A comparison of four scales predicting entrepreneurship.

Huefner, Jonathan C. ; Hunt, H. Keith ; Robinson, Peter B. 等

INTRODUCTION

Over the past 30 years a variety of scales and instruments have been used in the study of entrepreneurship. Some scales were intended primarily as predictors (Sexton & Bowman, 1984, 1986), others were intended to provide understanding (Boyd & Gumpert, 1984; Singh, 1989; Welsh & White, 1981). Some scales were entrepreneurship specific (Dandridge & Ford, 1987; Hornaday & Vesper, 1982; Scherer, Brodzinski, & Wiebe, 1990), others were broad measures of general characteristics (Begley & Boyd, 1987; Fagenson & Marcus, 1991; Hornaday & Aboud, 1971; Sexton & Bowman, 1983). Virtually all of these studies used only one scale (Sexton & Bowman, 1985 is a notable exception to this). This raises the research question of whether prediction and understanding of entrepreneurship might be enhanced by using several different types of scales in a multi-scale study. Such a comparison would allow entrepreneurship researchers to see which of the scales best discriminates between an entrepreneurial group and other groups. To the extent that the scales are different from each other, such a comparison could guide researchers in selecting the scale(s) most appropriate to their specific research question.

Four scales are used in this study. The Entrepreneurial Quotient[c] (EQ) and the Entrepreneurial Attitude Orientation[c] (EAO) were designed and validated to discriminate between entrepreneurs and non-entrepreneurs. The Myers-Briggs Type Indicator[c] (MBTI[TM]) and the Herrmann Brain Dominance Instrument[c] (HBDI[TM]) were designed and validated as general indicators of a person's preferred ways of thinking and behaving.

The EQ was selected because it was short, had face validity, and was specifically developed to measure entrepreneurship. The EAO was selected because it was specifically developed through rigorous scale development procedures to measure entrepreneurship based on attitude rather than personality theory. The MBTI was selected because it is so widely used across a diverse range of research situations. It has been estimated that over 1.7 million people a year in the United States take the MBTI and that "the MBTI is the most popular 'self-insight, insight into others' instrument in use today" (Druckman & Bjork, 1991, p.96). While the MBTI has been used to predict entrepreneurship (Hoy & Carland, 1983; Wortman, 1986), the authors of the scale made no specific statement regarding its prediction of entrepreneurship. An instrument in the area of brain dominance was selected because of occasional mentions that entrepreneurship was a "right brain" activity (Kao, 1991, p. 160; Timmons, 1985, p. 34; Williams, 1981). The HBDI is a commonly used measure of this type that is based on a brain dominance metaphor. The HBDI has been specifically proposed for entrepreneurship research (Winslow & Solomon, 1989), and Herrmann (1988) made specific statements regarding the HBDI's prediction of entrepreneurs.

In addition, several trait tests have been used in entrepreneurship research. Sexton and Bowman (1986) have done a series of studies using a modified version of the Jackson Personality Inventory and Personality Research Form-E (JPI/PRF-E). Hornaday and Aboud (1971) and DeCarlo and Lyons (1979) used the Edwards Personal Preference Schedule (EPPS). McClelland (1961) and Wainer and Rubin (1969) used the Thematic Apperception Test (TAT).

Based on the reviews of trait research (Brockhaus, 1982; Brockhaus & Horwitz, 1986; Gartner, 1988), it was our expectation that the trait approach to entrepreneurship would not discriminate between entrepreneurs and non-entrepreneurs. Therefore, we did not include any standard "trait" scales such as the EPPS or JPI/PRF-E. Instead, we included two entrepreneurship specific scales and two general scales. The EQ is a test of how one's self-perception and personal characteristics compare with those of "successful entrepreneurs." The EAO is based on tripartite attitude theory. The MBTI is based on Jung's personality types. The HBDI is a measure of preference for certain activities.

OVERVIEW OF THE FOUR SCALES USED IN THIS STUDY

The Entrepreneurial Quotient (EQ) is a paper and pencil instrument created by John Caspari, a Northwestern Mutual Life Insurance Company (NMLIC) employee. NMLIC felt it was to their advantage to hire agents who were "entrepreneurial." Through a literature search Caspari found 60-66 supposed characteristics of entrepreneurs. A university professor was hired to develop the scale using experimental and control groups, resulting in a final instrument of 22 questions with varying weightings. The EQ has been used since then by NMLIC in agent selection and in training and motivating current agents. The EQ has been offered widely to the general public as a self-scored instrument. Scores on the EQ range from -42 to +47 with entrepreneurs having positive scores. The more positive the score the more entrepreneurial the person is predicted to be.

The EQ items are based on many of the truisms surrounding entrepreneurship. Two examples follow. "Significantly high numbers of entrepreneurs are children of first generation Americans. If your parents were immigrants, score plus 1. If not score minus 1." "Entrepreneurial personalities seem to be easily bored. If you are easily bored, add 2. If not, subtract 2" (Northwestern Mutual Life Insurance Company, 1985). Thirteen of the 22 EQ items make it perfectly clear in the question itself how an entrepreneur would answer. Of these, 2 are based on life history (e.g. immigrant parents), while the other 11 are subject to personal interpretation (e.g. easily bored). The remaining 9 items of the EQ, while not explicit in stating how an entrepreneur would answer, are also subject to personal interpretation (e.g. "If you were daring, add 4 more"). As can be seen from these examples, all the EQ in essence is doing is asking subjects many times if they are an entrepreneur, and is highly susceptible to demand characteristics. The advantage to this might be that those who are entrepreneurs know they are entrepreneurs probably better than anyone. This approach of just asking them in different ways if they are entrepreneurs may be a reliable predictor. There does not appear to be any published research that has tested the EQ's ability to predict entrepreneurship.

The Entrepreneurial Attitude Orientation (EAO) is a paper and pencil instrument developed to predict entrepreneurship based on the tripartite model of attitude rather than on demographics or personality theory (Robinson, 1987, Robinson, Stimpson, Huefner & Hunt, 1991). The tripartite model states that cognition, affect, and conation are the fundamental components of attitude. The attitude components are included in the EAO in order to increase the content validity of each subscale. There is a single score for each subscale. The four EAO subscales are: "1) Achievement in Business (ACH) referring to concrete results associated with the start-up and growth of a business venture. 2) Innovation in Business (INN) relating to perceiving and acting upon business activities in new and unique ways. 3) Perceived Personal Control of Business Outcomes (PC) concerning the individual's perception of control and influence over his or her business. 4) Perceived Self-Esteem in Business (SE) pertaining to the self-confidence and perceived competency of an individual in conjunction with his or her business affairs" (Robinson, Stimpson, Huefner & Hunt, 1991, p. 19). The range of the four EAO subscale scores is 10 to 100. On each subscale the higher the value the more entrepreneurial the individual is predicted to be.

The EAO was created specifically to measure entrepreneurship and has successfully discriminated between entrepreneurs and non-entrepreneurs in several research contexts (Hunt, Huefner, Voegele, & Robinson, 1989; Robinson, Huefner, & Hunt, 1991; Robinson, Stimpson, Huefner & Hunt, 1991).

The Myers-Briggs Type Inventory (MBTI) is a paper and pencil instrument based on Carl Jung's personality theory that individual behavior is due to individual differences in perception and judgement. "Perception involves all the ways of becoming aware of things, people, happenings or ideas. Judgement involves all the ways of coming to conclusions about what has been perceived" (Myers & McCaulley, 1989, p. 1). The MBTI is based on four bipolar subscales: Extroversion-Introversion, Sensing-Intuition, Thinking-Feeling, and Judging-Perceiving, each a basic personality type (Myers & McCaulley, 1989). The MBTI is widely used in organizations in both employee selection and personnel development and also in academic research.

Extroversion-Introversion (E-I) is a measure of an individual's preferred way of interacting with people and things in the environment. Extraverts focus on the people and objects around them. Introverts focus on the inner world of concepts and ideas.

Sensing-Intuition (S-N) is a measure of an individual's preferred way of perceiving. A sensing orientation focuses on the moment and the information coming through the five senses. An intuiting orientation focuses more on insight coming from meanings, relationships, and possibilities.

Thinking-Feeling (T-F) is a measure of an individual's preferred way of judgment. A thinking orientation indicates inferences based on logic and analysis. A feeling orientation indicates inferences based on values and feelings of others.

Judging-Perceiving (J-P) is a measure of an individual's preferred way of dealing with the environment. A judging orientation indicates a desire for planning, order, and structure. A perceiving orientation indicates flexibility and a sensitivity to new information.

Based on Jung's theory, the MBTI identifies an individual's preference on each of the four dimensions. One pole of each dimension is preferred over the other, and each dimension is independent of the others. This is not to imply that if an individual is sensing that he or she never uses intuition, just which of the two is preferred. A classification for each of the four dimensions results in sixteen possible types (e.g. ENTP, ISFJ, etc).

The MBTI manual (Myers & McCaulley, 1989) makes no prediction of which preferences would be most common for entrepreneurs. Although the instrument is used as an occupational guidance tool and has extensive listings of occupations, entrepreneurs are not mentioned. There is a listing for "business-general, self-employed" and for "managers and administrators." Neither category is a close match with entrepreneur.

The MBTI has been used in entrepreneurship research. Hoy & Hellriegel (1982) in a study of small businesses managers found that the vast majority (70+%) of them were STs. Ginn and Sexton (1988; 1989; 1990) in their study of founders and cofounders of moderate and fast growing Fortune 500 firms found that I (53%), N (60%), T (80%), and J (54%) were the most common types. For each of these studies the strongest preference tendencies for business owner-managers were obtained for the SN and TF subscales. Based on these findings it is expected that the SN and TF subscales would differentiate between entrepreneurs and non-entrepreneurs.

The Herrmann Brain Dominance Instrument (HBDI) is a paper and pencil instrument developed by Ned Herrmann. It evolved from Herrmann's efforts to categorize individuals as having a left- or right-brain dominance by measuring their learning preference or style (Herrmann, 1988). Brain dominance is inferred from stated preferences. While some of its terminology and even its name bear evidence of its origin, its current form is preference oriented, not physiologically oriented (Ho, 1988).

Herrmann's theory underlying the HBDI is that people's behavioral tendencies can be seen as a combination of four categories of preferences. These four categories initially had brain-dominance names but now are known simply as quadrants A, B, C, and D. These four quadrants are based on the physiological metaphor of the human brain. Quadrant A, upper left, is typified by activities that are logical, analytical, and mathematical. Quadrant B, lower left, is typified by activities that are controlled, planned, and sequential. Quadrant C, lower right, is typified by activities that deal with emotion, are people oriented, or are spiritual in nature. Quadrant D, upper right, is typified by those activities that are imaginative, holistic, and require synthesis (Ho, 1988).

For each of the quadrant scores Herrmann (1988) states that a score of 67 or greater indicates a primary preference, a score between 34 and 66 indicates a secondary preference, and a score between 0 and 33 indicates a tertiary preference. A primary preference is where the individual actively pursues and prefers the activities. A secondary preference is where the individual neither prefers nor avoids the activities. A tertiary preference is where the individual actively avoids the activity.

The entrepreneurial profile is identified by Herrmann as featuring a very high score in the D quadrant, with moderate to strong scores in the other three quadrants (Herrmann, 1988, p. 104). In spite of the specific entrepreneur prediction of the HBDI, there does not appear to be any published research that has tested this relationship.

METHODOLOGY

Statement of Hypotheses

1. We expected to find that the EQ would discriminate between entrepreneurs, owner-managers, and non-entrepreneurs, and that the entrepreneur group mean would be significantly higher than the means for the other two groups.

2. We expected to find that the EAO subscales would discriminate between entrepreneurs, owner-managers, and non-entrepreneurs, and that the entrepreneur group means for each of the subscales would be significantly higher than the other two groups.

3. We expected to find that one or more of the MBTI subscales, especially the SN and TF, would discriminate between entrepreneurs, owner-managers, and non-entrepreneurs, and that the entrepreneur group means for the SN and TF subscales would be significantly more Intuitive and Thinking than the other two groups.

4. We expected to find that only the Quadrant D subscale of the HBDI would discriminate between entrepreneurs, owner-managers, and non-entrepreneurs, and that the entrepreneur group means for the Quadrant D subscale would be significantly higher than the means for the other two groups.

5. We expected to find combinations of the EQ, the EAO subscales, the MBTI subscales, and the HBDI subscales that would give greater discrimination between entrepreneurs and non-entrepreneurs than would any single scale or subscale.

One-hundred seventy-three students in four senior level and two graduate level business classes, participating in an ongoing class project, approached student and non-student friends, family, and acquaintances and solicited their help in answering four scales plus three sheets of demographic information. Respondents were undergraduate and graduate university students, their friends, family, and acquaintances.

We realize that this sample gives purists "sample heartburn." We relied on this sample because it seemed unlikely that uninvolved respondents would give the estimated two hours to complete all the questionnaires, and incomplete responses were worthless. The study was conducted without financial support, so had to rely on psychological involvement. Being asked by a student they knew to help that student in a class project provided enough involvement to obtain complete response sets from 335 respondents. These respondents were then classified into groups based on their own demographic information and self-categorization, not by the students who asked them to participate. Paying small business owner-managers and entrepreneurs and others for two hours of their time was prohibitive and had its own objections. Members of a trade association or small business association might have a biased frame of reference. While it was obvious from the requested demographic and self-categorization questions that the study had something to do with business ownership and entrepreneurship, no further elaboration was provided to the students or to the respondents. We did what we could to ameliorate the effects of a less than ideal sample. Bias and demand effects are seldom obvious, so disclaiming them is fruitless. However, we did at least get a variety of respondents, some in every classification, and each of the 335 provided a complete data set.

About 500 questionnaire packets were produced for distribution by students. Because participation was not required and the students themselves distributed the questionnaire packets, the total number of questionnaire packets actually given to potential subjects was unknown. Students received class project credit for completed, returned questionnaires. Three-hundred thirty-five subjects returned completed questionnaires (approximately 67% return rate). The age range for subjects was 16 to 72 (M = 32.6, s = 10.8). The years of education ranged from 8 to 20 (M = 16, s = 2.2).

A packet containing an instruction sheet, four scales and two demographic questionnaires was given to each subject. The EQ consisted of 22 forced choice items. The EAO consisted of 75 items that are rated on a 10-point strongly disagree to strongly agree scale. The MBTI consisted of 166 forced choice items. The HBDI consisted of 120 forced choice items. The first demographic questionnaire consisted of 13 questions about such topics as birth order, family background and economic status, education, and previous business experience. The second demographic questionnaire consisted of 9 questions dealing with previous, current, and expected future entrepreneurial experience. The demographic questions were answered on the questionnaire. The scales were answered on machine scorable answer sheets.

While the order of the questionnaires was the same for every packet, they were not connected, thus, subjects may have completed them in any order. When asked, most subjects reported that it took between 1.5 and 2.5 hours to complete all the questionnaires in the packet.

The packets were distributed by undergraduate and graduate students who were enrolled in entrepreneurship courses at Brigham Young University. Students were invited to fill out a packet themselves. Students were especially encouraged to give packets to people whey knew who they thought definitely were or were not entrepreneurs.

The instructions given to all subjects were that they were to go through the material quickly giving their first response. It was suggested that they not complete all the questionnaires in one sitting and that they could even do them over a period of days if they wanted. It was hoped that this instruction would help minimize fatigue. Subjects were told that their responses would be anonymous and that they could obtain a summary of their scores for each of the tests taken. Students were told their scores would be returned in class. Non-students were invited to provide a mailing address if they wanted their scores returned.

Subjects were told that the purpose of the project was to find family background, personal characteristics, brain dominance, and experience factors related to entrepreneurship. It was explained that the personal benefit of participating in this study would be that they would potentially gain some new personal insight, especially as it might relate to entrepreneurship.

Included in the packet was an addressed, postage paid envelope for subjects to mail back their completed questionnaire packets. The MBTI and HBDI were professionally scored, while one of the authors calculated the scores for the EQ and the EAO.

An explanation sheet and summary of all test scores were sent to each participant who provided a return address. Brief summaries of the MBTI and HBDI were returned to participants with the instruction that "Each of these instruments is meant to be administered by a professional and the output interpreted by a professional. If you want to use any of this output for guidance in your life we strongly suggest that you contact a professional psychologist for a thorough interpretation of your scores."

In our own previous work (Robinson et. al., 1991), we defined entrepreneurs as individuals who had started two or more businesses, the last within the past five years, using some form of innovation. In our current research efforts we again started with the expectation that entrepreneurs were those who had owned and managed one or more businesses. However, of those in our sample who had owned and managed one or more businesses (n = 148), in a question that asked if they were entrepreneurs, 27 individuals (18.2%) answered "no." This was contrary to our a priori approach to entrepreneurs and led us to reevaluate our operationalization.

This led us to conceptualize a 2 by 2 matrix illustrated in Figure 1. The first dimension is whether subjects currently owned and managed or had previously owned and managed businesses or other ventures (none versus 1 or more). The second dimension is how the subject answered "Are you an entrepreneur?" (yes or no). The number who qualified for each category is given in each cell.

Using the classification grid in Figure 1, non-entrepreneurs were those who said they were not entrepreneurs and had never owned and managed a business. Owner-managers were those who said they were not entrepreneurs but had owned and managed one or more business. Entrepreneurs were those who said they were entrepreneurs and had owned and managed one or more business. Potential entrepreneurs were those who said they were entrepreneurs but had never owned and managed a business.

[FIGURE 1 OMITTED]

We were conceptually troubled by exactly what a "potential entrepreneur" represents. We suspected at first that potential entrepreneurs were undergraduate students who thought they might be entrepreneurs in the future but had not yet owned and managed a business. However, in looking at the demographic characteristics of these 13 individuals (3 women and 10 men), the age ranged from 23 to 29 and education ranged from 16 to 19 years. These are people who have had time to start businesses, but have not done so. There was no way of knowing whether this was because they were still students, because the right opportunity had not yet presented itself, or because their assessment that they are entrepreneurs was incorrect. Because we couldn't clearly identify the entrepreneurial and business characteristics of this group, and because there were only 13 individuals in this group, they were dropped from the final analysis.

ANALYSIS

The EQ and EAO provide interval data, whereas the MBTI and HBDI result in categorizations based solely on interval data. While the express purpose of the MBTI and HBDI is to classify individuals into types, there cannot be any information in the categorical data that isn't more fully expressed in the interval data because the categorical information was derived from the interval data. The purpose of this research was to compare the relative efficacy of each of the instruments in predicting entrepreneurs. To be able to run parallel analyses for each of the four scales it was essential to use the interval subscale values from the MBTI and HBDI. The use of interval data is also a minimal requirement for MANOVA and discriminant analysis which were both used in this study.

The EQ, EAO, and HBDI produced interval scores directly usable in the statistical analyses. The subscale scores produced by the MBTI, however, were alphanumeric combinations. For example, the Introvert-Extravert subscale could be for one individual an "I 47" and for another individual an "E 47." To differentiate between these two identical numerical scores, there was a transformation done based on the "alpha" part of these subscales. To do that we multiplied one of the two dimensions of each subscale by a -1 to reverse the sign. For this analysis Extrovert, Sensing, Thinking, and Judging scores were all multiplied by a -1 while the Introvert, Intuitive, Feeling, and Perceiving scores remained the same. Thus, for the Extravert-Introvert subscale using the example above, the I/47 score became a -47 while the E/47 stayed a +47 on the new scale.

Sex was used as an experimental control variable in all analyses because it might interact with group for the dependent variables. Sexton and Bowman-Upton (1990) found personal characteristic differences for men and women, but concluded that it would have no impact on entrepreneurship success. The model for this study was a 2 x 3 (2 sexes by 3 groups) MANOVA.

The discriminant analysis used the "jackknife procedure" for the classification results. Internal classification schemes, when the same data set is used for both the calculation of the discriminant analysis and the classification results of that analysis 1) produce an artificially high correct hit rate (Huberty, 1984) and 2) produce tests of significance of difference between proportions that are ambiguous (Hsu, 1989). In this data set there were not enough subjects to run a split-half cross-validation procedure without significantly reducing the statistical power of the discriminant analysis. Instead, SAS Proc Discrim was used to crossvalidate the discriminant analysis using the jackknife classification procedure. The jackknife or "leave-one-out procedure" removes each subject one at a time, calculates the discriminant function based on all the other subjects, and then classifies the deleted subject according to that discriminant function. This process is carried out for every subject in the data set. The final classification table shows the tally of the classifications of the deleted subjects (Huberty, 1984). This approach generally provides a reasonable assessment of the performance of the discriminant function (McKay & Campbell, 1982b).

RESULTS

Table 1 shows the univariate F-scores and probabilities for sex, group, and the sex by group interaction for the EQ scale and the EAO, MBTI, and HBDI subscales. The last two rows show the MANOVA results and degrees of freedom for sex, group, and the sex by group interaction. There was an overall statistically significant effect for both sex (F (13, 304) = 5.91, p < .0001) and group (F (26, 608) = 2.86, p < .0001). There was not an overall effect for the sex by group interaction (F (26, 608) = 1.16, p = .26). The results for the univariate Fs for the sex and group effects are presented in separate sections for each of the scales below.

Table 2 shows the means, standard deviations, and statistical probabilities for the main effect for sex for the EQ and each of the subscales. Table 3 shows the means, standard deviations, and statistical probabilities for the main effect for group for the EQ and each of the subscales. Table 4 shows the means, standard deviations, and statistical probabilities for the sex by group interaction for the EQ and each of the subscales. While there are two significant univariate Fs for the sex by group interaction, there was not an overall MANOVA for this effect, and so these interactions will not be discussed. An examination of the univariate Fs for the sex and group effects is presented in separate sections for each of the scales below.

While the overall sex effect was statistically significant, this variable was included not because it comprises a point of primary relevance to the study, but because of the possibility that it might interact with the group effect. While the information for sex and sex by group interaction is included in Tables 2 and 4, the statistical results for each scale will focus primarily on the group effect.

As shown in line 1 of Table 2, there was a significant sex effect for the EQ (F (1, 316) = 9.35, p=.0024). The mean for men (M = 11.03) was significantly higher than the mean for women (M = .65). There was no significant sex by entrepreneur interaction for the EQ (Table 4 line 1) . As shown in line 1 of Table 3, there was a significant difference between the entrepreneur, owner-manager and the non-entrepreneur groups (F (2, 316) = 32.83, p = .0001). Sheffe's S showed that entrepreneurs (M = 18.24) were significantly higher than both non-entrepreneurs (M = 1.81) and owner-managers (M = -.63) on this scale. There was no statistically significant difference between non-entrepreneurs and owner-managers for this scale. The [[member of].sup.2] (eta-squared) for the group effect was 20.8% for the EQ. Eta-squared is a measure of the proportion of population variance in the dependent measure attributable to treatment group membership and is a construct similar to that of [r.sup.2].

Of the four EAO subscales (Table 2 lines 2-5), the PC and ACH subscales were not significantly different for men versus women. For the SE subscale men (M = 75.68) were significantly higher than women (M = 74.02). For the INN subscale men (M = 70.80) were significantly higher than women (M = 67.11). Three of the four EAO subscales were significantly different for the entrepreneur, owner-manager, and non-entrepreneur groups (Table 3 lines 2-5). The exception was the ACH subscale. The SE (F (2, 316) = 3.36, p = .0359), PC (F (2, 316) = 3.55, p = .0298), and INN (F (2, 316) = 9.69, p = .0001) subscales were significantly different for the entrepreneur, owner-manager, and non-entrepreneur groups. Sheffe's S showed the SE and PC means for entrepreneurs (M = 77.39 & 71.67) to be significantly higher than those means for non-entrepreneurs (M = 73.76 & 68.17), but not significantly higher than owner-managers (M = 74.20 & 69.16). For the INN subscale, Sheffe's S showed that entrepreneurs (M = 73.56) were significantly higher than both owner-managers (M = 68.06) and non-entrepreneurs (M = 67.16). The [[member of].sup.2] for the group effect was 2.1% for the SE subscale, 2.2% for the PC subscale, and 6.1% for the INN subscale.

There was no statistically significant difference between men and women for any of the MBTI subscales (Table 2 lines 6-9). Additionally, there was no statistically significant sex by group interaction (Table 4 lines 6-9). Of the four MBTI subscales for group effect (Table 3 lines 6-9), the Extravert-Introvert (EI) and Thinking-Feeling (TF) subscales were not significantly different for the entrepreneur, owner-manager, and non-entrepreneur groups. There was a significant difference between the entrepreneur, owner-manager, and non-entrepreneur groups on the Sensing-Intuitive (SN) subscale (F (2, 316) = 9.28, p = .0001).

Sheffe's S showed that entrepreneurs were significantly higher than both owner-managers and non-entrepreneurs on this subscale. The mean for entrepreneurs was clearly Intuitive (M = 5.05). The means for owner-managers (M = -11.07) and non-entrepreneurs (M = -9.28) were clearly Sensing. The standard deviations were large enough (s = 27.12, 25.38, and 27.37) that many entrepreneurs were Sensing (45.4%), and many owner-managers (33.3%) and non-entrepreneurs (36.8%) were Intuitive. There was also a significant difference between the entrepreneur, owner-manager, and non-entrepreneur groups on the Judging-Perceiving (JP) subscale (F (2, 316) = 4.48, p = .0120). Sheffe's S showed that entrepreneurs were significantly higher than both owner-managers and non-entrepreneurs on this subscale. The mean for entrepreneurs was Perceiving (M = 2.82). The means for owner-managers (M = -18.63) and non-entrepreneurs (M = -8.61) were clearly Judging. The standard deviations were large enough (s = 28.01, 22.54, and 26.31), however, that many entrepreneurs were Judging (48.8%), and many owner-managers (18.5%) and non-entrepreneurs (35.1%) were Perceiving. The ,2 for the group effect was 5.9% for the Sensing-Intuition (SN) subscale, and 2.8% for the Judging-Perceiving (JP) subscale.

Of the four HBDI subscales, the Quadrant B and Quadrant D subscales were not significantly different for men and women (Table 2 lines 10-13). For the Quadrant A subscale men (M = 73.89) were significantly higher than women (M = 54.97). For the Quadrant C subscale women (M = 76.60) were significantly higher than men (M = 60.13). There was one significant sex by entrepreneur interaction for the HBDI Quadrant B subscale (Table 4 line 11). A means comparison test showed male entrepreneurs (M = 67.09) were significantly lower than the other five groups (F (1,316) = 14.35, p = .0002). None of the other interaction means for Quadrant B were significantly different from the others. This suggests that male entrepreneurs, while not low on the scale in the absolute sense, are lower than the other groups in activities that are controlled, planned, and sequential. For the group effect of the four HBDI subscales (Table 4 lines 10-13), the Quadrant A, B, and C subscales were not significantly different for the entrepreneur, owner-manager, and non-entrepreneur groups. On the Quadrant D subscale there was a significant difference between entrepreneurs, owner-managers, and non-entrepreneurs (F (2, 316) = 7.92, p = .0004). Scheffe's S showed the mean for entrepreneurs (M = 74.08) to be significantly higher than the means for owner-managers (M = 59.44) and non-entrepreneurs (M = 59.41). There was no significant difference between owner-managers and non-entrepreneurs. This suggests that the entrepreneurs were higher for those activities that are imaginative, holistic, and require synthesis. The ,2 for the group effect was 5.0% for the Quadrant D subscale.

THE DISCRIMINANT ANALYSES

Our final research question dealt with whether we could find combinations of the EQ, the EAO subscales, the MBTI subscales and the HBDI subscales that would give greater discrimination between entrepreneurs and non-entrepreneurs than would any single scale or subscale. A series of discriminant analyses were run for each of the scales and subscales, singly and in combination, where the dependent measure was the entrepreneur, owner-manager, and non-entrepreneur categorization. The adequacy of the discriminant functions were evaluated based on the jackknife classification results produced for each analysis.

Following the suggestion of McKay and Campbell (1982a), virtually every possible combination of scales and subscales was run for a total of 39 different discriminant analyses. Because this is far too much information to present in a paper of this type and because so many of the analyses produced less satisfactory classification results, only the four scales and a select few of the best combinations of scales and subscales are presented here. It is important to make clear that the objective in the evaluation of the classification tables was to maximize the number of entrepreneurs that were correctly identified as entrepreneurs and minimize the number of non-entrepreneurs and owner-managers that were incorrectly identified as entrepreneurs. This combined consideration was seen as the best overall correct classification for entrepreneurs. For the remainder of this paper, consistent with the terminology of signal detection theory, a correct hit is when entrepreneurs were identified as entrepreneurs and an false hit was when either non-entrepreneurs or owner-managers were identified as entrepreneurs.

Table 5 presents the classification results for the four major scales used in this study. The column labeled "entrepreneur" gives the percent of entrepreneurs, business-owners, and non-entrepreneurs classified as entrepreneurs. The highest overall correct classification for entrepreneurs was obtained for the EQ (highest percent of entrepreneurs correctly classified as entrepreneurs and the lowest percent of business-owners and non-entrepreneurs incorrectly classified as entrepreneurs). What is noteworthy with the EQ is that it did not produce an especially high number of correct hits (53.3%), but produced by far the lowest number of false hits (8.1%). The analysis for the EAO produced a higher correct hit rate (60.6%), but also produced a much higher false hit rate (32.2%). Both the correct hit rate and the false hit rate for the MBTI (57.7% and 30.4%) and HBDI (58.4% and 29.1%) were slightly lower than for the EAO and both rates were substantially higher than for the EQ.

Table 6 presents the classification results for only those combinations of the four scales which produced the best results. The best combination of scales was the EQ/EAO/MBTI, with a correct hit rate of 62.8% and a false hit rate of 28.2%, which is superior to all single scales with the possible exception of the EQ. The next highest combination of scales was the EAO/MBTI which produced approximately the same results (62.8% and 29.8%). It is interesting to note that in combining all four scales, the EQ/EAO/MBTI/HBDI, the correct hit rate dropped (59.1%) and the false hit rate remained relatively high (28.9%).

The last line of Table 6 also presents the classification results for the EQ with two subscales, the EAO's INN subscale and the MBTI's EI subscale. This combination gave the highest correct hit rate, 67.2%, of any of the classification results; unfortunately, it also gave the highest false hit rate of 36.4%. Table 7 summarizes the correct hit rates and false hit rates for each of the classification results presented in Tables 5 and 6.

DISCUSSION

It seemed likely that male and female entrepreneurs might differ significantly in terms of the scales used in this study. Because of this it was important to test for the sex by group interaction. The MANOVA showed a significant overall sex effect. The HBDI Quadrant A and Quadrant C, the EAO SE and INN, and the EQ univariate ANOVAs were all significant.

The MANOVA results for the sex by group interaction did not show a significant overall effect. There were, however, two significant univariate ANOVAs: the EAO SE and the HBDI Quadrant B scores. Given the lack of an overall significant sex by group interaction, the effect of sex per se was of limited interest in this analysis.

It does bear mentioning, however, that for the EQ the mean difference between men and women was not only statistically significant but also large (M = 11.03 for men versus M = .65 for women). Given that the EQ is solely intended as a measure of entrepreneurship, this large difference suggests the strong possibility that the EQ is sex-biased (see Webb, 1991 for a discussion of the problems of sex bias in job related psychometric testing). The differences for the HBDI Quadrant A and Quadrant C were also large. However, the HBDI subscales are intended to measure broader behavioral patterns, not just entrepreneurship, so in the context of this research the sex difference on these subscales is less critical.

There was a significant sex effect for the EQ. Men scored substantially higher (M = 11.03) on this scale than did women (M = .65). The EQ is presented as an indicator of potential for entrepreneurial success. As there is no basis for supposing that men and women, in general, differ in their potential for entrepreneurial success, the EQ needs to be either revised or sex normed if it is going to be used as a general measure of entrepreneurship. There was no significant sex by entrepreneur interaction. For the EQ, entrepreneurs scored significantly higher than owner-managers and non-entrepreneurs, which supports hypothesis 1. The ANOVA result for the EQ was the strongest of any scale (the [[member of].sup.2] was 20.8%). This is not too surprising for a scale specifically designed to measure entrepreneurship.

The EQ, however, has heavy demand characteristics. Thirteen of the 22 questions indicate in the question how an entrepreneur would answer. The operationalization of entrepreneur used in this study was an individual who had started one or more businesses or other ventures and who said they were an entrepreneur. It is not surprising that people who indirectly say they are an entrepreneur on thirteen questions, also say they are an entrepreneur when asked "are you an entrepreneur?" Given that the demand characteristics of the EQ are in sum the same as asking subjects if they are entrepreneurs, this may account for this scale's discriminatory power. This doesn't argue against this scale's usefulness as an indicator of who might be an entrepreneur. It does, however, indicate that the EQ's discriminatory power might be due to the demand characteristics of the scale rather than the relevance of the material in the EQ's questions. Despite the EQ's shortcomings of obvious sex bias and serious demand characteristics, it is still one of the best predictors of entrepreneurship (see Table 7). Future development of the EQ, if any, needs to directly address these problems.

There was a significant sex effect for the ACH subscale. It is interesting to note that, perhaps contrary to stereotype, it was women who were significantly higher than men on this subscale. This may have something to do with the sample being taken from a highly educated population. There were no significant sex by entrepreneur interactions, however, so this finding for sex is of limited relevance to the central research question.

Three of the EAO subscales, SE, PC, and INN, were significantly different for entrepreneur versus owner-manager and non-entrepreneur groups. This should not be surprising given the fact that the EAO scale was specifically designed to measure entrepreneurship. The expectation was that all four of the EAO subscales would be statistically different for the group effect (Hypothesis 2). It was surprising that the ACH subscale, which is a measure of concrete results associated with the start-up and growth of a business venture, was not significant for the group effect. While entrepreneurs were significantly higher in their sense of self-esteem, their desire for personal control, and their innovation in business, all three groups reported about equally strong desire for concrete business results (Table 3 line 5).

Based on the [[member of].sup.2], of the EAO subscales the INN subscale was the strongest predictor (6.1%). PC and SE scores, while statistically significant, had relatively small statistical effects ([[member of].sup.2] = 2.2% and 2.1% respectively).

The non-significant results for ACH were similar to those found for the step-wise discriminant analysis results in earlier research using the EAO (Robinson, Stimpson, Huefner, & Hunt, 1991). The step-wise results indicated that of the four subscales, the effect of ACH was the weakest and was redundant once the other subscales were taken into account. This seems to suggest that, despite the folk-knowledge that entrepreneurs have a greater need for achievement than others, this dimension was not one that differentiates entrepreneurs from the population at large. Table 3 line 5 shows that the ACH means for all three groups were the highest means for any of the EAO subscales, with essentially no difference between them. This suggests that this dimension may be endemic to American culture.

In comparing the EAO results of this study to earlier research (Robinson, Stimpson, Huefner, Hunt, 1991), it is important to recognize that the current study used a somewhat different operationalization of entrepreneur. In the earlier study many of the current owner-manager group would have been classified as entrepreneurs. The entrepreneur / owner-manager distinction is important because for all of the statistically significant subscales for the group variable, the owner-managers were significantly different from the entrepreneurs.

There was no significant sex or sex by group interaction effect for any of the MBTI subscales (possibly because this scale is sex normed). In MBTI theory, the Extravert-Introvert subscale is the most dominant. There was not a significant difference between the groups for the Extravert-Introvert subscale, although it was very close to the conventional .05 level of significance (F (2, 316) = 2.85, p = .059). The mean for entrepreneurs was clearly Extravert, while the means for the other groups were slightly Introvert. In MBTI theory, the second most dominant scale is the Sensing-Intuitive, and the groups were significantly different for this scale. The mean for entrepreneurs was moderately Intuitive, while owner-managers and non-entrepreneurs were clearly Sensing (Table 3 line 7). There was also a significant difference for the groups for the Judging-Perceiving subscale. The mean for entrepreneurs was slightly Perceiving, while the means for the owner-managers and non-entrepreneurs were clearly Judging (Table 3 line 9).

While there was no significant difference for the Thinking-Feeling subscale, slightly more of the entrepreneur group were classified as Thinking rather than Feeling (Table 3 line 8). This is consistent in direction with earlier research findings (Ginn & Sexton, 1988, 1989, 1990; Hoy & Hellriegel, 1982), although the effect was not as strong. The non-significant findings, although in the right direction, do not support the hypothesized Thinking-Feeling difference between entrepreneur and non-entrepreneur groups (Hypothesis 3).

The Sensing-Intuition and the Judging-Perceiving subscales were significantly related to the group variable, with entrepreneurs being significantly more Intuitive and Perceiving. However, substantial proportions of entrepreneurs were the opposite: Sensing (45.4%) and Judging (48.8%). Thus, caution needs to be exercised in categorizing all entrepreneurs as "Intuitive-Perceiving" types.

The theory behind the MBTI is explicit in stating that only the type categorization, not the subscale scores, should be used. Based on this approach, the Intuitive to Sensing ratio for entrepreneurs was 54.6% / 45.4%. It seems unlikely that categorical analysis would have been as sensitive to this difference as was the ANOVA using the interval scores. This supports the use of the MBTI interval subscale scores especially in research settings.

Despite the statistical advantages of using the MBTI interval subscale scores, some mention of how the groups would have been classified bears mentioning. In Table 3 lines 6-9 a negative score indicates Extravert, Sensing, Thinking, or Judging while a positive score indicates Introvert, Intuitive, Feeling, or Perceiving. It is noteworthy that for each of the subscales the entrepreneur group means made them ENTP, which was the opposite of the owner-manager and non-entrepreneur groups' means (ISFJ). The type attributes of entrepreneurs in this study were also substantially different from those given in the MBTI manual (Myers & McCaulley, 1989) for business-general, self employed (ESTJ; E = 56.8%, I = 43.2%; S = 54%, N 46%; T = 58.4%, F = 41.6%; J = 60%, P = 40%), and managers and administrators (ESTJ; E = 56.7%, I = 43.3%; S = 56.3%, N = 43.7%; T = 61.6%, F = 38.4%; J = 69.3%, P = 30.1%), which were the categories most similar to entrepreneur listed in the MBTI manual. These categories from the MBTI manual have the E and T in common with the entrepreneur group (ENTP) and the S and J in common with the owner-manager and non-entrepreneur groups (ISFJ). It is possible that the MBTI norms for owner-managers may have included both the owner-manager category and the entrepreneurship category, confounding the differences found between the two groups. Not having found any previous entrepreneur norm per se for the MBTI, this study may be the first norming of the MBTI for entrepreneurs.

There were significant sex effects for the HBDI Quadrant A and Quadrant C subscales. Men were significantly higher than women on the Quadrant A subscale and men were significantly lower than women on the Quadrant C subscale. The Quadrant A subscale is intended to be a measure of activities that are logical, analytical, and mathematical. The Quadrant C subscale is intended to be a measure of activities that deal with emotion, are people oriented, or are spiritual in nature. Neither of these subscales was hypothesized as relevant to entrepreneurship nor significant for the group variable.

A significant sex by entrepreneur interaction was obtained for the HBDI Quadrant B score, but, because of the lack of an overall sex by group finding on the MANOVA, this finding could well be due to chance alone. Given this word of caution, for this subscale male entrepreneurs were significantly lower than the other groups. The Quadrant B subscale is intended to be a measure of activities that are controlled, planned, and sequential. However, it is important to note that the mean for the male entrepreneur group of 67.09 still indicates a primary preference for these activities, while the other group means are even higher in the 'primary' preference range for this subscale.

The only significant group difference for the HBDI subscales was for Quadrant D. The profile score for Quadrant D (upper-right) was significantly higher for entrepreneurs than for the other two groups. For the Quadrant D subscale (the right brain metaphor) the highest mean was for the entrepreneur group and the lowest means were for the owner-manager and non-entrepreneur groups. This Quadrant D score indicates a preference for activities that are imaginative, holistic, and require synthesis, preferences often seen as central to entrepreneurship. This is in line with the theoretical position stated by Herrmann (1988) that "the entrepreneurial profile features a very strong D quadrant preference" (p. 104), and confirms Hypothesis 4.

DISCRIMINANT ANALYSES

One of the central questions of this study was whether combinations of the scales and the subscales would discriminate between entrepreneurs, owner-managers, and non-entrepreneurs better than do single scales or subscales. There are any number of ways of framing discriminant analysis. The objective used here was to maximize the correct hit rate for entrepreneurs, while minimizing the number of non-entrepreneurs and owner-managers who were misclassified as entrepreneurs. These results are summarized in Table 7.

The correct hit and false hit percentages in Table 7 show that, with the exception of the EQ alone, there were not large differences for the various scale and subscale combinations when classifying subjects as entrepreneurs. To conserve on space, only the best classification results were reported in this paper. The EQ's low misclassification of owner-managers and non-entrepreneurs is probably due to those who said they were not an entrepreneur and had not owned a business (the operationalization of non-entrepreneur used here), also saying no they were not an entrepreneur when asked 22 more times in another context (the EQ). In answering the EQ very few non-entrepreneurs "said" they were entrepreneurs.

For the most part there was a tradeoff--more correct hits were usually accompanied by more false hits. Increasing the number of entrepreneurs correctly identified also increased the number of owner-managers and non-entrepreneurs that were misclassified as entrepreneurs. Because of this there was no clear recommendation for the best scale or combination of scales. The gain of a few percentage points in correct hits and fewer false hits resulting from some of the combinations (e.g. the improvement of the EAO/MBTI combination over the EAO alone), comes at the price of a much longer test battery.

In terms of overall efficiency, the EQ produced the best results because of its exceptionally low misclassification of owner-managers and non-entrepreneurs as entrepreneurs, rather than on having the highest correct hit rate for entrepreneurs. The best of the rest of the scales and subscale combinations were fairly similar in efficiency. The best combination of scales was the EQ/EAO/MBTI. This produced a higher correct hit rate for entrepreneurs (62.8%), but also produced a much higher false hit rate (28.2%) than the EQ alone. Perhaps the most interesting combination was the EQ/EAO-INN/MBTI-Extravert-Introvert, which produced both the highest correct hit rate (67.2%) and the highest false hit rate (36.4%).

CONCLUSIONS AND LIMITATIONS

The two primary limitations in this study are the non-random sample and possible order effects. The sample consisted of undergraduate and graduate university students, their friends, family, and acquaintances. This sample was relied on because it seemed unlikely that uninvolved respondents would give the estimated 2 hours to complete all the questionnaires. As it stands, those who did complete the questionnaires did it either as a class assignment or as a favor to a student they knew well. As the average subject was a 32-year-old with four years of college, the sample obviously did not just consist of undergraduate and graduate students.

The focus of this study was on the relative efficacy of four psychological scales in differentiating between entrepreneurs, owner-managers, and non-entrepreneurs. It was not intended that the results reported here be understood as being descriptive of the population in general. It is expected, however, that the scales that differentiated between the groups in this study would also differentiate between these groups if gathered through a random sampling process. Again, the reason a random sample was not used in this study was the improbability of uninvolved subjects taking the time to complete all the questionnaires used.

The second limitation involves potential order effects due to fatigue and tedium in completing the questionnaire packet. In the packet provided to each subject the questionnaires were always stapled independently and clipped together in the same order. However, once the clip was removed any order of completion was possible. Most of the packets were probably answered in the original order. The tedium of completing the full set of questionnaires may have had an effect on the way that subjects responded to certain scales. Thus, there may have been some order effects. Several of the subjects complained that the process took too long.

This study has clearly established that all four scales discriminated between entrepreneurs, owner-managers, and non-entrepreneurs. There were clear cut differences between entrepreneurs, owner-managers, and non-entrepreneurs on the various psychological measures used. The extent to which these differences impact the behavior of entrepreneurship remains to be seen. Future research will have to identify those psychological characteristics which predict not only who will be an entrepreneur, but also the type of venture selected and their likelihood of success.

The discriminant analysis showed that each of the scales, with the exception of the EQ, was about equally effective in discriminating entrepreneurs from owner-managers and non-entrepreneurs. Given near equal effectiveness, the scale of choice depends on what other research questions are being considered. The MBTI and HBDI might be best if the study deals with personality types or preferences in a context broader than just that of entrepreneurship. If the study being considered deals only with entrepreneurship, then perhaps the EAO is the best measure. If the goal is to minimize the number of non-entrepreneurs who are classified as entrepreneurs, then the EQ would be the scale of choice.

This study was not designed to discover if psychological characteristics make a difference in entrepreneurial success. With so many powerful influencing factors, even if psychological characteristics do play an important role, they could be easily swamped by the other factors (e.g. market pressure, national economy, funding availability, etc.). In spite of all this, if psychological characteristics do play even a small role in the expression of entrepreneurship, understanding the role they play needs to be a part of the general theory of entrepreneurship. This research supports the continued use of psychological scales in understanding and predicting entrepreneurship.

REFERENCES

Begley, T. M., & Boyd, D. P. (1987). Psychological characteristics associated with performance in entrepreneurial firms and smaller business. Journal of Business Venturing, 2, 79-93.

Boyd, D. P., & Gumpert, D. E. (1984). The loneliness of the start-up entrepreneur. In J. A.

Hornaday, F. Tarpley, Jr., J. A. Timmons, & K. H. Vesper, (Eds.), Frontiers of Entrepreneurship Research, 1984, pp. 478-487. Wellesley, MA: Babson College.

Brockhaus, R. H. (1982). The psychology of the entrepreneur. In C. A. Kent, D. L. Sexton, & K. H. Vesper, (Eds.), Encyclopedia of entrepreneurship, 39-56. Englewood Cliffs, NJ: Prentice Hall.

Brockhaus, R. H., & Horwitz, P. S. (1986). The psychology of the entrepreneur. In D. L. Sexton, & R. W. Smilor (Eds.), The art and science of entrepreneurship, 25-48. Cambridge, MA: Ballinger.

Dandridge, T. C., & Ford, M. A. (1987). Use of a self-assessment profile to differentiate among professional groups on entrepreneurial characteristics. In N. C. Churchill, J. A.

Hornaday, B. A. Kirchhoff, O. J. Krasner, & K. H. Vesper (Eds.), Frontiers of Entrepreneurship Research, 1987, 220-236. Wellesley, MA: Babson College.

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, 17, 22-29.

Druckman, D., & Bjork, R. A. (Eds.)(1991). In the mind's eye: Enhancing human performance. Washington, D. C.: National Academy Press.

Fagenson, E. A., & Marcus, E. C. (1991). Perceptions of the sex-role stereotypic characteristics of entrepreneurs: Women's evaluations. Entrepreneurship Theory and Practice, 15, 33-47.

Gartner, W. B. (1988). "Who is an entrepreneur?" is the wrong question. American Journal of Small Business, 12, 11-32.

Ginn, C. W., & Sexton, D. L. (1988). Psychological types of Inc. 500 founders and their spouses. Journal of Psychological Type, 16, 3-12.

Ginn, C. W., & Sexton, D. L. (1989). Growth: A vocational choice and psychological preference. In R. H. Brockhaus, Sr., N. C. Churchill, J. A. Katz, B. A. Kirchhoff, K.

H. Vesper, & W. E. Wetzel, Jr. (Eds.), Frontiers of Entrepreneurship Research, 1989, 1-12. Wellesley, MA: Center for Entrepreneurship Studies, Babson College.

Ginn, C. W., & Sexton, D. L. (1990). A comparison of the personality type dimensions of the 1987 Inc. 500 company founders/CEO's with those of slower growth firms as measured by the Myers-Briggs Type Indicator. Journal of Business Venturing, 5(3), 140-150.

Herrmann, N. (1988). The creative brain. Lake Lure, NC: Brain Books.

Ho, K. (1988). The dimensionality and occupational discriminating power of the Herrmann Brain Dominance Instrument. Dissertation Abstracts International, 49/06B, 2409. (University Microfilms No. 88-11716)

Hornaday, J. A., & Aboud, J. (1971). Characteristics of successful entrepreneurs. Personnel Psychology, 24, 141-153.

Hornaday, J. A., & Vesper, K. H. (1982). Entrepreneurial education and job satisfaction. In K. H. Vesper (Ed.), Frontiers of Entrepreneurship Research, 1982, 526-539. Wellesley, MA: Babson College.

Hoy, F., & Carland, J. W., Jr. (1983). Differentiating between entrepreneurs and small business owners in new venture formation. In J. A. Hornaday, J. A. Timmons, & K. H. Vesper, (Eds.), Frontiers of Entrepreneurship Research, 1983, pp. 157-166. Wellesley, MA: Babson College.

Hoy, F., & Hellriegel, D. (1982). The Killmann and Herden model for organizational effectiveness criteria for small business managers. Academy of Management Journal, 25, 308-322.

Hsu, L. M. (1989). Discriminant analysis: A comment. Journal of Counseling Psychology, 36, 244-247.

Huberty, C. J. (1984). Issues in the use and interpretation of discriminant analysis. Psychological Bulletin, 95, 156-171.

Hunt, H. K., Huefner, J. C., Voegele, C., & Robinson, P. B. (1989). The entrepreneurial consumer. In, G. E. Hills, R. W. LaForge, & B. J. Parker, (Eds.), Research at the Marketing/Entrepreneurship Interface, (pp. 175-184). Chicago: University of Illinois at Chicago.

Kao, J. J. (1991). The entrepreneur. Englewood Cliffs, NJ: Prentice Hall.

McClelland, D. (1961). The achieving society. Princeton, NJ: Van Nostrand.

McKay, R., & Campbell, N. (1982a). Variable selection techniques in discriminant analysis I: Description. British Journal of Mathematical and Statistical Psychology, 35, 1-29.

McKay, R., & Campbell, N. (1982b). Variable selection techniques in discriminant analysis II: Allocation. British Journal of Mathematical and Statistical Psychology, 35, 30-41.

Myers, I. B., & McCaulley, M. H. (1989). A Guide to the development and use of the Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists Press.

Northwestern Mutual Life Insurance Company (1985). What's your E.Q.? Milwaukee, WI.: Northwestern Mutual Life Insurance Company.

Robinson, P. B. (1987). Prediction of entrepreneurship based on an attitude consistency model. Dissertation Abstracts International, 48/09B, 2807. (University Microfilms No. 87-19822)

Robinson, P. B., Huefner, J. C., & Hunt, H. K. (1991). Entrepreneurial research on student subjects does not generalize to real world entrepreneurs. Journal of Small Business Management, 29, 42-50.

Robinson, P. B., Stimpson, D. V., Huefner, J. C., & Hunt, H. K. (1991). An attitude approach to the prediction of entrepreneurship. Entrepreneurship Theory and Practice, 15, 13-31.

Scherer, R. F., Brodzinski, J. D., & Wiebe, F. A. (1990). Entrepreneur career selection and gender: A socialization approach. Journal of Small Business Management, 28, 37-44.

Sexton, D. L., & Bowman, N. B. (1983). Comparative entrepreneurship characteristics of students: Preliminary results. In J. A. Hornaday, J. A. Timmons, & K. H. Vesper, (Eds.), Frontiers of Entrepreneurship Research, 1983, pp. 213-232. Wellesley, MA: Babson College.

Sexton, D. L. & Bowman, N. B. (1984). Personality inventory for potential entrepreneurs: Evaluation of a modified JPI/PRF-E test instrument. In J. A. Hornaday, F. Tarpley, Jr., J. A. Timmons, & K. H. Vesper, (Eds.), Frontiers of Entrepreneurship Research, 1984, pp. 513-528. Wellesley, MA: Babson College.

Sexton D. L., & Bowman, N. (1985). The entrepreneur: A capable executive and more. Journal of Business Venturing, 1, 129-140.

Sexton, D. L., & Bowman, N. B. (1986). Validation of a personality index: Comparative psychological characteristics analysis of female entrepreneurs, managers, entrepreneurial students and business students. In R. Ronstadt, J. A. Hornaday, R. Peterson, & K. H. Vesper (Eds.), Frontiers of Entrepreneurship Research, 1986, 40-51. Wellesley, MA: Babson College.

Sexton, D. L., & Bowman-Upton, N. (1990). Female and male entrepreneurs: Psychological characteristics and their role in gender-related discrimination. Journal of Business Venturing, 5, 29-36.

Singh, S. (1989). Personality characteristics, work values, and life styles of fast- and slow-progressing small-scale industrial entrepreneurs. The Journal of Social Psychology, 129, 801-805.

Timmons, J. A. (1985). New venture creation: A guide to entrepreneurship, 2nd ed. Homewood, IL: Irwin.

Wainer, H. A., & Rubin, I. M. (1969). Motivation of R&D entrepreneurs: Determinants of company success. Journal of Applied Psychology, 53, 178-184.

Webb, J. (1991). The gender relations of assessment. In J. Firth-Cozens & M. A. West (Eds.), Women at work: Psychological and organizational perspectives, pp. 13-25. Philadelphia: Open University Press.

Welsh, J. A., & White, J. F. (1981). Converging on the characteristics of entrepreneurs. In K. H. Vesper, (Ed.), Frontiers of Entrepreneurship Research, 1981, pp. 504-515. Wellesley, MA: Babson College.

Williams, E. E. (1981). Innovation, entrepreneurship, and brain functioning. In K. H. Vesper, (Ed.), Frontiers of Entrepreneurship Research, 1981, pp. 516-536. Wellesley, MA: Babson College.

Winslow, E. K., & Solomon, G. T. (1989). Further development of a descriptive profile of entrepreneurs. The Journal of Creative Behavior, 23, 149-161.

Wortman, M. S., Jr. (1986). A unified framework, research typologies, and research prospectuses for the interface between entrepreneurship and small business. In D. L. Sexton, & R. W. Smilor (Eds.), The art and science of entrepreneurship, pp. 273-331. Cambridge, MA: Ballinger.

Jonathan C. Huefner, Health Care International

H. Keith Hunt, Brigham Young University

Peter B. Robinson, University of Calgary

Table 1

Fs And Probabilities For The Sex,
Group, And Sex by Group Interaction Anovas and Manova

 Sex * Group
 Sex (1,316) Group (2,316) (2,316)

 F p F p F p

1) EQ 9.35 .0024 32.83 .0001 .19 .8279
2) EAO-SE 5.69 .0176 3.36 .0359 3.58 .0289
3) EAO-PC .36 .5500 3.55 .0298 .04 .9649
4) EAO-INN 6.28 .0127 9.69 .0001 1.15 .3180
5) EAO-ACH 3.22 .0737 1.05 .3523 .55 .5797
6) MBTI-EI <.01 .9866 2.85 .0595 .54 .5808
7) MBTI-SN .24 .6270 9.28 .0001 .64 .5300
8) MBTI-TF 3.19 .0750 1.23 .2945 .01 .9927
9) MBTI-JP .76 .3857 4.48 .0120 1.47 .2314
10) HBDI-Quadrant A 33.64 .0001 .56 .5705 .45 .6408
11) HBDI-Quadrant B 1.49 .2236 1.95 .1434 3.37 .0354
12) HBDI-Quadrant C 32.76 .0001 .07 .9356 .76 .4670
13) HBDI-Quadrant D .02 .9027 7.92 .0004 .94 .3929

MANOVA 5.91 .0001 2.86 .0001 1.16 .2629

 (df=13, 304) (df=26, 608) (df=26,608)

Table 2

Means, Standard Deviations, Fs, And Probabilities For The Sex
Variable for the Eq Scale and Eao, Mbti, and Hbdi Subscales

 Males Females

 Mean SD Mean SD

1) EQ 11.03 15.65 .65 16.50
2) EAO-SE 75.68 7.82 74.02 9.40
3) EAO-PC 69.46 9.40 69.79 10.14
4) EAO-INN 70.80 8.56 67.11 8.99
5) EAO-ACH 78.26 8.00 80.48 8.28
6) MBTI-EI -2.58 28.11 -3.38 26.61
7) MBTI-SN -2.10 27.29 -8.31 29.02
8) MBTI-TF 1.23 22.45 8.94 23.94
9) MBTI-JP -3.01 27.46 -9.85 26.96
10) HBDI-Quadrant A 73.89 22.98 54.97 21.00
11) HBDI-Quadrant B 73.17 19.19 75.80 17.66
12) HBDI-Quadrant C 60.13 20.04 76.60 21.08
13) HBDI-Quadrant D 65.71 24.85 63.21 21.68

 f(1,316) P-level

1) EQ 9.35 .0024
2) EAO-SE 5.69 .0176
3) EAO-PC .36 .5500
4) EAO-INN 6.28 .0127
5) EAO-ACH 3.22 .0737
6) MBTI-EI <.01 .9866
7) MBTI-SN .24 .6270
8) MBTI-TF 3.19 .0750
9) MBTI-JP .76 .3857
10) HBDI-Quadrant A 33.64 .0001
11) HBDI-Quadrant B 1.49 .2236
12) HBDI-Quadrant C 32.76 .0001
13) HBDI-Quadrant D .02 .9027

Table 3

Means, Standard Deviations, Fs, and Probabilities for the
Group Variable for the Eq Scale and Eao, Mbti, and Hbdi Subscales

 Owner-
 Entrepreneurs managers

 Mean SD Mean SD

1) EQ 18.24 13.23 -.63 14.96
2) EAO-SE 77.39 7.81 74.20 10.00
3) EAO-PC 71.67 8.82 69.16 8.68
4) EAO-INN 73.56 8.11 68.06 8.53
5) EAO-ACH 79.82 7.42 79.95 6.87
6) MBTI-EI -8.16 27.28 .11 28.95
7) MBTI-SN 5.05 27.12 -11.07 25.38
8) MBTI-TF -.01 23.58 3.52 23.93
9) MBTI-JP 2.82 28.01 -18.63 22.54
10) HBDI-Quadrant A 69.82 25.28 68.26 24.68
11) HBDI-Quadrant B 68.79 18.38 76.30 15.04
12) HBDI-Quadrant C 62.93 21.01 67.26 21.80
13) HBDI-Quadrant D 74.08 25.88 59.44 17.88

 Non-
 entrepreneurs

 Mean SD F(1,316) P-level

1) EQ 1.81 15.28 32.83 .0001
2) EAO-SE 73.76 8.19 3.36 .0359
3) EAO-PC 68.17 10.08 3.55 .0298
4) EAO-INN 67.16 8.46 9.69 .0001
5) EAO-ACH 78.20 8.75 1.05 .3523
6) MBTI-EI .43 27.22 2.85 .0595
7) MBTI-SN -9.28 27.37 9.28 .0001
8) MBTI-TF 6.21 22.54 1.23 .2945
9) MBTI-JP -8.61 26.31 4.48 .0120
10) HBDI-Quadrant A 66.61 23.04 .56 .5705
11) HBDI-Quadrant B 77.25 18.78 1.95 .1434
12) HBDI-Quadrant C 66.64 22.18 .07 .9356
13) HBDI-Quadrant D 59.41 21.27 7.92 .0004

Table 4

Means, Standard Deviations, Fs, and Probabilities for the Sex
by Group Interaction for the Eq Scale and for the Eao, Mbti,
and Hbdi Subscales

 Males Females

 Mean SD Mean SD

1) EQ Entrepreneurs 19.17 12.30 14.05 16.49
 Owner-managers 3.43 13.29 -5.00 15.92
 Non-entrepreneurs 4.56 15.25 -2.70 14.34
2) EAO-SE Entrepreneurs 77.73 7.44 75.86 9.35
 Owner-managers 78.11 9.20 69.98 9.37
 Non-entrepreneurs 73.49 7.44 74.20 9.33
3) EAO-PC Entrepreneurs 71.52 8.87 72.32 8.78
 Owner-managers 68.89 8.95 69.44 8.74
 Non-entrepreneurs 67.65 9.63 69.02 10.79
4) EAO-INN Entrepreneurs 74.36 8.20 69.94 6.78
 Owner-managers 70.13 4.50 65.84 11.20
 Non-entrepreneurs 67.61 8.04 66.42 9.11
5) EAO-ACH Entrepreneurs 79.65 7.51 80.61 7.11
 Owner-managers 78.69 5.76 81.31 7.91
 Non-entrepreneurs 76.94 8.51 80.28 8.80
6) MBTI-EI Entrepreneurs -7.61 27.85 -10.64 24.98
 Owner-managers -3.23 23.60 3.92 34.38
 Non-entrepreneurs 2.15 28.30 -2.39 25.30
7) MBTI-SN Entrepreneurs 4.48 27.27 7.64 26.94
 Owner-managers -9.14 26.11 -13.15 25.46
 Non-entrepreneurs -7.20 26.36 -12.67 28.83
8) MBTI-TF Entrepreneurs -1.24 23.03 5.55 25.78
 Owner-managers .71 21.38 6.54 26.95
 Non-entrepreneurs 3.57 22.00 10.55 22.91
9) MBTI-JP Entrepreneurs 5.14 27.46 -7.64 28.70
 Owner-managers -19.86 19.18 -17.31 26.43
 Non-entrepreneurs -8.30 26.23 -9.12 26.64
10) HBDI- Entrepreneurs 73.97 24.37 51.14 20.85
Quadrant A Owner-managers 78.93 24.59 56.77 19.73
 Non-entrepreneurs 73.16 21.56 55.89 21.44
11) HBDI- Entrepreneurs 67.09 18.62 76.46 15.42
Quadrant B Owner-managers 73.93 13.80 78.85 16.44
 Non-entrepreneurs 78.64 18.76 74.99 18.72
12) HBDI- Entrepreneurs 59.35 20.03 79.00 17.91
Quadrant C Owner-managers 56.57 21.15 78.77 16.36
 Non-entrepreneurs 61.31 20.01 75.38 22.94
13) HBDI- Entrepreneurs 75.06 26.29 69.68 24.02
Quadrant D Owner-managers 57.79 21.09 61.23 14.29
 Non- 58.17 20.87 61.44 21.92
 entrepreneurs

 F(1,316) P-level

1) EQ Entrepreneurs .19 .8279
 Owner-managers
 Non-entrepreneurs
2) EAO-SE Entrepreneurs 3.58 .0289
 Owner-managers
 Non-entrepreneurs
3) EAO-PC Entrepreneurs .04 .9649
 Owner-managers
 Non-entrepreneurs
4) EAO-INN Entrepreneurs 1.15 .3180
 Owner-managers
 Non-entrepreneurs
5) EAO-ACH Entrepreneurs .55 .5797
 Owner-managers
 Non-entrepreneurs
6) MBTI-EI Entrepreneurs .54 .5808
 Owner-managers
 Non-entrepreneurs
7) MBTI-SN Entrepreneurs .64 .5300
 Owner-managers
 Non-entrepreneurs
8) MBTI-TF Entrepreneurs .01 .9927
 Owner-managers
 Non-entrepreneurs
9) MBTI-JP Entrepreneurs 1.47 .2314
 Owner-managers
 Non-entrepreneurs
10) HBDI- Entrepreneurs .45 .6408
Quadrant A Owner-managers
 Non-entrepreneurs
11) HBDI- Entrepreneurs 3.38 .0354
Quadrant B Owner-managers
 Non-entrepreneurs
12) HBDI- Entrepreneurs .76 .4670
Quadrant C Owner-managers
 Non-entrepreneurs
13) HBDI- Entrepreneurs .94 .3929
Quadrant D Owner-managers
 Non-
 entrepreneurs

Table 5

Classification Table Results for Each of the Scales in Predicting
Group Membership

 Non-
Scale Actual Group Entrepreneur Owner-manager entrepreneur

EQ Entrepreneur 53.3% 13.1% 33.6%
 Owner-manager 3.7% 59.3% 37.0%
 Non-entrepreneur 12.6% 53.7% 33.7%
EAO Entrepreneur 60.6% 19.7% 19.7%
 Owner-manager 37.0% 11.1% 51.9%
 Non-entrepreneur 27.4% 35.4% 37.1%
MBTI Entrepreneur 57.7% 28.5% 13.9%
 Owner-manager 29.6% 33.3% 37.0%
 Non-entrepreneur 31.4% 40.0% 28.6%
HBDI Entrepreneur 58.4% 22.6% 19.0%
 Owner-manager 29.6% 22.2% 48.2%
 Non-entrepreneur 28.6% 36.6% 34.9%

Table 6

Classification Results for the Best Combination of Scales
and Subscales in Predicting Group Membership

 Owner- Non-
Scale Actual Group Entrepreneur manager entrepreneur

EQ, EAO Entrepreneur 61.3% 19.7% 19.0%
 Owner-manager 37.0% 14.8% 48.2%
 Non-entrepreneur 28.0% 33.7% 38.3%
EQ, EAO, Entrepreneur 62.8% 21.2% 16.1%
MBTI Owner-manager 29.6% 29.6% 40.7%
 Non-entrepreneur 26.9% 39.4% 33.7%
EAO, MBTI Entrepreneur 62.8% 21.9% 15.3%
 Owner-manager 33.3% 25.9% 40.7%
 Non-entrepreneur 26.3% 40.0% 33.7%
EQ, EAO, Entrepreneur 59.1% 21.9% 19.0%
MBTI, HBDI Owner-manager 33.3% 22.2% 44.4%
 Non-entrepreneur 24.6% 40.6% 34.9%
EQ, INN, EI Entrepreneur 67.2% 7.3% 25.6%
 Owner-manager 40.7% 14.8% 44.4%
 Non-entrepreneur 32.0% 18.3% 49.7%

Table 7
Summary Table of the Percent of Correct Hits and False
Hits for the Entrepreneur Group for the Classification
Results in Tables 5 Through 13

 Classified as Entrepreneur

Scale or Subscale Correct Hits False Hits

EQ 53.3% 8.1%
EAO 60.6% 32.2%
MBTI 57.7% 30.5%
HBDI 58.4% 29.1%
EQ/EAO 61.3% 32.5%
EAO/MBTI 62.8% 29.8%
EQ/EAO/MBTI 62.8% 28.2%
EQ/EAO/MBTI/HBDI 59.1% 28.9%
EQ/INN/EI 67.2% 36.4%