You are on page 1of 19

2nd Reading

February 14, 2018 8:47:59am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Journal of Information & Knowledge Management


Vol. 17, No. 1 (2018) 1850008 (19 pages)
.c World Scienti¯c Publishing Co.
#
DOI: 10.1142/S0219649218500089

Relationships Among Knowledge Management,


Organisational Innovativeness and Performance:
Covariance-Based Versus Partial Least-Squares
Structural Equation Modelling

Jose Roberto Frega


School of Business, Universidade Federal do Parana (UFPR)
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

CEP 80210-170, Jardim Bot^ anico, Curitiba, PR, Brazil


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

Alex Antonio Ferraresi and Carlos Olavo Quandt


School of Business
Pontif{cia Universidade Catolica do Paran
a (PUCPR)
CEP 80215-901, Prado Velho, Curitiba, PR, Brazil

Claudimar Pereira da Veiga


School of Business, Universidade Federal do Parana (UFPR)
CEP 80210-170, Jardim Bot^ anico, Curitiba, PR, Brazil
claudimar.veiga@gmail.com

Published 15 February 2018

Abstract. The relationships among e®ective knowledge management (KM), organisational innova-
tiveness (OI), market orientation (MO) and organisational performance (OP) have been explored in the
literature. These constructs are generally analysed in pairs, such as the in°uence of KM on OI, or KM on
OP, and other combinations, but the relationships among the full set of constructs in question are not fully
understood yet. In the extant literature, the relationships among them are analysed for the most part with
covariance-based structural equation modelling (CB-SEM). Partial least-squares (PLS) path modelling is
a component-based approach to SEM that is not as widely used as CB-SEM, but it has the potential to
allow increased °exibility in handling various modelling problems in comparison with CB models,
particularly for predictive and exploratory purposes. This paper aims to verify whether the PLS method
could con¯rm or reject the results of the more restrictive covariance-based method in modelling the
relationships among KM, OI, MO and OP. The results indicate that both methods yielded convergent and
discriminant validity for the constructs, displaying stability across model analysis and depuration. The
PLS model revealed the in°uence of KM on MO, OI and OP. It also shows that OI is the main driving
factor for OP. KM seems to have a direct e®ect on OP, which is greatly magni¯ed when mediated by
OI. The sample size, although borderline adequate for the CB method, was more than adequate for PLS,
yielding excellent model stability.

Keywords: Knowledge management; innovativeness; performance; PLS; market orientation.

1. Introduction
The analysis of relationships among latent variables is a well-established ¯eld of
study, and one of the most widely applied methodologies is covariance-based

1850008-1
2nd Reading
February 14, 2018 8:47:59am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

structural equation modelling (CB-SEM). Taking the model proposed by Ferraresi


et al. (2012) as a starting point, this study approaches the issue with a partial least-
squares (PLS) method and compares its ¯ndings with the CB method that was used
originally. PLS path modelling is a component-based approach to SEM that allows
increased °exibility in handling various modelling problems, thus providing a
potentially more capable alternative to the more restrictive, traditionally used
covariance-based method (Hair et al., 2011). The goal is to verify whether the PLS
method could con¯rm or reject the results of the more restrictive covariance-based
method in modelling the relationships among e®ective knowledge management
(KM), organisational innovativeness (OI), market orientation (MO) and organisa-
tional performance (OP). Ferraresi et al. (2012) found signi¯cative relationships
among the above-mentioned constructs, which are now tested using the PLS
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

approach.
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

The objective of this study is to identify the in°uence of KM on OI, MO and OP,
using partial least-squares structural modelling. These constructs are generally
analysed in pairs, such as the in°uence of KM on OI, or KM on OP, and other
combinations, but the relationships among the full set of constructs in question are
not fully understood yet. In the extant literature, relationships among part of them
have been analysed, predominantly with CB-SEM. For example, Liao and Wu
(2010) analysed the impact of KM on organisational learning and organisational
innovation using CB-SEM, whereas Abdi and Senin (2015) analysed the same
constructs using PLS path modelling, and Shih and Chou (2012) also used PLS for
an exploratory analysis of the e®ect of KM on organisational performance. Zheng
et al. (2010) explored the role of KM on organisational culture, structure, strategy
and organisational e®ectiveness using CB-SEM, while Tajeddini et al. (2006) ex-
amined the e®ect of market orientation on innovativeness, and Rapp et al. (2008)
analysed the in°uence of market orientation on e-business innovational performance,
also using CB-SEM.
In sum, this study was motivated by two issues: the ¯rst is a potential contri-
bution for a more comprehensive understanding of the relationships among KM,
OI, MO and OP. The second is to explore the potential of the PLS approach in
tackling various modelling problems, particularly for predictive and exploratory
purposes, given the preponderance of CB-SEM in the social science disciplines.
Although both approaches were developed at about the same time, the initial
dominance of CB-SEM is likely linked to the LISREL software, which was already
available in the 1970s (Sarstedt et al., 2014). Hence, CB-SEM established itself as
the primary method for estimating and testing structural equation modelling, and its
methodological advances and applications expanded quickly, in contrast with the
PLS-SEM alternative. Indeed, a study by Richter et al. (2016b) of international
business articles from 1990 to 2013 shows that, among studies that applied structural
equation modelling, 379 utilized CB-SEM and only 45 used PLS-SEM. Furthermore,
they point out that the choice of PLS-SEM is still chie°y due to sampling and data

1850008-2
2nd Reading
February 14, 2018 8:48:00am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

measurement issues, as opposed to its bene¯ts in terms of predictive and exploratory


purposes.
Although the use of PLS-SEM in strategic management research remains
relatively low, Hair et al. (2012, p. 323) point out that it \has grown linearly as a
function of time, which is typical for early introduction stages of a new research
technique's di®usion." A study of leading marketing journals (Hair et al., 2017)
between 1986 and 2015 shows that PLS use has gained momentum relative to factor-
based SEM in recent years. Hence, taking into account the exploratory nature of this
study and the relative complexity of the relationships among the constructs, as well
as the relatively small sample size, the PLS approach emerges as a valuable alter-
native technique for such purpose, since its results could con¯rm or reject previous
¯ndings using covariance-based models.
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

2. Research Constructs
The relationship between knowledge management and organisational performance
has been the object of various studies, as in Massey et al. (2001), Almashari et al.
(2002), Darroch (2003), Zheng (2005), Shih and Chou (2012) and Rezaei et al.
(2017). It has been pointed out that the primary goal of knowledge management is to
improve the ¯rm's innovation capacity, therefore increasing its competitiveness in a
turbulent environment with rapid changes (Nonaka and Takeuchi, 1995; Probst
et al., 2000; Liao and Wu, 2010; Abdi and Senin, 2015). According to other authors
(Botha, 2001; Hooley et al., 2001; Arthur, 1999), knowledge management can in-
crease the ability to monitor the marketplace, its nuances and changes and thus to
enable companies to anticipate those changes, and consequently to respond more
e®ectively to customer demands, and also to open new business opportunities.

2.1. Knowledge management


According to Quinn et al. (1998), the development of knowledge-based factors is
imperative for value creation in products and services. Those factors include cus-
tomer, process and market knowledge. A learning process oriented to creativity and
innovation is necessary to achieve this goal. Such reasoning is closely related to the
concept of market orientation (Day, 1999). A market-oriented ¯rm acquires skills to
attract and retain clients by developing an internal culture that promotes the un-
derstanding of superior value and commitment to it. Kohli and Jaworski (1990,
1993) proposed a view of market orientation as the creation of intelligence, its
dissemination and application throughout the organisation. The focus of this intel-
ligence is the analysis of the factors that a®ect behaviours and trends that may
in°uence customer needs and desires, in order to target the anticipation of those
needs. Slater and Narver (1995) related market orientation directly with organisa-
tional learning, thus incorporating implicitly the notion of knowledge management.

1850008-3
2nd Reading
February 14, 2018 8:48:00am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

In this study, e®ective knowledge management can be assessed in its three


proposed dimensions, which are: (i) knowledge creation and acquisition;
(ii) knowledge sharing, conversion, organisation and dissemination; and (iii) the use
and application of knowledge in organisational processes. In this study, the variables
and scales to measure this construct are adapted from the model developed by Gold
et al. (2001). After adaptation and pre-testing, the scale designed to measure the KM
construct included 11 questions on the knowledge acquisition process, 10 questions
related to knowledge conversion and sharing and nine questions on knowledge
application.

2.2. Market orientation


by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

Market orientation stems from a culture that e®ectively promotes the necessary
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

behaviours for superior value generation for the client and, consequently, sustainable
superior performance for the business (Narver and Slater, 1990; Gao, 2017). Such
strategic orientation provides a superior ability to compete (Zhou et al., 2005).
Market orientation also provides a strong \positional advantage" albeit within the
context of other capabilities, namely entrepreneurship, innovativeness and organi-
sational learning, as demonstrated by Hult and Ketchen (2001). Following that
perspective, MO is identi¯ed by its three dimensions: (a) customer orientation;
(b) competitor orientation; and (c) inter-function coordination. The ¯rst two deal
with information acquisition, its conversion into knowledge and the spreading of
this knowledge all over the organisation. The last one applies the knowledge to
coordinate the organisational e®orts to create superior value for the customer.
The established MKTOR scale (Narver and Slater, 1990) was used in this survey.
It comprises six items for customer orientation, four related to competitor orienta-
tion and four related to inter-functional coordination, totalling 14 questions.

2.3. Organisational innovativeness


In this study, organisational innovativeness is treated as a proxy for innovation,
as the former is de¯ned as the organisation's will and ability to adopt, adapt or
implement new technologies, processes and ideas, so it is able to o®er new and unique
products and services before its competitors. It has been argued that such disposition
is based largely on its organisational culture, as a system of beliefs and behaviours
that leads to this attitude (Hurley and Hult, 1998; McDonald, 2002; Pučėtaitė et al.,
2016). In this sense, organisational innovativeness refers to the ability or propensity
of the company to innovate or develop new products (Garcia and Calantone, 2002;
Andreassi and Sbragia, 2004; Lynch et al., 2010). Innovativeness means to be open to
new ideas, making this behaviour an intrinsic part of organisational culture (Hurley
and Hult, 1998). There are several studies about the factors that lead to innova-
tiveness in the organisation, yet there is still no consensus on how it can be related to
the organisational performance (Tajeddini, 2011; Tajeddini et al., 2006). For the OI
construct, this study adopted the ¯ve-item scale introduced by Capon et al. (1992).

1850008-4
2nd Reading
February 14, 2018 8:48:00am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

The scale is focussed on the \openness to innovation" aspects, where cultural values
and beliefs of innovativeness are formed and acted on to achieve strong long-term
performance, as noted by Hult and Ketchen (2001).

2.4. Organisational performance


As noted above, organisational performance has been discussed together with one or
more of the constructs involved in the present research by several authors, including
Kohli and Jaworski (1990), Narver and Slater (1990), Deshpande et al. (1993), Slater
and Narver (1995), Pelhan and Wilson (1996), Almashari et al. (2002), Darroch
(2003), Tajeddini et al. (2006), Rapp et al. (2008), Liao and Wu (2010), Zheng et al.
(2010), Tajeddini (2011), Abdi and Senin (2015), among others. Following these
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

authors, the methods to assess the organisational performance or results can be


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

divided into two groups: one that uses objective ¯nancial criteria and another ap-
proach using non-¯nancial (or non-objective) criteria. Some studies use mixed cri-
teria. The measures include market share, percentage of sales from new products or
services and rate of return on investments (ROI), as applied by Tajeddini et al.
(2006), among many others. OP also includes measures to assess the organisation's
internal aspects, such as those related to the streamlining of internal processes and
the reduction of response time to market changes, as proposed by Gold et al. (2001)
in their measures of organisational e®ectiveness.

3. Method
PLS modelling has gained much acceptance recently, despite the lack of overall
¯t indexes regarding the proposed models and the actual observed data. PLS con-
siders only aspects such as average variance extracted and R-square index to
assess the impact of exogenous constructs over the endogenous ones, and the ade-
quacy of the manifest variables (indicators) as a valid measure of the constructs
(Chin, 1995).
PLS estimation has this name because \it studies a system of linear relationships
between latent variables by solving blocks (combinations of theoretical constructs
and measurements), one at a time (partial), through the use of interdependent
ordinary least squares (OLS) regressions" (Ferreira et al., 2010). Table 1 shows some
relevant di®erences between the CB approach and the component-based approach
(PLS) to structural models estimation.
One of the relevant aspects of PLS analysis is its relative immunity to the lack of
multivariate normality of the variables, which can have a severe impact on
the results of structural equation modelling methods based on covariance matrices
estimation, such as AMOS and LISREL (Monecke and Leisch, 2012; McIntosh
et al., 2014).
Another advantage is related to sample size, which is a much less critical issue in
PLS estimation as in covariance estimation SEM. Several studies report that,

1850008-5
2nd Reading
February 14, 2018 8:48:00am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

Table 1. Comparisons of covariance-based and component-based SEM techniques.

Feature Covariance-based approach Component-based approach

Distributional assumptions Multivariate normal None


distribution
Purpose Theory-oriented Prediction-oriented
Parameter estimates Consistent Consistency at large
Hypothesis testing Available Inference requires Jackknife
or bootstrapping
Sample requirement Large Small
Parameter identi¯cation problems Not convergence or Few
improper solutions
Latent variables scores Factor indeterminacy Explicitly estimated
Re°ective and formative relations Only re°ective relations Both
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

Source: Hsu et al. (2006).


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

for PLS, a rule of thumb of 10 observations per variable provides adequate results,
considering the PLS construct with the greatest number of indicators (Chin, 1995).
However, this number is still under debate among researchers, as noted by
Marcoulides and Saunders (2006), despite the consensus that the PLS estimation is
much less demanding in terms of number of cases per variable.
As for the results of the estimation, it should be noted that a comparison made by
Ferreira et al. (2010) revealed that the structural equations (assessed by R 2 ) are
more signi¯cant with a covariance estimation method (LISREL) than with PLS, and
that the correlations between the indicators (also named manifest variables) and
their constructs (also named latent variables) are, in the majority of cases, stronger
with PLS than with the covariance estimation method (LISREL).

3.1. Population and sample


The target population consists of a list of 6,509 manufacturing and service compa-
nies with more than 200 employees, located in Brazil. The list originated from a
study called Demography of Companies, conducted by the Brazilian Institute of
Geography and Statistics (IBGE). Data were collected through a self-administered
electronic questionnaire. The respondents were directors and managers from
departments such as marketing, R&D, general management, ¯nancial and human
resources, who were targetted as the professionals with better access to the infor-
mation needed for the study. The respondents were invited to participate in the
survey by email and once they accepted the invitations, they were given a password
to access the questionnaire.
The questionnaire consisted of 54 questions using a 10-point Likert-type ordinal
scale, in which \1" corresponded to total disagreement with the question (assertion)
and \10" corresponded to full agreement with the statement in question. The
¯nal dataset consisted of 241 complete and valid questionnaires, the same used

1850008-6
2nd Reading
February 14, 2018 8:48:00am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

by Ferraresi et al. (2012), so the results of the present study may be compared with
the ¯ndings of the original research.

3.2. Construct de¯nition and measurement


The measurements and structural model are depicted in Fig. 1. Each of the proposed
constructs is evaluated by a set of indicator variables, as follows:
F1 : First sub-dimension of knowledge management
. A3 : There are processes to acquire knowledge about our suppliers.
. A7 : There are processes to acquire knowledge about new products/services in
our industry.
. A9 : There is a performance benchmarking process.
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

F2 : Second sub-dimension of knowledge management


. A11 : There are processes to facilitate internal knowledge exchange among
people.
. A2 : There are processes to create or generate new knowledge based on existing
knowledge.
. A4 : Results of ¯nished projects (feedback) are used to learn from errors and
develop better projects.

Fig. 1. Proposed structural model (Ferraresi et al., 2012).

1850008-7
2nd Reading
February 14, 2018 8:48:02am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

F4 :
. A224 : There are processes to absorb individual knowledge in the Organisation.
. A225 : There are processes to absorb knowledge from business partners in the
organisation.
. A227 : There are processes to integrate di®erent sources and types of
knowledge.
. A229 : There are processes to replace obsolete knowledge.

F5 :
. A31 : There are processes to apply knowledge learnt from experience.
. A32 : There are processes to use knowledge to develop new products and
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

services.
. A35 : Knowledge is used to improve e±ciency.
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

F6 :
. B1 : Our business goals are oriented towards client satisfaction.
. B3 : Our strategy to gain competitive advantage is based on our knowledge
about the needs of our customers.
. B5 : Satisfaction of our customers is systematically and frequently measured.

F7 :
. B10 : The upper echelon of our company discusses in a regular basis the
strengths and strategies of our competitors.
. B7 : Our salespeople bring and share information on the strategies of our
competitors.
. B8 : We respond to external actions, taken by other organisations, that
threaten ours.
F8 :
. B11 : The upper echelon of our company visits our present customers, as well as
the potential ones already prospected.
. B12 : Our experiences (good or bad) are communicated to the functional areas
of our company.
. B14 : All managers understand how each one can contribute to create value for
our customers.
F9 : Innovativeness
. C1 : The company is the ¯rst to launch new products and services.
. C2 : Managers actively seek innovative ideas.
. C4 : People are not punished for new ideas that do not work.

F10 : Organisational performance


. D3 : Investment return rate of our company has been greater than our
competitors'.

1850008-8
2nd Reading
February 14, 2018 8:48:02am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

. D4 : We optimize internal processes.


. D5 : We shorten the response time to market.
F11 : Knowledge management
Encompasses the constructs F1 and F2 (via F3 ), F4 and F5 .
F12 : Market orientation
Encompasses the constructs F6 , F7 and F8 .

3.3. Convergent validity


Each construct, or latent variable, has a property called average variance extracted
(AVE) that represents the amount of shared variance among its indicators (Hair
et al., 1998). According to Fornell and Larcker (1981) and Ferreira et al. (2010), the
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

amount of AVE of a construct has to be greater than 0.5, assessing its convergent
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

validity. If a construct fails to reach at least this value, it must be reassessed,


probably dropping one or more of its o®ending indicators, which are those that have
smaller loads on the construct. Any item (manifest variable) which has an estimated
loading less than 0.5 shall be removed from the construct (scale), since there is less
shared variance between the construct and its measure than error variance. In other
words, this item has no individual reliability (Hulland, 1999).

3.4. Second-order construct representation in PLS


The representation of the second-order constructs in the model for PLS estimation
followed the recommendations of Wold (1982), Lohm€oller (1989) and Wetzels et al.
(2009), which is that all the ¯rst-order constructs' indicators were used as re°ective
indicators for the associated second-order constructs (Richter et al., 2016a, 2016b).
For the sake of clarity and simplicity, the second-order constructs' indicators are not
represented in the structural models depicted in this study. It is also known as the
superblock approach (Tenenhaus et al., 2005) and is the most popular approach for
higher-order construct model estimation (Pelhan and Wilson, 1996; Zhang et al.,
2006).

3.5. Discriminant validity


In order to ensure that the introduction of a second-order construct is valid, the
model must be assessed in terms of discriminant validity, and it may be accom-
plished by verifying if the square root of the AVE of each one of its ¯rst-level
constructs is greater than any of the correlations between this ¯rst-level construct
and the others, as noted by Fornell and Larcker (1981). Those authors use
the comparison of the AVE with the correlations squared, which is equivalent to the
more convenient operational procedure adopted here. In this case, for an N  N
matrix structure, there are fewer elements with AVE to extract the square roots (N)
than elements with correlations to square [N  ðN  1Þ]. Usually, for the sake of
clarity, data is collected and put in a matrix form to test the discriminant validity of

1850008-9
2nd Reading
February 14, 2018 8:48:02am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

each construct. In the matrix structure, each square root of the AVE must be the
greatest value line-wise or column-wise.

3.6. Construct composite reliability


Chin (1995) recommends that the composite reliability of each construct must
exceed the value of 0.7 to ensure its adequacy to the PLS estimation.

3.7. Tenenhaus et al.'s proposed overall ¯t index


Tenenhaus et al. (2005) proposed a goodness-of-¯t (GoF) index to measure the
overall model performance by calculating the geometric mean between the average
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

R 2 (structural model adequacy) and the average AVE (measurement model


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

adequacy) of all constructs. These can be represented by the following set of equa-
tions, where n is the number of endogenous constructs m is the number of all
constructs and AVE, R 2 and GoF are as described before:

1 Xm
AVEavg ¼ AVE 2i ; ð1Þ
m i¼1

1 Xn
R 2avg ¼ R 2; ð2Þ
n i¼1 i
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
GoF ¼ AVEavg  R 2avg : ð3Þ

For the recommended values for GoF, Wetzels et al. (2009), suggest the value of
0.36 as high, for the social sciences and behavioural sciences.

3.8. Assessment of path coe±cient signi¯cances


For evaluating the signi¯cance of each path coe±cient, i.e. the in°uence one
construct makes over another, one procedure well known and available in
SmartPLS (Ringle et al., 2005) is bootstrapping, de¯ned as a \form of resampling
in which the original data are repeatedly sampled with replacement for model
estimation. Parameter estimates and standard errors are no longer calculated
with statistical assumptions but instead are based on empirical observations"
(Hair et al., 1998).
The bootstrap samples are of the same size as the original sample and it is
recommended that the number of resamples exceed 100, preferably exceeding 200
(Tenenhaus et al., 2005), or even \several thousands" (Hair et al., 1998). For the
current research, the original sample size is n ¼ 241 and the number of resamples
adopted is N ¼ 2; 000.

1850008-10
2nd Reading
February 14, 2018 8:48:03am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

4. Analysis and Discussion


The theoretical model developed and validated by Ferraresi et al. (2012) was
analysed using SmartPLS (Ringle et al., 2005; Rigdon, 2016), and the results are
shown as follows. The original model is depicted in Fig. 1, representing the
proposed structure with latent variables F10 ¼ organisational performance, F9 ¼
organisational innovativeness, F11 ¼ knowledge management and F12 ¼ market
orientation, and their manifest variables as described previously in Sec. 3.2.

4.1. Model adequacy and goodness-of-¯t


The analysis of model adequacy and goodness-of-¯t presented below was based on
the procedures and criteria developed in Sec. 3.
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

Convergent validity: As displayed in Fig. 2, there are no indicator loadings with


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

values below 0.5, thus the convergent validity of the model is assessed; reinforcing
this, all composite reliability (CR) coe±cients are well above the minimum recom-
mended of 0.7 (F1 ¼ 0:8341, F10 ¼ 0:8342, F2 ¼ 0:8616, F3 ¼ 0:8661, F4 ¼ 0:9236,
F5 ¼ 0:9126, F6 ¼ 0:9043, F7 ¼ 0:8826, F8 ¼ 0:8689, F9 ¼ 0:8107, KM ¼ 0:9397
and MO ¼ 0:9157), as prescribed by Fornell and Larcker (1981) and Tenenhaus
et al. (2005).

Fig. 2. Proposed model with calculated loads and R2 values.

1850008-11
2nd Reading
February 14, 2018 8:48:07am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

Discriminant validity: Table 2 shows that there are no correlations between


¯rst-order constructs that exceed in magnitude the square root of the AVE of the
related constructs, ensuring the discriminant validity and validating the second-
order factors. In addition, Table 3 shows that the AVE of each construct exceeds
the minimum recommended value of 0.5, demonstrating the good quality of the
constructs.
GoF: As proposed by Tenenhaus et al. (2005), the overall GoF was calculated and
it reached a value of 0.6978, far above the minimum recommended of 0.36, which
indicates an excellent overall ¯t of the model. All the individual GoF indices
(F1 ¼ 0:6982; F10 ¼ 0:624; F2 ¼ 0:7462; F3 ¼ 0:6608; F4 ¼ 0:8027; F5 ¼ 0:805;
F6 ¼ 0:755; F7 ¼ 0:7137; F8 ¼ 0:7537; F9 ¼ 0:5629 and MO ¼ 0:5878) also are well
above the recommended minimum.
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

Table 2. Model adequacy indexes.

AVE Composite reliability R2 Cronbach's alpha Communality Redundancy

F1 0.6271 0.8341 0.7774 0.7025 0.6271 0.4845


F10 0.6331 0.8342 0.6151 0.6945 0.6331 0.235
F2 0.6758 0.8616 0.8239 0.7582 0.6758 0.5563
F3 0.5213 0.8661 0.8376 0.8132 0.5213 0.4360
F4 0.7518 0.9236 0.8571 0.8893 0.7518 0.6437
F5 0.7768 0.9126 0.8342 0.8562 0.7768 0.6475
F6 0.7593 0.9043 0.7508 0.8406 0.7593 0.5699
F7 0.7151 0.8826 0.7123 0.8008 0.7151 0.5050
F8 0.6888 0.8689 0.8248 0.7734 0.6888 0.5674
F9 0.5925 0.8107 0.5347 0.6522 0.5925 0.2771
KM 0.5495 0.9397 0.9294 0.5495
MO 0.5483 0.9157 0.6301 0.8962 0.5483 0.34

Table 3. Correlation values between ¯rst-order constructs.

F10 F1 F2 F4 F5 F6 F7 F8 F9

F10 0.7957
F1 0.5715 0.7919
F2 0.6224 0.6023 0.8221
F4 0.6134 0.6191 0.7129 0.8671
F5 0.6661 0.6087 0.7302 0.7973 0.8814
F6 0.6472 0.4837 0.6097 0.6137 0.6954 0.8714
F7 0.5479 0.4720 0.5136 0.5061 0.5884 0.5559 0.8456
F8 0.6495 0.5663 0.6350 0.7341 0.7423 0.6884 0.6865 0.8299
F9 0.7050 0.5345 0.6303 0.6327 0.6953 0.5489 0.5467 0.6192 0.7697

Note: Diagonal elements replaced by the square root of the construct AVE.

1850008-12
2nd Reading
February 14, 2018 8:48:10am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

Table 4. Path coe±cients after bootstrap analysis (n ¼ 240; N ¼ 2; 000).

Original Sample Standard Standard error T -statistics


sample (O) mean (M) deviation (STDEV) (STERR) (IO/STERRI)

F3 ! F1 0.8817 0.8819 0.0238 0.0238 36.9925


F3 ! F2 0.9077 0.9091 0.0140 0.0140 64.8901
F9 ! F10 0.3515 0.3590 0.0910 0.0910 3.8608
KM ! F10 0.2063 0.1995 0.0978 0.0978 2.1101
KM ! F3 0.9152 0.9154 0.0168 0.0168 54.6187
KM ! F4 0.9258 0.9255 0.0154 0.0154 60.0187
KM ! F5 0.9133 0.9131 0.0126 0.0126 72.4846
KM ! F9 0.5287 0.5282 0.0978 0.0978 5.4051
KM ! MO 0.7938 0.7933 0.0336 0.0336 23.6585
MO ! F10 0.3131 0.3142 0.0777 0.0777 4.0296
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

MO ! F6 0.8665 0.8665 0.0233 0.0233 37.2193


MO ! F7 0.8440 0.8445 0.0319 0.0319 26.4691
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

MO ! F8 0.9082 0.9082 0.0142 0.0142 63.9718


MO ! F9 0.2371 0.2387 0.0985 0.0985 2.4080

4.2. Path analysis


For the assessment of the magnitude and signi¯cance of the path coe±cients, a
bootstrap analysis was performed, generating 2,000 di®erent resamples, each one
with 240 observations, as recommend by the literature. The results are shown in
Table 4. It is observed that every path reached a signi¯cance level of 0.05 (t-statistics
of 1.96). In other words, there are evidences, at a 0.05 level of signi¯cance, that all
path coe±cients listed in Table 4 are signi¯cantly di®erent from zero.

5. Conclusions
Table 5 shows the main results of the direct, indirect and total e®ects observed by
Ferraresi et al. (2012) with CB-SEM, compared with the current PLS analysis of the

Table 5. Direct, indirect and total e®ects among the constructs.

Relations Covariance-based PLS

Direct Indirect Total Direct Indirect Total


e®ects e®ects e®ects e®ects e®ects e®ects

KM ! MO 0.928 b 0.928 a 0.794 b 0.794 b


F11 (KM) ! F10 (IO) 0.475 c 0.381 c 0.855 b 0.529 b 0.188 d 0.717 b
F11 (KM) ! F10 (OP) 0.014 c 0.812 a 0.798 b 0.206 a 0.501 d 0.707 b
F12 (MO) ! F9 (IO) 0.410 c 0.410 c 0.237 a 0.237 a
F12 (MO) ! F10 (OP) 0.562 c 0.139 c 0.701 a 0.313 b 0.083 d 0.396 b
F9 (IO) ! F10 (OP) 0.340 c 0.340 c 0.352 b 0.352 b

Notes: a Signi¯cant at p < 0:05; b Signi¯cant at p < 0:01; c not substantial; d not reported
by software.

1850008-13
2nd Reading
February 14, 2018 8:48:10am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

same model using the same sample. The ¯rst remarkable di®erence is the lack of
information of the indirect e®ects for the PLS analysis, as this ¯gure is not reported
by the SmartPLS software. The indirect e®ects were calculated by hand, but the
signi¯cance ¯gure is still lacking in the results.

5.1. Comparison of results


This sub-section presents the structural coe±cients, as well as the explained variance
of each construct, so that comparisons are possible between the CB and PLS
analyses.
Structural coe±cients: The main goal, which was to assess the in°uence of
knowledge management (F11 ) on OP (F10 ), was ful¯lled by the analysis, and the
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

result is comparable to its counterpart calculated via CB analysis. Even though


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

the direct e®ect of KM on OP is not signi¯cant when calculated with CB and is


signi¯cant at p < 0:05 when calculated using PLS, the total e®ects have both
intensity and signi¯cance in the same order of magnitude, with 0.798 (p < 0:01) for
CB analysis and 0.707 (p < 0:01) for PLS analysis. In general, all structural coe±-
cients calculated with the CB method are slightly higher than the corresponding
ones calculated by PLS, similarly to what has been veri¯ed in previous studies. The
exception is the path from OI to OP, which is slightly higher (0.352 with p < 0:01)
when calculated via PLS than the corresponding one (0.340, non-signi¯cant) when
calculated with the CB method.
As for the overall results, the PLS analysis indicates that knowledge management
acts as a precursor for organisational performance, showing a weak (0.206) e®ect
when taken directly, but presenting a much stronger e®ect (0.707) when mediated by
the other structural paths. Knowledge management also presents a strong e®ect on
market orientation (0.794) and a strong e®ect (0.529 direct, 0.717 total) on orga-
nisational innovativeness.
Market orientation presents a weak in°uence in terms of direct e®ects
on organisational innovativeness (0.237) and a noticeable although weak
in°uence on organisational performance (0.313 direct, 0.396 total). This in°uence
shows up as weaker in PLS analysis than in CB analysis, suggesting the need
for further studies speci¯cally on this matter. Innovativeness presents a notice-
able, although weak, in°uence on organisational performance (0.352), completing
the structural coe±cients assessment and comparison of the CB and PLS
methods.
Explained variance (Fi ): As shown in Table 6, the explained variances for the
latent variables, assessed by their R 2 , are slightly higher for the CB method than for
PLS, as pointed out in Sec. 2.
Indicator loads: For the sake of parsimony, the indicator loads are not depicted
here, but, as pointed out in previous studies, the ¯gures presented in PLS analysis
are systematically higher than the ¯gures obtained in the CB method.

1850008-14
2nd Reading
February 14, 2018 8:48:10am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

Table 6. R 2 comparison between


CB and PLS methods.

R2

Construct CB method PLS method

F10 0.733 0.615


F9 0.755 0.535
MO 0.861 0.630

5.2. Concluding remarks


The results of this study indicate that PLS modelling can provide consistent results
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

when compared with a covariance-based analysis. The model's behaviour, with re-
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

gard to the magnitude of the path coe±cients, R 2 , and indicator loadings, was
consistent with the conclusions of previous studies. The sample size, although bor-
derline adequate for the CB method as reported by Ferraresi et al. (2012), was more
than adequate for PLS, yielding excellent model stability along the bootstrap
analysis.
One of the biggest limitations of the PLS modelling is the lack of a strong indi-
cator of model ¯t, as there are a plethora of indices in any CB software. Aside from
the GoF index advanced by Tenenhaus et al. (2005), which assesses the overall
quality of the model, more robust indicators of model adequacy are still lacking,
making it more di±cult to compare rival models and, perhaps, try to improve an
existing model. In this case, the model's stability could be further investigated along
sub-samples of the original 241 cases, testing the downside limit of the sample size
relative to the numbers of constructs and indicators used. The extant literature
points out that a ratio of 10:1 between sample size and the largest number of for-
mative indicators used to measure one construct is adequate for PLS models
(Barclay et al., 1995) but there is still controversy on this matter.
More generally, while CB-SEM remains the more popular method, PLS-SEM has
gained acceptance throughout many business disciplines for its ability to handle
problematic modelling issues such as non-normal data and complex models that are
common in empirical business and social sciences data. In addition, when large
samples are not available, or data is di±cult to obtain, the application of PLS is a
viable attempt to advance knowledge in such areas (Hair et al., 2017). However, it is
an evolving approach to structural equation modelling, and there have been criti-
cisms about inadequate handling of formative indicators and the criteria that are
used in the evaluation of formatively measured constructs, as well as constraints
related to assessment of model ¯t, consistency of the parameter estimates and am-
biguity in the evaluation of mediating e®ects. Nevertheless, it should be noted that
several alleged weaknesses of PLS (R€ onkk€o and Evermann, 2013) are not inherent in
the method, but are rather the result of the limitations of questionable research
designs (Henseler et al., 2014).

1850008-15
2nd Reading
February 14, 2018 8:48:10am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

Recent debates on these issues (see e.g. Hsu et al., 2006; Hair et al., 2011; Monecke
and Leisch, 2012; R€ onkk€o and Evermann, 2013; Davcik, 2014; Hair et al., 2014;
McIntosh et al., 2014; Hair et al., 2017) highlight both the potential and the pitfalls
of PLS-SEM, as this distinctive method gains ground over the more popular and
traditional CB-SEM approaches in business research. As Richter et al. (2016a,
2016b) point out, the PLS-SEM approach is particularly useful for theory devel-
opment, and, as new ¯t measures for this method are developed, they will enable
researchers to further exploit the method's explanatory capabilities for theory
testing, in combination with its ability to deal with complex hierarchical component
or second-order models.
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

Acknowledgements
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

We are also grateful to our anonymous reviewers for the tips contributions and
recommendations.

References
Abdi K and AA Senin (2015). The impact of knowledge management on organizational
innovation: An empirical study. Asian Social Science, 11(23), 153–168.
Almashari, M, M Zairi and A Alathari (2002). An empirical study of the impact of knowledge
management on organizational performance. Journal of Computer Information System,
42(5), 74–82.
Andreassi, T and R Sbragia (2004). Fatores determinantes do grau de inovatividade
das empresas: Um estudo utilizando a tecnica de an
alise discriminante. Working Paper 01/
004, FEA/USP, Brazil.
Arthur, L (1999). Managing in organizations that learn/organizational learning: The
competitive advantage of the future. Management Learning, 30(2), 251–254.
Barclay, D, C Higgins and R Thompson (1995). The partial least squares approach to causal
modeling: Personal computer adoption and use as illustration. Technology Studies, 2(2),
285–309.
Botha, D (2001). A conceptual framework for the management of knowledge in a knowledge-
based enterprise. South African Journal of Business Management, 31(4), 141–149.
Capon, N, JU Farley, DR Lehmann and J Hulbert (1992). Pro¯les of product innovators
among large US manufacturers. Management Science, 38(2), 157–169.
Chin, WW (1995). Partial least squares is to LISREL as principal components analysis is to
common factor analysis. Technology Studies, 2, 315–319.
Darroch, J (2003). Developing a measure of knowledge management behaviors and practice.
Journal of Knowledge Management, 7(5), 41–54.
Davcik, N (2014). The use and misuse of structural equation modeling (SEM) in management
research: A review and critique. Journal of Advances in Management Research, 11(1),
47–81.
Day, GS (1999). The Market Driven Organization: Understanding, Attracting, and Keeping
Valuable Customers. New York: The Free Press.
Deshpande, R, J Farley and F Webster, Jr (1993). Corporate culture, customer orientation,
and innovativeness in Japanese ¯rms: A quadrad analysis. Journal of Marketing, 57, 23–37.

1850008-16
2nd Reading
February 14, 2018 8:48:10am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

Ferraresi, A, C Quandt, S Santos and J Frega (2012). Knowledge management and strategic
orientation: Leveraging innovativeness and performance. Journal of Knowledge Manage-
ment, 16(5), 688–701.
Ferreira, I, J Cabral and P Saraiva (2010). An integrated framework based on the ECSI
approach to link mould customers' satisfaction and product design. Total Quality Man-
agement & Business Excellence, 21(12), 1383–1401.
Fornell, C and DF Larcker (1981). Evaluating structural equation models with unobservable
variables and measurement error. Journal of Marketing Research, 18(1), 39–50.
Gao, Y (2017). Business leaders' personal values, organisational culture and market orienta-
tion. Journal of Strategic Marketing, 25(1–2), 49–64.
Garcia, R and R Calantone (2002). A critical look at technological innovation typology and
innovativeness terminology: A literature review. The Journal of Product Innovation
Management, 19(4), 110–132.
Gold, A, A Malhotra and A Segars (2001). Knowledge management: An organizational
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

capabilities perspective. Journal of Management Information Systems, 18(1), 185–214.


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

Hair, J, R Anderson, R Tatham and W Black (1998). Multivariate Data Analysis, 5th edn.
New Jersey: Prentice Hall.
Hair, J, G Hult, C Ringle, M Sarstedt and K Thiele (2017). Mirror, mirror on the wall:
A comparative evaluation of composite-based structural equation modeling methods.
Journal of the Academy of Marketing Science, 45(5), 616–632.
Hair, J, C Ringle and M Sarstedt (2011). PLS-SEM: Indeed a silver bullet. Journal of Mar-
keting Theory and Practice, 19(2), 139–152.
Hair, J, M Sarstedt, T Pieper and C Ringle (2012). The use of partial least squares structural
equation modeling in strategic management research: A review of past practices and
recommendations for future applications. Long Range Planning, 45(5–6), 320–340.
Hair, J, M Sarstedt, L Hopkins and V Kuppelwieser (2014). Partial least squares structural
equation modeling (PLS-SEM): An emerging tool in business research. European Business
Review, 26(2), 106–121.
Henseler, J, TK Dijkstra, M Sarstedt, CM Ringle, A Diamantopoulos, DW Straub,
DJ Ketchen, JF Hair, GTM Hult and RJ Calantone (2014). Common beliefs and reality
about partial least squares: Comments on R€ onkk€o & Evermann (2013). Organizational
Research Methods, 17(2), 182–209.
Hooley, G, JA Saunders and NF Piercy (2001). Estrategia de Marketing e posicionamento
competitivo. São Paulo: Prentice Hall.
Hsu, SH, WH Chen and MJ Hsieh (2006). Robustness testing of PLS, LISREL, EQS and
ANN-based SEM for measuring customer satisfaction. Total Quality Management, 17(3),
355–371.
Hulland, J (1999). Use of partial least squares (PLS) in strategic management research:
A review of four studies. Strategic Management Journal, 20, 195–204.
Hult, G. and D Ketchen (2001). Does market orientation matter? A test of the relationship
between positional advantage and performance. Strategic Management Journal, 22, 899–906.
Hurley, R and G Hult (1998). Innovation, market orientation, and organizational learning:
As integration and empirical examination. Journal of Marketing, 62(7), 42–54.
Kohli, AK and BJ Jaworski (1990). Market orientation: The construct, research propositions,
and managerial implications. Journal of Marketing, 54(2), 1–17.
Kohli, AK and BJ Jaworski (1993). MARKOR: A measure of market orientation. Journal of
Marketing Research, 30(4), 467–477.
Liao, S-H and C Wu (2010). System perspective of knowledge management, organizational
learning, and organizational innovation. Expert Systems with Applications, 37(2), 1096–
1103.

1850008-17
2nd Reading
February 14, 2018 8:48:11am WSPC/188-JIKM 1850008 ISSN: 0219-6492

J. R. Frega et al.

Lohm€oller, J (1989). Latent Variable Path Modeling with Partial Least Squares. Heidelberg:
Physica-Verlag.
Lynch, P, M Walsh and D Harrington (2010). De¯ning and dimensionalizing organizational
innovativeness. In Proceedings of the International CHRIE Annual Summer Conference
and Marketplace. Available at http://scholarworks.umass.edu/refereed/CHRIE 2010/
Saturday/18. Accessed on 21 January 2018.
Marcoulides, GA and C Saunders (2006). PLS: A silver bullet? MIS Quarterly, 30(2), 3–9.
Massey, A, A Montoya-Weiss and T O'Driscoll (2001). Knowledge management in pursuit of
performance: Insights from Nortel Networks. MIS Quarterly, 26(3), 269–289.
McDonald, RE (2002). Knowledge entrepreneurship: Linking organizational learning and
innovation. Doctoral dissertation, School of Management, The University of Connecticut.
McIntosh, C, J Edwards and J Antonakis (2014). Re°ections on partial least squares path
modeling. Organizational Research Methods, 17(2), 210–251.
Monecke, A and F Leisch (2012). semPLS: Structural equation modeling using partial least
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

squares. Journal of Statistical Software, 48(3), 1–32.


J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

Narver, JC and SF Slater (1990). The e®ect of a market orientation on business pro¯tability.
Journal of Marketing, 54(4), 20–35.
Nonaka, I and I Takeuchi (1995). The Knowledge Creating Company. New York: Oxford
University Press.
Pelhan, AM and DT Wilson (1996). A longitudinal study of the impact of market structure,
¯rm structure, strategy, and market orientation culture on dimensions of small ¯rm per-
formance. Journal of the Academy of Marketing Science, 24(1), 27–43.
Probst, G, S Raub and K Romhardt (2000). Managing Knowledge: Building Blocks for
Success, London: Wiley & Sons.
Pučėtaitė, R, A Novelskaite, A-M Lämsä and E Riivari (2016). The relationship between
ethical organisational culture and organisational innovativeness: Comparison of ¯ndings
from Finland and Lithuania. Journal of Business Ethics, 139(4), 685–700.
Quinn, JB, P Anderson and S Finkelstein (1998). Managing professional intellect. In Harvard
Business Review on Knowledge Management. Boston: Harvard Business School Publishing.
Rapp, A, N Schillewaert and AO Hao (2008). The in°uence of market orientation on
e-business innovational performance: The rule of the top management team. Journal of
Marketing Theory and Practice, 16(1), 7–25.
Rezaei, G, H Gholami, A Shaharou, M Zameri, L Sadeghi and N Zakuan (2017). Shared
knowledge mediated correlation between cultural excellence and organisational perfor-
mance. Total Quality Management and Business Excellence, 28(3–4), 427–458.
Richter, NF, G Cepeda, JL Roldan and CM Ringle (2016). European management research
using partial least squares structural equation modeling (PLS-SEM). European Manage-
ment Journal, 34(6), 589–597.
Richter, NF, R Sinkovics, C Ringle and C Schlägel (2016). A critical look at the use of SEM in
international business research. International Marketing Review, 33(3), 376–404.
Rigdon, E (2016). Choosing PLS path modeling as analytical method in European manage-
ment research: A realist perspective. European Management Journal, 34(6), 598–605.
Ringle, CM, S Wende and A Will (2005). SmartPLS 2.0 M3. SmartPLS GmbH, Hamburg.
Available at http://www.smartpls.com. Accessed on 21 January 2018.
R€onkk€o, M and J Evermann (2013). Critical examination of common beliefs about partial
least squares path modeling. Organizational Research Methods, 16(3), 425–448.
Sarstedt, M, C Ringle and J Hair (2014). PLS-SEM: Looking back and moving forward
(Editorial). Long Range Planning, 47(3), 132–137.
Shih, C-P and H-F Chou (2012). The e®ect of knowledge management strategies and enablers
on knowledge creation process and organizational performance by using partial least

1850008-18
2nd Reading
February 14, 2018 8:48:11am WSPC/188-JIKM 1850008 ISSN: 0219-6492

Modelling of Relationships Among Knowledge Management, Organisational Innovativeness and Performance

squares regression method. International Journal of Knowledge and Systems Science, 3(4),
38–52.
Slater, SF and JC Narver (1995). Market orientation and the learning organization. Journal of
Marketing, 59(3), 63–75.
Tajeddini, K (2011). The e®ects of innovativeness and e±ciency. Education, Business and
Society: Contemporary Middle Eastern, Issues, 4(1), 6–18.
Tajeddini, K, M Trueman and G Larsen (2006). Examining the e®ect of market orientation on
innovativeness. Journal of Marketing Management, 22(6), 529–551.
Tenenhaus, M, VE Vinzi, Y-M Chatelin and C Lauro (2005). PLS path modeling. Compu-
tational Statistics & Data Analysis, 48(1), 159–205.
Wetzels, M, G Odekerken-Schroder and C van Oppen (2009). Using PLS path modeling for
assessing hierarchical construct models: Guidelines and empirical illustration. MIS Quar-
terly, 33(1), 177–196.
Wold, H (1982). Soft modeling: The basic design and some extensions. In Systems Under
by UNIVERSITY OF NEW ENGLAND on 02/16/18. For personal use only.

Indirect Observation: Causality, Structure and Prediction — Part II, K J€ oreskog and
J. Info. Know. Mgmt. Downloaded from www.worldscientific.com

H Wold (eds.), pp. 1–154. Netherlands: North-Holland.


Zhang, P, N Li and H Sun (2006). A®ective quality and cognitive absorption: Extending
technology acceptance research. In Proceedings of the 39th Hawaii International Confer-
ence on Systems Sciences. Hawaii: IEEE Computer Society.
Zheng, W (2005). The impact of organizational culture, structure, and strategy on knowledge
management e®ectiveness and organizational e®ectiveness. Doctoral dissertation, Faculty
of the Graduate School, University of Minnesota, USA.
Zheng, W, B Yang and G McLean (2010). Linking organizational culture, structure, strategy,
and organizational e®ectiveness: Mediating role of knowledge management. Journal of
Business Research, 63(7), 763–771.
Zhou, Z, K Yim and D Tse (2005). The e®ects of strategic orientations on technology- and
market-based breakthrough innovations. Journal of Marketing, 69(1), 42–60.

1850008-19

You might also like