Documente Academic
Documente Profesional
Documente Cultură
Recent Studies
Author(s): John Hulland
Source: Strategic Management Journal, Vol. 20, No. 2 (Feb., 1999), pp. 195-204
Published by: Wiley
Stable URL: http://www.jstor.org/stable/3094025
Accessed: 06-07-2017 14:10 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Wiley is collaborating with JSTOR to digitize, preserve and extend access to Strategic Management
Journal
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
Strategic Management Journal
Strat. Mgmt. J., 20: 195-204 (1999)
Advances in causal modeling techniques have made it possible for researchers to simultaneously
examine theory and measures. However, researchers must use these new techniques appropri-
ately. In addition to dealing with the methodological concerns associated with more traditional
methods of analysis, researchers using causal modeling approaches must understand their
underlying assumptions and limitations.
Most researchers are well equipped with a basic understanding of LISREL-type models. In
contrast, current familiarity with PLS in the strategic management area is low. The current
paper reviews four recent studies in the strategic management area which use PLS. The review
notes that the technique has been applied inconsistently, and at times inappropriately, and
suggests standards for evaluating future PLS applications. Copyright ? 1999 John Wiley &
Sons, Ltd.
Advances in causal modelling techniques have relationships; (3) they permit a more complete
made it possible for researchers to simultaneously representation of complex theories; and (4) they
examine theory and measures. Such techniques provide a formal framework for constructing and
can be thought of as superior to more traditional testing both theories and measures.
techniques (e.g., multidimensional scaling, factor The best-known causal modeling technique is
analysis) in that they permit: (1) the explicit LISREL (Joreskog and Sorbom, 1989; Hagedoor
inclusion of measurement error, and (2) an ability and Schakenraad, 1994). However, LISREL is
to incorporate abstract and unobservable con- poorly suited to deal with small data samples
structs (Fornell, 1982). Bagozzi (1980) suggests (Forell, 1982), and can yield nonunique or
that causal models provide researchers with four otherwise improper solutions in some cases
key benefits: (1) they make the assumptions, (Forell and Bookstein, 1982). An alternative
constructs, and hypothesized relationships in a causal modeling approach known as Partial Least
theory explicit; (2) they add a degree of precision Squares (PLS) has been developed to avoid some
to a theory, since they require clear definitions of these limitations (Wold, 1974, 1985), although
of constructs, operationalizations, and functional use of PLS requires its own set of assumptions.
PLS has been used both in other business disci-
plines (e.g., Duxbury and Higgins, 1991; Hulland
Key words: causal modeling; measurement issues; and Kleinmuntz, 1994; Smith and Barclay, 1997;
partial least squares; strategic management research is- Zinkhan, Joachimsthaler, and Kinnear, 1987) and
sues
in the strategic management area to examine risk-
* Correspondence to: John Hulland, Richard Ivey School of
return outcomes
Business , The University of Western Ontario, London, Onta-
(Cool, Dierickx, and Jemison,
rio N6A 3K7, Canada 1989), cooperative ventures (Fornell, Lorange,
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
196 J. Hulland
Conceptual model
and Roos, 1990), global strategy (Johansson and specification
Yip, 1994), and global integration (Birkinshaw,
Morrison, and Hulland, 1995).' The studies by Birkinshaw et al., Cool et al.,
While techniques such as LISREL andand PLSet al. each present and test a single
Fornell
can enrich existing methodological approaches to In contrast, Johansson and Yip
conceptual model.
conducting strategic management research,
specify (andthey
estimate) eight different conceptual
must be used appropriately. Most researchers
models which are
follow five seemingly different ba-
well equipped with a basic understanding
sic forms.
of There
LIS- is nothing inherently wrong in
making
REL-type models. In contrast, current use of alternative models. Indeed, in the
familiarity
with PLS in the strategic management area
early stages of is
theory refinement such compari-
sons often play
low, making it difficult for most researchers to a critical role. However, the
properly evaluate its use by others.approach used by Johansson and Yip in
Exacerbating
presenting
this problem of limited familiarity, existing their various models is largely ad
appli-
hoc-it is not area
cations of PLS in the strategic management clear why some models have been
have used the technique inconsistently, andothers
included while at have not. Furthermore, their
times inappropriately. eight models can all be represented as special
cases
To help shape future application of PLSofin
onethe
of the two more general models
shownreviews
strategy area, the following discussion in Figure 1.2 Thus, it would be more
use of the technique in four recent studies
appropriate to first estimate the two general mod-
(Birkinshaw et al., 1995; Cool et al.,els1989;
(i.e., 'unrestricted'
For- Models A and B), and to
nell et al., 1990; Johansson and Yip, 1994).
then This
estimate each of the more specific variants
discussion is organized around two (i.e.,
general sets models). Although R2 values and
'restricted'
of issues: conceptual and methodological. As from entirely different theoretical
path coefficients
Bagozzi (1984) noted, theory and measurement
models are not directly comparable, when one
model
are intimately intertwined, and both mustis athere-
subset of (or 'nested within') a more
generalcontext.
fore be considered in a causal modeling model the researcher can compare R2
values on the key endogenous construct(s) using
a Chow test. For Johansson and Yip's study,
CONCEPTUAL ISSUES this approach would permit direct comparisons
between Models 1, 2, 3, 4, and 6, and between
Models
Although measures and theory are both 5, 7, and 8.
important,
the process of model specification necessarily
begins by considering the theoretical model
Construct dimensionality
underlying a particular piece of research. That is,
the causal modeling process begins atThe
the concep- of constructs is a critical con-
dimensionality
tual level. Many of the conceptual issues
sideration
which
in the development of causal models.
apply to all empirical research are also relevant
In many cases, researchers can correctly assume
that their
in a causal modeling context. However, constructs are unidimensional. However,
we focus
here on three issues that play a particularly
when a particular construct is more properly con-
ceptualizedcon-
important role in causal modeling research: as multidimensional (e.g., business
ceptual model specification, construct dimen-
strategy according to Venkatraman, 1989),
sionality, and the distinction between constructs
and measures.
2 Specifically, Models 1, 2, 3, 4, and 6 in Johansson and Yip
(1994) are all nested within Model A shown in the figure.
For example, their Model 1 assumes that only paths 2, 7, 9,
and 10 are nonzero, Model 2 assumes that only paths 3, 6,
8, and 10 are nonzero, and Model 3 assumes that only paths
2, 3, 6, 7, 8, 9, and 10 are nonzero. Model 4 assumes that
'For an excellent introduction to the use of PLS in practical
applications, see Barclay, Higgins, and Thompson (1995).
paths 1, 3, 5, 6, 7, 8, and 10 are nonzero, while Model 6
Barclay et al. also provide a thorough description of the that paths 1, 3, 5, 6, 8, 10, 11, and 12 are nonzero.
assumes
Models 5, 7, and 8 are all nested within model B of the
objectives of PLS, the estimation process, sample size issues,
the relative strengths and weaknesses of both LISREL and Specifically, Model 5 assumes that paths 4, 5, and 6
figure.
PLS, and the availability of software. For a good discussion of Model 7 assumes that path 4 is zero, and Model 8
are zero,
the model structure employed by PLS, see Cool et al. assumes
(1989). that path 5 is zero.
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
Review of PLS Use in Strategic Management Research 197
Model A
Mode B
Figure 1. Two comprehensive conceptual models encompassing all eight models estimated in Johansson and
Yip
researchers' causal models should include sepa- these four drivers could have a different impact
rate constructs representing each of these dimen- on the other constructs (e.g., global strategy).
sions. However, in the empirical section of their paper,
In general, the four studies reviewed here Johansson and Yip ignore these distinctions, using
assume and use construct unidimensionality.instead a single, higher order (but conceptually
However, an interesting contrast in approaches toundefined) 'industry drivers' construct. In con-
this issue of dimensionality can be seen by com- trast, Birkinshaw et al. include three separate
paring the treatments of the industry structure 'driver' constructs in their model in order to
construct by Johansson and Yip and by Birkin- look at their individual effects on other strategic
shaw et al. Early discussion of their constructs constructs. This latter approach is more consistent
by Johansson and Yip (1994: 580-582) stronglywith existing theory while also yielding superior
suggests that industry structure, global strategy,empirical results.
organization structure, management processes,
and performance are all multidimensional. For Constructs versus measures
example, they clearly identify four separate and
distinct 'drivers' in their discussion of the indus- Although related to one another, constructs and
try structure construct (i.e., government drivers, measures are distinct entities (Bagozzi, 1984).
cost drivers, market drivers, and competitive
Much of the work in the four studies reviewed
drivers). Intuitively, it seems likely that each here
of properly separates the theoretical and empiri-
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
198 J. Hulland
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
Review of PLS Use in Strategic Management Research 199
only six exceed the 0.7 threshold. Similarly, for Broadly speaking, the convergent validities of
the models reported in Johansson and Yip (1994), the constructs used by Birkinshaw et al., Cool et
about one third of the loadings consistently fall al., and Fornell et al. appear to be acceptable. In
below 0.4, whereas just over half exceed 0.7.3 contrast, most of the constructs used by Johansson
Thus, while most of the measures used in and Yip exhibit poor convergent validity. For
these four studies appear to exhibit satisfactory Models 6 and 8 (their two most preferred
individual item reliabilities, two of the four stud-
models), only the measures used to assess per-
formance demonstrate sufficient internal consist-
ies retained a significant number of low-reliability
items in their final analyses. Since these low
ency to exceed Nunnally's 'modest' standard of
reliabilities can attenuate the estimated relation- 0.70. At the same time, two sets of measures
ships between constructs, results based on(those
the representing organizational structure and
retention of low-reliability items must be inter-
global strategy) exhibit extremely low internal
preted with caution. consistency (i.e., less than 0.5).
Low internal consistency can result from a
variety of underlying causes, including poor con-
Convergent validity
struct definition and/or construct multi-
When multiple measures are used for an individ-
dimensionality. In the first case, the faulty co
ual construct, the researcher should be concerned struct definition severely impairs the
not only with individual measurement item determination of relevant and appropriate meas-
reliability, but also with the extent to which the ures for the construct. In the latter case, if the
measures demonstrate convergent validity. Tra- underlying construct is actually multidimensional,
ditionally, researchers using PLS have generally but it is measured using items which are assumed
reported one or both of two measures of conver- to be linked to a unidimensional construct, the
gent validity (also referred to as composite measures as a group will demonstrate poor inter-
reliability): Cronbach's alpha and the internal nal consistency. Furthermore, while some of the
consistency measure developed by Forell and individual measurement items will have strong
Larcker (1981).4 Fomell and Larcker argue that loadings linking them closely with the construct,
their measure is superior to alpha since it uses others will have loadings which are close to zero,
the item loadings obtained within the nomological or even negative in sign. In such instances, the
network (or causal model). Nonetheless, the inter- researcher should consider either splitting the
pretation of the values obtained is similar, and original construct into new constructs (each with
the guidelines offered by Nunnally (1978) can its own set of measures), or eliminating items
be adopted for both. Specifically, Nunnally sug- until only a unidimensional construct remains.
gests 0.7 as a benchmark for 'modest' composite For example, the organizational structure
reliability, applicable in the early stages of construct-as measured by Johansson and Yip-
research.5 appears to be multidimensional. To rectify this
Although it is not possible to calculate Cron- problem, they could either drop one of their
bach's alphas for the individual constructs in any measures or split the construct into two distinct
of the models reviewed here, internal consistency subconstructs.
values were determined by using the reported
loadings and Forell and Larcker's formula.
Discriminant validity
These values are reported in column one of Table
1 for three of these studies. (Birkinshaw et al. The traditional methodological complement to
report internal consistency values in their Table convergent validity is discriminant validity, which
2). represents the extent to which measures of a
given construct differ from measures of other
constructs in the same model. In a PLS context,
3 In fact, one of the retained loadings reported by Johansson
and Yip is as low as 0.08. one criterion for adequate discriminant validity is
4 Internal consistency = ((X Ay,)2) / ((E Ay,)2 + E var (e,)). that a construct should share more variance with
5 Strictly speaking, this discussion of convergent validity and its measures than it shares with other constructs
the preceding discussion of item reliability can only be applied
to measures that are reflective, rather than formative. This in a given model. To assess discriminant validity,
issue is explored more fully in the next section of this paper. Fornell and Larcker (1981) suggest the use of
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
200 J. Hulland
Model 8
Industry drivers 0.54 0.46 0.68
Global strategy 0.38 0.27 0.52
Management processes 0.62 0.45 0.67
Organizational structure 0.40 0.46 0.68
Performance 0.84 0.72 0.85
Average Variance
Birkinshaw et al. report bothExtra
average variance
variance shared between a construct and its extracted (AVE) values and inter-construct corre-
measures).6 lations in their Table 3, providing clear evidence
This measure should be greater than the vari- of discriminant validity. For the other three stud-
ance shared between the construct and other con- ies, the table reports AVE and root AVE values
structs in the model (i.e., the squared correlationfor each construct in columns two and three.7
between two constructs). This can be demon- Although none of these studies report inter-
strated in a correlation matrix which includes the construct correlations, it is possible to roughly
correlations between different constructs in the assess the discriminant validity of the constructs
lower left off-diagonal elements of the matrix, using the estimated path coefficients. For
and the square roots of the average variance example, the root AVE values reported for the
extracted values calculated for each of the con- Cool et al. study are all larger than the path
structs along the diagonal. For adequate discrim-
coefficients they estimate. Likewise, the root AVE
inant validity, the diagonal elements should be
significantly greater than the off-diagonal
7Cool et al. report AVE as 'convergent validity' in their
elements in the corresponding rows and columns.
Table 1. They also report 'discriminant validity' for each
construct. However, this latter measure assesses the average
squared correlation of a particular construct with all other
6 Average variance extracted = 2 Ayi2 / constructs in the model. Since this value will vary consider-
ably depending on the other constructs included, it is not really
appropriate to compare such values to any fixed threshold. In
(I i2 + S var (,)). general, the approach described here is preferred.
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
Review of PLS Use in Strategic Management Research 201
values found for the Fomell et al. study all meaningful, rendering irrelevant traditional
exceed their reported path coefficients. Thus, it assessments of individual item reliability and con-
is possible to conclude that discriminant validity vergent validity.
is adequate in both of these studies. This does not give the researcher carte blanche
Such is not the case for the study by Johansson to arbitrarily link sets of measures to constructs,
and Yip. Consider the path coefficient between however. Cohen et al. (1990) suggest that when
global strategy and management processes in their the relationship is formative, researchers must be
Model 8 (0.70). If this is taken as a rough proxy careful to employ strong theory (which helps
of the correlation between the two constructs, to identify appropriate measures) and multiple
then the off-diagonal correlation between global measures (to ensure acceptable content validity).
strategy and management processes is greater than Bollen and Lennox (1991: 307) expand on this
the root AVE value for each of the constructs latter point by emphasizing that researchers 'need
considered on its own (i.e., 0.48, 0.68). This a census of indicators, not a sample. That is, all
implies that the constructs and their measures constructs that form [the underlying construct]
cannot be adequately discriminated, and it is
should be included.'
therefore entirely inappropriate to view them asThe choice between using formative or reflec-
distinct and separate theoretical entities.8 tive indicators for a particular construct can at
times be a difficult one to make. The researcher
needs to think carefully about whether it is more
Construct-measurement relationships
correct to think of the underlying construct as
The nature of the links between constructs and 'causing' the observed measures (i.e., a reflective
measures are referred to as epistemic relation-relationship) or of the measures as 'causing' (or
ships, or 'rules of correspondence' (Bagozzi,
defining) the construct (i.e., a formative
1984; Fornell, 1982). Two basic types of epis-
relationship). An example of the former might
temic relationships are relevant to causal mode-
be performance, whereas an example of the latter
might be social economic status (SES). From a
ling: reflective indicators and formative indicators.
In the first case, indicators (measures) are statistical perspective, use of formative indicators
believed to reflect the unobserved, underlying tends to increase the R2 value for the endogenous
construct, with the construct giving rise to (or (i.e., predicted) constructs, although this effect is
'causing') the observed measures. In contrast, usually not large. Thus, use of formative indi-
formative indicators define (or 'cause') the con- cators tends to eliminate the need for the exoge-
struct. A defined construct is completely deter- nous constructs, since all explanation is 'pushed'
mined by a linear combination of its indicators. towards the endogenous variables.
The discussion to this point has assumed thatWhether researchers use formative or reflective
all indicators are reflective. When unobservable, relationships in their models, their choice of a
underlying constructs are viewed as giving rise particular form of epistemic relationship should
to associated measures, it is appropriate to talk be both justified clearly and applied consistently.
about item reliability and convergent validity. The studies by both Birkinshaw et al. and Cool
However, this is not necessarily true for formative et al. fail to describe the epistemic relationships
indicators (Bollen and Lennox, 1991; Cohen et between measures and constructs. However, it
al., 1990; MacCallum and Browne, 1993). In appears quite likely that reflective relationships
fact, formative indicators of the same construct were used exclusively in both studies. Conse-
'can have positive, negative, or no correlation' quently, the earlier discussions of reliability and
with one another (Bollen and Lennox, 1991: 307). validity should hold in both cases.
Consequently, observed correlations among the Forell et al. use both formative and reflective
measures associated with a construct may not relationships
be in their model. They provide a clear
argument for choosing one form of epistemic
relationship over the other for each of their four
8 In fact, a quick look at the two items used by Johansson
and Yip to measure management processes-global budgeting constructs. Although it is possible to question
and global group meetings-suggests that they are likely whether
to or not Fornell et al.'s choice of formative
be closely related to the three items used to measure global
strategy-standardized products, integrated competitiveindicators is sufficiently complete for all of their
moves, and overall global strategy. constructs, the natures of the relationships studied
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
202 J. Hulland
Model goodness-of-fit
LISREL and other covariance structure analysis CONCLUSION
modeling approaches involve parameter esti-
mation procedures which seek to reproduce asBy combining and confronting theory with d
closely as possible the observed covariance (Forell, 1982), and by forcing researchers
matrix. In contrast, PLS has as its primary objec- be explicit about both their measurement
tive the minimization of error (or, equivalently, theoretical assumptions (Bagozzi, 1980, 198
the maximization of variance explained) in all causal models such as PLS can help strateg
endogenous constructs. The degree to which any management researchers to achieve new insigh
particular PLS model accomplishes this objective
can be determined by examining the R2 values
'0 Even if the standards used to evaluate covariance structure
for the dependent (endogenous) constructs.
analysis (e.g., LISREL) models could be appropriately applied
One consequence of this difference in objec- to models estimated using PLS, the results presented by
tives between LISREL and PLS is that no proper Johansson and Yip are questionable in at least two respects.
overall goodness-of-fit measures exist for modelsFirst, the choice of the Bentler and Bonett (1980) normed fit
index (NFI) as the sole arbiter of model fit is a poor one,
estimated using the latter. Although existing PLSsince NFI is easily influenced by sample size and therefore
algorithms report goodness-of-fit statistics such as often biased. After carefully reviewing a wide variety of
the Bentler-Bonett normed fit index (Bentler and goodness-of-fit measures, Gerbing and Anderson (1993) rec-
ommended use of several alternative measures, but not NFI.
Second, although exact standards for evaluating goodness-of-
fit measures have not yet been formally established, various
9 Only the performance construct is modeled as reflective by pragmatic rules of thumb do exist. For example, for the NFI
Fornell et al. As the earlier discussion noted, this construct measure, researchers often suggest that a value of 0.95 or
exhibits acceptable convergent and discriminant validity. How- greater indicates a strong model, a value between 0.90 and
ever, one of the retained items has a very low loading (0.38), 0.949 indicates an adequate model, and a value of less than
and should arguably be dropped. The content validity of 0.90 indicates poor model fit (Hulland, Chow, and Lam,
some of Fornell et al.'s formative constructs (particularly 1996). Johansson and Yip report NFI values ranging from
stakeholder strength, which includes only two rather abstract 0.26 (Model 1) to 0.46 (Models 7 and 8), suggesting that
measures) can also be called into question. none of their models adequately fit the observed data.
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
Review of PLS Use in Strategic Management Research 203
Although PLS demands a level of rigor and dom on measurement: A structural equation perspec-
clarity beyond that required by more traditional tive', Psychological Bulletin, 110 (2), pp. 305-314.
Carmines, E.G. and R.A. Zeller, (1979). 'Reliability
methodological approaches, as the field of stra- and validity assessment'. Sage University Paper Ser-
tegic management continues to mature researchers ies on Quantitative Applications in the Social
need to increasingly rise to the challenge of Sciences, No. 07-017, Sage, Beverly Hills, CA.
meeting such demands. Cohen, P., J. Cohen, J. Teresi, M. Marchi and C. N.
However, researchers must use these new tech- Velez (1990). 'Problems in the measurement of
latent variables in structural equations causal mod-
niques appropriately. In addition to dealing with els', Applied Psychological Measurement, 14(2),
the methodological concerns associated with more pp. 183-196.
traditional methods of analysis, researchers using Cool, K., I. Dierickx, and D. Jemison (1989). 'Business
causal modeling approaches such as LISREL and strategy, market structure and risk-return relation-
ships: A structural approach', Strategic Management
PLS must understand their underlying assump-
Journal, 10(6), pp. 507-522.
tions and limitations. As the preceding review of Duxbury, L.E. and C.A. Higgins (1991). 'Gender
four recent uses of PLS in the strategic man- differences in work-family conflict', Journal of
agement area shows, the technique has been Applied Psychology, 76, pp. 60-74.
applied with considerable variability. The study Fornell, C. (1982). A Second Generation of Multivari-
by Cool et al. provides an excellent example for ate Analysis, Vol. 1. Praeger, New York.
Forell, C. and F. Bookstein (1982). 'Two structural
future applications of PLS. In contrast, some equations models: LISREL and PLS applied to con-
aspects of Johansson and Yip's use of PLS are sumer exit-voice theory', Journal of Marketing
seriously flawed, providing an important contrast Research, 19, pp. 440-452.
to the work by Cool et al. Forell, C. and D. F. Larcker (February 1981). 'Evalu-
ating structural equation models with unobservable
variables and measurement error', Journal of Mar-
keting Research, 18, pp. 39-50.
ACKNOWLEDGEMENTS Fornell, C., P. Lorange, and J. Roos (1990). 'The
cooperative venture formation process: A latent vari-
ablefinancial
The author gratefully acknowledges the structural modeling approach', Management
Science,and
support provided by the Social Sciences 36(10), pp. 1246-1255.
Gerbing, D. W. and J. C. Anderson (1993). 'Monte
Humanities Research Council of Canada, the
Carlo Ivey of goodness-of-fit indices for
evaluations
Business School, and the Barford Family.
structuralHe equation models'. In K. A. Bollen and
also appreciates the thoughtful comments of (eds.),
J. S. Long two Testing Structural Equation Mod-
els. Sage, Newbury Park, CA. pp. 40-65.
reviewers and the editor on an earlier version of
Hagedoor, J. and J. Schakenraad, (1994). 'The effect
this paper.
of strategic technology alliances on company per-
formance', Strategic Management Journal, 15(4),
pp. 291-309.
REFERENCES Hulland, J.S., Y.H. Chow and S. Lam (1996). 'Use
of causal models in marketing research: A review',
International Journal of Research in Marketing,
Bagozzi, R.P. (1980). Causal Models in Marketing.
Wiley, New York. 13(2), pp. 181-197.
Hulland, J. S. and D.N. Kleinmuntz (1994). 'Factors
Bagozzi, R.P. (1984). 'A prospectus for theory con-
influencing the use of internal summary evaluations
struction in marketing', Journal of Marketing, 48
versus external information in choice', Journal of
(Winter), pp. 11-29.
Barclay, D.W., C. Higgins and R. ThompsonBehavioral (1995). Decision Making, 7(2), pp. 79-102.
Johansson, J.K. and G. S. Yip (1994). 'Exploiting glob-
'The partial least squares (PLS) approach to causal
modeling: Personal computer adaptation and use as
alization potential: U.S. and Japanese strategies',
an illustration', Technology Studies, 2(2),Strategic pp. 285- Management Journal, 15(8), pp. 579-601.
309. Joreskog, K. A. and D. Sorbom (1989). LISREL 7
User's Reference Guide (1st ed.). Scientific
Bentler, P.M. and D.G. Bonett (1980). 'Significance
Software, Mooresville, IN.
tests and goodness of fit in the analysis of covariance
MacCallum, R. C. and M. W. Browne (1993). 'The use
structures', Psychological Bulletin, 88(3), pp. 588- of causal indicators in covariance structure models:
606.
Birkinshaw, J., A. Morrison and J. Hulland (1995). Some practical issues', Psychological Bulletin,
'Structural and competitive determinants of a global
114(3), pp. 533-541.
Nunnally, J. C. (1978). Psychometric Theory (2nd ed.).
integration strategy', Strategic Management Journal,
McGraw-Hill, New York.
16(8), pp. 637-655.
Bollen, K. and R. Lennox (1991). 'Conventional wis- Smith, J. B. and D. W. Barclay (January 1997). 'The
effects of organizational differences and trust on the
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms
204 J. Hulland
Wold, H. (1985).
effectiveness of selling partner relationships', Jour- 'Systems analysis by partial least
nal of Marketing, 61, pp. 3-21. squares'. In P. Nijkamp, L. Leitner, and N. Wrigley
Venkatraman, N. (1989). 'Strategic orientation (eds.), Measuring
of busi-the Unmeasurable. Marinus
ness enterprises: The construct, dimensionality, Nijhoff, Dordrecht,
and pp. 221-251.
measurement', Management Science, 35(8), Zinkhan, G. M., E. Joachimsthaler, and T. C. Kinnear
pp. 942-962. (May 1987). 'Individual differences and marketing
Wold, H. (1974). 'Causal flows with latent variables', decision support system usage and satisfaction',
European Economic Review, 5, pp. 67-86. Journal of Marketing Research, 24, pp. 208-214.
Copyright ? 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 195-204 (1999)
This content downloaded from 14.139.108.34 on Thu, 06 Jul 2017 14:10:44 UTC
All use subject to http://about.jstor.org/terms