perceptions of women and men as entrepreneurs: evidence for differential effects of attributional...

7
Journal of Applied Psychology 2001, Vol. 86, No. 5, 923-929 Copyright 2001 by the American Psychological Association, Inc. 0021-9010/01/$5.00 DO1: 10.1037//0021-9010.86.5.923 Perceptions of Women and Men as Entrepreneurs: Evidence for Differential Effects of Attributional Augmenting Robert A. Baron, Gideon D. Markman, and Azita Hirsa Rensselaer Polytechnic Institute It was hypothesized that perceptions of women who become entrepreneurs are enhanced by attributional augmenting because they adopt this role despite major obstacles to doing so. In contrast, attributional augmenting was expected to operate to a lesser degree for men who become entrepreneurs because they presumably face weaker obstacles. Three studies offered support for these hypotheses; all of these investigations used between-subjects designs in which women and men shown in standard-format photos were described to different groups of raters as being either entrepreneurs or managers. As predicted, raters assigned significantly higher scores to women, but not to men, when they were described as entrepreneurs. Despite progress toward equality in recent years, women con- tinue to experience less favorable outcomes than men in many business settings. They are concentrated in low-paying jobs and receive lower average salaries (e.g., Fisher, 1992). Further, when they succeed on various tasks or in various roles, their performance is often attributed to external factors (e.g., luck) rather than to their ability or effort (e.g., Heilman, Block, & Lucas, 1992). These negative effects are especially apparent when women seek or hold managerial-level positions; in this context, they re- ceive lower ratings than men as job applicants (e.g., Heilman, Martell, & Simon, 1988) and lower performance appraisals once they hold such positions (Heilman & Stopeck, 1985b). In addition, women appear to face more barriers than men in developing informal mentoring relationships (Ragins & Cotton, 1991) and benefit less from such relationships, especially if their mentors are also women (Ragins & Cotton, 1999). Perhaps most disturbing, such negative effects seem to occur even when women are closely matched to men in terms of experience, specific job duties, edu- cation, and several other factors (Lyness & Thompson, 1997). Under these conditions, differences between women and men in terms of outcomes experienced are reduced, but they continue to exist. Although many factors contribute to these persistent and dis- turbing disadvantages for women, considerable evidence suggests that negative gender stereotypes—widely shared beliefs about the attributes of men and women—play an important role (e.g., Heil- man, 1983; Ruble, Cohen, & Ruble, 1984). Despite major changes in many societies in recent decades, gender stereotypes remain largely unchanged (e.g., Heilman, Block, Martell, & Simon, 1989) and are generally less favorable for women than for men. For instance, in recent research, Hosada and Stone (in press) found that Robert A. Baron, Gideon D. Markman, and Azita Hirsa, Lally School of Management and Technology, Rensselaer Polytechnic Institute. Correspondence concerning this article should be addressed to Robert A. Baron, Lally School of Management and Technology, Rensselaer Poly- technic Institute, Troy, New York 12181-3590. Electronic mail may be sent to [email protected]. women and men continue to be viewed as different with respect to many traits that have long been included in traditional gender stereotypes. For example, participants in Hosada and Stone's study—more than 1,000 college students—reported that men are more forceful, assertive, aggressive, confident, and independent, whereas women are more nurturant, emotional, considerate, inde- cisive, and submissive. Perhaps most damaging to women is the fact that traits included in gender stereotypes for women appear to be less consistent with those required for managerial success than traits included in gen- der stereotypes for men. As a result, women are often seen as less qualified for managerial-level jobs (e.g., Van Vianen & Willem- sen, 1992). In sum, persistent gender stereotypes appear to work against women in many business settings, especially ones involv- ing managerial-level positions. To date, however, existing litera- ture on this important topic has not considered the following question: Do such negative effects occur in another business context as well—that is, with respect to female entrepreneurs? This appears to be an important issue because mounting evi- dence suggests that the activities of entrepreneurs are crucial to the economic growth and prosperity of modern societies (Shane & Venkataraman, 2000). Indeed, the new ventures started by entre- preneurs more than made up for the loss of millions of jobs in large, existing companies during the 1980s and 1990s and thus played a key role in the current strength of the U.S. economy. In this context, the fact that women are underrepresented among entrepreneurs and face daunting obstacles to assuming this role are disturbing. Thus, efforts to identify the factors that influence their decision to become entrepreneurs and their ultimate success seem both timely and important. The present research focused on one of these factors: perceptions of women who adopt the role of entrepreneur. At first glance, it might seem that women entrepreneurs would be subject to the same negative stereotypes and processes that adversely affect women in other business contexts. Recent findings (Baron, 1999) have indicated that current stereotypes of entrepre- neurs are heavily weighted toward traits traditionally viewed as "masculine." Entrepreneurs are perceived as being more assertive, achievement oriented, and confident than managers and as having 923

Upload: azita

Post on 06-Aug-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

Journal of Applied Psychology2001, Vol. 86, No. 5, 923-929

Copyright 2001 by the American Psychological Association, Inc.0021-9010/01/$5.00 DO1: 10.1037//0021-9010.86.5.923

Perceptions of Women and Men as Entrepreneurs:Evidence for Differential Effects of Attributional Augmenting

Robert A. Baron, Gideon D. Markman, and Azita HirsaRensselaer Polytechnic Institute

It was hypothesized that perceptions of women who become entrepreneurs are enhanced by attributionalaugmenting because they adopt this role despite major obstacles to doing so. In contrast, attributionalaugmenting was expected to operate to a lesser degree for men who become entrepreneurs because theypresumably face weaker obstacles. Three studies offered support for these hypotheses; all of theseinvestigations used between-subjects designs in which women and men shown in standard-format photoswere described to different groups of raters as being either entrepreneurs or managers. As predicted,raters assigned significantly higher scores to women, but not to men, when they were described asentrepreneurs.

Despite progress toward equality in recent years, women con-tinue to experience less favorable outcomes than men in manybusiness settings. They are concentrated in low-paying jobs andreceive lower average salaries (e.g., Fisher, 1992). Further, whenthey succeed on various tasks or in various roles, their performanceis often attributed to external factors (e.g., luck) rather than to theirability or effort (e.g., Heilman, Block, & Lucas, 1992).

These negative effects are especially apparent when womenseek or hold managerial-level positions; in this context, they re-ceive lower ratings than men as job applicants (e.g., Heilman,Martell, & Simon, 1988) and lower performance appraisals oncethey hold such positions (Heilman & Stopeck, 1985b). In addition,women appear to face more barriers than men in developinginformal mentoring relationships (Ragins & Cotton, 1991) andbenefit less from such relationships, especially if their mentors arealso women (Ragins & Cotton, 1999). Perhaps most disturbing,such negative effects seem to occur even when women are closelymatched to men in terms of experience, specific job duties, edu-cation, and several other factors (Lyness & Thompson, 1997).Under these conditions, differences between women and men interms of outcomes experienced are reduced, but they continue toexist.

Although many factors contribute to these persistent and dis-turbing disadvantages for women, considerable evidence suggeststhat negative gender stereotypes—widely shared beliefs about theattributes of men and women—play an important role (e.g., Heil-man, 1983; Ruble, Cohen, & Ruble, 1984). Despite major changesin many societies in recent decades, gender stereotypes remainlargely unchanged (e.g., Heilman, Block, Martell, & Simon, 1989)and are generally less favorable for women than for men. Forinstance, in recent research, Hosada and Stone (in press) found that

Robert A. Baron, Gideon D. Markman, and Azita Hirsa, Lally School ofManagement and Technology, Rensselaer Polytechnic Institute.

Correspondence concerning this article should be addressed to RobertA. Baron, Lally School of Management and Technology, Rensselaer Poly-technic Institute, Troy, New York 12181-3590. Electronic mail may be sentto [email protected].

women and men continue to be viewed as different with respect tomany traits that have long been included in traditional genderstereotypes. For example, participants in Hosada and Stone'sstudy—more than 1,000 college students—reported that men aremore forceful, assertive, aggressive, confident, and independent,whereas women are more nurturant, emotional, considerate, inde-cisive, and submissive.

Perhaps most damaging to women is the fact that traits includedin gender stereotypes for women appear to be less consistent withthose required for managerial success than traits included in gen-der stereotypes for men. As a result, women are often seen as lessqualified for managerial-level jobs (e.g., Van Vianen & Willem-sen, 1992). In sum, persistent gender stereotypes appear to workagainst women in many business settings, especially ones involv-ing managerial-level positions. To date, however, existing litera-ture on this important topic has not considered the followingquestion: Do such negative effects occur in another businesscontext as well—that is, with respect to female entrepreneurs?

This appears to be an important issue because mounting evi-dence suggests that the activities of entrepreneurs are crucial to theeconomic growth and prosperity of modern societies (Shane &Venkataraman, 2000). Indeed, the new ventures started by entre-preneurs more than made up for the loss of millions of jobs inlarge, existing companies during the 1980s and 1990s and thusplayed a key role in the current strength of the U.S. economy. Inthis context, the fact that women are underrepresented amongentrepreneurs and face daunting obstacles to assuming this role aredisturbing. Thus, efforts to identify the factors that influence theirdecision to become entrepreneurs and their ultimate success seemboth timely and important. The present research focused on one ofthese factors: perceptions of women who adopt the role ofentrepreneur.

At first glance, it might seem that women entrepreneurs wouldbe subject to the same negative stereotypes and processes thatadversely affect women in other business contexts. Recent findings(Baron, 1999) have indicated that current stereotypes of entrepre-neurs are heavily weighted toward traits traditionally viewed as"masculine." Entrepreneurs are perceived as being more assertive,achievement oriented, and confident than managers and as having

923

Page 2: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

924 BARON, MARKMAN, AND HIRSA

greater risk-taking propensity. In view of such stereotypes, it ispossible that women might be viewed as even more unsuited forthe role of entrepreneur than they are for the role of manager.However, one process may tend to alter this equation and tip it infavor of women entrepreneurs: the process of attributionalaugmenting.

Attributional augmenting refers to the following phenomenon:If causes that would be expected to facilitate a specific form ofbehavior and causes that would be expected to inhibit or prevent itare both present, yet the behavior actually occurs, then the facili-tating causes are assigned extra weight or importance—they areaugmented (e.g., Hansen & Hall, 1985; Kelley, 1973; McClure,1998). We reasoned that such augmenting may occur for womenentrepreneurs and can serve to enhance others' perceptions ofthem. Several considerations point to this possibility.

First, women who contemplate becoming entrepreneurs facedaunting obstacles. In recent years, for instance, women havereceived less than 2% of available venture capital funds (Thomas,1999). Second, women are underrepresented among entrepreneurs,especially in potentially lucrative new ventures in high-technologyfields (Zimmerer & Scarborough, 1997). Thus, fewer models ofsuch behavior are available to them. Third, the businesses headedby women tend to be smaller and to grow less quickly than thoseheaded by men, suggesting, again, that women face greater obsta-cles to obtaining success as entrepreneurs (e.g., Cliff, 1998; Coo-per, Gimeno-Gascon, & Woo, 1994). Fourth, as noted previously,entrepreneurs are perceived as possessing many stereotypically"masculine" traits (e.g., a propensity for risk taking, assertiveness,high achievement orientation; Baron, 1999). Thus, according toexisting gender stereotypes, women may appear, prima facie, to beunsuited for this role.

What happens when, despite these many obstacles (inhibitingcausal factors), women actually become entrepreneurs? One pos-sibility is that attributional augmenting occurs. Because thesewomen have become entrepreneurs despite the many barriers,potential facilitating causal factors may be augmented (i.e., per-ceived as more important or influential). Given the strong andgeneral tendency to perceive others' behavior as stemming largelyfrom internal causes (the correspondence bias or fundamentalattribution error; Gilbert & Malone, 1995; Van Overwalle, 1997),we reasoned that such facilitating factors would be perceivedprimarily as involving characteristics of the entrepreneurs, such astheir motivation, ability, desire for achievement, and so on. Thus,as a result of attributional augmenting, women who become en-trepreneurs may be perceived more favorably with respect to thesedimensions than women who occupy other business roles (e.g.,managerial roles). In addition, because traits such as high motiva-tion or desire for achievement are counter to traditional genderstereotypes, female entrepreneurs may also be perceived as lessstereotypically feminine than women occupying other businessroles. On the basis of this reasoning, we proposed the followinghypotheses:

Hypothesis I: Women who are described as entrepreneurs will berated significantly more favorably than women who are described asmanagers with respect to several personal traits (e.g., assertiveness,decisiveness, ambition, seriousness about their careers, etc.).

Hypothesis 2: Women who are described as entrepreneurs will berated significantly more favorably than women who are described as

managers in terms of attributions concerning the causes of theirsuccess (i.e., the extent to which such success is attributed to internalcauses, such as ability or effort, or external causes, such as luck).

Hypothesis 3: Women who are described as entrepreneurs will beperceived as less stereotypically feminine than women who are de-scribed as managers.

We should add that attributional augmenting may also occur forwomen who become managers, especially in fields that arestrongly sex-typed and perceived as more appropriate for men thanfor women (e.g., Heilman et al., 1988). However, given that theperceived obstacles to becoming an entrepreneur are generallygreater than those to becoming a manager, augmenting would beexpected to occur more strongly for women who choose to becomeentrepreneurs.

What about men who become entrepreneurs? As noted above,existing evidence suggests that they may well face weaker obsta-cles to adopting this role than do women (e.g., Cliff, 1998; Cooperet al., 1994). Thus, attributional augmenting would be expected tooperate to a lesser degree for men and to be less likely to enhanceperceptions of them with respect to personal traits and attributionsabout the causes of their success. On the basis of this reasoning, weproposed the following hypothesis:

Hypothesis 4: Men who are described as entrepreneurs will not berated significantly more favorably than ones described as managerswith respect to either personal-traits or attributions concerning thecauses of their success.

To test these hypotheses, we conducted three studies. In allthree, raters examined photos of strangers and were told (in abetween-subjects design) either that the photos showed entrepre-neurs or that they showed managers. In Studies 1 and 2 we usedphotos of women, and in Study 3 we used photos of both womenand men.

Method

Raters, Stimulus Materials, and Design

Raters. Raters for Studies 1 and 2 were 82 employed adult men andwomen ranging in age from their early 20s to their early 70s (Afag<.s = 41.5years in Study 1, 36.2 years in Study 2). In terms of education, most wereeither college graduates or had completed some college education (78.4%in Study 1, 83.6% in Study 2). In Study 1, 70% were women and 30% weremen; in Study 2, 65% were women and 35% were men. They held a widerange of jobs (e.g., secretarial: 20.5%; professional: 23.1%; technical:17.9%; government: 28.2%; other, 10.3%). Forty-two raters participated inStudy 1, and 40 additional raters participated in Study 2. Raters for Study 3were 52 adult women (51.9%) and men (48.1%), ranging in age from their20s to their 60s. Most had some college education (59.5%) or were collegegraduates (36.1%). They were employed in a wide range of fields: sales(8.5%), office/clerical (4.3%), technical (14.9%), government (8.5%), andprofessional (12.7%). Others were self-employed (12.7%).

Stimulus materials. All photos showed actual entrepreneurs. The pho-tos of women in Studies 1 and 2 were taken from a publication prepared bya national organization with which the entrepreneurs were affiliated; thispublication was widely distributed and was therefore in the public domain.The photos showed a head-and-shoulders view, and all were formal photosfor which the entrepreneurs had posed. A total of 97 photos was availablefrom this source. These were divided, by random selection, into two sets ofphotos, one for Study 1 and the other for Study 2. Raters in Study 1rated 46 photos, and those in Study 2 saw and rated 51 photos. (This small

Page 3: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

PERCEPTIONS OF WOMEN AND MEN AS ENTREPRENEURS 925

difference in number resulted from a minor error in reproducing the photos,which were inadvertently divided into slightly unequal sets during thecopying process.) Photos in both sets were presented in random order.Raters lived in a different state than the stimulus persons and wereunacquainted with them. The entrepreneurs varied considerably in age(from their 30s to their 60s) and appearance. Most were of Europeandescent (84%), but persons of Hispanic (4%), Asian (5%), and AfricanAmerican (7%) descent were also included.

In Study 3 we used 10 photos of men and 10 photos of women. Thephotos of women consisted of a random sample of 10 of the photos ofwomen used in Studies 1 and 2. The photos of men also showed entrepre-neurs. These photos were taken by the researchers or, in a few cases, wereprovided by the entrepreneurs themselves. Like the photos in Studies 1and 2, they showed a head-and-shoulders view. All the men shown wereCaucasian, but as was true for the photos in Studies 1 and 2, the entrepre-neurs differed considerably in terms of age and appearance.

Design. A mixed between-subjects/within-subjects factorial designwas used in which the description of the persons shown in the photos(entrepreneurs vs. managers) was a between-subjects factor, and the dif-ferent stimulus persons rated by participants were a within-subjects factor.That is, participants rated a number of different photos, but all photos in theset they evaluated were described as showing either entrepreneurs ormanagers. In addition, in Study 3 we used a complete factorial design inwhich two variables—gender of stimulus person and ascribed role (man-agers or entrepreneurs)—were systematically varied. The major predictionfor Study 3 was that there would be a significant interaction between thesetwo variables such that ascribed role would exert stronger effects on ratingsof women than on ratings of men.

Procedure

For Studies 1 and 2, participants were recruited from several localorganizations (e.g., a large public library, an insurance company, a retailchain). Participants were told that the study would involve examiningphotos of strangers and that on the basis of these photos, they would ratesome of the strangers' personal traits. Persons who agreed to participate(approximately 92% of those approached) were then given a bookletcontaining the photos they were to rate and a separate rating form. Therating form described the persons shown either as entrepreneurs or asmanagers; participants were randomly assigned to these two conditions.

The photos were presented in random order and were identified bynumber on the rating sheet. In Study 1, raters evaluated the stimuluspersons (women who were described to them as being either managers orentrepreneurs) on the following dimensions: career seriousness, decisive-ness, assertiveness, ambitiousness, sincerity, manipulativeness, mascu-linity-femininity, and attractiveness. These dimensions were chosen be-cause they have been used successfully and informatively in previous,related research (e.g., Heilman & Stopeck, 1985a, 1985b; Heilman et al.,1988). In Study 2, raters evaluated the stimulus persons on the followingdimensions: causes of career success (ability, motivation/effort, luck, socialskills), concern with one's appearance (vain, conceited), masculinity-femininity, and attractiveness. All ratings in both studies (and in Study 3)were made on 5-point scales (1 = very low, 5 = very high). Pilot researchindicated that rating all of the photos in each set on all the dimensionsnamed above was perceived by participants as too time consuming (e.g., itwould require rating 97 photos on 16 different dimensions). This was amajor consideration in our decision to collect different dependent measuresin the two studies.

Raters in Study 3 evaluated each photo on the following dimensions:attractiveness, masculinity-femininity, career seriousness, decisiveness,assertiveness, ambition, and perceived causes of career success (ability,effort, luck, social skills). We deemed it appropriate for raters in these latterstudies to rate each photo on a greater number of dimensions because thetotal number of photos was smaller. Stimulus persons in Study 3 were bothmen and women.

Results

As noted earlier, the photos used in all three studies showedindividuals who varied greatly in age and appearance. Clearly,such differences may influence ratings of the stimulus persons.However, this was not a focus of the present research, and noattempt was made to systematically vary the appearance of thestimulus persons. Rather, the major goal was to determine whetherdescribing these persons as entrepreneurs as opposed to managerswould influence the ratings they received. In view of this fact, wedeemed it appropriate to collapse ratings across stimulus persons.Thus, dependent variables were mean ratings on each dimensionfor all stimulus persons in the photos. We reasoned that shoulddifferences in ratings emerge between the entrepreneur and man-ager conditions (in which participants saw identical sets of photos),this would suggest that such results could be generalized across awide range of age and personal appearance.

Study 1: Ratings of Personal Traits

A multivariate analysis of variance (MANOVA) on the data forall trait measures yielded a significant effect of condition (man-agers vs. entrepreneurs), F(8, 89) = 4.28, p < .001 (rj2 = .28, byWilks's lambda criterion). Follow-up univariate tests indicated thatthe effects of condition were significant for decisiveness, F(l,96) = 5.94, p < .02 (T)2 = .06), and career seriousness, F(l,96) = 5.66, p < .02 (T)2 = .06). For both measures, the women inthe photos received significantly higher ratings when they weredescribed as being entrepreneurs than when they were described asbeing managers (see Table 1). These findings offer partial supportfor Hypothesis 1.

Follow-up univariate analyses also indicated that the effect ofcondition (entrepreneur vs. manager) was significant for ratings onthe masculinity-femininity dimension, F(l, 96) = 23.22, p < .001(T)2 = .20), and on the attractiveness dimension, F(l, 96) = 15.77,p < .001 (T/2 = .14). As shown in Table 1, participants rated thestimulus persons as less feminine but as more attractive when theywere described as entrepreneurs than when they were described asmanagers. The former finding is consistent with Hypothesis 3.

Table 1Mean Ratings on Trait Dimensions as a Function of Condition(Entrepreneur, Manager): Study 1

Description of stimulus persons

Managers

Dependent measure

Career seriousnessDecisivenessAssertivenessAmbitiousnessSincerityManipulativenessMasculinity— femininityAttractiveness

M

3.453.333.353.463.222.853.582.72

SD

0.330.280.270.280.350.450.440.49

Entrepreneurs

M

3.59**3.49**3.443.543.132.903.19**3.13**

SD

0.270.320.340.290.420.730.380.52

Note. There were 21 raters in each condition (managers, entrepreneurs).**p < .01.

Page 4: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

926 BARON, MARKMAN, AND HIRSA

Study 2: Attributions Concerning Success

Study 2 focused primarily on attributions concerning the causesof each stimulus person's success (ability, motivation/effort, luck,social skills). A MANOVA performed on all dependent measuresrevealed a significant effect for condition (managers vs. entrepre-neurs), F(8, 79) = 7.21, p < .001 (rj2 = .42, by Wilks's lambdacriterion). Follow-up univariate analyses indicated that the effectof condition was significant for the attributional items relating toluck, F(l, 86) = 17.89, p < .001 (r? = .18), and social skills, F(l,86) = 5.61, p < .02 (7j2 = .07). As shown in Table 2, the successof persons in the photos was attributed less to luck and more tospecific abilities (i.e., social skills) when they were described asentrepreneurs than when they were described as managers. Asimilar pattern was obtained with respect to effort (entrepreneurswere rated higher than managers), but the effect of condition wasnot significant for this measure or for any of the other dependentmeasures. Overall, there was partial support for Hypothesis 2.

Additional univariate tests revealed that the effect of conditionwas significant for ratings on three other dimensions: masculinity-femininity, F(l, 86) = 3.82,/? < .003 (rf = .05); attractiveness,F(l, 86) = 3.07, p-< .003 (-rj2 = .03); and vanity, F(l,86) = 11.07, p < .001 (tf = .11). As shown in Table 2, womenin the photos were rated as significantly less feminine and less vainbut as more attractive when they were described as entrepreneursthan as managers.

Study 3: Complete Factorial Design

The major prediction investigated in Study 3 was that gender ofstimulus person and ascribed role (manager, entrepreneur) wouldinteract in influencing ratings of the persons shown in the photos.Specifically, we hypothesized that ascribed role would exert stron-ger effects on ratings of women than on ratings of men. AMANOVA performed on the data for the trait dimensions (careerseriousness, decisiveness, assertiveness, ambition, attractiveness,masculinity-femininity) yielded a significant effect for gender ofstimulus person, F(6, 43) = 7.15, p < .01 (rj2 = .50), and for theinteraction between condition (ascribed role) and gender of stim-

Table 2Mean Ratings on Attribution Dimensions and Other DependentMeasures as a Function of Condition (Entrepreneur,Manager): Study 2

Description of stimulus persons

Managers

Dependent measure

AbilityEffortSocial skillsLuckVanityConceitMasculinity-femininityAttractiveness

M

3.383.433.422.722.862.743.472.78

SD

0.330.310.340.340.430.410.640.53

Entrepreneurs

M

3.403.483.65**2.55**2.69**2.713.32**2.89**

SD

0.680.260.370.330.380.730.780.51

Note. There were 20 raters in each condition (managers, entrepreneurs).** p < .01.

ulus person, F(6, 43) = 2.36, p < .03 (if = .25). Follow-upunivariate analyses indicated that the effect for gender of stimulusperson stemmed from the fact that women were rated as signifi-cantly more feminine than men (for women, M = 3.20; for men,M = 2.13), F(l, 48) = 27.10, p < .001 (?)2 = .36). In addition,women were rated as significantly more attractive than men (forwomen, M = 2.63; for men, M = 2.13), F(l, 48) = 10.92, p <.002 (-rf = .19).

Univariate analyses performed to examine the interaction be-tween condition and gender of stimulus person more closely indi-cated that this interaction was significant for two trait dimensions:assertiveness, F(l, 48) = 3.87, p < .05 (-r/2 = .04), and decisive-ness, F(l, 48) = 3.74, p < .05 (r)2 = .03). As predicted, condition(manager, entrepreneur) exerted a stronger effect on ratings ofwomen than on ratings of men (see Table 3) for both of thesedimensions. These findings provide support for Hypothesis 4.

A MANOVA performed on the attributional measures (theextent to which any success obtained by the stimulus persons wasdue to ability, effort, luck, or social skills) yielded a significantmean effect only for gender of stimulus person, F(4, 45) = 2.64,p < .05 (-r/2 = .18). Follow-up univariate analyses indicated thatthis effect stemmed from one dependent measure—the item relat-ing to effort. Men to a greater extent than women were seen assucceeding because of effort (for men, M = 3.70; for women,M = 3.42, p < .05). The multivariate effect for the interactionbetween gender of stimulus person and condition did not attainsignificance, F(4, 45) = 1.62, p > .10. However, univariateanalyses indicated that this interaction was significant for onedependent measure—the item relating to luck, F(l, 48) = 3.71,p < .05 (T)2 = .07). As predicted, women's success was seen asstemming less from luck when they were described as entrepre-neurs (see Table 3). Again, these findings are consistent withHypothesis 4.

Discussion

Results offered support for Hypotheses 1 and 2; as predicted,women shown in standard-format photos received higher ratingswhen they were described as being entrepreneurs than when theywere described as being managers. This pattern was obtained bothfor personal traits (Hypothesis 1) and for the perceived causes oftheir success (i.e., attributions—Hypothesis 2), and it was found inthree separate studies (Studies 1, 2, and 3). In more specific terms,the women were rated significantly higher on two traits—deci-siveness and career seriousness—when described as entrepreneurs(Hypothesis 1). Similarly, their success was attributed less to luckand more to specific skills (i.e., social skills—Hypothesis 2).Finally, as predicted by Hypothesis 3, the women shown in thephotos were viewed as less feminine when described as entrepre-neurs than as managers.

Results obtained in Study 3 offer support for the prediction thatthe effects of ascribed role (entrepreneur vs. manager) would beweaker for men than for women. In Study 3, which used acomplete factorial design, men did not receive significantly higherratings when described as entrepreneurs than when described asmanagers, but such effects did occur for women. Taken together,these findings suggest that women may benefit to a greater extentthan men from assuming entrepreneurial roles, at least with respectto how they are perceived by persons unacquainted with them.

Page 5: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

PERCEPTIONS OF WOMEN AND MEN AS ENTREPRENEURS 927

Table 3Mean Ratings on Trait Dimensions and Perceived Causes of Success as a Function of Condition(Entrepreneur, Manager) and Gender of Stimulus Person: Study 3

Gender of stimulus person

Female Male

Condition(manager, entrepreneur)

AssertivenessDecisivenessAmbitionMasculinity-femininityCareer seriousnessAbilityEffortLuckSocial skills

Managers

M

3.083.153.453.483.423.303.233.073.52

SD

0.340.100.270.280.160.380.430.520.55

Entrepreneurs

M SD

3.33* 0.443.41* 0.343.49 0.183.50 0.163.56 0.113.40 0.603.62* 0.542.75* 0.403.55 0.35

Note, n = 13 except for assertiveness in female entrepreneurs (n =* p < .05 (for effect of condition: managers vs. entrepreneurs).

Managers

M

3.443.403.482.723.603.413.653.063.47

12) and male

SD

0.500.400.120.160.080.530.380.680.68

managers

Entrepreneurs

M

3.303.463.482.443.613.493.752.893.41

(n = 14).

SD

0.160.230.110.150.070.540.500.530.69

One explanation for these findings is provided by the operationof attributional augmenting (e.g., Kelley, 1973; McClure, 1998).Reasoning based on the operation of this attributional mechanismleads to the prediction that women who become entrepreneurs willbe perceived more favorably than women who adopt other busi-ness roles—that is, ones associated with weaker perceived barriersto entry. Further, reasoning based on this mechanism also leads tothe prediction that such effects will occur more strongly forwomen than for men. Results of all three studies are, by and large,consistent with these predictions. Women received higher ratingswhen described as being entrepreneurs than as managers (Studies 1and 2), but similar effects did not occur for men (Study 3). Indeed,Study 3 yielded the interaction between ascribed role and genderof stimulus person predicted by attributional augmenting.

It is important to note, however, that even the findings ofStudy 3 do not provide conclusive evidence that the present resultsstem solely from the operation of attributional augmenting. On thecontrary, the possibility remains that other factors, too, playedsome role. Several such factors can be readily suggested. Oneinvolves the possible existence of a general positive stereotype forentrepreneurs. Such a positive entrepreneurial stereotype couldcertainly enhance ratings assigned to women described as entre-preneurs. However, it would also be expected to enhance ratings ofmen as well (see, e.g., Murphy, Jako, & Anhalt, 1993). Yet thiswas not the case in the present data: Only women benefitedsignificantly from being described as entrepreneurs. Thus, attribu-tional augmenting, which would be expected to occur morestrongly for women than for men, seems to offer a more reasonableinterpretation of the obtained findings than operation of a generalpositive stereotype for entrepreneurs.

Second, it is possible that the role of entrepreneur is less subjectto gender stereotyping than the role of manager. To the extent thatthis is true, it might tend to boost the ratings of female entrepre-neurs more than those of men who adopt this role. However, asnoted earlier in this article, recent findings (Baron, 1999) haveindicated that the role of entrepreneur is far from gender neutral; infact, current views of entrepreneurs appear to be heavily weighted

toward traits traditionally viewed as "masculine" (e.g., entrepre-neurs are perceived as being more assertive, achievement oriented,and confident than managers and as having greater risk-takingpropensity). Thus, this factor, too, does not seem to account for thepresent results.

Finally, it is possible that being an entrepreneur is viewed as amore exciting or adventurous occupation than being a manager andthat as a result, describing strangers as entrepreneurs generateshigher levels of positive affect among raters than describing thesame persons as managers. Higher levels of positive affect, in turn,could boost the ratings of persons to whom this label is attached(e.g., Forgas, 1995). But again, there is no clear reason why femaleentrepreneurs should induce higher levels of positive affect thanmen who adopt this role.

Certainly, this list of potential causal mechanisms is far fromexhaustive, and the possibility remains that factors other thanattributional augmenting could well contribute to the observedfindings. Only additional research designed to examine such mech-anisms can provide a conclusive answer to this question. All wewish to suggest here is that attributional augmenting offers arelatively parsimonious explanation for our results and in thisregard may be a useful working hypothesis for further investiga-tions of perceptions of women and men who adopt the role ofentrepreneur.

An additional finding of interest involves the fact that partici-pants rated the women shown in Studies 1 and 2 as simultaneouslyless feminine and more attractive when they were described asentrepreneurs than as managers. In a series of carefully conductedstudies, Heilman and her colleagues (e.g., Heilman & Saruwatari,1979; Heilman & Stopeck, 1985a, 1985b) have found that attrac-tiveness is often a negative factor for women in managerial-levelsettings. Attractive women receive lower ratings than less attrac-tive women as job applicants and also receive lower performanceappraisals (e.g., Heilman & Stopeck, 1985a, 1985b). In contrast,attractiveness generally produces beneficial effects for men in awide range of business contexts. Additional findings in thesestudies indicated that both effects—negative outcomes for women

Page 6: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

928 BARON, MARKMAN, AND HIRSA

and positive outcomes for men—stem from the fact that a highlevel of attractiveness elicits or intensifies gender stereotypes.Because such stereotypes are mainly negative for women in thecontext of managerial roles, adverse outcomes occur for attractivewomen. In contrast, because gender stereotypes for men are gen-erally positive with respect to managerial roles, positive effectsoccur.

The present findings suggest that being described as an entre-preneur simultaneously boosts women's apparent attractivenesswhile lowering their perceived femininity. Under these conditions,negative gender stereotypes for women may operate to a reduceddegree. As a result, attractiveness may not lead to the negativeeffects observed for women in previous studies (e.g., Heilman& Saruwatari, 1979). On the contrary, it might well yield thebeneficial effects frequently observed for men (e.g., Heilman &Stopeck, 1985a, 1985b). Further research in which attractiveness isvaried in a systematic manner, as was done in Heilman's previousstudies, is necessary to fully assess the accuracy of this reasoning.

At this point, important limitations of the present findingsshould be carefully considered. First, participants in this researchwere asked to rate stimulus persons along several dimensionssolely on the basis of their photos. In real-life situations, a personrarely bases important judgments about others on such limitedinformation. The restricted and somewhat unusual nature of thetask performed by raters may have influenced their responses inunpredictable ways; at the very least, use of this task raisesimportant issues concerning the degree to which the present find-ings can be generalized to more realistic circumstances.

Second, significant results were obtained for some of the ratingscales but not for others, and there were no strong theoreticalgrounds for predicting this specific pattern of findings. The traitdimensions included in this research were selected because theyhad been successfully used in previous, related research (e.g.,Heilman et al., 1988; Heilman & Stopeck, 1985a, 1985b). Thus,significant effects were anticipated for all of these measures. Yetin Study 1, female entrepreneurs received significantly higherratings than male entrepreneurs in terms of career seriousness anddecisiveness but did not receive significantly higher ratings forassertiveness or ambitiousness. In Study 3, female entrepreneurswere again rated significantly higher than male entrepreneurs indecisiveness and also in assertiveness, but they were not ratedsignificantly higher in career seriousness, as was found in Study 1.Careful examination of Tables 1 and 3 suggests that the pattern ofmean differences for three of these dependent measures—asser-tiveness, decisiveness, and career seriousness—was very similar inboth Studies 1 and 3. Specifically, female entrepreneurs were ratedhigher than female managers on all these dimensions in bothstudies. It seems possible, then, that differences in significance forthese measures across the two studies may stem, at least in part,from differences between the samples of participants. Whereasboth samples consisted of employed persons, they differed slightlyin terms of gender composition, occupation, and age, and suchdifferences may have contributed to the fact that different patternsof significant results were obtained in the two investigations. It isimportant to note, however, that such factors cannot explain theabsence of significant findings for ratings of ambitiousness in bothStudies 1 and 3. One possibility is that in contrast to the otherdependent measures, ambitiousness is viewed as something of a"mixed blessing." That is, very high levels of ambition may be

perceived as having potentially negative effects to a greater extentthan such traits as decisiveness or seriousness about one's career.To the extent that this is true, then raters may have been somewhatreluctant to assign higher ratings on this dimension to entrepre-neurs than to managers. This possibility can be examined in futurestudies that focus on the perceived benefits and costs of the varioustraits included in the present research.

A third and even more important limitation involves the fact thatthe present research compared only two occupational groups:entrepreneurs and managers. Thus, the question of whether similareffects would occur for other occupations remains open. For in-stance, would attributional augmenting also enhance perceptionsof women who are physicians, attorneys, or architects? The rea-soning behind our hypotheses suggests that to the extent women inthese fields face stronger obstacles to entering them than men do,this would be the case. Only additional data can address this issue.However, the fact that the proportion of women has risen sharplyin several professions in recent years suggests that perhaps attri-butional augmenting would occur to a lesser degree in these fieldsthan is true for entrepreneurs, where women are still, even today,greatly underrepresented.

Despite these limitations, the present findings appear to havepotentially important implications. First, they suggest that by be-coming entrepreneurs, women may reduce the tendency of othersto view them through the lens of negative gender stereotypes,which exert adverse effects on women's success in many differentcontexts. After they become entrepreneurs, women may benefitfrom attributional processes that enhance perceptions of their mo-tivation, decisiveness, and abilities and reduce the extent to whichthey are perceived as stereotypically feminine. Unfortunately,these processes may occur only after women have become entre-preneurs, so these processes would not be expected, in and ofthemselves, to increase the number of women who choose to starttheir own businesses. However, it seems possible that if womenwho are considering the possibility of becoming entrepreneurs aremade aware of the beneficial effects that may follow from doingso, such knowledge might actually encourage them to "take theplunge." Two of us (Robert A. Baron and Gideon D. Markman)currently teach courses on entrepreneurship, and we hope thatmaking this information available to women contemplating careersas entrepreneurs will encourage them to actually pursue thisoption.

Second, from a broader perspective, the present findings add tothe small but growing body of literature seeking to establish closerconceptual links between psychology and the rapidly expandingfield of entrepreneurship (e.g., Baron, 2000a, 2000b). The researchreported here adds to this literature by suggesting that basic prin-ciples of attribution and social cognition can shed new light onimportant questions relating to entrepreneurship—for instance,how female entrepreneurs are perceived by other persons. Becauseperceptions of entrepreneurs often influence important decisionsabout them by venture capitalists, potential customers, prospectiveemployees, and others (e.g., Shane & Venkataraman, 2000), suchperceptions may strongly affect entrepreneurs' success in estab-lishing new ventures. In the past, psychologists have devotedrelatively little attention to the study of entrepreneurs and thefactors that influence their success. Given the major contributionof entrepreneurs to modern economies, however, it seems impor-tant—and potentially very useful—to bring the tools, findings, and

Page 7: Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting

PERCEPTIONS OF WOMEN AND MEN AS ENTREPRENEURS 929

principles of modern psychology to bear on this topic. The re-search reported here represents one step in the direction of encour-aging this process.

ReferencesBaron, R. A. (1999). Perceptions of entrepreneurs: Evidence for a positive

stereotype. Unpublished manuscript, Rensselaer Polytechnic Institute.Baron, R. A. (2000a). Counterfactual thinking and venture formation: The

potential effects of thinking about "What might have been." Journal ofBusiness Venturing, 15, 79-92.

Baron-, R. A. (2000b). Psychological perspectives on entrepreneurship:Cognitive and social factors in entrepreneurs' success. Current Direc-tions in Psychological Science, 9, 15-18.

Cliff, J. E. (1998). Does one size fit all? Exploring the relationship betweenattitudes toward growth, gender, and business size. Journal of BusinessVenturing, 13, 523-542.

Cooper, A., Gimeno-Gascon, F. J., & Woo, C. (1994). Initial human andfinancial capital as predictors of new venture performance. Journal ofBusiness Venturing, 9, 371-395.

Fisher, A. B. (1992, September 21). When will women get to the top?Fortune, pp. 44-56.

Forgas, J. P. (1995). Mood and judgment: The affect infusion model(AIM). Psychological Bulletin, 117, 39-66.

Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psycho-logical Bulletin, 117, 21-38.

Hansen, R. D., & Hall, C. A. (1985). Discounting and augmenting infacilitatory and inhibitory forces: The winner takes almost all. Journal ofPersonality and Social Psychology, 49, 1482-1493.

Heilman, M. E. (1983). Sex bias in work settings: The lack of fit model. InB. M. Staw & L. I. Cummings (Eds.), Research in organizationalbehavior (Vol. 5, pp. 269-298). Greenwich, CT: JAI Press.

Heilman, M. E., Block, C. J., Martell, R. F., & Simon, M. C. (1989). Hasanything changed? Current characterizations of men, women, and man-agers. Journal of Applied Psychology, 74, 935-942.

Heilman, M. E., Block, J. E., & Lucas, J. A. (1992). Presumed incompe-tent? Stigmatization and affirmative action efforts. Journal of AppliedPsychology,'77, 536-544.

Heilman, M. E., Martell, R. F., & Simon, M. C. (1988). The vagaries of sexbias: Conditions regulating the undervaluation, equivaluation, and over-valuation of women job applicants. Organizational Behavior and Hu-man Decision Processes, 41, 98-110.

Heilman, M. E., & Saruwatari, L. R. (1979). When beauty is beastly: Theeffects of appearance and sex on evaluations of job applicants formanagerial and nonmanagerial jobs. Organizational Behavior and Hu-man Performance, 23, 360-372.

Heilman, M. E., & Stopeck, M. H. (1985a). Attractiveness and corporatesuccess: Different causal attributions for men and women. Journal ofApplied Psychology, 70, 379-388.

Heilman, M. E., & Stopeck, M. H. (1985b). Being attractive: Advantage ordisadvantage? Performance-based evaluations and recommended per-sonnel actions as a function of appearance, sex, and job type. Organi-zational Behavior and Human Decision Processes, 32, 202-215.

Hosada, M., & Stone, D. L. (in press). Current gender stereotypes and theirevaluative content. Perceptual and Motor Skills.

Kelley, H. H. (1973). The processes of causal attribution. American Psy-chologist, 28, 107-128.

Lyness, K. S., & Thompson, D. E. (1997). Above the glass ceiling? Acomparison of matched samples of women and men executives. Journalof Applied Psychology, 82, 359-375.

McClure, J. (1998). Discounting causes of behavior: Are two reasons betterthan one? Journal of Personality and Social Psychology, 74, 7-20.

Murphy, K. R., Jako, R. A., & Anhalt, R. L. (1993). Nature and conse-quences of halo error: A critical analysis. Journal of Applied Psychol-ogy, 78, 218-225.

Ragins, B. R., & Cotton, J. (1991). Easier said than done: Gender differ-ences in perceived barriers to gaining a mentor. Academy of Manage-ment Journal, 34, 939-951.

Ragins, B. R., & Cotton, J. (1999). Mentor functions and outcomes: Acomparison of men and women in formal and informal mentoringrelationships. Journal of Applied Psychology, 84, 529-550.

Ruble, T. L., Cohen, R., & Ruble, D. M. (1984). Sex stereotypes: Occu-pational barriers for women. American Behavioral Scientist, 27, 339-356.

Shane, S., & Venkataraman, S. (2000). The promise of entrepreneurship asa field of research. Academy of Management Review, 25, 217-226.

Thomas, P. (1999, February 25). When Venus talks to Mars. Wall StreetJournal, pp. Bl, B8.

Van Overwalle, F. (1997). Dispositional attributions require the jointapplication of the methods of difference and agreement. Personality andSocial Psychology Bulletin, 23, 974-980.

Van Vianen, A. E. M., & Willemsen, T. M. (1992). The employmentinterview: The role of sex stereotypes in the evaluation of men andwomen job applicants in the Netherlands. Journal of Applied SocialPsychology, 22, 471-491.

Zimmerer, T. W., & Scarborough, N. M. (1997). Entrepreneurship andnew venture formation. Upper Saddle River, NJ: Prentice-Hall.

Received January 18, 2000Revision received October 20, 2000

Accepted November 3, 2000 •