engine
crew that is off duty). However, we assume that the inspectors, for various reasons, do not
manage to reveal all cases that are not in accordance with the regulations.
Fatigue, lookout, and safe manning are highly interrelated with one another as well as
with safety in general. IMO, in addition to the industry in general, is familiar with this
problem. In 1993, a joint ILO/IMO group of experts on fatigue drafted a report on fatigue and
its contribution to maritime casualty and accidents (International Maritime Organization). In
21
order to improve overall safety, such experience data are fed back into the original quadrant
of “design”, resulting in the “redesign” phase. Based on our data, two different approaches
towards redesign were identified: (1) organisations where global rules are redesigned
detached from operational systems and (2) organisations where global rules are redesigned by
frontline personnel. Research identifies shipping as an industry with an extensive use of
procedures and checklists; when redesigning the SMS system, new procedures or checklists
are likely to be added (Oltedal, 2010). However, data suggest that crews are more satisfied
with operational SMS, procedures, and checklists when given the opportunity to influence
them and when sharp end experiences are taken seriously. Snook (2000) also pointed out that,
for organisational leaders, it is easy to get carried away on control trips with the misuse of
procedures and checklists. When real situations reflect the fact that control is not efficient,
one solution might be to control even more. Sometimes such well-intentioned efforts to
prevent failure through tight control often produce just the opposite effect. The original
problem of practical drift became a mismatch between situational couplings and rationale of
logic, with rules being designed for a tightly coupled situation. As a result of this mismatch,
we argue that a better solution is to acknowledge the world as loosely coupled, thereby
minimising the use of global rules and being open for local adaption.
6 Conclusion
A paradox is inherent in safety management by local control strategies. New safety

measurements are often a result of deviations (errors, incidents, or accidents) resulting in
some disturbance of the production process—namely, material damage or injury (Oltedal and
Engen, 2009; Rasmussen, 1997). This implies that, in order to enhance safety, an unsafe
deviation has to happen first. Thus, to be safe, you need to be unsafe. Moreover, resulting
safety measures are then used as a standard for finding explanatory cause(s) towards future
similar deviations. Reality is compared with the organisational safety standards and deviations
attributed to human error. This is analogous to having a nautical chart that does not show the
sandbank the ship has just grounded on and then blaming the seabed for not corresponding to
the nautical chart.
As previously mentioned, the shipping industry’s approach is often person orientated.
Safety measures tend to aim at controlling human actions, often in the form of excess use of
procedures and checklists. These measures are standardised to fit all; it could be a fleet of 5
vessels or 100 vessels. However, in actual work, no operation is ever the same. The vessels
are different as well as the people, constellations of people, power figurations, weather, and so
on. A standardised measurement will therefore never align with reality. Despite this, human
actions and deviations are compared to the standard and found to be erroneous. In the search
for control, new and even more detailed measurements may be developed. This is a vicious
cycle created by the anxiety of not being in control. Snook (2000), Reason (2001), Hollnagel
(2009), and others have all suggested that organisations should leave such person-oriented
approach in their search for causal and influencing factors.
Reason (2001) also warned about what he calls “anxiety-avoidance”, which fits the
previous narrative. Anxiety-avoidance describes an organisation that has discovered a
technique to reduce risk and repeats it over and over again regardless of its effectiveness, like
constantly adding yet another procedure in response to unwanted incidents. Thus, the
organisations working towards safety in this manner do not use the safety system theory as
suggested, but abuse it due to the anxiety of not being in control or for other reasons. Reason
(2001) and Stacey (2007) further mentioned anxiety related to not being in control as an
important reason for overreliance on structure and systems. Moreover, Reason (2001)
22
emphasised taking local experience into account through a safety information system and
reporting. During the research, we also encountered companies that let the crew have a major
control over the development of the safety system, with the shore management’s views being
overridden by the sailors. In these situations, the sailors demonstrated a more positive attitude
towards, and experience with, the procedural system. Thus, when the SMS is perceived as
relevant, when applied to the premises for practical drift, it is attenuated.
However, when it comes to actual behaviour, Reason (2001) recognised that both human
attitude and behaviour are extremely difficult to change. This shifts the focus towards
organisational structures and practices, is supposed to be more controllable, and should guide
and arrange for a certain kind of behaviour through conditions at the local workplace.
Furthermore, human error can never be eliminated, and the future is unpredictable. It might be
provoking for some, but we suggest that organisations, managers, and planners accept that
work and life imply risks and uncertainties; no matter how much we try, unwanted events of
some kind will always happen. The real difference in safety management is how we approach
such events. Staying with human error as the main explanatory factor may result in greater
efforts in order to control all human actions. Alternatively, it is important to understand the
underlying processes of practical drift and its interrelationship with mechanisms of control,
situation, and human rationality. Yet however much we try, accidents like the Herald of Free
Enterprise in 1987, the Exxon Valdez in 1989, the Scandinavian Star in 1990, and the Estonia
in 1999—and more—will always happen.
Snook (2000) suggested that any organisation, when working with safety management,
should pay attention to three areas: (1) looking beyond individual error by framing puzzling
behaviour in complex organisations as individuals struggling to make sense; (2) following the
basic design principles of high performance teams and thinking twice about chasing the
advantages of social redundancy; and (3) treating organisational states of integration and
reliability with chronic suspicion. The important thing is to recognise them for what they are:
constant outcomes of dynamic systems and ongoing accomplishments that require active
preventive maintenance. This article has also underlined how the individual character of
single units must keep a certain degree of flexibility while simultaneously being guided by
superior regulations. If a complex and heterogeneous sector such as international shipping is
overloaded with standardised regulations, local actions will be hampered in solving everyday
incidents. The consequence will be a lid on local flexibility, which may increase risk rather
than reduce it. The challenge is to strike a balance between a centrally developed design that
is accepted and followed by the local actors and an environment in which these regulations
can open up for pragmatic solutions, when required. Such a regulatory framework requires
intelligent designers and continuous revision when confronted with an endless stream of real
incidents. An effective feedback loop of information between the local actors and the central
designers may thus create a pathway between central design and local pragmatism that can
prevent the system from moving from practical drift into failure and catastrophe.
References
Adams, J., 1995. Risk, UCL Press, London.
Anderson, P., 2003. Cracking the code: The relevance of the ISM code and its impact on
shipping practices, Nautical Institute, London.
Bertalanffy, L.v., 1971. General system theory: Foundations, development, applications,

Allen Lane, London.
23
Clarke, L.B., 1999. Mission improbable: using fantasy documents to tame disaster, University
of Chicago Press, Chicago.
Cooper Ph.D., M.D., 2000. Towards a model of safety culture. Safety Science 36 (2), 111-
136.
Corbett, A., 9 January 2009. Thousands of ships slip through inspection net. Tradewinds.
Danish Maritime Authority, 2009. The grounding of MIRABELLE on 16 Desember 2008,

The division for Investigation of Maritime Accidents, Denmark.
Department of Transport, 1987, Report of Court No. 8074. mv Herald of Free Enterprise.
Formal investigation.
DiMaggio, P., 1994. Culture and Economy. In: Smelser, N.J., Swedberg, R. (Eds.), The
Handbook of economic sociology. Princeton University Press, Princeton, pp. 835.
European Maritime Safety Agency (EMSA), 2010. Maritime Accident Review 2009,
Hair, J.F., 1998. Multivariate data analysis, Prentice Hall, Upper Saddle River, N.J.
Heinrich, H.W., Roos, N., Petersen, D., 1980. Industrial accident prevention: a safety
management approach, New York.
Hollnagel, E., 2009. The ETTO Principle: Efficiency-Thoroughness Trade-Off Why Things
That Go Right Sometimes Go Wrong, Ashgate.
International Maritime Organization (IMO), 2006. IMO Sub-Committee on Standards of

Training and Watchkeeoing (STW) 38/13/7, International Maritime Organization, London.
International Maritime Organization (IMO), 2005. Study on the impact of the ISM Code and
its effectiveness in the enhancement of safety of life at sea and. protection of the marine
environment, International Maritime Organization, London.
International Maritime Organization (IMO), MSC/Circ.621, International Maritime

Organization, London.
International Maritime Organization (IMO), 1999. Resolution A.890(21): Principles of Safe

Manning, International Maritime Organization, London..
Justis- og politidepartementet, 1991. Scandinavian Star - ulykken, 7 april 1990, Norway,

Oslo.
Kjellén, U., 2000. Prevention of accidents through experience feedback, Taylor & Francis,
London.
Knudsen, F., 2004. If you are a good leader I am a good follower. Research Unit of Maritime
Medicine, Denmark.
Kuo, C., Nautical Institute, 2007. Safety management and its maritime application 2, 288.
24
MacDonald, R., 2006. Safe Manning of Ships - Yesterday, Today and Tomorrow, WMU
Journal of Maritime Affairs 5 (2), 143-151.
Marine Accident Investigation Branch (MAIB), 2004. Bridge Watchkeeping Safety Study,
United Kingdom.
Maritime International Secretariat Services Ltd (MARISEC), 2008. Shipping Industry Flag
State Performance Table, United Kingdom, London.
National Transportation Safety Board, 1990. Grounding of U.S. Tankship EXXON VALDEZ
on Bligh Reef, Prince William Sound Near Valdez, AK March 24, 1989.
O'Cinneide, E., 16. January 2009. Death ship drama, Tradewinds.
Oltedal, H.A., Engen, O.A., 2010. Tanker versus dry cargo - The use of safety management
systems within Norwegian dry cargo shipping. In: Ale, Papazoglou, Zio (Eds.), Reliability,
Risk and Safety, Taylor & Francis Group, London, pp. 2118-2125.
Oltedal, Helle, Wadsworth, Emma, 2010. Risk perception in the Norwegian shipping industry
and identification of influencing factors. Maritime Policy & Management 37 (6), 601-623.
Oltedal, H.A., Engen, O.A., 2009. Local management and its impact on safety culture and
safety within Norwegian Shipping. In: Artorel, M., Guedes, C., Barnett, J. (Eds.), Safety,
Reliability and Risk Analysis. Theory, Methods and Applications, Taylor & Francis
Group, London, pp. 1423-1430.
Oltedal, H., 2010. The use of safety management systems within the Norwegian tanker
industry - do they really improve safety? In: Bris, Guedes Soares, Martorell (Eds.),
Reliability, Risk, and Safety: Theory and Applications, Taylor & Francis Group, London.
Perrow, C., 2004. A Personal Note on Normal Accidents, Organization & Environment. 17
(1), 9-14.
Perrow, C., 1999. Normal accidents: living with high-risk technologies, Princeton University
Press, Princeton, N.J.
Pett, M.A., Lackey, N.R., Sullivan, J.J., 2003. Making sense of factor analysis: The use of
factor analysis for instrument development in health care research, Sage, Thousand Oaks,
Calif.
Rasmussen, J., 1997. Risk management in a dynamic society: A modelling problem. Safety
Science 27(2-3), 183-213.
Reason, J., 2001. Managing the risks of organizational accidents, Ashgate, Aldershot.
Rothblum, A.R., 2000. Human Error and Marine Safety.
Scott, W.R., Davis, G.F., 2007. Organizations and organizing: Rational, natural, and open
system perspectives, Pearson Education, Upper Saddle River, N.J.
25
Smith, A., 2007. Adequate Crewing and Seafarers' Fatigue: The International Perspective,
Centre for Occupational and Health Psychology, Cardiff University.
Snook, S.A., 2000. Friendly fire: the accidental shootdown of U.S. Black Hawks over
Northern Iraq, Princeton University Press, Princeton, N.J.
Soma, T., 2008. Are the accidental losses increasing in shipping - what can be done?, DNV
Maritime Solutions.
Stacey, R., 2005. Local and global processes in organizational life. In: Stacey, R.D. (Eds.),
Experiencing emergence in organizations: Local interaction and the emergence of global
pattern, Routledge, London, pp. 17-47.
Stacey, R.D., 2007. Strategic management and organisational dynamics: The challenge of
complexity to ways of thinking about organisations. FT Prentice Hall, Harlow.
Standard P&I Club 2010. 16 September-last update, Standard Safety, September issue
[Homepage of Standard P&I Club], [Online]. Available: http://www.standard-
club.com/docs/14468_STANDARD_SEPTEMBER_LR-2.pdf .
Studio Apertura, 2004. Sikkerhetskulturer i transportsektoren. Metoder for kartlegging av

sikkerhetskultur: Evaluering av noen eksisterende verktøy, Norway.
Weick, K.E., 1976. Educational Organizations as Loosely Coupled Systems, Administrative

Science Quarterly 21 (1), 1-19.
Weick, K.E., Sutcliffe, K.M., 2007. Managing the unexpected. Jossey-Bass, San Francisco,
Calif.
www.ParisMOU.org, 2010. 21. July-last update [Homepage of Paris MOU], [Online].

Available: www.ParisMOU.org.
26
Appendix 1: Survey questionnaire and letter of
introduction

Dear crew member
In this survey I want to assess how safety is handled on Norwegian-controlled bulk cargo and tanker
vessels and how the crew onboard perceive safety. This is part of my PhD work, which is aimed at
improving our understanding of how various safety-related circumstances work in practice. The results
of the survey will help your shipping company make safety-related decisions, enabling your safety to be
better ensured.
All information that is obtained through the survey is anonymous. It will not be possible to trace any
answers to individuals, shipping companies or vessels. My research has been financed by
Stord/Haugesund University College and I am therefore acting independently of all shipping companies,
public authorities and other organisations and interests. As a PhD student and researcher, my work will
be carried out in line with Norwegian guidelines for research ethics, which among other things protect
your right to be anonymous. Only I will handle the completed forms or have other access to the data.
To optimise the quality of the survey, it is important that as many people as possible complete the form.
It will take about 30 minutes to complete the form. It is also important that the questions are answered as
frankly as possible. When you have completed the form, place it in the enclosed envelope, seal the
envelope and deliver it directly to the vessel’s safety delegate or the chosen contact person onboard for
this survey.
- As regards the answering of the questions on the form itself, please relate your answers to the
circumstances onboard this particular vessel.
- As regards comments concerning questions and suggestions for improvements, you can relate
these to the experience you have of the sector in general.
This type of survey is very common in Norway. Both Norwegian companies and authorities want the
safety of employees to be given the highest priority. From similar surveys we have learned that safety-
related matters often do not function as intended, and that the reason can be traced back to weaknesses
linked to the company's management or other organisational factors. Some of the questions may be
difficult to answer, but the aim is not to place the blame on individuals. We know for example that
procedures can be broken without the blame resting on the individual who breaks the procedure but on
other levels within an organisation. In such cases I want to identify the organisational reasons for
breakdowns in procedure. Possible reasons here include insufficient involvement in the development of
procedures, the existence of too many procedures or the adoption of dangerous procedures.
Participation in and completion of the questionnaire is voluntary. If you choose not to take part in the
survey, please fill in the enclosed green sheet and return it in the same way as the questionnaire.
If you have any questions concerning the survey, you can contact me, Helle Oltedal, on telephone no.
(+47) 93 82 61 87 or (+47) 52 70 26 44.
Thank you for taking the time to fill in the questionnaire!
Yours sincerely,
Helle Oltedal,
Stord/Haugesund University College
Safety Culture Survey
Please indicate your answer by crossing off a box for each question like this: _. Mark wrong answers
like this: J Where it is not possible to select an answer, please provide an alternative in the space
provided for other or fill inn.
PLEASE WRITE ALL COMMENTS OR SUGGESTIONS IN ENGLISH

1. Background Information
A Gender:
Female 1
Male 2
B Job position/title
Captain 1
Mate 2
Engineer 3
AB /Seaman 4
Electrician 5
Catering 6
Apprentice 7
Other (Fill in):_________________
C Vessel class/cargo
Bulk 1
Combined 2
Shuttle tanker 3
Gas 4 General Cargo 5
Chemical 6
Oiltanker 7 Other (Fill in)_______________
D Age Under 31 years 31-40 years 2
41-50 years 3
51-60 years 4 Over 60 years 5
1
E Nationality
Norwegian 1
Polish 2
Filipino 3 Other (Fill in):______________
F How long have you
been working within
shipping? 0-2 year(s) 1
3-5 years 2
6-10 years 3
11-20 years 4 Over 20 years 5
G
How long have you been working for this shipping company?Fill in:____________
H
All in all, how long have you been working at this vessel? Fill in:____________
I Which kind of employment contract do you have?
Permanent employee 1
9 months duration 2
6 months duration 3
3 months duration 4
Other (Fill in):
J
How long is your ordinary work /sailing schedule at the ship? (for contract (Fill in)__________________
workers this will often be the same as the contract duration)
K How long do you usually stay at home or on shore between each work / (Fill in)__________________
sailing period?
L Do you normally work at the same vessel at every work / sailing period?
Very seldom /never 1
Sometime 2
Very often / always 3
MHow is your ordinary watch system, without overtime?(for example

Dayman, or 4-8-4 meaning 4 hours work – 8 hours off – 4 hours work)
(Fill in)__________________
1 English version 65
2. Top Management’s Safety Priorities
Strongly Disagree Not sure Agree Strongly Don’t
Only select one answer per question disagree agree know
A The on-shore top management in my company
1
2
3
4
5
6
prioritizes safety before economy.

B I experience conflicting requirements from my
1
2
3
4
5
6
company and the captain.

C I experience that safety work is more a facade than a
1
2
3
4
5
6
real priority area.

D I am familiar with the company’s safety goal. 1
2
3
4
5
6
Comments and suggestions:
3. Local Management
Please state your evaluation of your closest superior’s attitude toward safety. If you are the
captain, relate the questions to the closest on-shore manager.
A All in all, for how long have you been working with your closest superior?(Fill inn) ___________________________
Very seldom Seldom Sometime Often Very often Don’t
Only select one answer per question /never /always know
B Is your closest superior clear in his engagement to
1
2
3
4
5
6
ensuring his co-workers’ safety?

C Does your closest superior follow up to ensure that
1
2
3
4
5
6
all work on board is done in a safe manner?

D Is your closest superior a good role model when it
1
2
3
4
5
6
comes to attending to his own and others’ safety?

Once a Twice a Once a Once Once More
week month month every 2nd every 6 seldom
month moths
E How often do you participate in meetings with your 1
2
3
4
5
6
closest superior where safety is a topic?
disagree agree know
F My closest superior appreciates that the crew is
1
2
3
4
5
6
willing to discuss safety-related conditions.

G My closest superior is not afraid of admitting his
1
2
3
4
5
6
own mistakes.
H My closest superior has too little confidence in his
1
2
3
4
5
6
co-workers.
I My closest superior is supportive if safety is 1
2
3
4
5
6
prioritized in all situations
4 Procedures and Guidelines
A Due to the company’s demand for efficiency we 1

2
3
4
5
6
sometimes have to violate procedures.

B Due to the captain’s demand for efficiency we
1
2
3
4
5
6
sometimes have to violate procedures.

C I have received good training in the company’s
1
2
3
4
5
6
procedures.
D I feel that it is difficult to know which procedures
1
2
3
4
5
6
are applicable.
E The procedures are helpful in my work. 1
2
3
4
5
6
F The procedures are difficult to understand or are

1
2
3
4
5
6
poorly written.
G We have the opportunity to influence and form the
1
2
3
4
5
6
procedures.
If you don’t follow the procedures in a specific situation, what may be the reason? Please don’t
mark more than three options.
A The work will be done faster. 1
B The rest of the crew does it. 2
C I feel pressured because I am overloaded with work. 3
D It improves the quality of my work. 4
E I am not familiar with the applicable procedures. 5
F The procedures do not work as intended. 6
G There are too many procedures. 7
H Others (please specify):
5. Interaction
In relation to following questions safety will be any issue or condition that you feel may threaten or
cause any injury or damage to yourself, your co-workers or the vessel.
A Do you normally work with the same team members
1
2
3
4
5
6
within your working area / working group?

B Do you discuss safety issues with your co-workers? 1
2
3
4
5
6
C Do you ever feel forced to continue your work even

1
2
3
4
5
6
if safety may be threatened?
D Does the crew get positive feedback when they raise
1
2
3
4
5
6
safety issues?
E Can you tell the captain to ”stop”/”time out” if you
1
2
3
4
5
6
feel that safety is threatened?

F Can you say ”stop”/”time out” to the company if you
1
2
3
4
5
6
feel that safety is threatened?

disagree agree know
G The working environment on board is characterized
1
2
3
4
5
6
by openness and dialog.

H We solve problems and conflicts in a good manner. 1
2
3
4
5
6
I We receive sufficient safety-related information

1
2
3
4
5
6
when we start a new watch.

J We receive sufficient safety-related information
1
2
3
4
5
6
when we sign on / start a new sailing period.

K I am confident that my company always prioritizes
1
2
3
4
5
6
the crew’s safety.

L I am confident that the captain always prioritizes the
1
2
3
4
5
6
crew’s safety.
M If I ask for help I will appear incompetent. 1
2
3
4
5
6
N We usually speak up to a co-worker if we notice that

1
2
3
4
5
6
he is doing his work in a risky manner.

O We usually speak up to the ship management if we
notice that a co-worker is doing his work in a risky 1
2
3
4
5
6
manner.
P I stop work if I am not sure that safety is
1
2
3
4
5
6
satisfactorily ensured.
Q I feel appreciated by my co-workers. 1
2
3
4
5
6
R I feel appreciated by the ship management. 1

2
3
4
5
6
S My co-workers do their jobs in a way that makes me

1
2
3
4
5
6
feel safe.
T My co-workers can communicate effectively in
1
2
3
4
5
6
English.
U Different languages on board may represent a safety
1
2
3
4
5
6
risk.
V Different national cultures on board may represent a
1
2
3
4
5
6
safety risk.
6. Work Situation
A Do you have the possibility to prioritize safety first
1
2
3
4
5
6
in your daily work?

B Do you carry out a “Safe Job Analysis”/“Risk
1
2
3
4
5
6
Analysis” before high-risk operations?

C Do you carry out a safety evaluation before new
1
2
3
4
5
6
working methods, tools, or routines are introduced?

D Have you experienced situations where you need to
1
2
3
4
5
6
expose your self to danger to get the work done?

E Do you take a ”time-out” when unforeseen situations
1
2
3
4
5
6
occur?
F Do you use protective equipment in situations when
1
2
3
4
5
6
it is mandatory?
G Do you feel sufficiently rested to carry out your
1
2
3
4
5
6
tasks in a safe manner on your shift?

H Is the safety documentation you need readily
1
2
3
4
5
6
available?
I Is the safety documentation you need up to date? 1
2
3
4
5
6
J Is anyone ever intoxicated/drunk on board? 1

2
3
4
5
6

disagree agree know
K The number of crewmembers is not sufficient to
1
2
3
4
5
6
ensure safety on board.

L The management doesn’t care how we do our work
1
2
3
4
5
6
as long as the work gets done.

MI miss feedback on the work I do. 1
2
3
4
5
6
N The on-board maintenance is sufficient to ensure

1
2
3
4
5
6
safety.
O I am familiar with the on-board safety goals. 1
2
3
4
5
6
P I have to work much overtime to get the work done 1

2
3
4
5
6
7. Competence
A I have received the training that is necessary in
1
2
3
4
5
6
order to work safely.

B I have received the education that is necessary in
1
2
3
4
5
6
order to work safely.

C I have received the training that is necessary in
1
2
3
4
5
6
order to handle critical or hazardous situations.

D I have received the education that is necessary in
1
2
3
4
5
6
order to handle critical or hazardous situations.

E New crew members get a thorough introduction to
1
2
3
4
5
6
safety-related issues.
F On our vessel we frequently carry out drills in safety
1
2
3
4
5
6
procedures.
G What we learn in courses is not relevant in practice. 1
2
3
4
5
6
H Some of my co-workers lack experience. 1

2
3
4
5
6
8. Responsibility & Sanctions

A In my day-to-day work there is no doubt about who
1
2
3
4
5
6
is responsible for the different tasks.

B When an undesirable incident has occurred, people
are more preoccupied with placing blame than 1
2
3
4
5
6
finding the cause of the incident.

C If I violate the safety regulations, there will be
1
2
3
4
5
6
negative consequences for me.

D Vague responsibilities on board contribute toward
1
2
3
4
5
6
creating hazardous situations.

E In my opinion, the consequences for violating the
1
2
3
4
5
6
company’s safety regulations are fair.

F Responsibility for the safety of others is a
1
2
3
4
5
6
motivational factor in the performance of my work.

G I know which tasks I am responsible for if a critical
1
2
3
4
5
6
or hazardous situation should occur.
9. Working Environment
A I enjoy my job. 1
2
3
4
5
6
B I feel sure that I will not loose my job. 1

2
3
4
5
6
C I feel that the work we do on board is too little 1

2
3
4
5
6
appreciated by the company.
D This company is a good employer compared to 1
2
3
4
5
6
others.
E I have too little influence on my working situation. 1
2
3
4
5
6
F The working situation is less physically challenging 1

2
3
4
5
6
than 2 years ago.
G The working situation is less mentally challenging 1
2
3
4
5
6
than 2 years ago.
H The safety delegates have an important role in 1
2
3
4
5
6
ensuring safety at my work site.
I I feel certain that I will not be exposed to an injury/ 1
2
3
4
5
6
accident at my work site.
10. Learning from Incidents

If accidents or severe incidents happen on board, I believe they happen because….
Do not select more than 3 alternatives.
A The crew has a large work load. 1
B The crew does not feel enough responsibility for their tasks. 2
C The crew lacks knowledge and experience in relation to the job they are doing. 3
D There is no tradition for speaking up when someone is working in a hazardous manner. 4
E There are too many interruptions in the work. 5
F Procedures/best practice is not followed. 6
G There are inadequate instructions for using technical equipment. 7
H There are mistakes or deficiencies in the procedures. 8
I There is bad maintenance. 9
J There is defective equipment. 10
K Others (please specify):
Only select one answer per question
A During the last 2 years, have you been involved in a serious
incident/accident?
Yes No
1 2
B During the last 2 years, have any of your co-workers been

involved in a serious incident/accident?
Yes No I don’t know
1 2 3
If yes, please comment the last incident / accident that happened:
C During the last 2 years, have you been involved in what

was almost a serious incident/accident?
Yes No
1 2
D During the last 2 years, have any of your co-workers been

involved in what was almost a serious incident/accident?
Yes No I don’t know
1 2 3
If yes, please comment the last episode that happened:

E Do minor incidents get reported in writing? 1
2
3
4
5
6
F Do close calls get reported in writing? 1

2
3
4
5
6
G Are reports of undesirable incidents ever “fixed up”

1
2
3
4
5
6
to cover mistakes?
H Do you receive constructive feedback from the
1
2
3
4
5
6
company on the conditions you report?

I Do you receive constructive feedback from the
1
2
3
4
5
6
captain on the conditions you report?

J Do you get information from incidents/accidents on
1
2
3
4
5
6
other vessels?
disagree agree know
K Here it is seldom improvements are made before
1
2
3
4
5
6
something has gone wrong.

L Reporting is important to prevent the recurrence of
1
2
3
4
5
6
accidents or incidents.
M Most of all, I report incidents because I have to. 1
2
3
4
5
6
N In my company, they are more preoccupied with the

statistics than the human consequences of an 1
2
3
4
5
6
incident.
O Reporting in itself take too much time. 1
2
3
4
5
6
Assume that you were involved in an incident. Would some of the following conditions stop you
from reporting the incident? Do not mark more than 3 alternatives.
A There is no tradition for reporting all incidents that happen. 1
B No improvements ever happen based on the reports. 2
C The incident didn’t have any serious consequences. 3
D I am afraid that the information will be used against me. 4
E I am afraid that the information will be used again my co-

5
workers.
F This could cause the company to loose contracts. 6
I There could be negative reactions from my co-workers. 7
J I don’t feel comfortable discussing my actions/mistakes. 8
K We have too much to do and don’t have time to write reports. 9
L Mistakes I make don’t concern anyone but me. 10
M I don’t know how to report an incident.

11

N Other (please specify):
11. Description of the Organization

How would you describe this organization? Please provide your estimate based on the statements
below. Select only one box per statement.
The work is
The work is
characterized by
characterized by 1
2
3
4
5
6
7
8
9
10
flexibility and
control in detail and
democracy/influence.
overall control.
It is important to It is important to
do what we are 1
2
3
4
5
6
7
8
9
10 be creative and
told. original.
The work is The work is usually

usually performed 1
2
3
4
5
6
7
8
9
10
performed as a
on an individual team.
basis.
12. Risk perception
How would you assess the risk involved with your work? Please provide your estimate based on
the statements below. Select only one box per statement..
A) All in all, how would you assess the safety in your working situation?
Very 1
2
3
4
5
6
7
8
9
10 Very
bad good
B) All in all, how do feel that the level of safety has developed over the last two years?
It is 1
2
3
4
5
6
7
8
9
10
It is
much much
poorer better
C) All in all, how likely is it that you will have an accident on the vessel during the next 12
months?
Very 1
2
3
4
5
6
7
8
9
10 Very
likely unlikely
D) All in all, how likely is it that any of the other crew members will have an accident on the vessel
during the next 12 months?
Very 1
2
3
4
5
6
7
8
9
10 Very
likely unlikely
E) How safe do you feel when you consider the risk involved with your work on board?
Very 1
2
3
4
5
6
7
8
9
10 Very
safe unsafe
F) How much do you worry when you consider the risk involved with your work on board?
Very 1
2
3
4
5
6
7
8
9
10 Very
much little
13. Suggestions for New Safety Actions or Other Comments
View publication stats

Safety Culture and Safety Management Within The Norwegian-Controlled Shipping Industry State of Art, Interrelationships, and in Uencing Factors

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Safety Culture and Safety Management Within The Norwegian-Controlled Shipping Industry State of Art, Interrelationships, and in Uencing Factors

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Safety culture and safety management within the Norwegian-controlled

Thesis · October 2011

RISKOP Managing Risk in Offshore Operations View project

The user has requested enhancement of the downloaded file.

Thesis submitted in fulfillment of

Faculty of Social Sciences

©2011 Helle A. Oltedal

My interest in the shipping industry started while working on a project in

Haugesund, April 2011, Helle A. Oltedal

This research focuses attention on safety challenges within the Norwegian

The study results indicate several deficiencies in all parts of a traditional

The most prominently identified organizational influential factors are the

Oltedal, H. & Wadsworth, E. (2010). Risk perception in the Norwegian

Oltedal, H. A. (2010). The use of safety management systems within the

Oltedal, H. & McArthur, D. (2010). Reporting practices in merchant shipping,

Oltedal, H. A., & Engen, O. A. (2010). Safety Management in Shipping—

List of articles in thesis: ............................................................................... viii

1 General background and introduction ..................................................1

1.1 Safety culture and safety management within shipping.................... 2

1.2 Research aims and research questions .............................................. 5

1.3 Aims of articles ................................................................................. 7

2 Safety responsibilities in maritime industry ........................................11

2.1 The International Safety Management (ISM) Code ........................ 13

2.2 Maritime administrations and responsibilities ................................ 17

The flag state ........................................................................................... 18

The Port State Control ............................................................................ 20

The company and crew management ...................................................... 23

2.4 Safety responsibilities in the maritime industry—a summary ........ 27

3.1 Safety culture as an organizational integrated concept ................... 33

3.2 Organizations, management and cultural change ........................... 37

3.3 Organizational culture and safety management .............................. 41

The Man Made Disaster model ............................................................... 43

Normal Accident Theory ........................................................................ 45

High Reliability Organizations Theory ................................................... 47

Managing risk and safety culture ............................................................ 49

The theory of Practical Drift – why organizations fails .......................... 51

3.4 Safety culture and measurable outcome variables .......................... 53

4 Research methodology ...........................................................................56

4.1 Quantitative research and questionnaire survey.............................. 58

Questionnaire development .................................................................... 60

Survey sample and respondents’ demographics ..................................... 61

Validity and reliability through theoretical conceptualization ................ 62

Validity and reliability through factor analysis....................................... 66

Validity and reliability through scale analysis ........................................ 68

Causal relationship through structural equation modeling ..................... 70

Document study ...................................................................................... 71

Case studies............................................................................................. 73

Participatory observation ........................................................................ 74

Participation in maritime forums ............................................................ 75

4.3 Applied methods and statistics in articles ....................................... 76

5 Research results .....................................................................................78

5.1 Summary and results of article 1 .................................................... 78

5.2 Summary and results of article 2 .................................................... 80

5.3 Summary and results of article 3 .................................................... 83

5.4 Summary and results of article 4 .................................................... 85

5.5 Summary and results of article 5 .................................................... 87

5.6 Summary and results of article 6 .................................................... 89

5.7 Causal relationships between components of safety culture ........... 91

5.8 Summing up and presentations of main conclusions ...................... 98

7 Concluding remarks ............................................................................107

7.2 Theoretical limitations .................................................................. 109

7.3 Future research .............................................................................. 109