Professional Documents
Culture Documents
Meta-Analytic Research in
International Business and
International Management
Peter J. Buckley a, Timothy M. Devinney a, and Ryan W. Tang b
a
Centre for International Business, University of Leeds; and b University of Technology, Sydney
Introduction
Reprinted with permission from Emerald Group Publishing Limited, All rights
reserved. Philosophy of Science and Meta-Knowledge in International Business and Manage-
ment Advances in International Management (2013) 26, 263–297.
100
P. J. Buckley, The Multinational Enterprise and the Emergence of the Global Factory
© Palgrave Macmillan, a division of Macmillan Publishers Limited 2014
Meta-Analytic Research in International Business and International Management 101
the narrative literature reviews and the primary studies by integrating extant
research with a particular focus or goal. Meta-analytic techniques can also be
powerful tools to improve IB/IM research by highlighting where construct
and model validity is weak (Farley & Lehmann, 2001). Hence, IB/IM is not
just an area where meta-analytic approaches can add great value, but meta-
analytic approaches are arguably critical to filling in the contextual modera-
tors that are the hallmark of IB/IM scholarship.
This chapter aims to (1) review meta-analyses in IB/IM discipline, (2) out-
line major issues for evaluating meta-analyses and (3) examine methodologi-
cal issues in the extant IB/IM meta-analyses. To achieve these goals, the next
section takes an overview of meta-analytic studies in the IB/IM discipline.
MIR, 3
JIBS, 6
IBR, 2
JIM, 2
JWB, 2
1
No meta-analyses were published between 2000 and 2003 in the top five journals.
Before 2000, there was only one published meta-analysis: Peterson and Jolibert (1995).
104 Peter J. Buckley et al.
4 1.40%
1.20%
3
1.00%
0.80%
2
0.60%
0.40%
1
0.20%
0 0.00%
2004 2005 2006 2007 2008 2009 2010 2011 2012
Culture
The effect of cultural factors is not only a prevalent IB/IM research topic but
also a crucial meta-construct in IB/IM research per se, reflecting as it does
the international environment in which MNCs reside.
Tihanyi et al. (2005) and Magnusson et al. (2008) examined how culture
influenced the MNC’s entry mode choice and performance. These two
studies found similar results – the effects of cultural factors on entry mode
and performance are not statistically significant, but MNC’s home country,
industry sector, time and operationalization of cultural difference/distance
significantly moderated the effective relationship. For example, the patterns
of US-based and Europe-based MNCs’ entry strategies showed a substantial
Meta-Analytic Research in International Business and International Management 105
Multinational Performance
MNC performance has been investigated in relationship to other variables.
Besides Tihanyi et al. (2005) and Magnusson et al. (2008), who examined
the relationship between performance and cultural values, Bausch and Krist
(2007) and Yang and Driffield (2012) integrated findings about the relation-
ship between an MNC’s globalization and its performance via testing various
moderator effects. Among the moderators, MNC’s country of origin is seen as
a significant factor influencing the globalization-performance relationship.
Likewise, van Essen, Heugens, Otten, and van Oosterhout (2012) examined
the relationship between MNC performance and executive compensation.
These authors found a moderately positive but considerately significant
relationship between MNC performance and compensation, although this
relationship varied dramatically across countries, arguably due to the dif-
ferent institutional structures operating in those countries. Reus and Rottig
(2009) analysed factors that may determine an MNC’s performance in
different host countries and found unique features of MNCs in China, with
106 Peter J. Buckley et al.
Entry Mode
Tihanyi et al. (2005) and Magnusson et al. (2008) analysed how entry
mode choice was impacted by cultural factors, while Zhao et al. (2004) and
Morschett et al. (2010) studied the determination of entry strategies. Zhao
et al. (2004) conducted a meta-analysis based on transaction cost theory
that helps the study construct six dimensions of determinants (e.g., cultural
distance, country risk, international experience and advertising intensity).
Morschett et al. (2010) was not constrained by a single theoretical paradigm
and examined 13 external determinants of entry mode choice (e.g., cultural
distance, country risk, market growth and volatility of demand). However,
they did not include internal factors (e.g., international experience and
advertising intensity) and thus any generalization of the results is limited.
Obviously, a strategic decision related to international entry is impacted
by its experience (e.g., Delios & Beamish, 1999) and constrained by its
marketing ability (e.g., Dikova & van Witteloostuijn, 2007).
Remaining Studies
Except for the preceding topics, the three remaining meta-analytic stud-
ies selected different research focuses. García-Meca and Sánchez-Ballesta
(2006) found significant moderator effects of country, measurement, and
time by synthesizing studies on factors that influence the accuracy of
financial analysts’ forecasts. Meyer and Sinani (2009) statistically reviewed
the literature regarding positive spillovers of foreign direct investment.
The authors revealed a curvilinear relationship between spillovers and the
development levels of host countries. Judge, McNatt, and Xu (2011) inte-
grated previous research on causes as well as effects of national corruption.
The results suggest the political/legal effects had the strongest relationship
with corruption.
The 15 meta-analytic studies published in the top 5 IB/IM journals in the
21st century reveal a limited overview and synthesis of the domain to date.
Nevertheless, these studies are a significant improvement in the application
of meta-analytic approaches compared to earlier work. For instance, the first
published meta-analysis in our review (i.e., Zhao et al., 2004) aggregated
data from only 38 primary studies, whereas Steel and Taras (2010) inte-
grated data from 508 primary studies and Taras et al. (2012) summarized
findings from 451 primary studies. Furthermore, in more recent years, IB/
IM meta-analysts have applied more deliberative and sophisticated methods
to model the data compiled from more sizeable literatures. For example,
Fischer and Mansell (2009) and van Essen et al. (2012) applied multivariate
modelling techniques to synthesize the findings, while the studies published
Meta-Analytic Research in International Business and International Management 107
in the early years of the last decade (e.g., García-Meca & Sanchez-Ballesta,
2006) employed only very basic statistical approaches (e.g., distribution and
relationship) for disclosing patterns of integrated data. The preceding exam-
ples hopefully underscore the fact that meta-analytic methods are gradually
being applied to answer more complex questions and explain more deep-
seated phenomena requiring more sophisticated and nuanced approaches.
Yet, one must acknowledge that larger sample sizes and/or more com-
plicated modelling do not pari passu lead to a better meta-analytic study.
In particular, as mentioned previously and discussed subsequently, meta-
analyses in IB/IM still have considerable space in which to make progress.
Furthermore, IB/IM researchers need to harken back to the criteria for
conducting a valid meta-analysis if these studies are going to be useful in
truly synthesizing and integrating what we do and do not know. The next
section provides nonstatistical guidelines for assessing meta-analytic studies
in efforts to outline the evaluation of meta-analytic studies. These guidelines
are synthesized from major methodological literature with considerable
references to widely applied and commonly accepted practices published in
major journals. Since the statistical details regarding the calibrations (e.g.,
formula, calculation, critical value, etc.) are not in the scope of this chapter,
IB/IM researchers are referred to meta-analytic literature – such as Cooper et al.
(2009), Hunter and Schmidt (2004), Lipsey and Wilson (2001), Rosenthal
(1991) and Hedges and Olkin (1985) – for more detailed general discussion.
Research Specification
Similar to any empirical study, meta-analysis requires clearly defined vari-
ables that can represent the questions that the study will answer, the phe-
nomena the study will explain or/and the research topic that the study
will address. The variables should be able to reflect the theoretical and
conceptual frameworks that underlie the arguments of the meta-analytic
study. Also, the research topic should be stated explicitly to validate the
variables defined in the study. All of the 15 IB/IM meta-analyses addressed
the 4 aspects: explicitly stated research topic, the explanation of concept in
the theoretical and/or practical context, well-defined variable, and the appli-
cability of the research question to meta-analysis. For example, Taras et al.
(2012) meta-analysed national cultural indices within Hofstede’s framework
(e.g., Hofstede, 1980; Hofstede, Hofstede, & Minkov, 2010) by focusing on
the four culture dimensions (i.e., power distance, individualism, masculinity
and uncertainty avoidance) that had been investigated for three decades and
where the work appeared in around 100 journals.
However, 2 of the 15 meta-analyses did not set up the theoretical and
practical contexts (i.e. García-Meca & Sánchez-Ballesta, 2006; Yang &
Driffield, 2012). The theoretical and practical context that serves as the
basis of the meta-analysis has to be verified in order to establish meaning-
ful research questions and build connections between concepts (i.e., vari-
ables) and concrete events (i.e., the proxies or operationalizations of the
variables). Nevertheless, García-Meca and Sánchez-Ballesta (2006) failed to
embed their meta-analysis into a context that demonstrates the rationale
for searching effect sizes (i.e., six categories of correlation coefficients)
and moderators (e.g., measurements of variables), although they briefly
mentioned the importance of their research and reviewed extant literature
related to variables of interest. Yang and Driffield (2012) applied an attrac-
tive effect size metric (i.e., estimate of coefficient in a linear model) that
may disclose causality. However, these authors did not verify the logic for
choosing account-and market-based firm performance and controlling firm
characteristics and business cycle effects. They also fail to mention how
they define the degree of multinationality, which is a notable exclusion
given that their research was aimed at examining the multinationality–
performance relationship.
Literature Retrieval
The first concern regarding literature retrieval is the sources of primary
studies. As the availability of studies varies across databases, a well-
conducted meta-analysis often searches potential studies in multiple data-
bases for broader and more complementary literature. Other important
sources include journals that concentrate on a certain area, references cited
by articles reviewing previous literature, and researchers who are expert in
a related research field. Multiple-source searching is always prior to single
110 Peter J. Buckley et al.
one, because any one or two sources may not provide comprehensive and
exhaustive primary studies. An unrepresentative literature pool resulting
from limited sources may bias the conclusions of a meta-analysis (Devinney &
Tang, this volume).
The studies examined were quite limited in terms of their breadth and
depth of literature examined. Only three articles (i.e., Stahl et al., 2010;
Steel & Taras, 2010; van Essen et al., 2012) took into account multiple
databases, major journals and published reviews simultaneously. Of the 12
remaining meta-analyses, 5 took advantage of two data sources (i.e., mul-
tiple databases and major journals, or both journals and published reviews
in a research area), while four drew from only one of the three major data
sources (i.e., multiple databases, specific journals or published reviews).
Three meta-analyses searched potential extant literature only in one single
individual database but not multiple databases (i.e., Judge et al., 2011; Reus &
Rottig, 2009; Yang & Driffield, 2012).
Similar to data sources, key terms and their combinations (e.g., AND,
OR) are important for literature retrieval. The implication of this impor-
tance is threefold. First, a good meta-analysis reports what key terms are
utilized. Second, the terms precisely represent the research topic of inter-
est. Third, the terms and their combinations are related to the inclusion
of primary studies. In IB/IM meta-analyses, both pieces of information are
reported simultaneously in only one study (i.e., Stahl et al., 2010). Key
terms but not combinations are provided in four other meta-analyses,
while the remaining 10 IB/IM meta-analyses did not mention what key
terms were searched in the data sources. Authors of the 10 articles might
have employed some keywords in literature searching, but the fact that
they did not disclose the key terms used makes it difficult for readers to
know how the eligible primary studies were retrieved. In particular, other
researchers who want to examine those meta-analytic studies will not be
able to replicate them.
In addition, other strategies for searching previous studies are employed
by many authors of meta-analyses because of the complementarity of com-
mon searching approaches. For example, e-mail listservers and contacting
authors are ways for collecting unpublished studies from potential research-
ers. Browsing candidate literature selected in the previous steps may find
missing studies in eligible bibliography. In addition, although the majority
of academic research is published or written in English, non-English studies
may also investigate research topics of meta-analyst’s interest.
These strategies were applied in meta-analytic studies published by one
journal more frequently than the other. Specifically, in the JIBS-published
meta-analyses all were comprehensive in concentrating on the authors of
publication. On the other hand, they never mentioned e-mail listservers,
and all the top five IB/IM journals published meta-analytic articles that
Meta-Analytic Research in International Business and International Management 111
Dataset Preparation
Meta-analysis requires a common effect size (i.e., effect size metric, effect size
estimation) to represent the quantitative findings that can be standardized
across studies for meaningful and numerical analyses. Three types of effect
size metrics are commonly used: standardized mean difference (i.e., Hedges’
g, Cohen’s d, and Glass’ gGlass), correlation (e.g., Pearson correlation r, partial
correlation rxy,z) and odds ratio (i.e., i or o) (Devinney & Tang, this volume).
Judgement on the selection of effect size metric is linked to the research
questions specified initially. In other words, a good meta-analysis selects the
effect size metric that can demonstrate the essential concept of research topic
(i.e., the phenomenon, association or causality). The 15 IB/IM meta-analyses
applied typical effect size metrics (i.e., mean, correlation and odds ratio) as well
as some uncommon ones (e.g., t-statistic in Meyer & Sinani, 2009). These effect
size metrics were generally well-justified for answering the research questions.
However, inclusion criteria were not explicitly disclosed in all IB/IM
meta-analyses. Four studies did not explain what criteria they used to screen
potential literature. Inclusion criteria are important because they define
the eligible studies for meta-analysis and explain what the meta-analysis is
about. The criteria are also important for other researchers who may want to
replicate the meta-analysis and understand the analysing scope.
Coding procedure was more frequently explained by IB/IM meta-analysts
than inclusion criteria, because the procedure collects data from eligible
studies according to coding protocol that illustrates the definition and
scope of effect size and moderator. This coding information helps users of
the meta-analysis understand what data are included as well as excluded
(Devinney & Tang, this volume). An entire protocol may be too long to
be enclosed in a journal article, but a brief introduction on how the meta-
analyst codes the variables (i.e., what measurements or operationalizations
represent a variable) is helpful for understanding the results, as well as
assessing quality, of the meta-analysis. In addition, multiple-coder cod-
ing is an important feature of a good meta-analysis because it reduces the
bias associated with a single coder. The reliability (e.g., agreement rate and
Cohen’s ϕ) of coders for the same literature pool is a crucial value to evaluate
a meta-analytic study.
In IB/IM meta-analyses, authors of 12 of the meta-analyses gave coding
information and half of these articles employed multiple coders to collect
data from primary studies. In particular, the method of multiple coding was
applied more often in studies published after 2009. Specifically, six of the
112 Peter J. Buckley et al.
seven multi-coded meta-analyses were published after 2009, but only one
before 2009. It seems to suggest that the reliability of data collection has
gradually improved in recent IB/IM meta-analyses.
Furthermore, correction of artefacts aims to adjust the imperfect raw
data collected in primary studies for desirable statistic power (e.g., from
correlation r to Fisher’s z) and accurate estimation (e.g., correcting artificial
dichotomization by the numbers and proportions). Only five meta-analyses
published after 2009 corrected for effect sizes, while none did so before. In
particular, the meta-analyses published in JIBS are more likely to correct for
effect sizes (i.e., three out of six did this).
In addition, the published meta-analytic studies in the top five IB/IM jour-
nals indicate that most IB/IM meta-analysts were not aware of the outliers
in their datasets. Those values potentially result in a skewed distribution of
effect size and potentially bias meta-analytic findings. However, only one
IB/IM meta-analysis discussed outliers. There might be three reasons for the
majority of authors not doing so. First, their meta-analytic datasets did not
have any outliers. Nevertheless, a perfect raw-dataset hardly exists. Despite
having a perfect dataset it is still practical to let readers know there are no
outliers. Second, the outliers had been trimmed or recoded, but not men-
tioned in the articles. The adjustment of original datasets by trimming or
recoding is important information for other researchers to judge and evalu-
ate the data collection process, and therefore should be disclosed. Lastly,
some IB/IM meta-analysts might think it was not necessary to take outliers
into account. In fact, the distribution of an original dataset is always skewed
by extreme values, but the typical purpose of meta-analysis is to summarize
the overall findings in a research domain (Lipsey & Wilson, 2001).
Data Analysis
The 15 IB/IM meta-analyses described effect sizes by the mean, standard devi-
ation and confidence intervals. These statistics can demonstrate the distribu-
tion of effect size, but only mean effect size was presented in every article.
Standard deviations were available in 10 meta-analytic studies. Confidence
intervals were only provided by four meta-analyses. In contrast to the
authors who did not provide these essential statistics, some meta-analysts
computed extra statistical data to facilitate their discussions. For example,
Fischer and Mansell (2009) and Steel and Taras (2010) provide the correlation
matrices of effect sizes and moderators to infer the relationships of interest.
An important item of meta-analytic data analysis is homogeneity analysis.
It provides evidence of the variability of studies by computing the prob-
ability that the variance of effect size is only due to sampling error. This is
crucial for meta-analysts to figure out whether significant moderator effects
exist in effect sizes (Devinney & Tang, this volume). However, only eight
IB/IM meta-analyses present this item. Most of them (i.e., five out of eight)
were published in JIBS, with only 1 IB/IM meta-analysis not presenting
Meta-Analytic Research in International Business and International Management 113
In Taras et al. (2012), additional data were collected from public databases
(e.g., the World Bank and the United Nations) and provided extra evidence
on how national cultural values evolved over the decades. Likewise, when
integrating the institution-based view into the meta-analysis of contracting-
theory-based literature, van Essen et al. (2012) collected external data on
institutional variables that were not available in the examined literature.
Their meta-analytic findings extended the previous understanding on
firm performance–executive compensation relationship from a theoretical
framework to another paradigm.
Analytic Report
Major results of meta-analytic studies are presented by stating major
findings, generality and limitations of the meta-analysis. If the statements
are facilitated by figures, graphs and tables, the meta-analytic report can be
more efficient and effective. Another critical item for meta-analyst’s report-
ing is the included studies. Users of meta-analysis need to know which
primary studies led to final conclusions. If inappropriate primary studies
are included, the meta-analysis based on them may not present to valid or
generalizable findings. In particular, the reference list of included studies is
evidence of the validity of the inclusion criteria.
In the top 5 IB/IM journals, all meta-analyses reported findings, generality of
conclusions, and included studies by illustrating them in tables, while only 8 of
15 applied figures and graphs. Among the eight studies, seven were published
after 2009. Furthermore, three among the 15 meta-analyses did not explicitly
report the methodological limitations of their meta-analytic procedures (i.e.,
Magnusson et al., 2008; van Essen et al., 2012; Yang & Driffield, 2012).
The preceding discussion is summarized in Table 5.2. It is not intended
to form an exclusive checklist for calibrating all facets of these IB/IM meta-
analyses, because this review evaluates the meta-analytic methodologies
used rather than the specifics of the IB/IM implications. In addition, both
the IB/IM discipline and meta-analytic methodologies are still developing.
Progress may update the criteria for judgement. Finally, the procedures
underlying meta-analyses vary. This chapter is not able to cover all the
details of all types of meta-analyses, but only capable of highlighting the
important and general issues.
Findings
Bausch & Magnusson Fischer & Meyer & Reus & Rottig
Krist (2007) et al. (2008) Mansell (2009) Sinani (2009) (2009)
Y Y Y Y Y
Y Y Y Y Y
Y Y Y Y Y
Y Y Y Y Y
Y Y
Y
Y Y
Y Y
Y
Y
Y Y
Y Y Y Y Y
Y Y Y Y
Y Y Y Y Y
Y
Y Y
Y
Y Y Y Y Y
Y Y
Y Y
Y Y Y
Y Y Y Y
Y Y Y Y Y
Y Y Y Y
Y
Y Y
Y Y Y Y Y
Y Y Y Y Y
Y Y Y Y
Y Y Y
Y Y Y Y Y
Y Y Y (a.r.) Y Y
20 17 23 23 21
(continued)
118
Judge et al. (2011) van Essen et al. (2012) Taras et al. (2012) Yang & Driffield (2012)
The antecedents The relationship The national cultural The relationship between
and effects between firm indices multinationality and
of national performance performance
corruption and executive
compensation
Y Y Y Y
Y Y Y
Y Y Y Y
Y Y Y Y
Y Y
Y Y
Y
Y Y
Y
Y Y
Y Y Y Y
Y Y
Y Y Y
Y Y Y
Y Y
Y Y Y Y
Y Y
Y
Y
Y Y Y Y
Y Y
Y
Y Y
Y Y Y Y
Y Y Y Y
Y Y
Y Y
Y Y Y Y
Y Y Y Y
19 24 20 13
Table 5.3 Summary of major issues in meta-analyses published in the Top-Five IB/IM Journals
120
Total (15) JIBS (6) JWB (2) JIM (2) IBR (2) MIR (3)
Count Ratio Count Ratio Count Ratio Count Ratio Count Ratio Count Ratio
Section 1: Stated research topic 15 100% 6 100% 2 100% 2 100% 2 100% 3 100%
Research Theoretical and 13 87% 6 100% 2 100% 2 100% 1 50% 2 67%
specificaion practical context
Well-defined variable 15 100% 6 100% 2 100% 2 100% 2 100% 3 100%
Applicable to 15 100% 6 100% 2 100% 2 100% 2 100% 3 100%
meta-analysis
Section 2: Multiple databases 8 53% 3 50% 1 50% 1 50% 2 100% 1 33%
Literature Search in specific 9 60% 4 67% 1 50% 2 100% 1 50% 1 33%
retrieval journals
Published reviews 6 40% 4 67% 0 0% 2 100% 0 0% 0 0%
Key terms 5 33% 4 67% 1 50% 0 0% 0 0% 0 0%
Combination of 1 7% 1 17% 0 0% 0 0% 0 0% 0 0%
key terms
E-mail listserver 2 13% 0 0% 0 0% 1 50% 0 0% 1 33%
Contacting author 5 33% 5 83% 0 0% 0 0% 0 0% 0 0%
Selected literature 5 33% 1 17% 1 50% 1 50% 1 50% 1 33%
Non-English works 0 0% 0 0% 0 0% 0 0% 0 0% 0 0%
Section 3: Appropriate 15 100% 6 100% 2 100% 2 100% 2 100% 3 100%
Dataset effect size metric
preparation Inclusion criteria 11 73% 4 67% 2 100% 2 100% 2 100% 1 33%
Coding information 12 80% 5 83% 2 100% 1 50% 2 100% 2 67%
Multiple coders 7 47% 4 67% 2 100% 1 50% 0 0% 0 0%
References
Bausch, A., & Krist, M. (2007). The effect of context-related moderators on the interna-
tionalization-performance relationship: Evidence from meta-analysis. Management
International Review, 47, 319–347.
Bausell, R. B., Li, Y.-F., Gau, M.-L., & Soeken, K. L. (1995). The growth of meta-analytic
literature from 1980 to 1993. Evaluation & the Health Professions, 18, 238–251.
Bohlin, I. (2012). Formalizing syntheses of medical knowledge: The rise of meta-
analysis and systematic reviews. Perspectives on Science, 20, 273–309.
Brouthers, K. D., & Hennart, J.-F. (2007). Boundaries of the firm: Insights from inter-
national entry mode research. Journal of Management, 33, 395–425.
Buckley, P. J. (2002). Is the international business research agenda running out of
steam? Journal of International Business Studies, 33, 365–373.
Canabal, A., & White, G. O., III. (2008). Entry mode research: Past and future.
International Business Review, 17, 267–284.
Card, D., & Krueger, A. B. (1995). Time-series minimum-wage studies: A meta-analysis.
The American Economic Review, 85, 238–243.
Cooper, H. M., & Hedges, L. V. (1994). The handbook of research synthesis. New York,
NY: Russell Sage Foundation.
Cooper, H. M., Hedges, L. V., & Valentine, J. C. (2009). The handbook of research
synthesis and meta-analysis. New York, NY: Russell Sage Foundation.
Delios, A., & Beamish, P. W. (1999). Ownership strategy of Japanese firms:
Transactional, institutional, and experience influences. Strategic Management
Journal, 20, 915–933.
Dikova, D., & van Witteloostuijn, A. (2007). Foreign direct investment mode choice:
Entry and establishment modes in transition economies. Journal of International
Business Studies, 38, 1013–1033.
Farley, J. U., & Lehmann, D. R. (2001). The important role of meta-analysis in inter-
national research in marketing. International Marketing Review, 18, 70–79.
Fischer, R., & Mansell, A. (2009). Commitment across cultures: A meta-analytical
approach. Journal of International Business Studies, 40, 1339–1358.
García-Meca, E., & Sánchez-Ballesta, J. P. (2006). Influences on financial analyst fore-
cast errors: A meta-analysis. International Business Review, 15, 29–52.
Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational
Researcher, 5, 3–8.
Harzing, A.-W. (2012). Journal quality list, 2012. Retrieved from http://www.harzing.
com/jql.htm
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL:
Academic Press.
Hofstede, G. H. (1980). Cultures consequences: International differences in work-related
values. Beverly Hills, CA: Sage.
Hofstede, G. H., Hofstede, G. J., & Minkov, M. (2010). Cultures and organizations:
Software of the mind: intercultural cooperation and its importance for survival (3rd ed.).
New York, NY: McGraw-Hill.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: correcting error and bias
in research findings. Newbury Park, CA: Sage.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias
in research findings. Beverly Hills, CA: Sage.
ISI (2012). Journal citation reports. Thomson Reuters, 2012. Retrieved from http://admin-
apps.webofknowledge.com.ezproxy.lib.uts.edu.au/JCR/JCR?SID=U1j22FJ1D56
cj4CcJlb&locale=en_US
124 Peter J. Buckley et al.
Judge, W. Q., McNatt, D. B., & Xu, W. (2011). The antecedents and effects of national
corruption: A meta-analysis. Journal of World Business, 46, 93–103.
Kirca, A. H., & Yaprak, A. (2010). The use of meta-analysis in international business
research: Its current status and suggestions for better practice. International Business
Review, 19, 306–314.
Kulik, J. A., Cohen, P. A., & Ebeling, B. J. (1980). Effectiveness of programmed instruc-
tion in higher education: A meta-analysis of findings. Educational Evaluation and
Policy Analysis, 2, 51–64.
Kulik, J. A., Kulik, C.-L. C., & Cohen, P. A. (1980). Effectiveness of computer-based
college teaching: A meta-analysis of findings. Review of Educational Research, 50,
525–544.
Lau, J., Antman, E. M., Jimenez-Silva, J., Kupelnick, B., Mosteller, F., & Chalmers, T. C.
(1992). Cumulative meta-analysis of therapeutic trials for myocardial infarction.
New England Journal of Medicine, 327, 248–254.
Lee, W.-L., Bausell, R. B., & Berman, B. M. (2001). The growth of health-related meta-
analyses published from 1980 to 2000. Evaluation and the Health Professions, 24,
327–335.
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-Analysis. Thousand Oaks, CA:
Sage.
Magnusson, P., Baack, D. W., Zdravkovic, S., Staub, K. M., & Amine, L. S. (2008). Meta-
analysis of cultural differences: Another slice at the apple. International Business
Review, 17, 520–532.
Malhotra, N. K., Agarwal, J., & Ulgado, F. M. (2003). Internationalization and
entry modes: A multi-theoretical framework and research propositions. Journal of
International Marketing, 11, 1–31.
Meyer, K. E., & Sinani, E. (2009). When and where does foreign direct investment
generate positive spillovers? A meta-analysis. Journal of International Business Studies,
40, 1075–1094.
Morschett, D., Schramm-Klein., H., & Swoboda, B. (2010). Decades of research on
market entry modes: What do we really know about external antecedents of entry
mode choice? Journal of International Management, 16, 60–77.
Peterson, R. A., & Jolibert, A. J. P. (1995). A meta-analysis of country-of-origin effects.
Journal of International Business Studies, 26, 883–900.
Reus, T. H., & Rottig, D. (2009). Meta-analyses of international joint venture perfor-
mance determinants: Evidence for theory, methodological artifacts and the unique
context of China. Management International Review, 49, 607–640.
Rosenthal, R. (1979). The file drawer problem and tolerance for null results.
Psychological Bulletin, 86, 638–641.
Rosenthal, R. (1984). Meta-analytic procedures for social research. Beverly Hills, CA: Sage.
Rosenthal, R. (1991). Meta-analytic procedures for social research (Rev. ed.). Newbury
Park, CA: Sage.
Schmidt, F. L., & Hunter, J. E. (1977). Development of a general solution to the prob-
lem of validity generalization. Journal of Applied Psychology, 62, 529–540.
Slangen, A., & Hennart, J.-F. (2007). Greenfield or acquisition entry: A review
of the empirical foreign establishment mode literature. Journal of International
Management, 13, 403–429.
Stahl, G. K., Maznevski, M. L., Voigt, A., & Jonsen, K. (2010). Unraveling the effects
of cultural diversity in teams: A meta-analysis of research on multicultural work
groups. Journal of International Business Studies, 41, 690–709.
Stanley, T. D., & Jarrell, S. B. (1989). Meta-regression analysis: A quantitative method
of literature surveys. Journal of Economic Surveys, 3, 161–170.
Meta-Analytic Research in International Business and International Management 125
Steel, P., & Taras, V. (2010). Culture as a consequence: A multi-level multivariate meta-
analysis of the effects of individual and country characteristics on work-related
cultural values. Journal of International Management, 16, 211–233.
Takahashi, K., Ishikawa, J., & Kanai, T. (2012). Qualitative and quantitative studies
of leadership in multinational settings: Meta-analytic and cross-cultural reviews.
Journal of World Business, 1–9. In press.
Taras, V., Steel, P., & Kirkman, B. L. (2012). Improving national cultural indices using
a longitudinal meta-analysis of Hofstedes dimensions. Journal of World Business, 47,
329–341.
Thomas, D. C., Cuervo-Cazurra, A., & Brannen, M. Y. (Eds.). (2011). Journal of
International Business Studies, (42, pp. 1073–1078).
Thompson, S. G., & Sharp, S. J. (1999). Explaining heterogeneity in meta-analysis:
A comparison of methods. Statistics in Medicine, 18, 2693–2708.
Tihanyi, L., Griffith, D. A., & Russell, C. J. (2005). The effect of cultural distance on
entry mode choice, international diversification, and MNE performance: A meta-
analysis. Journal of International Business Studies, 36, 270–283.
van Essen, M., Heugens, P. P., Otten, J., & van Oosterhout, J. H. (2012). An institution-
based view of executive compensation: A multilevel meta-analytic test. Journal of
International Business Studies, 43, 396–423.
Yang, Y., & Driffield, N. (2012). Multinationality-performance relationship.
Management International Review, 52, 23–47.
Zhao, H., Luo, Y., & Suh, T. (2004). Transaction cost determinants and ownership-
based entry mode choice: A meta-analytical review. Journal of International Business
Studies, 35, 524–544.
Appendix: Meta-analyses in top 5 IB Journals between 2004 and 2012
126
Panel A
Research Topic Theoretical paradigm Databases Journal searched Restriction Key terms
Zhao et al. JIBS Ownership-based Transaction ABI/Inform JIBS, MIR, AMJ, No n.r.
(2004) entry mode choice Cost Theory SMJ
Tihanyi et al. JIBS The effect of cultural n.r. ABI/Inform; JSTOR AMJ, ASQ, JIBS, No n.r.
(2005) distance on entry mode JM, JMS, MIR,
choice, international MS, SMJ
diversification, and MNE
Performance
García-Meca & IBR The factors that n.r. ScienceDirect; EJS JAR, JAE, TAR, No n.r.
Sánchez influence the accuracy Ebsco; SSRN; ABI JBR, JB, FA, AQS,
Ballesta (2006) of financial analysts’ Inform IBR, JFE
predictions
Bausch & MIR the relationship between Internalization; Business Source Premier; SMJ, AMJ, JIBS, No n.r.
Krist, 2007 internationalization and Learning; RBV; EconLit; ABI/Inform JBV, MIR, IBR
firm performance Eclectic
Magnusson IBR The effects of cultural Internationalization; ABI/Inform; Business n.r. No n.r.
et al. (2008) differences on MNE’s transaction cost Source Premier
entry strategy and theory
performance
Fischer & JIBS Employee’s commitment Commitment and PsycINFO n.r. No ‘Organizational
Mansell (2009) across cultures cultural theory commitment’
Meyer & JIBS The local spillovers Competitive Dynamics EconLit n.r. No ‘spillovers from
Sinani (2009) of FDI Theory (awareness- technology
motivation-capability transfer’,
framework) productivity
FDI spillovers’
Reus & MIR The determinants The Agency Theory, the ABI/Inform n.r. No n.r.
Rottig (2009) of IJV’s performance behavioural perspective
Stahl JIBS The cultural diversity Similarity-attraction, ABI/INFORM, Business Relevant books No ‘Team’, ‘group’;
et al. (2010) in work teams social identify & Source Premier, EconLit, and research ‘culture’,
categorization, PsychInfo, Science journals ‘diversity’;
information- Direct, and the Social ‘multicultural’,
processing Science Citation Index ‘international’,
‘multinational’
Morschett JIM The external antecedents Learning perspective Business Source JIBS, JIM, IBR, No n.r.
et al. (2010) entry mode choice of organizational Premier JBR, MIR
capabilities;
Transaction Cost
Theory
Steel & JIM The factors shaping Ecological Inference; EBSCO; PsycINFO; All relevant in No n.r.
Taras (2010) culture, the explanation Divergence & ERIC; ProQuest; and 1980–2006
of cultural variation, and Convergence Theory; ProQuest Digital
the relationship between Hofstede’s paradigm Dissertations;
individual and national
cultural values
Judge JWB The antecedents and Institutional choice ABI/Inform n.r. No ‘corruption’;
et al. (2011) effects of national perspective ‘many
corruption countries’
van Essen JIBS The relationship Institutional Based ABI/INFORM Global, 25 journals No ‘compensation’,
et al. (2012) between firm View EconLit, Google Scholar, ‘incentives’,
performance and JSTOR, SSRN, and ISI ‘pay’,
executive compensation Web of Knowledge ‘remuneration’,
‘salary’, ‘stock
option’
Taras JWB The national cultural Hofstede’s dimensions All major electronic Almost 100 No n.r.
et al. (2012) indices databases journals of
1980–2010
Yang & MIR The relationship between n.r. n.r. n.r. No n.r.
Driffield multinationality and
127
(2012) performance
(continued)
Appendix Continued
Panel A
128
Terms Other efforts Effect size Criteria for Coding Correcting for Identification
Combination metric inclusion reliability effect size of outliers
studies
(continued)
Appendix Continued
Panel B
130
Terms Other efforts Effect size Criteria for Coding Correcting for Identification
Combination metric inclusion reliability effect size of outliers
Judge JWB n.r. n.r. Correlation r Conceptualized Two coders; By the mean n.r.
et al. (2011) corruption; national interrater reliability;
level; multiple reliability
countries; empirical
studies; direct
relationship
van Essen JIBS n.r. Reviews and Pearson n.r. Two coders; Fisher’s z n.r.
et al. (2012) meta-analyses; correlations; subsample;
reference part partial Cohen’s κ
search (forward- correlation
tracing) of coefficients
selected articles;
contacting
authors;
including
unpublished
Taras JWB n.r. Reference parts Standardized Commensurability; Multiple n.r. n.r.
et al. (2012) of selected Cohen’s d instruments based coders;
studies; citation (mean) on Hofstede; test-retest
checking by empirical support for reliability
Google Scholar; convergent validity
mailing list-serves
of AIB and AOM;
unpublished
papers;
contacting
authors
Yang & Driffield MIR n.r. n.r. Estimate n.r. n.r. n.r. n.r.
(2012) of linear
coefficient
Panel B
Zhao JIBS Sample size weighted mean; Q statistics Category analyses n.r. n.r. 38
et al. (2004) standard error; z-value;
p-value; % of variance;
Tihanyi JIBS Sample size weighted mean; n.r. n.r. n.r. n.r. 66
et al. (2005) variances
García-Meca & IBR Sample size weighted mean; Q statistics Category analysis If other statistics Fail-safe N 38
Sánchez-Ballesta variances; confidence are available, then
(2006) intervals transfer
Bausch & MIR Sample size weighted mean; 75% rule Category analysis n.r. File-drawer 41
Krist (2007) confidence & credibility analysis;
intervals; z-statistics; calculation of
average residual variance fail-safe N(x)
Magnusson IBR Sample size weighted n.r. Category analysis Contacting authors n.r. 74
et al. (2008) mean
Fischer & JIBS Sample size weighted; Q statistics Category analysis; Missing standard n.r. 164 (means)
Mansell (2009) intercorrelation and three-level deviation is handled and 37
rankorder correlation hierarchical linear by the means; if (correlations)
modelling missing country
data, take region
Meyer & JIBS Mean, standard deviation, Heterogeneity Meta-analytic Missing data of a n.r. 66
Sinani (2009) correlation matrix test regression year replaced by the
nearest year
Reus & MIR Sample size weighted mean; 75% rule; Category analysis Missing reliability n.r. 66
Rottig (2009) variances; chi-square Q statistics handled by an
artifact distribution
method
131
(continued)
Appendix Continued
Panel B
132
Stahl JIBS Sample size, mean, Q statistics Sub-sample Eliminating the Fail-safe N 108
et al. (2010) confidence interval, range, comparison by missing categories
variance Z-statistics
Morschett et al. JIM Variance and sample-size Q statistics; Sub-sample n.r. n.r. 72
(2010) weighted mean forest plot comparison on
vote counting
scale
Steel & JIM Sample size weighted n.r. Hierarchical Missing GDP data n.r. 508
Taras (2010) mean, standard deviation, linear model were assumed
correlation matrix by that prior to
publication
Judge JWB Sample size weighted mean; n.r. Category analysis Contacting author; n.r. 42
et al. (2011) confidence and credibility missing reliability
intervals replaced by
estimate of National
Academy of Science
van Essen JIBS Inverse variance weighted Q statistics; Hierarchical linear Contacting author n.r. 332
et al. (2012) mean, standard error I index model, weighted
least squares-
based regression
Taras JWB Mean n.r. Category analysis n.r. n.r. 451
et al. (2012)
Yang & MIR Mean, standard deviation, n.r. Meta-analytic n.r. Regressing 67
Driffield (2012) sample size regression t-ratio of each
estimate
Reference Journal Other Information
(continued)
Appendix Continued
Panel B
134
Journals in IB are defined by Journal Qualtity List (47th ed.) (Harzing, 2012).
The rank is made according to Total Cites and Impact Factor of Journal Citation Reports (Web of Knowledge, 2011).
The above listed are published between 2000 and 2012; only one before 1995 by Peterson & Jolibert in JIBS; no meta-analyses were published between
2000 and 2003.
‘n.r.’ means not reported.
AIB – The Academy of International Business; AOM – the International Management Division of the Academy of Management; JIBS – the Journal of Inter-
national Business Studies; MIR –Management International Review; AMJ – Academy of Management Journal; SMJ – Strategic Management Journal; ASQ – Admin-
istrative Science Quarterly; AOS – Accounting, Organization and Society; JM – Journal of Management; FA – Finance and Accounting; JB – Journal of Business; JMS
– Journal of Management Studies; MS – Management Science; JAR – Journal of Accounting Research; JAE – Journal of Accounting and Economics; TAR – The
Accounting Review; JBR – Journal of Business Research; IBR – International Business Review; JFE – Journal of Financial Economics; JBV – Journal of Business Venturing.
n is the number of studies (i.e., the sample size of a meta-analysis).
k is the number of effect sizes, which may or may not equal n. When using correlation r, for instance, if a study reports 2 r’s regarding interested
variables, the n will be smaller than k.
N is the number of cumulative samples size, which is the total amount of sample sizes in the primary studies included in a meta-analysis.