Cornell University ILR School

DigitalCommons@ILR
Articles and Chapters

ILR Collection

2004

The Relationship of Achievement Motivation to

Entrepreneurial Behavior: A Meta-Analysis
Christopher J. Collins
Cornell University, [email protected]

Paul J. Hanges
University of Maryland

Edwin A. Locke
University of Maryland

Follow this and additional works at: http://digitalcommons.ilr.cornell.edu/articles

Part of the Entrepreneurial and Small Business Operations Commons, Human Resources
Management Commons, and the Labor Relations Commons
Thank you for downloading an article from DigitalCommons@ILR.
Support this valuable resource today!
This Article is brought to you for free and open access by the ILR Collection at DigitalCommons@ILR. It has been accepted for inclusion in Articles
and Chapters by an authorized administrator of DigitalCommons@ILR. For more information, please contact [email protected].

The Relationship of Achievement Motivation to Entrepreneurial Behavior:

A Meta-Analysis
Abstract

Entrepreneurship is a major factor in the national economy; thus, it is important to understand the
motivational characteristics spurring people to become entrepreneurs and why some are more successful than
others. In this study, we conducted a meta-analysis of the relationship between achievement motivation and
variables associated with entrepreneurial behavior. We found that achievement motivation was significantly
correlated with both choice of an entrepreneurial career and entrepreneurial performance. Further, we found
that both projective and self-report measures of achievement motivation were valid. Finally, known group
studies yielded a higher validity coefficient than did individual difference studies.
Keywords

entrepreneurship, motivation, decision-making, performance

Disciplines

Entrepreneurial and Small Business Operations | Human Resources Management | Labor Relations
Comments

Suggested Citation
Collins, C. J., Hanges, P. J., & Locke, E. A. (2004). The relationship of achievement motivation to entrepreneurial
behavior: A meta-analysis [Electronic version]. Retrieved [insert date], from Cornell University, ILR School
site: http://digitalcommons.ilr.cornell.edu/articles/x
Required Publisher Statement
Taylor & Francis. Reprinted with permission. All rights reserved. Final version published as: Collins, C. J.,
Hanges, P. J., & Locke, E. A. (2004). The relationship of achievement motivation to entrepreneurial behavior:
A meta-analysis. Human Performance, 17(1), 95-117.

This article is available at DigitalCommons@ILR: http://digitalcommons.ilr.cornell.edu/articles/831

The Relationship of Achievement Motivation to

Entrepreneurial Behavior: A Meta-Analysis
Christopher J. Collins
Cornell University
Paul J. Hanges
University of Maryland
Edwin A. Locke
University of Maryland
Entrepreneurship is a major factor in the national economy; thus, it is important to
understand the motivational characteristics spurring people to become entrepreneurs and
why some are more successful than others. In this study, we conducted a meta-analysis of the
relationship between achievement motivation and variables associated with entrepreneurial
behavior. We found that achievement motivation was significantly correlated with both
choice of an entrepreneurial career and entrepreneurial performance. Further, we found that
both projective and self-report measures of achievement motivation were valid. Finally,
known group studies yielded a higher validity coefficient than did individual difference
studies.
Nearly all supporters of the capitalistic system would agree that the encouragement
of successful innovations and entrepreneurship is beneficial for the continued health and
long-term growth of a nations economy (Schumpeter, 1961). Further, small businesses
account for over 50% of private sector employment and U.S. gross domestic product (Small
Business Administration Office of Advocacy, 2003). Because entrepreneurship is a
dominant factor in the economy, researchers have examined a number of factors that may
explain entrepreneurial activity. Though a good deal of recent research has tended to focus
on the characteristics of the business and industry environment or the characteristics of
the entrepreneurial opportunity itself (Kaufmann & Dant, 1998), our understanding of
entrepreneurship will not be complete unless we understand the motivation of the
individuals involved (Venkataraman, 1997). Indeed, recent research suggests that
motivational traits are an important factor in entrepreneurial activity and success (Baum,
Locke, & Smith, 2000; Stewart & Roth, 2001).
Research on the motivational traits of entrepreneurs seems especially promising for
helping to identify those individuals that might be best suited for identifying and exploiting
entrepreneurial opportunities in the marketplace (Shane & Venkataraman, 2000). For
example, individual traits and motivations may be used by schools, career counselors,
investors, government agencies, and so on to identify individuals that may be suited to
undertake and succeed in entrepreneurial ventures. Despite the potential importance of

individual characteristics, there are still many unanswered questions regarding the role
that motivation and personal characteristics have on entrepreneurial activity (Shane &
Venkataraman, 2000). Therefore, in this article, we seek to better understand how one
important motivating factor, need for achievement, is related to both the choice of
entrepreneurship as a career and performance in entrepreneurial roles.
The concept of need for achievement (nAch) was formulated in the 1950s
(McClelland, Clark, Roby, & Atkinson 1958). McClelland and his colleagues argued that
high-nAch people are more likely than low-nAch people to engage in energetic and
innovative activities that require planning for the future and entail an individuals
responsibility for task outcomes. McClelland (1961) argued that high-nAch people should
also prefer tasks that involve skill and effort, provide clear performance feedback, and
were of moderate challenge or risk. He also argued that entrepreneurial positions have
more of these characteristics than other types of positions. Hollands (1985) vocation
choice model suggests that individuals will be most attracted to careers that offer the
environmental characteristics that match their personality. Holland (1985) also argued
that performance and career satisfaction are higher when there is a good fit between work
environment characteristics and personality. Therefore, as was suggested by McClelland
(1961), it seems likely that individuals high in nAch should be attracted to and perform
well in entrepreneurial jobs.
Although there is a considerable body of literature in the area of achievement,
motivation and entrepreneurship, there are important issues that have yet to be addressed.
For example, Johnson (1990) conducted a traditional review of this literature and
concluded that there is evidence of a relationship between achievement motivation and
entrepreneurial activity. However, a close examination of Johnsons review raises some
questions about the strength of this conclusion. First, because of the narrative nature of this
review, the magnitude of the relationship between achievement motivation and
entrepreneurial activity is unclear. In fact, many of the studies reviewed by Johnson
showed non-significant relationships between achievement motivation and
entrepreneurial activity. Johnsons review also indicated that there was considerable
variance in the results across studies, suggesting that there may be moderators of the
achievement-motivationentrepreneurial activity relationship. In contrast to traditional
reviews, meta-analysis has the potential to address some of these issues by quantitatively
combining the results of individual studies to estimate the magnitude of the relationship
(Hunter & Schmidt, 1990) and by allowing formal tests of potential moderators of these
relationships.
This study has three objectives. First, we assess the degree of support for
McClellands claim by conducting a meta-analytic review of the achievement motivation
and entrepreneurship research that is sensitive to level of analysis issues. Second, we
identify and test several potential moderators of the relation between achievement
motivation and entrepreneurial action. Third, we compare the validity of achievement
motivation measures as a function of measurement methodology (e.g., projective measure
vs. self-report questionnaire measures). We discuss the implications of our findings for
future research and for organizations and institutions interested entrepreneurship.

Background and Hypotheses

Much of the early work on achievement motivation was conducted at the macrolevel of analysis. McClelland (1961) hypothesized that countries with a higher mean level of
nAch would show more entrepreneurial activity and economic growth than those countries
with a lower mean level of nAch. He found a statistically significant relationship between a
countrys average level of nAch (as measured by childrens stories) and the subsequent
economic growth of that country. He also found evidence for the relationship between
nAch and entrepreneurial activity based on historical records of earlier societies. However,
McClellands claims have not gone unchallenged (e.g., Finison, 1976; Frey, 1984; Mazur &
Rosa, 1977). Both Finison (1976) and Mazur and Rosa (1977) failed to replicate
McClellands results when they corrected for statistical errors and used 1950 nAch data to
predict economic growth through 1971, and Frey (1984) found problems with McClellands
historical tests of nAch.
Because it is very difficult to isolate one causal factor at the societal level of analysis
when so many causal factors are involved (Siehl & Martin, 1990), it would seem better to
test McClellands thesis at a more micro-level. Indeed, there are a number of studies that
have been conducted that examine differences in achievement motivation between both
individuals and known groups. In individual differences studies, researchers have typically
measured the need for achievement of individual entrepreneurs and correlated that
measure with individual performance (e.g., firm growth, successfully gaining funding).
Regarding studies exploring differences in achievement motivation between groups,
researchers have typically utilized the known-group research approach discussed in the
construct validity literature. That is, researchers identified two or more preexisting groups
of individuals (e.g., entrepreneurs versus managers, scientists, and professionals) and
tested for mean differences on some dependent variable among these groups. For example,
in terms of career choice, researchers have examined mean differences in achievement
motivation to test McClellands hypotheses. Regarding job performance, researchers have
assessed achievement motivation differences between high and low performing groups of
entrepreneurs. In contrast to societal level studies, individual and known group studies
allow cleaner assessment of McClellands original hypotheses.
Whereas some of these micro-level studies have shown a significant relationship
between achievement motivation and entrepreneurial behavior, previous reviews of this
literature have uncovered three major problems. First, these reviews consistently point to
the low internal consistency and testretest reliability of the preferred measure of nAch,
the TAT (Thematic Apperception Test; i.e., both types of reliability are often less than .60;
Fineman, 1977). Second, although alternative achievement motivation measures (i.e., selfreport) have been developed and tend to have high levels of internal consistency and test
retest stability (Fineman, 1977; Spangler, 1992), these self-report measures agree neither
with projective achievement motivation measures (Fineman, 1977; Spangler, 1992) nor
with each other (Fineman, 1977). Spangler (1992) reported a meta-analytic average
correlation of 0.09 between the TAT and the self-report measures. Furthermore, he found
that the different achievement motivation measures correlated differently with other
variables and were differentially affected by moderators. Given the aforementioned two
points, we decided to include both TAT and questionnaire measures in our analysis. Third,
previous reviews have combined studies using different micro-(individual and group)

levels of analysis in their reviews. However, inappropriate conclusions can be drawn when
concepts developed at one level of analysis are tested at a different level of analysis (Ostroff
& Harrison, 1999). Thus, in our analysis we separated these two levels. 1
Since Spangler (1992) conducted a previous meta-analysis of achievement
motivation studies, some key differences between his study and ours need to be
mentioned. First, in Spanglers (1992) study, only 5 out of 115 or 4% of the studies he used
involved entrepreneurs or entrepreneurial activity. In addition, Spanglers studies included
a very heterogeneous array of dependent variables, including IQ test scores, attitudes,
sensation seeking, enrollment in an academic honors program, performance on a ring toss
game, GPA, and industrial output. This meta-analysis was restricted to only studies
involving entrepreneurs and entrepreneurial performance. When testing the predictive
validity of achievement motivation, at least as conceived by McClelland (1961), it seems
best to restrict validity analysis to the domain where it is asserted to be most directly
applicable. For example, McClelland (1965) explicitly argued that college performance was
not an appropriate domain for achievement motivation (1965). Finally, Spangler did not
test the homogeneity of effect size assumption (Hedges & Olkin, 1985). We included this
test, because failure to assess this assumption can lead to the aggregation of effect sizes
from multiple populations resulting in an average effect size estimate that does not
generalize to any particular population (Hedges & Olkin, 1985).
By sticking closely to the most appropriate theoretical domain, 2 using different
measures of achievement motivation, conducting a test of homogeneity of effects, and
controlling for differences between known group and individual studies, we believe that
this meta-analysis has the best chance of accurately identifying the relationship between
achievement motivation and performance.
Definitions of Entrepreneur

One major criticism of research in the area of entrepreneurship is that researchers

have not consistently defined and operationalized what they mean by entrepreneurs
(Kaufmann & Dant, 1998); thus, it is difficult to compare findings across studies. Mill
(1848) provided one of the earliest definitions of entrepreneurship and suggested that risk
bearingwas the major feature that separated entrepreneurs from managers. McClelland
(1961) also argued that risk is an essential element faced by entrepreneurs, and he argued
that the definition of entrepreneurship should center around the responsibilities of
initiating and being accountable for business decisions.
McClelland (1961) argued that entrepreneurs should include both the owners of small
businesses and other managers with decision-making responsibilities. Following these
arguments, some researchers have included both small business founders and managers
when examining the role of achievement motivation.
In contrast to Mill and McClelland, Schumpeter (1934) argued that both
entrepreneurs and managers face risk and instead differ in terms of their emphasis on
innovation. The key notion behind this definition of entrepreneurship is the idea of growth
and creation of a new business opportunity where one did not exist previously.
Researchers that follow this definition tend to identify entrepreneurs as those who have
initiated or guided the process of new venture creation (Shane & Venkataraman, 2000).
Typically entrepreneurs are operationalized as founders under this later definition.

Because entrepreneurship researchers have tended to study either a broader (both

founders and managers) or narrower (founders only) set of individuals, it is unclear if need
for achievement is important for all types of entrepreneurial roles. Baum et al. (2000)
argued that founders and managers of small businesses may face different levels of
ambiguity and, thus, require different competencies. For example, founders may need
higher opportunity skills whereas individuals that assume the top role of an existing small
company may need more leadership and managerial skills. Thus, we expect that there may
be differences in the relationship between need for achievement and entrepreneurial
activity depending on how researchers defined entrepreneurs in their study.
H1. The relationship between need for achievement and entrepreneurial
activity will be moderated by the definition of entrepreneurship.

Known Group Versus Individual Differences Studies

Ostroff and Harrison (1999) have recently argued that it is critical to consider the
levels of analysis of the original empirical studies when conducting meta-analyses. It is
inappropriate to pool studies from multiple levels of analysis because the meta-analytically
derived average correlation will not be representative of any population parameter
(Ostroff & Harrison, 1999). This is a particular problem in our review because, as
previously indicated, researchers have examined the effects of achievement motivation
differences between both individuals and known groups. Further, the complex nature of
achievement motivation makes it important to examine potential differences between
known group and individual studies. Achievement motivation has been conceived of as a
complex set of values associated with entrepreneurial activity (McClelland, 1961).
However, though values can affect action, they are too abstract by themselves to say what a
person would do in a specific situation. For example, Eisenstadt (1963) argued that the
social and organizational setting in which people exist will affect the extent to which people
exhibiting high levels of achievement motivation will act in an entrepreneurial manner.
Consistent with this perspective, Holland (1985) and Schneider (1987) suggested
that when individuals are free to select themselves into certain groups (e.g., jobs,
occupations, careers), their attraction and willingness to remain in certain jobs and careers
is a function of their personality, attitudes, and values. Further, the homogeneity among
people who remain in a particular job or career tend to increase as people with different
personality, attitudes, and values leave these jobs or careers (Schneider, 1987). Thus, effect
sizes for analyses conducted on known groups should be enhanced because the
homogenization that occurs as a result of individuals staying in their careers that fit their
personality and leaving careers that are a mismatch will increases the size of the difference
between groups over time (e.g., entrepreneurs vs. non-entrepreneurs). 3
H2. In relation to levels of analysis, achievement motivation will have a
stronger association with entrepreneurial activity in known group studies
than in individual difference studies. 4

Need for Achievement and Career Choice

Holland (1985), in his theory of vocational choice, argued that individuals are most
likely to be attracted to those career roles that have work characteristics that match their
personalities, needs, and values. McClelland (1961) argued that individuals high in
achievement motivation are more likely to pursue occupations that allow for more control
over outcomes, provide more direct and immediate feedback on performance, and offer
moderate levels of risk. McClelland (1961, 1965) further argued that entrepreneurial
environments are the most likely job environments to meet these characteristics. Because
entrepreneurial occupations provide more opportunities to take advantage of the
characteristics associated with high achievement motivation, individuals who are high in
need for achievement will be more likely to be attracted to these roles than other types of
professions (Holland, 1985). Therefore, we would expect to find a significant effect size in
studies examining differences in need for achievement between entrepreneurs and
individuals pursuing other types of professions.
H3. There will be a significant association between achievement motivation
and choice of an entrepreneurial occupation.

To show how entrepreneurs differ from other individuals on achievement

motivation scores, studies have compared entrepreneurs to a wide variety of other groups
ranging from managers (e.g., Cromie & Johns, 1983; Waddell, 1983) to the general
population (e.g., Hornaday & Aboud, 1971) to politicians (e.g., Vincente-Wiley, 1979).
Brockhaus (1982) argued that a potential reason for the discrepancies in findings between
studies is the diverse set of comparison groups used in entrepreneurial studies; therefore,
it is possible that the type of comparison sample used may be moderating the findings in
the achievement motivation entrepreneurial activity literature.
Based on McClellands (1961) theory of nAch, one would expect that entrepreneurs
should have more characteristics in common (i.e., work settings, types of challenges, etc.)
with managers than they do with the general population or other types of professions,
because the settings for both business managers and entrepreneurs tend to be changing at
a fairly rapid pace and are open to individual initiative. In fact, McClelland (1961) identified
some managers (e.g., top level) as entrepreneurs. Therefore, it is likely that individuals high
in achievement motivation might also be more attracted to management positions within a
business setting than to other types of positions that do not involve management in a
business setting (e.g., politics, secretaries, engineers). We, therefore, predict that
H3a. In relation to career choice, achievement motivation measures will
differentiate better between entrepreneurs and non-managerial roles than
between entrepreneurs and managers.

Need for Achievement and Performance

Holland (1985), in his theory of vocational choice, also suggested that the
interaction of work environment and personality may affect performance in a career.
Specifically, he argued that higher levels of fit between the personality and work

environment characteristics will result in higher performance in that role. McClelland

(1961, 1965) similarly argued that need for achievement will be related to successful
performance in an entrepreneurial role. Individuals who are high in achievement
motivation are more likely to engage in the instrumental activities that are necessary for
success in an entrepreneurial situation than are individuals low in achievement motivation
(McClelland, 1965). For example, entrepreneurs who are high in achievement motivation
are more likely to overcome obstacles, utilize resources for help, compete, and improve
their skills. Therefore, we would expect to find a significant effect size in studies examining
differences in achievement motivation for high performance entrepreneurs versus lowperformance entrepreneurs.
H4. There will be a significant association between achievement motivation
and entrepreneurial performance.

Career Choice Versus Performance Studies

Though achievement motivation may significantly predict both career choice and
performance, it is not necessarily the case that the effect sizes for the two distinct outcomes
would be the same. McClelland (1965, 1972) suggested that nAch would be more likely to
predict career choice than performance in an entrepreneurial position. This argument
seems logical on statistical grounds alone. If it is true that those high in achievement
motivation will be more likely to choose an entrepreneurial career, then there should be a
restriction in the range of achievement motivation for those in entrepreneurial jobs. Thus,
achievement motivation should have less power to predict which entrepreneurs would be
successful and which would not be successful. In a review of achievement motivation,
Johnson (1990) found that 20 of 23 studies showed a significant relationship between
achievement motivation and entry into entrepreneurial positions. However, none of the 23
studies found significant ties between achievement motivation and performance of
entrepreneurs. As discussed earlier, Johnson did not use any meta-analytic technique to
statistically test the combined results of these studies, and we are still not clear on whether
achievement motivation is related to both career choice and performance. Thus, we predict
that
H5. Achievement motivation will predict entrepreneurial choice better than
entrepreneurial performance.

Measures of Need for Achievement

As we noted earlier, the achievement motivation literature suggests that there are
distinct differences between the types of tests used to measure achievement motivation. In
his formulation of the concept of nAch, McClelland (1961, 1972) contended that the
underlying causes of nAch are rooted in the subconscious and were developed during early
childhood. Projective tests are the only method to accurately assess nAch, because it is a
subconscious trait (McClelland, 1972). According to McClelland, questionnaire methods are
inappropriate for measuring nAch, because they measure an individuals conscious value

for achievement (McClelland, 1972). Therefore, McClelland (1980) argued that the TAT will
be a better predictor of entrepreneurial behavior than questionnaire measures.
Though it might be expected that the TAT would perform worse because projective
tests tend to have much lower reliability than questionnaire measures, there is some
support for McClellands hypothesis. For example, Spangler (1992) found that the TAT was
a better predictor of outcome criterion than were questionnaires. However, as discussed
earlier, Spangler did not address the issue of achievement motivation and
entrepreneurship. Thus, it is still a question whether the TAT predicts performance or
distinguishes between entrepreneurs and other types of individuals (i.e., predicts career
choice) better than other types of measures.
This is an especially important question to answer when considering questions
regarding the validity and reliability of the TAT that we raised earlier. For example,
researchers and practitioners might consider using questionnaire tests that have
demonstrated higher reliability if such tests are valid predictors and the TAT is not.
Similarly, researchers might consider other projective tests (besides the TAT) if those tests
have shown equal validity and/or higher reliability. The Miner Sentence Completion Scale
is one projective test that has been used to predict entrepreneurial performance and has
shown higher reliability than the TAT (Smith & Miner, 1984). To see if the TAT is a better
predictor than the other types of measures of achievement motivation, we explored the
difference in predictive power of the different measures of achievement motivation (i.e.,
TAT, questionnaire measures, and the Miner Sentence Completion Scale) in our metaanalysis. Based on the arguments of McClelland and prior reviews of this literature, we
predict that

Method

H6. The strength of the relationship between the TAT measure of nAch and
(a) occupational choice and (b) entrepreneurial performance will be
significantly higher than for other projective measures or questionnaire
measures of achievement motivation.

Search for Primary Data

An extensive literature search was conducted for individual level studies that
involved achievement motivation and entrepreneurship. The first step in locating potential
studies was to review original research studies cited in a previous meta-analysis (Spangler,
1992) and in previous review articles (Johnson, 1990; McClelland, 1980). Next, we
conducted a computer search using Psychlit, ProQuest, and Dissertation Abstracts
International databases. Inquiries were also made of prominent researchers in the
achievement motivation area to uncover any additional studies not located through the
other methods. We wrote David McClelland and the McBerCorporation to see if we had
missed any studies, but they could not provide us with any additional published or
unpublished studies. Finally, we used the Social Sciences Citation Index to identify articles
that cited key articles on achievement motivation.
Our original search resulted in a potential set of 47 studies. Twenty-one of these
studies used the TAT, six used the Miner Sentence Completion Scale, and 20 used various

types of questionnaire-based measures. We kept only those studies that conformed to the
following three criteria. First, the study had to involve findings linking entrepreneurs and
achievement motivation. Second, the study needed to provide either an effect size or
sufficient statistical information to calculate an effect size. Finally, the study needed to
provide sufficient background information and description of the variables of interest so
that the study could be coded according to one or more of the potential moderators
identified earlier.
Based on these criteria a total of six studies were dropped from further analysis.
Four of these studies were dropped because the authors failed to provide standard
deviations with reported mean scores (i.e., failed to meet the second criterion). Two other
studies were dropped when we realized that the author had reported the same findings in
multiple journals. For example, Singh reported the same findings in two studies (Ray &
Singh, 1980; Singh, 1979) so we retained only one of the studies in our sample. All six of the
dropped studies were from the TAT set of studies. The final set used for analysis comprised
41 studies.
Coding of Studies

Coding of studies was based on the potential moderators discussed earlier. Because
the studies were all field studies and did not involve any direct manipulations, the
moderators clearly fell into these very broad categories (see Table 1 for examples of
coding). The first step in coding the studies was to develop a coding manual based on
several potential mediating variables (discussed later). Next, the first author and three
research assistants independently coded the research articles. The first author and two of
the three research assistants agreed on the coding for all of the studies included in the
meta-analysis. The first author and the third research assistant disagreed on coding for two
studies, but these discrepancies were resolved after further discussion.
A sample of coded studies can be found in Table 1. First, we coded studies based on
grouping of participants (individual or known group). Studies were coded as individual
level studies when researchers examined mean differences based on individuals responses
and coded as known group studies when researchers examined mean differences on some
dependent variable among these groups. Second, we coded studies based on dependent
variable (performance or career pattern). Studies were coded as performance when the
dependent variable included some form of entrepreneurial performance and as career
choice when the studied compared mean differences in achievement motivation scores
between entrepreneurs and some other career (e.g., scientist, manager). Note that
performance measures varied across studies from secondary source measures (e.g., sales
growth, stock growth) to scaled measures of performance. Type of performance measure
was not a statistically significant moderator; therefore, all studies with performance as a
dependent variable were included together in tests of hypotheses.
Third, we coded studies based on type of career pattern comparison (entrepreneur vs.
manager, non-founder or entrepreneur vs. all other types of professions). Fourth, we coded
studies based on the type of test used (TAT, MSCSt [Miner Sentence Completion ScaleForm t], or questionnaire). Finally, we coded studies based on the definition of
entrepreneurship. We coded studies as following a broad definition if they included both

founders and non-founding managers and coded them as following a narrow definition
when they included only founders.
Several studies provided tests on multiple independent groups, used more than one
type of achievement motivation test, or looked at more than one outcome. These
independent sets of data were treated as individual data points if they could not be
meaningfully collapsed. For example, Tracy, Locke, and Renard (1997) used both the TAT
and the MSCSt to measure achievement motivation and provided effect sizes for both
tests. We felt that it was legitimate to treat these data points individually to test hypotheses
about potential moderators. The consequence of including the multiple coefficients from a
single study is that we decrease our ability to find moderators (i.e., it is a conservative bias
for the test of moderators). However, including multiple correlations from the same single
study may result in a liberal bias in the significance test for the general achievement
motivationactivity relationship. 5 To guard against the liberal bias, we replicated our
results for the overall achievement motivationactivity relationship using only one average
correlation per study. The results of this analysis did not significantly differ from our
analyses with multiple correlations per study in terms of effect sizes or conclusions
regarding hypotheses. Therefore, we have only reported the findings from the larger
sample in which we included multiple coefficients from the single studies.
When accounting for studies with multiple predictors or multiple criteria, we had 69
individual data points (observed relationships between one achievement motivation
measure and one outcome variable) and a total sample size of 9,032 participants. The TAT
data included 24 data points and a total sample size of 2,711 participants. The MSCSt data
were made up of 15 data points and a total sample size of 1,166. For the questionnaire data,
there was a total of 30 data points and a total sample size of 5,155 participants.
Transformation of Raw Data

The studies in our sample reported results with a wide range of statistics, including
F tests, t tests, means and standard deviations, and chi-square statistics. The raw statistics
were all converted to correlations using the formulas provided by Hunter and Schmidt
(1990) and Wolf (1986). Several studies reported their findings in multiple ways. When
this occurred, both sets of statistics were converted and compared as a means of testing the
accuracy of our conversions. In all cases, the correlations generated were statistically the
same.
Analyses

We used the random effects meta-analysis method developed by Hedges and Olkin
(1985) to test our hypotheses. This method treats studies as heterogeneous instead of
assuming generalizability from the start as fixed effects models of meta-analysis assume.
The Hedges and Olkin method first tests the overall sample for homogeneity of variance. If
the homogeneity of variance hypothesis is rejected, then the sample is believed to contain
true between study differences, and a weighted least-squares regression is used to test
potential moderators that are identified through theory. Differences between studies that
significantly explain between study variance in a regression equation are significant
moderators of the relationship in question.

Results
The correlations for all samples and measures are shown in Table 2. The test for
homogeneity of variance (Hedges & Olkin, 1985) indicated that there was significant
between study variability (Q = 22.08, p < .001). Thus, the results indicated that moderators
are operating in the relationship between achievement motivation and overall
entrepreneurial activity.
H1 predicted that the relationship between achievement motivation and
entrepreneurial behavior will be moderated by the definition of entrepreneurship. As
shown in Table 2, the mean correlation for studies defining entrepreneurs as founders was
.21 and the mean correlation for studies defining entrepreneurs as business leaders was
.19. Further, as shown in Table 3, the weighted least squares analysis did not show support
for our hypothesis, F(1, 58)=.41, ns. Therefore, it does not appear that the relationship
between achievement motivation and entrepreneurial activity varies depending on how
researchers have defined entrepreneurs in their study.
H2 predicted that achievement motivation will have a have a stronger association
with entrepreneurial activity in known group studies than in individual studies. When all
need for achievement studies were included (see Table 2), the respective mean
correlations were .18 in individual studies and .25 in known group studies. As shown in
Table 3, the weighted least squares analysis showed support for this hypothesis, F(1, 67) =
3.42, p < .05. When we examined this relationship with only the performance studies, we
again found a large difference in effect sizes across levels of analysis. The mean correlation
between achievement motivation and performance was .46 in known group studies and .18
in individual studies. As shown in Table 3, these mean correlations were significantly
different, F(1, 26) = 24.60, p < .01. Thus, H2 is supported; it appears that the relationship
between achievement motivation and entrepreneurial activity is stronger in known group
studies than in individual studies.
H3 predicted that achievement motivation would be significantly related to choice
of an entrepreneurial occupation. Note that career choice has only been studied using the
known group design. The analyses showed significant correlations between achievement
motivation and career choice (r = .21, p < .001). The lower bound of the 95% confidence
interval was .18 and the upper bound was .24. Because the confidence interval did not
contain zero, H3 was supported. It appears that individuals that pursue entrepreneurial
careers are significantly higher in achievement motivation than individuals that pursue
other types of careers.
H3a predicted that the relationship between achievement motivation and career
choice will be moderated by the comparison sample that is used. Specifically, we predicted
that achievement motivation will have a weaker relationship with career choice when the
sample compared entrepreneurs and managers than when the sample compared
entrepreneurs and non-managers. The correlation between achievement motivation and
career choice was lower when the sample included only managers (mean r = .14) than
when it included all other professions (mean r = .35). The regression equation showed a
significant difference between the two types of studies, F(1, 39) = 14.78, p < .01. Therefore,

it appears that the sample of participants involved in the study moderates the achievement
motivationcareer choice relationship.
H4 predicted that achievement motivation would be significantly related to
performance. As noted before, the analyses for known group studies showed a significant
mean correlation between achievement motivation and performance (r = .46, p < .01) with
a lower bound of .42 and an upper bound of .49. The analyses of individual studies also
showed a significant correlation between achievement motivation and performance (r =
.18, p < .01) with a lower bound of .15 and an upper bound of .22.Becauseneither of these
confidence intervals included zero,H4is supported. 6

H5 predicted that achievement motivation would predict entrepreneurial choice

better than entrepreneurial performance. The results with known group studies showed
that the respective mean correlations were .21 (choice) and .46 (performance). Further, the
regression results showed a significant difference between the relationships of
achievement motivation and the two outcome variables, F(1, 47) = 15.58, p < .01.
Therefore, there does appear to be a significant difference in the strengths of the
relationships between need for achievement and the two different types of dependent
variable; however, the difference was in the direction opposite of what we predicted. In
known group studies, it appears that achievement motivation is better able to explain
performance differences than career choice differences. This hypothesis could not be tested
for individual difference studies, because, as indicated earlier, we found no career choice
studies conducted with this method.
H6 predicted that the type of achievement motivation measure would moderate the
correlation of achievement motivation with the outcome criterion. Specifically, we
predicted that the strength of the relationship between the TAT measure of achievement
motivation and entrepreneurial activity would be significantly higher than it is for either
other projective measures or questionnaire measures of achievement motivation. The
analyses showed mean correlations of .16 (TAT), .19 (questionnaire), and .20 (MSCSt) for
the performance studies at the individual level of analysis and mean correlations of .20
(TAT), .20 (questionnaire), and .23 (MSCSt) for career choice studies at the group level of
analysis. As shown in Table 3, there were no significant differences in the mean
correlations between the measurement typesperformance studies, F(2, 17)=.47, ns;
career choice studies, F(2, 38)=.58, ns. Thus, it appears as though there is no difference in
the relationship between achievement motivation and entrepreneurial activity across the
different measures of achievement motivation. 7

Discussion
Overall, our results support McClellands theory that achievement motivation is
significantly related to both occupational choice and performance in an entrepreneurial
role. Although these results superficially agree with prior reviews (e.g., Spangler, 1992) in
showing that need for achievement is related to action, our results were based on
entrepreneurship studiesthe domain that is directly relevant to achievement motivation.
Further, we found a stronger relationship between need for achievement and
entrepreneurial activity in known group studies (mean r = .25) than in individual level
studies (mean r = .19). Consistent with Holland and Schneider, achievement motivation
appears to be an important characteristic for entrepreneurs; thus, the correlations from
known group studies were stronger because of the increasing similarity among these
entrepreneurs over time.
We were surprised that we failed to find support for McClellands hypothesis that
achievement motivation strength would predict career choice better than it predicted
performance. For the known group studies, we found the exact opposite to hold. The mean
r for performance was significantly higher than the mean r for career choice (.46 vs. .21).
Although both correlations were significant, this finding suggests that achievement
motivation may be particularly potent at differentiating between successful and
unsuccessful groups of entrepreneurs. Thus, achievement motivation may be particularly
useful for selecting entrepreneurs that may be more likely to successfully take advantage of
entrepreneurial financing and other supporting activities. However, we note that the mean
r for career choice may be lower than that for performance because individuals high in
achievement motivation may be attracted to several different types of careers. For
example, our results suggested that managers are much more similar to entrepreneurs in
terms of their achievement motivation than are individuals attracted to other, less similar
careers such as engineers or scientists. Thus, achievement motivation may be a useful
predictor of career choice for a number of types of positions that have job characteristics
and responsibilities similar to entrepreneurial positions.
Also contrary to McClellands arguments, we found that the TAT showed no greater
validity than other projective tests or than self-report questionnaire measures in predicting
performance or career choice. When we examined the results for performance at the
individual level, the mean correlation for the TAT studies is .16compared to .20 for MSCSt
studies and .20 for the questionnaire studies. For predicting career choice at the group
level, the mean correlations are again very similar (.21 for the TAT, .23 for the MSCSt, and
.22 for the questionnaire). Our failure to find a higher mean validity r for the TAT
contradicts the findings of Spanglers (1992) meta-analysis and those of more traditional
reviews (e.g., Johnson, 1990). In a review of the raw data, it appears that recent studies
using the TAT have reported much lower correlations with entrepreneurial activity than
did the earlier studies conducted by McClelland. Thus, it is possible that the mean r for TAT
studies may have been affected by undertraining on the intricacies of evaluating the TAT or
other errors that may have led to incorrect scoring by later researchers who did not have
the same level of experience with the test as McClelland. We were unable to test this post
hoc speculation because we could not find data to test for reliability differences between
McClelland and other researchers. However, if our speculation is true, it suggests that

practitioners should rely on questionnaire measures of achievement motivation that are

easier to administer and accurately score.
A second practical implication based on our findings is that practitioners may
consider using both self-report and projective measures to assess achievement motivation,
because previous studies have shown consistently that TAT, MSCSt, and questionnaire
measures of achievement motivation are valid but not significantly correlated or correlated
at a very low level (e.g., Fineman, 1977). These low correlations suggest that the different
measures of achievement motivation must be measuring different aspects of achievement
motivation. As noted, McClelland has argued that TAT measures get at subconscious
motives whereas questionnaires get at conscious motives. The question arises, then,
whether using multiple types of measures will yield greater total validity than using only
one alone. Future studies should test this hypothesis.
Our results were consistent with McClellands prediction that individuals high in
achievement motivation are more likely to be attracted to occupations that offer high
degrees of control over outcomes, personal responsibility, feedback on performance, and
moderate degree of risk. As predicted, we found a significant overall effect size in studies
comparing achievement motivation of individuals that pursued entrepreneurial versus
other types of careers. However, just as Brockhaus (1982) argued, we found greater
differences in achievement motivation between entrepreneurs and people in all other
professions than between entrepreneurs and managers. As we indicated previously, this
result is not surprising because managers jobs often have some entrepreneurial elements,
and entrepreneurial jobs typically entail management activities. Organizations in highpaced environments that are looking for people to take on entrepreneurial roles (i.e., those
that involve personal responsibility, feedback on performance, and a moderate degree of
risk) to help drive innovations within the firm may wish to use achievement motivation as
one of several selection criteria.
Although one might question the utility of a concept that accounts for so little
variance (R2 at the group level = .06, R2 at the individual level = .04), it is important to
remember that the influence of achievement motivation on behavior is not a one-time
influence. Rather, there are repeated opportunities for the achievement motivation of an
individual to affect their decisions and behavior over the span of a lifetime. Recent
simulations by Martell, Lane, and Emrich (1996) have demonstrated that interventions that
appear to have small effects (e.g., effect size of 1% explained variance) can have dramatic
consequences when individuals repeatedly encounter the intervention over time. Thus,
even though our meta-analysis found that only 4% to 6% of the variance in entrepreneurial
behavior was accounted for by achievement motivation, the implication is that
achievement motivation can have substantial consequences because there are multiple
opportunities for an individual to exhibit achievement oriented behavior over his or her
lifetime.
Despite the merits of meta-analysis as a means of summarizing existing findings, we
feel that it is important to note several limitations of our study based on the nature of the
studies that composed our sample. First, there is a bias within the field of management to
publish papers with significant rather than null findings. If this bias exists in the literature
on achievement motivation and entrepreneurship, then our findings may potentially
overstate the true relationships between need for achievement and entrepreneurial
activities. We feel that this issue may be somewhat mitigated based on the nature of the

studies in our sample. First, we included a number of unpublished doctoral dissertations

that should not have been affected by this bias. Second, a large portion of our studies
examined a number of predictor variables besides achievement motivation; thus, these
studies were more likely to be published even if need for achievement was a nonsignificant predictor.
A second limitation is that the original studies were unlikely to have data on
ventures that failed; thus, the sample of entrepreneurs was constrained to those that
managed or owned surviving businesses. If need for achievement does predict
entrepreneurial success, then this limitation potentially constrains the range of variability
on need for achievement. In other words, this limitation in the original studies should limit
our ability to find a relationship between need for achievement and entrepreneurial
activity; our study may be a conservative test of this relationship. A final limitation is that
the original studies were correlational rather than predictive in nature, limiting our ability
to assess the direction of causation. It is possible that entrepreneurs may be socialized into
becoming high nAch because of the goals, deadlines, and accomplishments that are part of
the role of entrepreneurs. Indeed, the few predictive studies in the literature on need for
achievement and entrepreneurship provide mixed findings. Early studies by McClelland
(1965) and Durand and Shea (1974) found that scores on achievement motivation
significantly predicted entrepreneurial activity years later. A later study by McClelland
(1972) did not confirm this relationship. Therefore, additional research is needed to
explore the direction of the relationship between the achievement motivation and
entrepreneurial behavior.
In summary, we found that achievement motivation does significantly predict
entrepreneurial activity (both choice of an entrepreneurial occupation and performance in
that role) across the studies that were included in this meta-analysis. Further, we argued
that the relatively small variance in entrepreneurial activity that was explained by
achievement motivation might have significant implications because of the multiple
opportunities for an individual to exhibit achievement- oriented behavior. Further, our
findings suggest that achievement motivation may be particularly potent in predicting
outcomes at particular levels of analysis and in specific situations. Of course, it is simplistic
to think that personality alone is the key to entrepreneurship. Many other factors are
involved in the success of new business venture and venture growth. Baum et al. (2000),
for example, found that five different domains of variables predicted venture growth:
personality traits, situationally specific motivation (goals, self-efficacy, vision), skills,
business strategies, and environmental factors. Furthermore, some of these domains had
direct effects, others had indirect effects, and others had both. Thus, trait factors should be
viewed as only one causal factor in a complex model of factors that promote venture
success. Additional research is needed to determine whether achievement motivation
would still retain a significant role in such a multivariate model.
References

(References marked with an asterisk indicate studies included in the meta-analysis.)

*Ahmed, S. U. (1985). nAch, risk-taking propensity, locus of control and entrepreneurship.

Personality and Individual Differences, 6, 781782.

*Babb, E. M., & Babb, S. V. (1992). Psychological traits of rural entrepreneurs. The Journal of
Socio-Economics, 21, 353362.
Baum, R., Locke, E. A., & Smith, K. G. (2000). A longitudinal model of a multi-dimensional
model of venture growth. Academy of Management Journal, 44, 292303.
*Begley, T. M. (1995). Using founder status, age of firm, and company growth rate as the
basis for distinguishing entrepreneurs form managers of smaller businesses. Journal of
Business Venturing, 10, 249263.

*Begley, T. M., & Boyd, D. P. (1987). A comparison of entrepreneurs and managers of small
business firms. Journal of Management, 13, 99108.

Brockhaus, R. H. (1982). The psychology of the entrepreneur. In C. A. Kent, D. L. Sexton, & K.

H. Vesper (Eds.), Encyclopedia of entrepreneurship (pp. 3957). Englewood Cliffs, NJ:
Prentice-Hall.

*Carland, J. W. (1982). Entrepreneurship in a small business setting: An exploratory study.

Unpublished doctoral dissertation, University of Georgia.

*Cromie, S., & Johns, S. (1983). Irish entrepreneurs: Some personal characteristics. Journal
of Occupational Behaviour, 4, 317324.
*DeCarlo, J. F., & Lyons, P. R. (1979). A comparison of selected personal characteristics of
minority and non-minority female entrepreneurs. Journal of Small Business Management,
12, 2229.

*Dorer, L. (2001). Motivational differences between entrepreneurs and employees: A

comparison of three models. Unpublished doctoral dissertation, Virginia-CommonwealthUniversity.
*Durand, D., & Shea, D. (1974). Entrepreneurial activity as a function of achievement
motivation and reinforcement control. The Journal of Psychology, 88, 5763.

Eisenstadt, S. N. (1963). The need for achievement. Economic Development and Cultural
Change, 11, 420431.

Fineman, S. (1977). The achievement motive construct and its measurement: Where are we
now? The British Journal of Psychology, 68, 122.
Finison, L. J. (1976). The application of McClellands national development model to recent
data. The Journal of Social Psychology, 98, 5559.
Frey, R. S. (1984). Need for achievement, entrepreneurship, and economic growth: A
critique of the McClelland thesis. The Social Science Journal, 21, 125134.

*Hansemark, O. C. (1999). Predictive validity of TAT and CMPS on the entrepreneurial

activity, start of a new study: A longitudinal study. Journal of Managerial Psychology, 15,
634654.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Academic.

*Hines, G. H. (1973). Achievement motivation, occupations, and labor turnover in New

Zealand. Journal of Applied Psychology, 58, 313317.
Holland, J. L. (1985). Making vocational choices. Englewood Cliffs, NJ: Prentice Hall.

*Hornaday, J. A.,& Aboud, J. (1971). Characteristics of successful entrepreneurs. Personnel

Psychology, 24, 141153.

*Hull, D. L., Bosley, J. J., & Udell, G. G. (1980). Renewing the hunt for heffalump: Identifying
potential entrepreneurs by personality characteristics. Journal of Small Business, 18, 1118.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis. Newbury Park, CA: Sage.

Johnson, B. R. (1990). Toward a multidimensional model of entrepreneurship: the case of

achievement motivation and the entrepreneur. Entrepreneurship Theory and Practice, 14,
3954.
Judd, C. M., Smith, E. R., & Kidder, L. H. (1991). Research methods in Social Relations (6th
ed.). Orlando, FL: Harcourt Brace.
Kaufmann, P. J., & Dant, R. P. (1998). Franchising and the domain of entrepreneurship
research. Journal of Business Venturing, 14, 516.

*Lee, D. Y., & Tsang, E.W. K. (2001). The effects of entrepreneurial personality, background
and network activities on venture growth. Journal of Management Studies, 38, 584602.

*Lee, J. (1997). The motivation of women entrepreneurs in Singapore. International Journal

of Entrepreneurial Behavior and Research, 3, 93110.
*Lynn, R. (1969). An achievement motivation questionnaire. British Journal of Psychology,
60, 529534.
Martell, R. F., Lane, D. M., & Emrich, C. (1996). Male-female differences: A computer
simulation. American Psychologist, 51, 157159.

Mazur, A., & Rosa, E. (1977). An empirical test of McClellands achieving society theory.
Social Forces, 55, 769774.
*McClelland, D. C. (1961). The achieving society. Princeton, NJ: Van Nostrand.

*McClelland, D. C. (1965). Toward a theory of motive acquisition. American Psychologist, 20,

321333.
McClelland, D. C. (1972). Opinions predict opinions: So what else is new? Journal of
Consulting and Clinical Psychology, 38, 325326.

McClelland, D. C. (1980). Motive dispositions: The merits to operant and respondent

measures. In L. Wheeler (Ed.), Review of personality and social psychology (Vol. 1, pp. 10
41). Beverly Hills, CA: Sage.

McClelland, D. C., Clark, R. A., Roby, T. B., & Atkinson, J. W. (1958). The effect of the need for
achievement on thematic apperception. In J.W. Atkinson (Ed.), Motives in fantasy, action and
society (pp. 6482). Princeton, NJ: Van Norstrand.
Mill, J. S. (1848). Principles of political economy with some of their applications to social
philosophy. Boston: C. C. Little and Brown.

*Miner, J. B., Smith, N. R., & Bracker, J. S. (1989). Role of entrepreneurial task motivation in
the growth of technologically innovative firms. Journal of Applied Psychology, 74, 554560.

*Miner, J. B., Smith, N. R., & Bracker, J. S. (1994). Role of entrepreneurial task motivation in
the growth of technologically innovative firms: Interpretations from follow-up data. Journal
of Applied Psychology, 79, 627630.

*Morris, J. L., & Fargher, K. (1974). Achievement drive and creativity as correlates of
success in small business. Australian Journal of Psychology, 26, 217222.

*Nandy, A. (1973). Motives, modernity, and entrepreneurial competence. The Journal of

Social Psychology, 91, 127136.

Ostroff, C., & Harrison, D. A. (1999). Meta-analysis, levels of analysis, and best estimates of
population correlations: Cautions for interpreting meta-analytic results in organizational
behavior. Journal of Applied Psychology, 84, 260270.
*Ray, J. J., & Singh, S. (1980). Effects of individual differences on productivity among
farmers in India. The Journal of Social Psychology, 112, 1117.

Robbins, N. E. (1986). Entrepreneurial assessment: Characteristics which differentiate

entrepreneurs, intrapreneurs and managers. Unpublished doctoral dissertation, University
of Minnesota.
*Rodriguez, L. T. (2000). The wind beneath their wings: The moderating effects of social
support on the entrepreneur and entrepreneurial performance. Unpublished doctoral
dissertation, University of Nebraska.

Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological
Bulletin, 86, 638641.
Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437454.

Schumpeter, J. A. (1934). The theory of economic development: An inquiry into profits,

capital, credit, interest, and the business cycle. Cambridge, MA: Harvard University Press.
Schumpeter, J. A. (1961). The theory of economic development. Cambridge, MA: Harvard
University Press.

Shane, S.,&Venkataraman, S. (2000). The promise of entrepreneurship as a field of research.

Academy of Management Review, 25, 217226.

Siehl, C., & Martin, J. (1990). Organizational culture: A key to financial performance? In B.
Schneider (Ed.), Organizational climate and culture (pp. 241281). San Francisco: JosseyBass.
*Singh, S. (1969). nAch among successful-unsuccessful and traditional progressive
agricultural entrepreneurs of Delhi. Journal of Social Psychology, 79, 145149.

*Singh, S. (1970). nAch among agricultural and business entrepreneurs of Delhi. The
Journal of Social Psychology, 81, 145149.
*Singh, S. (1979). Relationships among projective and direct verbal measures of
achievement motivation. Journal of Personality Assessment, 43, 4549.

*Singh, S., & Ray, J. J. (1980). Modernization and development among Indian farmers: A
modern proof of some old theories. Economic Development and Cultural Change, 80, 509
521.

Small Business Administration Office of Advocacy. (2003). Small business share of

economic growth. Office of economic research: Research publications 2002. Retrieved April
5, 2003 from http://www.sba.gov.advo
*Smith, N. R., & Miner, J. B. (1984). Motivational considerations in the success of
technologically innovative entrepreneurships. In J. A. Hornaday, E. B. Shils, J. A. Timmons, &
K. H. Vesper (Eds.), Frontiers of entrepreneurship research (pp. 482488). Wellesley, MA:
Babson College.
*Smith, N. R., & Miner, J. B. (1985). Motivational considerations in the success of
technologically innovative entrepreneurships: Extended sample findings. In J. A. Hornaday,
E. B. Shils, J. A. Timmons, & K. H. Vesper (Eds.), Frontiers of entrepreneurship research (pp.
482488). Wellesley, MA: Babson College.

Spangler, W. D. (1992). Validity of questionnaire and TAT measures of need for

achievement: Two meta-analyses. Psychological Bulletin, 112, 140154.

*Stewart, W. H., Watson, W. E., Carland, J. C., & Carland, J.W. (1995). A proclivity for
entrepreneurship: A comparison of entrepreneurs, small business owners, and corporate
managers. Journal of Business Venturing, 14, 189214.

Stewart, W. H., & Roth, P. L. (2001). Risk propensity differences between entrepreneurs and
managers: A meta-analysis review. Journal of Applied Psychology, 86, 145153.

*Sutcliffe, C. R. (1974). Achievement motivation and economic development among

peasants: An exploration of measurement problems. Rural Sociology, 39, 238246.
*Tracy, K. B., Locke, E. A., & Renard, M. K. (1997). Conscious goal setting versus
subconscious motives: Longitudinal and concurrent effects on the performance of
entrepreneurial firms. Unpublished paper, University of Maryland at College Park.

Venkataraman, S. (1997). The distinctive domain of entrepreneurship research: An editors

perspective. In J. Katz & R. Brockhaus (Eds.), Advances in Entrepreneurship, firm emergence,
and growth (Vol. pp. 119138). Greenwich, CT: JAI.
*Vicente-Wiley, L. (1979). Achievement values of Filipino entrepreneurs and politicians.
Economic Development and Cultural Change, 79, 467483.

*Waddell, F. T. (1983). Factors affecting choice, satisfaction, and success in the female selfemployed. Journal of Vocational Behavior, 23, 294304.

*Wainer, H. A., & Rubin, I. M. (1969). Motivation of research and development

entrepreneurs: Determinants of company success. Journal of Applied Psychology, 53, 178
184. Wolf, F. M. (1986). Meta-analysis: Quantitative methods for research synthesis.
Newbury Park, CA: Sage.

For a further discussion of the problems of ecological fallacies, see Judd, Smith, and Kidder (1991).
McClelland does not consider nAch to be a predictor of performance in a traditional, bureaucratic firm; rather he uses
his leadership motive pattern (which involves a certain relationship between nPower, power inhibition, and nAffiliation).
3 It should be noted that our expected enhanced effect of nAch and performance at the group level of analysis is different
from McClellands societial level hypotheses because our hypothesis depends on people freely selecting themselves into
and out of groups. Clearly, society level of analyses include more people than just those who freely select themselves into
that society.
4 As suggested by Ostroff and Harrison (1999), we began by examining whether need for achievement studies can be
examined across levels of analysis. If this hypothesis is supported, then all of the following hypotheses will be tested at
both the group and individual levels of analysis where appropriate.
5 Though inclusion of multiple correlations from the same sample will not affect the average correlation estimate, it will
affect the variance in correlations across studies.
6 We performed Rosenthals (1979) file-drawer analysis and found that 86 studies would be needed to change our results.
Given the extensive nature of our search to find studies focusing on entrepreneurs, these results appear to be stable.
7 We corrected our average correlations for measurement error by applying the correction for attenuation formula. Using
Fineman (1977), we estimated the reliability of the TAT at .54, the reliability of questionnaire methods to be .62, and the
reliability of the MSCS to be .65. After averaging the meta-analysis correlations for each measure, the results did not
significantly differ.
1
2