Professional Documents
Culture Documents
Wahyuni, Dina 2012, The research design maze: understanding paradigms,cases,methods and
methodologies, Journal of applied management accounting research, vol. 10, no. 1, Winter
2012, pp. 69-80.
http://hdl.handle.net/10536/DRO/DU:30057483
Introduction
Research Note
Research paradigms address the philosophical
The Research Design dimensions of social sciences. A research
paradigm is a set of fundamental assumptions
Maze: Understanding and beliefs as to how the world is perceived
Paradigms, Cases, which then serves as a thinking framework that
guides the behaviour of the researcher (Jonker
Methods and and Pennink 2010).
Methodologies
Although the philosophical backgrounds
Dina Wahyuni* usually remain implicit in most research, they
affect the practice of research. Some writers
(e.g. Berry and Otley 2004; Creswell 2009;
Abstract Saunders, Lewis and Thornhill 2009; Neuman
2011) emphasise that it is important to initially
The present paper introduces the logical question the research paradigm to be applied in
choices available in research conducting research because it substantially
methodologies; which enable the drawing influences how one undertake a social study
of correct inferences to answer the various from the way of framing and understanding
research questions that are asked by social phenomena. Following this suggestion,
accounting researchers. various research paradigms are discussed
below to enable a justification of the
It starts with an overview of research theoretical assumptions and fundamental
paradigms as fundamental beliefs that beliefs underpinning a social research.
affect the ways to conduct social research,
including the choice of a particular Philosophical Dimensions
research methodology. The paper then
details the elements of case study design, The two main philosophical dimensions to
including the justification to choose case distinguish existing research paradigms are
organisations. The sections that follow ontology and epistemology (Laughlin 1995;
present an overview of the required data Kalof, Dan and Dietz 2008; Saunders, Lewis
and collection methods and discuss the and Thornhill 2009). They relate to the nature
methods used to analyse the collected of knowledge and the development of that
data. Considerations regarding research knowledge, respectively. Ontology is the view
quality are also presented. of how one perceives a reality.
69
JAMAR Vol. 10 · No. 1· 2012
of values in the research and the researcher’s research process in the context of particular
stance in relation to the subject studied. The paradigm. These basic beliefs as they relate to
latter refers to a model for undertaking a research paradigms are outlined in Table 1.
Research Paradigms
Fundamental Positivism Postpositivism Interpretivism Pragmatism
Beliefs (Naïve realism) (Critical Realism) (Constructivism)
Ontology: the External, Objective. Exist Socially External,
position on the objective and independently of constructed, multiple, view
nature of reality independent of human thoughts and subjective, may chosen to best
social actors beliefs or knowledge change, multiple achieve an
of their existence, but answer to the
is interpreted through research question
social conditioning
(critical realist)
Epistemology: Only Only observable Subjective Either or both
the view on what observable phenomena can meanings and observable
constitutes phenomena can provide credible social phenomena and
acceptable provide credible data, facts. phenomena. subjective
knowledge data, facts. Focus on explaining Focus upon the meanings can
Focus on within a context or details of provide
causality and contexts situation, the acceptable
law-like reality behind knowledge
generalisations, these details, dependent upon
reducing subjective the research
phenomena to meanings and question. Focus
simplest motivating on practical
elements actions applied research,
integrating
different
perspectives to
help interpret the
data
Axiology: the Value-free and Value-laden and etic Value-bond and Value-bond and
role of values in etic emic etic-emic
research and the Research is value
researcher’s Research is laden; the researcher Research is Values play a
stance undertaken in a is biased by world value bond, the large role in
value-free way, views, cultural researcher is interpreting the
the researcher is experiences and part of what is results, the
independent of upbringing being researcher
the data and researched, adopting both
maintains an cannot be objective and
objective stance separated and so subjective points
will be of view
subjective
Research Quantitative Quantitative or Qualitative
Quantitative and
Methodology: qualitative qualitative
the model (mixed or multi-
behind the method design)
research
process
Based on Saunders et al.(2009, p.119),Guba and Lincoln (2005), and Hallebone and Priest (2009)
70
JAMAR Vol. 10 · No. 1· 2012
71
JAMAR Vol. 10 · No. 1· 2012
The importance of and the debate over analyse data. However, a research method is
research paradigms has been recorded in the a-theoretical (Sarantakos 2005), that is, it is
accounting literature (see, for instance, independent from methodologies and
Laughlin 1995; Ahrens 2008; Kakkuri- paradigms. Therefore, a research method, e.g.
Knuuttila, Lukka and Kuorikoski 2008; an interview, can be used in different research
Kaidonis, Moerman and Rudkin 2009; Lukka methodologies. In other words, a method is a
2010). Accounting issues, which typically deal practical application of doing research whereas
with number crunching, are predominantly a methodology is the theoretical and
studied within the positivist paradigm. This ideological foundation of a method. A research
strand of strongly numerical studies forms the design then becomes important to connect a
basis of the so-called mainstream research in methodology and an appropriate set of research
accounting. Using an objective viewpoint and methods in order to address research questions
the hypothetico-deductive method, such and or hypotheses that are established to
research analyses large archival data sets in examine social phenomena.
order to provide explanations for social
phenomena. Research Designs
The shift from positivist to subjective thinking Research purpose and research questions are
in accounting research (Laughlin 1995), the suggested starting points to develop a
including from quantitative to qualitative research design because they provide
research (Parker 2012), became apparent in important clues about the substance that a
1970s. In particular, the application of researcher is aiming to assess (Berry and Otley
qualitative research has gained popularity in 2004; Saunders, Lewis and Thornhill 2009;
the management accounting arena (Parker Yin 2012).
2012). By investigating the subjective
meanings of social phenomena and the A research method that facilitates a deep
resultant behaviour, interpretive accounting investigation of a real-life contemporary
researchers have sought to provide a better phenomenon in its natural context is a case
understanding of functioning accounting study (Woodside 2010; Yin 2012). There are
practices. This category is called non- three ideal conditions for conducting a case
mainstream or alternative research. study advocated by Yin (2009) in comparison
to the other research methods in social
Research Methodologies and sciences: experiment, survey, archival analysis
Methods and history. First, the form of asking a research
question is in the form of why or how. The
At this point it should be noted that research remaining two conditions are that no control is
methodology and research method are required over behavioural events being studied
distinctive concepts. Analogically, a and the study focus on contemporary events.
methodology is a domain or a map, while a Thus, a case study should be of a
method refers to a set of steps to travel contemporary event as opposed to a historical
between two places on the map (Jonker and one.
Pennink 2010). A methodology refers to a
model to conduct a research within the context Ideally case study research should use a
of a particular paradigm. It comprises the multiple case study design involving
underlying sets of beliefs that guide a multi-sites to be studied and using multiple
researcher to choose one set of research methods to analyse the collected data. The
methods over another. Because methodologies rationale behind the choice of a multiple case
are closer to research practice than the study over a single case study is to enable
philosophical concepts found in paradigms, comparisons between the observed practices
many researchers commonly state that they are by subjects studied in order to obtain a more
conducting ‘qualitative’ instead of comprehensive understanding of these
‘interpretivist’ research (Sarantakos 2005). practices.
72
JAMAR Vol. 10 · No. 1· 2012
Multi-method1 qualitative research refers to Gaining approval from the targeted firms to
using more than one data collection techniques join as participants in a case study research
and applying multiple methods to analyse these appeared to be one of the most challenging
data using non-numerical (qualitative) tasks in this research. Often, an email
procedures to answer the research question. containing an official letter from the supervisor
Here, the investigation should result in both a requesting access turns out to be unsuccessful.
descriptive model as to how the world is, and Sensitivity of the information to be studied
prescriptive suggestions as to how the world seemed to be the factor that makes these firms
should be. Thus, case study research should hesitant to talk to the researcher. Informal links
ideally be performed through a two-stage case should instead be chosen. Often such links are
study with the use of both qualitative and more favourable and may in fact yield referrals
quantitative data collected from multiple to more firms to be studied.
sources, and predominantly qualitative analysis
procedures applied in a sequential design. This practice of obtaining a studied subject via
a referral system is known as snowball or
Selection of Case Organisations networking method. As noted by Bryman
(2012), qualitative research has practically
It is very important to be mindful about the applied more than one method of sampling, for
trap of selecting a representative case solely to instance by selecting a sample purposively
provide a basis for generalisation (Scapens which is followed by using the snowball
2004). Unlike positivist research that uses technique to obtain studied subjects.
frequencies and statistical generalisation to
relate its findings to a larger population, an Data Collection
interpretive case study focuses on analytical
generalisation to develop and extend theory. Data is collected in form of primary and
Therefore, the selection process should be secondary data. The primary data is usually
driven by the research question as it provides collected using semi-structured interviews with
the characteristics of the cases to be studied. the experts in the observed topic from the case
organisations. As suggested by Parker (2003),
Following the suggestion by Scapens (2004), a qualitative researchers should get involved in a
critical case or an extreme case should be communication with the practitioners in the
sought for a representative case. The former is organisational coal-face in order to better
a case in which the social phenomena being understand the current state of
observed are some critical events that cause the real-world practices. The secondary data
research questions asked to become important constitute internal publications provided by
in the organisation. The latter is mainly chosen participants to the researchers and publicly
to test a theory or to broaden the application available data which are relevant to the topic
into a wider range of circumstances. For being observed. This method of collecting data
instance, a case study that tests if a theory from multiple sources, termed data
works on a very small sized firm is an example triangulation (Patton 2002), assists the
of an extreme case. This method of choosing a researcher not only to collect more
case according to pre-determined reasons is comprehensive relevant information but also to
known as non-probability sampling. Here, a cross-check their consistency in order to
purposive sampling is exercised to specifically enhance the robustness of findings.
pick information-rich cases on the basis of
their matched criteria to the ones required to Semi-Structured Interviews
answer the research questions being asked The main feature of an interview is to facilitate
(Bloor and Wood 2006). the interviewees to share their perspectives,
stories and experience regarding a particular
social phenomena being observed by the
interviewer. The participants, who are the
1
Multi-method approach is different from mixed-
practitioners in their field, will pass on their
method research, which uses both quantitative and
qualitative data collection and analysis procedures
knowledge to the researcher through the
that are applied either in a concurrent or in a conversations held during the interview
sequential design (Saunders, Lewis & Thornhill process (Boeije 2010). The interview method
2009). is most often selected as the main method for
73
JAMAR Vol. 10 · No. 1· 2012
collecting empirical data of the relevant and unexpected thoughts provided by the
practices. The interview procedures, interviewees. The probes should be used
encompassing all procedures from 1) designing (prepared ideally in advance) not only to keep
the interview questions and developing the the discussion flowing, but also to clarify some
interview guides, to 2) the process of discussion points by asking for more details or
interviewing itself, are discussed below. examples of what had been said. Once the
Design and development of interview questions had been determined, a peer who has
questions: A semi-structured interview, also considerable experience in conducting
known as the non-standardised or qualitative interviews should be asked to check them to
interview (Saunders, Lewis and Thornhill ensure that there are no wording questions that
2009), is a hybrid type of interview which lies might could lead to predetermined answers.
in between a structured interviews and an in-
depth interviews. Therefore, it offers the merit Prior to conducting the formal interview, the
of using a list of predetermined themes and researcher should hold mock interviews with
questions as in a structured interview, while colleagues to fine-tune the research instrument.
keeping enough flexibility to enable the As a result, some expressions and words could
interviewee to talk freely about any topic be changed to make the questions clearer.
raised during the interview. The use of an Often the structure of the main questions is
in-depth qualitative interview is considered as reordered to improve the flow of the discussion
the appropriate format for case study research during the planned interviews. This instrument
because in-depth questions cannot be answered should now be submitted for ethics clearance
briefly. It is anticipated that the researcher from the researcher’s organisation.
would need to ask for examples or more
explanation on the answer given in order to
gain a deep understanding of the issues. The Interviews: After obtaining the ethics
clearance from the Human Research Ethics
Rubin and Rubin (2005) provide a guidance to Committee at the researcher’s organisation, the
develop the interview questions and interviews can be conducted. As suggested by
procedures. They have created a model of an Kvale and Brinkmann (Kvale and Brinkmann
in-depth qualitative interviewing, termed 2009), the interview should be framed by a
‘responsive interviewing’ (Rubin and Rubin briefing before the interview commenced, and
2005, p.20), which is heavily underpinned by a debriefing afterwards. Ideally, a research
the interpretive research philosophy. This information package should be provided in the
approach emphasises the importance of introductory meeting or sent beforehand in an
keeping the research design and questioning email. The researcher starts off the interview
flexible and adaptive in order to facilitate new by briefly explaining the aim of the interview
information to emerge or to adapt to an and emphasizing the confidentiality,
unexpected direction. anonymity and the voluntary nature of the
study. The interviewee is then given a consent
Responsive interviewing concentrates on form which should be signed off by both this
obtaining a deep understanding, rather than person and the researcher. With the
breadth, about the investigated topic (Rubin participant’s permission, each interview should
and Rubin 2005). The interview questions are be recorded.
structured to include open-ended main
questions, follow-up questions and probes. The Besides recording the interview, the researcher
main questions should be carefully developed should also take notes during and soon after
based on the research problem and the research each interview to record additional information
questions in such a way that there are separate in the form of research memos. There are three
interview questions for each part. Relevant types of memos or notes that can be recorded
articles, webpage publications, and industrial for an interview: observational,
research should be explored to gather ideas methodological and theoretical (Schatzman
about relevant practices to be included in the and Strauss 1973). Observational memos, also
interview questions. known as field notes, are used to describe the
situation during the interview. Methodological
Follow-up questions should be developed to memos are the records of any issues and
explore the particular themes, concepts, ideas concerns regarding the methods used.
74
JAMAR Vol. 10 · No. 1· 2012
Theoretical memos focus on what themes and Transcribing Recorded Interviews: The
findings emerged from the interview process. transcribing task is often outsourced to a
professional transcriber. Once transcribed,
Each interview should last no more than one each one and a half hour interview is usually
and a half hours. After each interview, a transformed into around 35 pages of text.
debriefing should be performed during which After receiving the transcription output, it
the practitioners are given the opportunity to should be checked against the voice recording
ask questions, make comments or add any for accuracy. The researcher will be interested
information that was not discussed during the in the content of the interview, so checking the
interview. During this visit, the researcher accuracy of the transcript’s content is
should ask for relevant documents by the considered to be crucial. The parts of
practitioners, such as company magazines and transcripts containing linguistic details, such as
sustainability reports. Such secondary data is laughter, should be deleted. This approach
obtained to triangulate findings in order to which concentrates more on content and less
answer the research questions. Relevant data on actual expressions is termed denaturalised
will usually be both qualitative and transcription (Oliver, Serovich and Mason
quantitative in nature that includes a wide 2005). Because the researcher needs to decide
range of relevant data for the case what is to be included and excluded in
organisations. transcription and becomes familiar with the
content of the interview, transcribing can be
Data Analysis seen as the initial step in data analysis (Miles
and Huberman 1994; Kvale and Brinkmann
Data analysis involves the drawing of 2009).
inferences from raw data. Data analysis can
involve multi-methods that are applied Cleaning Data: Considering the ethics
sequentially. Multi-method application in concerns about anonymity and confidentiality,
conducting research is called methodological all information that can identify both the
triangulation (Patton 2002). Each of these steps practitioners and the case organisations that
is discussed below. they represent should be omitted. The data will
be identified by a specific coding, e.g. Firm 1
Data Preparation (F1), to the information provided by Firm 1.
Raw data, which is the format as they are This identifiable information is intended to be
generated, need to be managed so that they are used for data analysis only, to enable a
ready to be analysed (Boeije 2010). Different comparison of the findings between firms
from data generated from quantitative research using the constant comparative analysis
which is mainly numerical; data collected in method. In most case research, replacing the
qualitative research are primarily text-based. identity of case organisations with unique
Data management in such qualitative research codes is particularly important because the
involves three important aspects: data storage, interview texts would be read by a peer in a
transcribing audio sources, and cleaning the later stage of data analysis to check the
data. consistency of coding.
75
JAMAR Vol. 10 · No. 1· 2012
findings requires interpretation of empirical regrouped based on their relevant content into
data. categories. This categorising step is termed
axial coding. Finally, selective coding was
A common approach to the interpretation of conducted by making logical connections
meanings from textual data is using content between the core categories to make sense of
analysis. This technique has been used by both understanding what has been really happening
quantitative and qualitative researchers in the in the observed practices.
social sciences, including accounting (Milne
and Adler 1999; Sarantakos 2005). However, Ideally, coding is also performed iteratively.
they apply different focuses. Quantitative Coding is initially conducted right after
researchers transform qualitative information removing case organisations’ identity from
into numerical numbers. They establish a set of text. Then, the textual data is re-examined for
categories and then count the number of developing further interview questions for each
instances that utterances fall into each subsequent meeting with case organisations.
category. The characteristics of categories need All findings from each coding process should
to be defined clearly in order to allow other be recorded in a codebook as a part of interim
researcher to draw similar results from the summaries. As suggested by Saunders et al.
texts, e.g. annual reports (Guthrie and Parker (2009), interim summaries are used to record
1990). Conversely, qualitative contents the progress to achieve conclusions, including
analysis concentrates on portraying reality by what had been found so far and what needed
discovering meanings from the textual data to be done to improve the quality of findings
(Silverman 2011). Consistent with the or to find alternative explanations. The use of a
paradigm used, qualitative content analysis is draft codebook is inspired by Boyatzis (1998)
applied in this study from the perspective of and Neuendorf (2002). Here, each code is
case organisations (Sarantakos 2005), which given a label, a definition or description to
means using the emic or an insider’s approach guide how to apply the code, and an example
to view the practices of the case organisation. of the texts. Hierarchical numbering should be
assigned to codes to show the relationships
Qualitative content analysis which is done between codes. Applying coding analysis
through identifying patterns and themes within technique more than one times to the similar
data is termed thematic analysis (Given 2008). texts results in the refinement of codes.
The constant comparative method follows
similar principles to thematic analysis in In addition, peer debriefing (Long and Johnson
extracting themes from within texts. However, 2000) could be performed to test code
the constant comparative method focuses more reliability. The researcher should then discuss
on describing variation in different emerging findings and how the coding
circumstances of social phenomena (Boeije assigned in the text with a colleague who has
2010). It provides a more systematic way to experience in using the constant comparative
identify any difference that emerges in method. All the concerns and
empirical data (see Boeije 2002). Therefore, recommendations from peer debriefing are
the constant comparative method is preferred if then used as a basis to further refine the
the objective is to reveal important concepts, codebook. This method of using more than one
processes, and the overarching professional evaluator in examining one particular topic is
experiences between the case organisations. termed investigator triangulation (Patton
2002). Keeping the progress records of interim
In practice, qualitative content analysis uses a summaries and of a codebook is often found
coding method. Coding simply means very useful in drawing conclusions in case
labelling. It refers to the assignment of a code study research.
representing the core topic of each category of
data. As applied in the grounded research Research Quality
approach, coding in usually undertaken on
three levels: open coding, axial coding and The rigour of qualitative research has been
selective coding (Boeije 2010). Open coding is subject to continuous discussion in the
conducted by dismantling texts and literature (see, for example Denzin and Lincoln
distinguishing different themes and concepts 2005; Boeije 2010). Qualitative research has
found in the data. These pieces of data are then been criticised as lacking generalisibility by its
76
JAMAR Vol. 10 · No. 1· 2012
77
JAMAR Vol. 10 · No. 1· 2012
78
JAMAR Vol. 10 · No. 1· 2012
Guba, E.G. and Lincoln, Y.S. (2005), Critical Effectiveness in Nursing, 4(1): pp. 30-
“Paradigmatic Controversies, Contradictions, 37.
and Emerging Influences”, in Denzin, N.K.
and Lincoln, Y.S. (eds), The Sage Handbook of Lukka, K. (2010), “The roles and effects of
Qualitative Research, 3rd edn, Sage paradigms in accounting research”,
Publications, London. Management Accounting Research,
21(2): pp. 110-115.
Guthrie, J. and Parker, L.D. (1990), “Corporate
social disclosure practice: A comparative Miles, M.B. and Huberman, A.M. (1994),
international analysis”, Advances in Public Qualitative Data Analysis: An Expanded
Interest Accounting, 3: pp. 159-175. Sourcebook, 2nd edn, Sage Publications,
Thousand Oaks, California.
Hallebone, E. and Priest, J. (2009), Business
and Management Research: Paradigms and Milne, M.J. and Adler, R.W. (1999),
Practices, Palgrave Macmillan, New York. “Exploring the reliability of social and
environmental disclosures content analysis”,
Hennink, M.M., Hutter, I. and Bailey, A. Accounting, Auditing and Accountability
(2011), Qualitative Research Methods, Sage Journal,
Publications, London. 12(2): pp. 237-256.
Jonker, J. and Pennink, B. (2010), The Essence Neuendorf, K.A. (2002), The Content Analysis
of Research Methodology: A Consice Guide Guidebook, Sage Publications, Thousand
for Master and PhD Students in Management Oaks, California.
Science, Springer, Heidelberg.
Neuman, W.L. (2011), Social Research
Kaidonis, M., Moerman, L. and Rudkin, K. Methods: Qualitative and Quantitative
(2009), “Paradigm, paradox, paralysis: An Approaches, 7th edn, Pearson/Allyn and
epistemic process”, Accounting Forum, Bacon, Boston.
33(4): pp. 285-289.
Oliver, D.G., Serovich, J.M. and Mason, T.L.
Kakkuri-Knuuttila, M.L., Lukka, K. and (2005), “Constraints and opportunities with
Kuorikoski, J. (2008), “No premature closures interview transcription: Toward reflection in
of debates, please: A response to Ahrens”, qualitative reseach”, Social Forces, 48(2): pp.
Accounting, Organizations and Society, 1273-1298.
33(2–3): pp. 298-301.
Parker, L.D. (2003), “Qualitative research in
Kalof, L., Dan, A. and Dietz, T. (2008), accounting and management: The emerging
Essentials of Social Research, McGraw-Hill, agenda”, Journal of Accounting and Finance,
New York. 2: pp. 15-30.
Laughlin, R. (1995), “Empirical research in Patton, M.Q. (2002), Qualitative Research and
accounting: alternative approaches and a case Evaluation Methods, 3rd edn, Sage
for “middle-range” thinking”, Accounting, Publications, Thousand Oaks, California.
Auditing and Accountability Journal,
8(1): pp. 63-87. Rubin, H.J. and Rubin, I. (2005), Qualitative
Interviewing: The Art of Hearing Data, 2nd
Lincoln, Y.S. and Guba, E.G. (1985), edn, Sage Publications, Thousand Oaks,
Naturalistic Inquiry, Sage Publications, California.
Beverly Hills, California.
Sarantakos, S. (2005), Social Research, 3rd
Long, T. and Johnson, M. (2000), “Rigour, edn, Palgrave Macmillan, New York.
reliability and validity in qualitative research”,
79
JAMAR Vol. 10 · No. 1· 2012
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.