Evidential value of the success factors for entrepreneurs

(1)

Bachelor thesis

Evidential value of the success factors for entrepreneurs

Author: Keanu Tan

Student number: 10269800

Date: June 19th, 2014

Bachelor’s programme: Economics and Business Specialization: Business studies

(2)

1

Abstract

This study did test the evidential value of the non-psychological success factors for

entrepreneurs: education, experience, gender, ethnic group, entrepreneurial

experience of relatives and close friends, and initial financial capital, with help of the

tool p-curve. There has been found evidential value in the studies for education,

experience, gender, and initial financial capital. A lack of evidential value was found for

the factor entrepreneurial experience of relatives and close friends, and at a 10%

confidence interval it could be even argued that there has been intensely p-hacked.

For the success factor ethnic group, this study was inconclusive because of a too small

amount of p-values. Taken all the factors together in one test, did result in evidence for

the claim that the studies contained evidential value.

Acknowledgements

I would like to start this thesis by expressing my gratitude to the people who supported

me.

First of all I would like to thank prof. Dr. P.D. Koellinger for supervising this thesis

project.

Furthermore, I would like to thank my friends and family, with special reference to

Ananda Jhinkoe-Rai and Dolf Starreveld, who were of great supporting value during

my education career and this Bachelor thesis.

(3)

2

Section

Page number

Introduction

3 Literature review

5 Research design

9 Methodology



Study selection rules



P-curve selection rules



Search criteria

11

12

14 Data

15 Results

23 Discussion



Interpreting the results



Implications



Practical implications



Limitations



Suggestions for further research

30

31

31 Conclusion

33 Appendix

34 Reference list

37

(4)

3

Introduction

In the academic literature there has been done much research to the factors which differentiates the successful entrepreneurs form the less successful ones. Nevertheless, the question remains if these claimed effects of success factors do really exist or are artificial, as there are reports of academics manipulating the results of studies. Hwang Woo Suk, Marc Hauser and Diederik Stapel are examples of academics who were uncovered faking data and results (Bhattacharjee, 2013). Another bias of the academic literature is the “file-drawer problem”. This phenomenon is described by Rosenthal (1979). He stated, “The "file drawer problem", is the possible problem that journals are filled with the 5% of the studies that show Type I errors, while the file drawers back at the lab are filled with the 95% of the studies that show non-significant (e.g., p > .05) results” (p. 638). This can result in claimed effects which simply do not exist in the reality. Moreover, file-drawing happens also on a lower level, instead of not publishing studies, are analyses that do not produce significant results deleted from studies. Simonsohn, Nelson, and Simmons (2013) explained how this work, “While collecting and analyzing data, researchers have many decisions to make, including whether to collect more data, which outliers to exclude, which measure(s) to analyze, which covariates to use, etc. “ (p. 2). This combined with the findings of Kunda (1990), that researchers tend to be biased to find results confirming their own hypotheses, results in p-hacking (only publishing analyses that confirm the hypothesis). Simmons, Nelson, and Simonsohn (2011) found that selective reporting can easily create significant results for non-existing effects. In other words, not all the effects reported in journals are existing in reality.

Fortunately, have Simonsohn et al. (2013) developed the ‘p-curve’, a tool to test whether claimed effects in the literature are true effects or are only the result of selective reporting. Since, this tool is relative new it has not yet been used in all the fields of science. Until now were the claimed success factors for entrepreneurs left untested as well. Therefore, the objective of this thesis is to test the evidential value (rule out selective reporting) of, in the academic literature described, success factors for entrepreneurs. First, there needs to be made a decision of the scope of this thesis. Because of the limited time available to complete this thesis, it may not be feasible to study the evidential value of all the factors ever proposed in the literature. Therefore, the focus of this thesis will only be on the non-psychological characteristics of the entrepreneurs themselves. The factors which will be studied in this thesis will be mentioned now and discussed more extensively in the Literature review.

The starting point of the search to non-psychological success factors for entrepreneurs will be the book ‘A general theory of entrepreneurship: The individual-opportunity nexus’ of Shane (2003), a well-respected book in the field of entrepreneurship (more than 2000 times cited according to google scholar). According to this book, do non-psychological success factors for entrepreneurs include: marital status, education, experience, age, social position, and opportunity cost. To search for additional non-psychological success factors for entrepreneurs the database web of science is used with the searching criteria: ‘success’, ‘factors’, and ’entrepreneurs’, with the results sorted by times cited from highest to lowest. The first study in this list is the study of Cooper, Gimeno-Gascon, and Woo (1994). This study adds to the non-psychological factors: gender, ethnic group, parental background, partners (whether the entrepreneur starts with a team or alone), and initial financial capital. Robinson and Sexton (1994) provides an additional potential success factors namely (whether the entrepreneur has dependent) children.

These studies combined resulted in the claimed non-psychological success factors for entrepreneurs which can be found in table 1. Social position has been changed into social network, and parental background has been broaden to entrepreneurial experience of

(5)

4

relatives and close friends, to make the factors clearer and/or more comprehensive. To my knowledge this is a complete list of claimed non-psychological success factors for

entrepreneurs in the current academic literature. To keep the analyze manageable, not all these factors will be tested. However, to provide a complete overview of the current

knowledge of the non-psychological success factors for entrepreneurs all these factors will be discussed in the Literature review.

The remainder of the paper will be organized as follow: the Literature review will discuss the claimed success factors in more detail, in the Research design and Methodology part the tool ‘p-curve’ will be explained and rules for selection will be provided, the Data section will shortly discuss the included literature and presents the p-curve disclosure tables, the Results section will contain the p-curves, which will be examined in detail in the

Discussion, and this thesis will be completed with a conclusion at the end.

Table 1: List of non-psychological factors identified in the literature as of influence on entrepreneurial success.

Category Factors

Non-psychological factors  Marital status  Age

 Social network  Opportunity cost  Partners

 Children Non-psychological factors tested with

the p-curve

 Education  Experience  Gender  Ethnic group

 Entrepreneurial experience of relatives and close friends

(6)

5

Literature review

In this section the claimed non-psychological success factors for entrepreneurs will be discussed in more detail. These factors are: marital status, age, social network, opportunity cost, partners, children, education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital.

However, before discussing these factors it is of importance to provide the definition of an entrepreneur which will be used in this thesis. In the book ‘Successful entrepreneurship’ of Van Praag (2005) two definitions are used, “an entrepreneur is someone who indicates either that (s)he has started a business venture alone or with a group or that (s)he has acquired a (family) business, alone or with a group” and “an entrepreneur is someone who acknowledges that (s)he is being the sole owner of a corporation or is self-employed. Self-employed people do not derive main income from a wage or salary, but by exercising their profession or business on their own account and at their own risk” (p. 5). The self-employment part is taken often as a criterion for an entrepreneur, Therefore, the definition of an entrepreneur will be a combination of the above, in this thesis an entrepreneur is ‘someone who is an (co-)owner of a business or is self-employed’

Another important general definition if studying success factors for entrepreneurs is performance measurement. Murphy, Trailer and Hill (1996) did study the used performance measurements in the literature from 1987 until 1993, they found little consistency in performance measures across studies. Eight different performance dimensions were found: “efficiency, growth, profit, size, liquidity, success/failure, market share, and leverage” (p. 16). In total, 71 different measurements of performance were found (see table 2 in appendix), with most of them uncorrelated or even negatively correlated. Murphy et al. (1996) proposed to try, were possible, to include performance measurement of multiple dimension in studies. For the p-curve tool it is not of importance to use the same performance measurement. Therefore, will this study exclude none of the performance measurements which are used more than once, which enables this study to test the evidential value of the claimed success factors on the performance (in general) for entrepreneurs.

As the general definitions are discussed, each claimed non-psychological success factor for entrepreneurs can be discussed in some detail now.

Marital status

The reason that marital status can be of influence on the performance of entrepreneurs is because being married reduces uncertainty. Uncertainty is reduced in the personal life and the income of the wife or husband can be a buffer against a potential failure. Therefore, should married entrepreneurs be more willing to exploit entrepreneurial opportunities than single entrepreneurs (Shane, 2003, p. 67). The study of Brockhaus (1980) found similar results, a successful entrepreneur is more likely to be married than a less successful one. Brockhaus’ explanation for this is that married people do receive (personal and financial) support from their spouse.

There are also some studies which argue that marriage has a negative effect on the performance of female entrepreneurs. Hundley (2000) found a statistically significant negative effect of marriage on the earnings of female entrepreneurs, in contrast to a positive effect for male entrepreneurs. His explanation for these findings is that marriage gives women an extra burden in the form of more housework, were men get an extra incentive to work harder because of the larger demand for goods and services of the family which works motivating.

(7)

6 Age

The factor age is two-sided, initially an increase in age should be a success factor because entrepreneurs acquire the knowledge and skills necessary to be a successful entrepreneur with the years. However, at a certain point aging leads to uncertainty avoidance, which has a negative effect on performance (Shane, 2003). This inverted u-shaped age/performance graph is observed by various studies, one of these studies is the study of Preisendorfer and Voss (1990). They indeed found an inverted u-shaped age/performance relationship, confirming the initial positive and later negative effect of age on performance.

Social network

The social network of an entrepreneur could be a potential success factor as resources and information, of influence on performance, can be obtained from others. Therefore, it is argued that entrepreneurs with broader and more diverse social networks should perform better than entrepreneurs with smaller and less diverse networks (Shane, 2003). There can be made a difference in social network contacts between strong ties (relatives and close friends) and weak ties (contacts outside the regular group of contacts). In the section ‘entrepreneurial experience of relatives and close friends’ the focus will be on strong ties, where the focus of this factor is on weak ties. To illustrate, Hoang and Antoncic (2003) did study weak ties and found that the main measurement for weak ties is network size instead of specific characteristics of contacts or specific relationships with these contacts.

Opportunity cost

The time spent on being self-employed could be used alternatively also, for instance to work as an employee. Therefore, being an entrepreneur has an opportunity cost. To illustrate, having a higher income increases the costs and being unemployment lowers the costs. This implies that entrepreneurs with a higher opportunity cost should perform better than other entrepreneurs, as they need to compensate for the higher opportunity cost (Shane, 2003). This claim is indeed supported by various studies, which found that entrepreneurs with a higher opportunity cost did perform better than entrepreneurs with a lower opportunity cost (Cressy, 1996; Reid & Gavin, 1999).

Partners

It can also be argued that the skills and knowledge of two entrepreneurs together should be higher than that of a single entrepreneur. Cooper et al. (1994) confirmed this argumentation, ventures with partners had higher marginal survival and growth rates. They stated, “Benefits associated with the presence of partners include capital, functional expertise, and a broader range of management experience. There may also be benefits from the psychological support they can provide each other and from the lessened reliance upon a single entrepreneur’s drive and judgment” (p. 390). Furthermore, the study of Eisenhardt and Schoonhoven (1990) found a positive relationship between the numbers of founders and growth of the firm.

Children

Another factor, which is of influence on the performance of entrepreneurs is the presence of children. For males, having children should work stabilizing in their lives, which should improve their earnings. The contrasting effect is observed for females, the presence of children interrupts the women’s labor force participation, causing her human capital and productivity to decline and eventually also her earnings. This argument holds for the salary case and even twice as much for the self-employed (wo)men (Robinson & Sexton, 1994). The study of Lerner, Brush, and Hisrich (1997) found also supporting evidence, the age of children had a significant effect on the profitability of female entrepreneurs. They stated, “women with older children

(8)

7

have more time to devote to making their businesses successful, whereas a woman with younger children is expected to first fulfill her family responsibilities” (p. 333).

The previous discussed factors will not be tested with the p-curve and were only presented to provide a complete overview of the claimed non-psychological success factors. The reason for the exclusion of the previous factors is to ensure a certain degree of quality, including all the non-psychological success factors would make the amount of work unmanageable in the given time. Therefore, only the yet to be discussed success factors: education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital, will be tested with the p-curve. These factors will be discussed in more detail now. Education

Shane (2003) argued that, education should have a positive effect on the probability of becoming an entrepreneur and on the performance of an entrepreneur, as the knowledge and abilities obtained from education are of facilitating value. Shane (2003) stated, “people who have relevant information and skills should be more likely to exploit opportunities than people who lack these things” (p. 69). The study of Robinson and Sexton (1994) did indeed find supporting evidence for the claim that years of education increases the probability of becoming an entrepreneur and has a positive effect on the performance of an entrepreneur. Also, Jo and Lee (1996) found evidence for the positive impact of education on profitability. However, they could not find significant evidence of a relation between education and growth rates.

Experience

As previous explained can knowledge and skills be obtained from education. However, another source of knowledge and skills can be experience. There are four types of experience suggested in the literature which are of influence on performance: “general business experience, industry experience, start-up experience, and vicarious experience” (Shane, 2003). General business experience provides entrepreneurs with information and skills of the basic aspects of business, with industry experience is experience in the same industry meant, start-up experience is prior experience in being an entrepreneur, and vicarious experience is learning from experience from somebody else. All these forms of experience provide entrepreneurs with knowledge and skills which help exploiting opportunities successfully (Shane, 2003). Stuart and Abetti (1990) argued that education and age are also a form of experience.

All these forms of experience will be tested with this factor except for vicarious experience, education, and age because these forms of experience are divided in separate categories, under the success factors entrepreneurial experience of relatives and close friends, education, and age.

Gender

Cooper et al. (1994) argued that gender could have an effect on performance. They stated based on the study of Sexton and Robinson (1989): “we might expect women and minority entrepreneurs to have had fewer opportunities to develop relevant experience, to have fewer contacts who can provide assistance, and to have greater difficulty in assembling resources” (p. 376). Cooper et al. (1994) did indeed find evidence that female entrepreneur do perform less than male entrepreneurs. However, there are also studies arguing there is no performance difference between male or female entrepreneurs. For example, the study of Kalleberg and Leicht (1991) found that business of females were not more likely to fail, neither less successful, than business of males.

(9)

8 Ethnic group

With Ethnic group is meant that entrepreneurs from an ethnic minority are perhaps less successful than other entrepreneurs. The argumentation of Cooper et al. (1994) for this claim is the same as the argumentation for gender: it can be expected that ethnic minorities have fewer opportunities, fewer valuable contacts, and greater difficulty in gathering resources. Meyer (1990) studied entrepreneurship among the black population in the US (minority) and compared them to white entrepreneurs. He found big differences in performance, confirming Ethnic group as a potential success differentiator.

Entrepreneurial experience of relatives and close friends

As described at the experience section, one way how experience can influence the performance of an entrepreneur is learning from the experiences of others, for example entrepreneurial experience of parents. Therefore, it is expected that children whose parents were entrepreneur have greater entrepreneurial talent because they learned the necessary information and skills from their parents (Shane, 2003). The claim that entrepreneurs with self-employed parents are more likely to be successful than others, is confirmed by the study of Gimeno, Folta, Cooper, and Woo (1997).

Moreover, Davidsson and Honig (2003) argued that, not only parents with entrepreneurial experience are a source of entrepreneurial information and skills, also other family and close friends can be a valuable source.

Initial financial capital

With more money, many things do go a lot easier. This is also applicable to running a business. According to Cooper et al. (1994), should the probability of marginal survival and growth increase with the level of initial capital. They argued that the amount of starting capital influences the pursued strategy and can buy extra time to learn from and overcome the first problems. Duchesneau and Gartner (1990) found confirming evidence that a higher amount of initial capital does significantly improve the success of an entrepreneur. Furthermore, they argued that entrepreneurs entering a more mature market, compared to markets at the beginning of the life cycle, do need even more capital to become successful.

The text above did describe the claimed non-psychological success factors and the argumentation/evidence behind it. However, as previous mentioned not all the presented evidence should be taken for granted immediately. Therefore, the remainder of this paper will be dedicated to testing the evidential value of the factors: education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital.

(10)

9

Research design

To test the evidential value of the found success factors, the tool named “p-curve” will be used. The p-curve is a tool described by Simonsohn et al. (2013) which examines multiple studies reporting the same effect and uses the distribution of all the statistically significant p-values (p<0.05) of these studies. The statistical logic behind the p-curve is that when the Null is true (the effect does not exist), the p-values testing this effect should be uniformly distributed. To illustrate, the chance of getting a p-value, testing a non-existing effect, of less than 0.05 should be 5% and the chance of finding a p-value of more than 0.05 should be 95%. Only true effects will have a right-skewed p-value distribution. Meaning that the probability of getting a p-value, for an existing effect, lower than 0.025 is bigger than the probability of finding a p-value between 0.025 and 0.05. However, if there has been intensely p-hacked, there will be a peak around the 0.05 level. Implying that the p-value distribution will be left-skewed. This is expected because it is assumed that researches manipulating the data do not try to get an as low as possible p-value, but stop at the significance level (p=0.05). This will result in more p-values close to 0.05 (a left-skewed distribution) (Simonsohn et al., 2013).

To test whether the distribution is significantly skewed or not, the p-values are tested as if it are test statistics in two steps. First, is the p-value calculated of each p-value, “the probability of observing a significant p-value at least as extreme if the null were true” (p. 10). The second step is to test the skew using the Fisher’s method. The Fisher’s method uses a chi-squared test with a degrees of freedom as large as twice the number of p-values included in the p-curve.

Hereafter, another test is performed. For findings that are not significantly skewed, will the results be compared to the results of if the study was performed at a 33% power. A statistical power of 33% means that the study fails to draw the correct conclusion two of the three times. If the distribution is significant flatter than the 33%, it can be concluded that the effect lacks evidential value. If it is not significant flatter, the p-curve is inconclusive and more p-values are needed (Simonsohn et al., 2013, pp. 10-13).

When using the p-curve two aspects are of extremely importance: how to select the set of studies, and how to select the p-values to use. For selecting studies it is of importance to create and report rules for which studies to select and which not. Therefore, based on the discussion of the factors in the Literature review, rules will be made, for each factor to be analyzed, and reported in the Methodology section.

For selecting p-values Simonsohn et al. (2013) provided three criteria: “(1) test the hypothesis of interest, (2) have a uniform distribution under the null, and (3) be statistically independent of other p-values in p-curve” (p. 16). To meet these requirements Simonsohn et al. (2013) proposed to follow five steps: “Step 1) Identify researchers’ stated hypothesis and study design, Step 2) Identify the statistical result testing stated hypothesis, Step 3) Report the statistical results of interest, Step 4) Recompute the precise p-value(s) based on reported test statistics, Step 5) Report robustness results” (p. 32). Furthermore, they encourage the use of a p-curve disclosure table (table 3), which shows how the steps are implemented. Therefore, these steps will be followed and in the Data section the p-curve disclosure table for each factor will be presented.

Simonsohn et al. developed an online app, which does all the statistics. This app is available at: http://p-curve.com/. The input will be the findings of the studies to examine, and the output are the p-curve graph, the recalculated p-values, the pp-values, and the numbers which forms the p-curve graph. For this thesis the online app will be used as well.

(11)

10 Table 3: Example of a p-curve disclosure table (Simonsohn et al., 2013)

(12)

11

Methodology

The p-curve makes it possible to test the evidential value for each factor separate and of all the potential success factors together. However, as explained before it is of importance to develop and publish precise rules for selecting studies and p-values. Below are the rules for this study stated and enlightened.

Study selection rules

First, are the rules discussed which are of influence on the study selection of all the p-curve analyzes. Hereafter, are the factor specific rules formulated. All rules are based on the literature previous discussed in the Literature review.

General rules

1) All studies which examine the effect of one of the factors of interest on the performance of entrepreneurs will be included.

 ‘Factors of interest’ are the factors discussed in the Literature review (education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital).

 For the word ‘entrepreneur’ the following definition is used: ‘someone who is an (co-)owner of a business or is self-employed’

Education

2) All studies testing the relationship between the level of education and performance of an entrepreneur will be included.

Experience

3) Studies studying the effect of experience on the performance of entrepreneurs will be included in this study.

 With ‘experience’ is meant, general business experience, industry experience, and start-up experience.

 The forms vicarious learning, education, and age will be studied as separate success factors and will therefore not be included in this p-curve.

Gender

4) Studies which argue that gender is of influence on the performance of entrepreneurs will be used.

 There are also studies arguing there is no performance difference. However, if there is indeed no difference the p-curve should conclude that there is no evidential value.

Ethnic group

5) Studies testing the effect of being from a specific ethnic origin on the performance of the entrepreneur will be included in this analysis.

Entrepreneurial experience of relatives and close friends

6) All the studies testing the effect of having relatives or close friends with entrepreneurial experience on the performance of an entrepreneur, will be included in this p-curve. Initial financial capital

7) Studies providing evidence for the claim that the amount of initial financial capital has an effect on the performance of an entrepreneur will be included.

(13)

12 General p-curve

8) The studies included in the separate, single factor p-curves, will be included in the general p-curve as well.

 This general p-curve will test the evidential value of the non-psychological success factors: education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital, together.

P-value selection rules

General rules

a) The p-value, of a study meeting the selection rules, which controls for the most variables will be used.

 This is to get an as focused view as possible on the specific factors. b) Only one p-curve a sample can be selected

 One criterion of the p-curve is that the p-values are statistically independent of other p-values in the p-curve. Therefore, only one p-curve a sample can be selected. If multiple ‘performance’ measurements are used the priority of selecting will be as followed: First efficiency will be used (return on investment, return on equity, return on assets, return on net worth, gross revenues per employee, or average return on assets), followed by growth (change in sales, change in Employees, market share growth, change in net income margin, or change in CEO/owner compensation), profit (return on sales, net profit margin, gross profit margin, net profit level, net profit from operations, or pretax profit), size liquidity (sales level, cash flow level, number of employees, ability to fund growth, current ratio, or quick ratio), success/fail (discontinued business), market Share (respondent assessment), and as last will market share be used (debt to equity). This order is based on the frequency used found by the study of Murphy et al. (1996). If none of these performance measurements is used the study will be excluded.

c) Only p-values smaller than 0.05 will be included.

 The p-curve only works with p-values <0.05, including higher p-values will result in exclusion of these values, during the analyze, by the p-curve app.

d) If it is not possible to calculate the exact p-value because of a lack of information, the study will be excluded.

 For example: a paper stated a p-value of <0.05, but without other useful information, will be excluded.

Education

e) The factor controlling for the highest form of education will be used

 For example, if a study testing the effect of education uses the factors: obtained high school degree, obtained bachelor degree, and obtained master degree, to explain performance. Then, only the factor obtained master degree will be used in the p-curve.

Experience

f) If a study uses multiple experience measurements in one sample, only one p-curve will be used to ensure independence.

 The priority order will be: first start-up experience, followed by industry experience, and as last general business experience.

(14)

g) If a study (with one sample) is used for multiple p-curves, the priority of p-value selection for the general curve will be in order of sample size of the single factor p-curves, from low to high.

 To illustrate, assume that the study of Duchesneau and Gartner (1990) is used for the p-curves of the factors initial financial capital and gender. The p-curve of initial financial capital consists out of seven p-values and the p-curve of gender out of five p-values.

In this case, for the general p-curve, only the p-value of the factor gender will be used of the study Duchesneau and Gartner (1990) because the sample of the gender curve (five p-values) is smaller than the sample of the initial financial capital p-curve (seven p-p-values).  This is to fulfill the assumption of the p-curve that the p-values in a single p-curve are

independent of each other.

 Furthermore, will this selection rule ensure the contribution of each factor, even those with few studies. Without this rule, it would be possible that the general p-curve exists out of only a couple of factors with big samples, as all the studies testing the other factors (with smaller samples) can already be included under the factors with a big sample. As the purpose of the general p-curve is to test the evidential value of the success factors education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital, together, it is preferable to indeed include all these factors in the general p-curve.

(15)

14

Search criteria

The starting point of the search to studies to include in the p-curve analyzes will be the literature described in the Literature review. After including or excluding these studies, following the rules above, the references of these studies will be checked as well. In addition, specific and general search criteria (see table 3) will be used to find the final relevant studies. To complete the search, the references of all the included studies will be checked as well.

Table 3: Search criteria

General search criteria  ‘success’  ‘factors’  ‘entrepreneur’  ‘founder’  ‘income’  ‘duration’  ‘survival’  ‘performance’  ‘start up’  ‘self-employment’  ‘human capital’  ‘market entry’  ‘profit’  ‘new venture’ Education  ‘education’  ’schooling’  ‘human capital’ Experience  ‘experience’  ‘human capital’  ‘entrepreneurial  experience’  ‘start-up experience’  ‘business experience’ Gender  ‘male’  ‘female’  ‘gender’  ‘gender difference’

Ethnic group  ‘race’

 ‘ethnic group’  ‘minority’  ‘immigrants’  ‘Hispanic’  ‘black’  ‘white’  ‘Chinese’  ‘Afro-American’ Entrepreneurial experience of relatives and close friends

 ‘background’  ‘parents’  ‘relatives’  ‘spouse’  ‘friends’  ‘entrepreneurial experience’  ‘start-up experience’ Initial financial capital  ‘initial financial capital’  ‘initial investment’

(16)

15

Data

In this section the data included in the p-curves will be presented in the disclosure tables and shortly discussed. This will happen separately for each p-curve.

Education

The education sample is, with its fourteen p-values, the biggest of all the factors. All the studies found a positive relationship between education and the performance of entrepreneurs, implying that the p-curve will test if there is enough evidential value to conclude that education has a positive effect on the performance of an entrepreneur. As can be seen in table 4, most studies tested performance by profit. However, there is little difference in the results among the different performance measurements.

Table 4: p-curve disclosure table (education)

Original paper Quoted text from original paper indicating prediction of interest to researchers Study design Key statistical result

Quoted text from original paper with (statistical) results Results Bates (1990) Multivariate regression Regression coefficient (education vs. profit)

The most conventional measure-years of education-is the strongest human capital variable for identifying business continuance. t(4416)= 4.481 p=.00001 Bates (1995) Multivariate log-logistic regression Regression coefficient (education vs. survival)

The surviving firms active in late 1991 are disproportionately those headed by highly educated owners who worked full-time in the business

t(19445)= 7.6923 p=.00000 Boden Jr and Nucci (2000) Multivariate regression Regression coefficient (education vs. survival)

The two owner attributes that have positive, significant influences on men’s women’s businesses’ survival prospects are education and years of prior, paid work experience t(1797)= 3.06361 p=.00222 Borjas and Bronars (1989) (Blacks sample) Multivariate regression Regression coefficient (education vs. profit) t(992)= 2.75 p=.00607 Borjas and Bronars (1989) (Hispanics sample) Multivariate regression Regression coefficient (education vs. profit) t(1003)= 1.99 p=.04686 Borjas and Bronars (1989) (Whites sample) Multivariate regression Regression coefficient (education vs. profit) t(2036)= 4.36 p=.00001 Brüderl and Preisendörfer (1998) Bivariate probit model Regression coefficient (education vs. sales growth)

Founders with more years of schooling start less failure-prone businesses, but their firms do not have a stronger growth potential t(956)= -2.31 p=.02110 Brüderl, Preisendörfer, and Ziegler (1992)

We expect more schooling to improve a

firm's survival chances

Multivariate log-logistic regression Regression coefficient (education vs. survival)

More years of schooling and work experience of the founder significantly improve the survival chances of a new business t(1594)=-3.18 p=.00150 Cooper, Gimeno-Gascon, and Woo (1994) H1: Probabilities of marginal survival and growth increase with levels of education.

Multivariate regression Regression coefficient (education vs. survival)

Overall, performance appeared to be enhanced by level of education.

t(1041)= 4.04747 p=.00006 Gimeno, Folta, Cooper, and Woo (1997)

Human capital will be positively related to the economic performance of the venture Multivariate regression Regression coefficient (education vs. profit)

As expected education is positively related to the economic performance of the venture t(903)= 2.089 p=.03699 Meyer (1990) Multivariate regression Regression coefficient (education vs. profit)

Other interesting coefficients include a significantly higher return to education in self-employment than in wage and salary jobs,

t(1592)= 6.51092 p=.00000

(17)

16 (Continued) Table 4: p-curve disclosure table (education)

Quoted text from original paper with (statistical) results

Results

Robinson and Sexton (1994)

H3: The relationship between years of formal education and success of the self-employed will be positive and significant.

Multivariate regression Regression coefficient (education vs. profit)

The beta coefficients represent the slope of a vector representing the relationship between education and earnings, in each case the slope represents a significant positive relationship as tested using a t-test between the slope and a slope of 0 (All self-employed t = 23.67, male self-employed t = 19.14, female self-employed t = 6.48, all wage and salaried t = 90.22, male wage and salaried t = 75.2, and female wage and salaried t = 39.07, p < .0001 for all groups). t(21343)= 23.67 p=.00000 Taylor (1996) Multivariate regression Regression coefficient (education vs. profit) t(436)=2.46 p=.01428 Van Praag (2005) Multivariate regression Regression coefficient (education vs. number of employees) t(248)=9.98 p=.00000 Experience

With eleven p-values the experience sample is the second largest. Again, there is little difference observed in the different performance measurements. Moreover, all the studies except for the study of Jo and Lee (1996) found a positive effect of experience on entrepreneurial performance. Jo and Lee (1996) provided as explanation for this negative found result that experience can oppose change when change is actually required and experience could result in overconfidence. However, the mixed results imply that the p-curve will test if there is enough evidential value to claim that experience has an effect on performance.

Table 5: p-curve disclosure table (experience)

Quoted text from original paper with (statistical) results Results Brüderl and Preisendörfer (1998) Bivariate probit model Regression coefficient (industry experience vs. sales growth)

Industry-specific experience is the most important human capital variable; founders lacking this experience clearly have restricted prospects

t(956)= 2.49 p=.01294 Brüderl, Preisendörfer, and Ziegler (1992)

We expect work experience to show a decreasing payoff

Multivariate log-logistic regression Regression coefficient (industry experience vs. survival)

More years of schooling and work experience of the founder significantly improve the survival chances of a new business t(1594)= -3.53 p=.00043 Cooper, Gimeno-Gascon, and Woo (1994) H9: Probabilities of marginal survival and growth are higher for ventures similar to the previous organizations of the entrepreneur. Multivariate regression Regression coefficient (industry experience vs. growth)

Business similarity turned out to be a significant determinant of both marginal survival and growth.

t(1041)= 2.48724 p=.01303 Duchesneau and Gartner (1990)

We posit that successful entrepreneurs are likely to obtain a broad range of managerial experiences Factorial analysis of variance F-ratio (general business experience vs. success/failure)

Comparisons between successful and unsuccessful entrepreneurs indicated that successful entrepreneurs were more likely to have a broad range of previous managerial experience.

F(1,25)= 5.93 p=.02236

(18)

17 (Continued): Table 5: p-curve disclosure table (experience)

Quoted text from original paper with (statistical) results Results Dyke, Fischer, and Reuber (1992) (computer services sample) Multivariate regression Regression coefficient (start-up experience vs. profit)

Start-up experience was positively related to the profitability of food manufacturing and computer service firms r(96)=.46 p=.00000 Dyke, Fischer, and Reuber (1992) (food wholesale sample) Multivariate regression Regression coefficient (start-up experience vs. sales) r(64)=.30 p=.01439 Dyke, Fischer, and Reuber (1992) (food manufacturing sample) Multivariate regression Regression coefficient (general business and industry experience vs. profit)

Start-up experience was positively related to the profitability of food manufacturing and computer service firms r(55)=.68 p=.00000 Gimeno, Folta, Cooper, and Woo (1997)

Human capital will be positively related to the economic performance of the venture Multivariate regression Regression coefficient (start-up experience vs. profit) t(903)= 2.355 p=.01874 Lerner, Brush, and Hisrich (1997) H4d: The influence of related occupational experience will be positively related to business performance. Pearson correlations Regression coefficient (industry experience vs. profit)

The entrepreneur's involvement as a founder was significantly associated with gross revenues, whereas previous experience in the industry was similarly correlated (r = 0.28,p < .01) r(206)=.28 p=.00004 Jo and Lee (1996) Factorial analysis of variance F-ratio (start-up experience vs. ROE)

start-up experience has a negative impact on performance F(2,29)= 5.78 p=.00772 Robinson and Sexton (1994) H4: The relationship between experience and self- employment success will be positive and significant Multivariate regression Regression coefficient (general business experience vs. profit)

Hypothesis four (H4: Experience will have a similar relationship to self-employment as does education with a weaker overall impact) is supported in Tables 3, 4, and 5.

t(21343)= 12.21 p=.00000

Gender

This sample is small, consisting out of only five p-values. Probably this will be just enough to get significant results, however, a bigger sample could improve the reliability. One explanation for the small sample is that some studies do tests with a male sample and a female sample and discuss the differences between the two samples, but do not test these differences. Because of the small sample it is also not possible to say something about the difference between performance measurements. However, all five studies found a disadvantage for female entrepreneurs compared to male entrepreneurs. Therefore, will the p-curve test if being a male is a success factor for entrepreneurs.

(19)

18 Table 6: p-curve disclosure table (gender)

Quoted text from original paper with (statistical) results Results Brüderl and Preisendörfer (1998) Bivariate probit model Regression coefficient (female founder vs. sales growth)

Compared with businesses of male founders, businesses of female founders have a 2.5 percentage point higher probability of survival, a 15.2 percentage point lower probability of an employment growth, and a 12.9 percentage point lower probability of a sales growth (at the mean of all other covariates

t(956)= -3.08 p=.00213 Cooper, Gimeno-Gascon, and Woo (1994) H2: Probabilities of marginal survival and growth are lower for female entrepreneurs. Multivariate regression Regression coefficient (gender vs. growth)

Gender was significant only in the growth equation. t(1041)= 2.73589 p=.00633 Du Rietz and Henrekson (2000) Multivariate regression Regression coefficient (gender vs. sales)

We may conclude that our multivariate tests show female underperformance in the sales variables but not in any of the other three variables, profitability, employment and orders

t(1126)= 2.63629 p=.00850 Storey and Wynarczyk (1996) It is therefore possible to construct a gender variable female. A priori the direction of the impact upon firm survival is not specified

Multivariate regression Regression coefficient (gender vs. survival) t(167)= 2.01149 p=.04588 Van Praag (2005) Multivariate regression Regression coefficient (gender vs. number of employees) t(248)= -21.22 p=.00000 Ethnic group

The ethnic group category is the smallest sample in this study. With a sample of only four p-values this category alone is probably too small to get a significant answer form the p-curve. Furthermore, the results are also mixed about the claim if being from a minority origin is of positive or negative effect on performance. Therefore, will the p-curve test if there is enough evidential value to claim there is an effect between the ethnic origin of an entrepreneur and his performance.

Table 7: p-curve disclosure table (ethnic group)

Quoted text from original paper with (statistical) results Results Bates (1994) Factorial analysis of variance F-ratio (ethnic group vs. survival) F(1,1613)= 5.39 p=.02038 Cooper, Gimeno-Gascon, and Woo (1994) H3: Probabilities of marginal survival and growth are lower for minority entrepreneurs. Multivariate regression Regression coefficient (ethnic group vs. growth)

Racial minority was linked to lower probabilities of both marginal survival and growth. t(1041)= 2.06357 p=.03931 Fairlie and Meyer (1996) Multivariate regression Regression coefficient (immigrant vs. profit) t(14847)= -2.57736 p=.00997 Haganti, Watts, Chaganti, and Zimmerman-Treichel,(2008)

The presence of ethnic-immigrant member(s) in a new venture's founding team will be positively associated with venture performance Multivariate regression Regression coefficient (ethnic group vs. growth)

Ethnic-immigrants on founding teams, however, did make a significant

difference to new venture performance in founding teams that had relatively young members.

t(46)= 2.25 p=.02928

(20)

19 Entrepreneurial experience of relatives and close friends

This sample is of medium size with seven p-values. As can be seen in table 8 only in the furniture manufacturing sample of the study of Dyke, Fischer, and Reuber (1992) there was found a negative effect of entrepreneurial experience of relatives and close friends on entrepreneurial performance, compared to the positive effect found in the other six studies/samples. However, this still implies that there are mixed results. Therefore, will the p-curve test if there is evidential value for the claim that entrepreneurial experience of relatives and close friends has an effect on the performance of an entrepreneur.

Table 8: p-curve disclosure table (entrepreneurial exp. of relatives and close friends)

Quoted text from original paper with (statistical) results Results Cooper, Gimeno-Gascon, and Woo (1994) H4: Probabilities of marginal survival and growth are higher for entrepreneurs whose parents have owned a small business. Multivariate regression Regression coefficient (entrepreneurial parent vs. survival)

Having parents who had owned a business contributed to marginal survival, but not to growth.

We posit that successful entrepreneurs are likely to have entrepreneurial parents

Factorial analysis of variance F-ratio (entrepreneurial parent vs. success/failure)

Lead entrepreneurs in successful firms were more likely to have entrepreneurial parents. F(1,25)= 4.28 p=.04905 Dyke, Fischer, and Reuber (1992) (food manufacturing sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. growth number of employees)

For food manufacturing, there was a significant and positive relationship with growth in employees r(55)=.30 p=.02337 Dyke, Fischer, and Reuber (1992) (food retail sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. number of employees)

For two industries (food retail and food wholesale) there was a significant and positive relationship with the number of full-time employees, indicating that owners with

entrepreneurial parents had larger firms in these two industries

r(66)=.24 p=.04869 Dyke, Fischer, and Reuber (1992) (furniture manufacturing sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. sales)

However, for furniture manufacturers there was a significant and negative relationship with annual total sales.

r(75)=-.23 p=.04419

Gimeno, Folta, Cooper, and Woo (1997)

High psychic income from entrepreneurship should decrease the entrepreneur's threshold level of

performance.

As a result, higher levels of psychic income from entrepreneurship should be negatively related to the likelihood of exit. Multivariate regression Regression coefficient (entrepreneurial parent vs. survival)

It seems that entrepreneurs who are more intrinsically motivated and have a family history in entrepreneurship are simply more likely to accept a lower level of economic performance to remain in business. t(1514)= -2.143 p=.03227 Van Praag (2005) Multivariate regression Regression coefficient (entrepreneurial parent vs. number of employees)

We find that entrepreneurial talent is higher if an individual comes from an entrepreneurial family.

t(248)= 3.20 p=.00155

Initial financial capital

This sample as well, has seven p-values included. As can be seen in table 9 there is no observable difference between different performance measurements. Furthermore, all the studies found a positive relationship between initial financial capital and the performance of an entrepreneur, implying that the p-curve can test the evidential value of the positive effect of initial financial capital on performance.

(21)

20 Table 9: p-curve disclosure table (initial financial capital)

Quoted text from original paper with (statistical) results Results Bates (1995) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

The surviving firms began operations with greater owner financial capital investments t(19445)= 8.8 p=.00000 Boden Jr and Nucci (2000) (female sample) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

For the 1982 cohort, using $5,000 or more in start-up capital has a positive effect on the survival prospects of both men’s and women’s businesses

t(2164)= 2.40088 p=.01644 Boden Jr and Nucci (2000) (male sample) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

t(1797)= 2.52752 p=.01157 Brüderl and Preisendörfer (1998) Bivariate probit model Regression coefficient (initial capital vs. sales growth)

higher start-up capital improve the probability of success t(956)= 3.07 p=.00220 Cooper, Gimeno-Gascon, and Woo (1994) H10: Probabilities of marginal survival and growth increase with the level of initial capital. Multivariate regression Regression coefficient (initial capital vs. growth)

The level of capitalization also contributed to marginal survival and growth. t(1041)= 2.31528 p=.02079 Duchesneau and Gartner (1990)

Successful new ventures are likely to have started at higher levels of capitalization. Factorial analysis of variance F-ratio (initial capital vs. success/failure)

Higher levels of initial capital were clearly associated with firm success, with the mean investment of $123,000 for successful new ventures versus $54,000 for unsuccessful firms.

F(1,25)= 4.52 p=.04356 Gimeno, Folta, Cooper, and Woo (1997)

Initial capital investment may provide a liquidity buffer for the firm to survive under conditions of low performance. Multivariate regression Regression coefficient (initial capital vs. profit)

Initial capital had significant positive effects (two-tailed Wald test: p < .001) on economic performance, suggesting that better capitalized, larger, and older firms were better performers

t(1514)= 3.887 p=.00011

General P-curve

The general p-curve exists out of a total of 28 different p-values, six p-values from the category education, six p-values from experience, four from gender, four from ethnic group, five from entrepreneurial experience of relatives and close friends, and three from the category initial financial capital. However, as a result of the exclusion of p-values higher than 0.05 a lot of other studies could not be included, which is a limitation of the p-curve. In table 10 it can be seen that there is no notable difference between the different performance measurements. As this p-curve exists out of different success factors and mixed results the p-curve will test if there is evidential value for the success factors: education, experience, gender, ethnic group, entrepreneurial experience of relatives and close friends, and initial financial capital, together. This makes it possible to say something of the presence of p-hacking in the study to non-psychological success factors.

Table 10: p-curve disclosure table (general P-curve)

Quoted text from original paper with (statistical) results Results Bates (1990) Multivariate regression Regression coefficient (education vs. profit)

The most conventional measure-years of education-is the strongest human capital variable for identifying business continuance. t(4416)= 4.481 p=.00001 Bates (1994) Factorial analysis of variance F-ratio (ethnic group vs. survival) F(1,1613)= 5.39 p=.02038 Bates (1995) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

The surviving firms began operations with greater owner financial capital investments

t(19445)= 8.8 p=.00000

(22)

21 (Continued) Table 10: p-curve disclosure table (General P-curve)

Quoted text from original paper with (statistical) results Results Boden Jr and Nucci (2000) (female sample) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

t(2164)= 2.40088 p=.01644 Boden Jr and Nucci (2000) (male sample) Multivariate log-logistic regression Regression coefficient (initial capital vs. survival)

t(1797)= 2.52752 p=.01157 Borjas and Bronars (1989) (Blacks sample) Multivariate regression Regression coefficient (education vs. profit) t(992)= 2.75 p=.00607 Borjas and Bronars (1989) (Hispanics sample) Multivariate regression Regression coefficient (education vs. profit) t(1003)= 1.99 p=.04686 Borjas and Bronars (1989) (Whites sample) Multivariate regression Regression coefficient (education vs. profit) t(2036)= 4.36 p=.00001 Brüderl and Preisendörfer (1998) Bivariate probit model Regression coefficient (female founder vs. sales growth)

Compared with businesses of male founders, businesses of female founders have a 2.5 percentage point higher probability of survival, a 15.2 percentage point lower probability of an employment growth, and a 12.9 percentage point lower probability of a sales growth (at the mean of all other covariates

t(956)= -3.08 p=.00213 Brüderl, Preisendörfer, and Ziegler (1992)

We expect work experience to show a decreasing payoff

Multivariate log-logistic regression Regression coefficient (industry experience vs. survival)

More years of schooling and work experience of the founder significantly improve the survival chances of a new business t(1594)= -3.53 p=.00043 Cooper, Gimeno-Gascon, and Woo (1994) H3: Probabilities of marginal survival and growth are lower for minority entrepreneurs. Multivariate regression Regression coefficient (ethnic group vs. growth)

Racial minority was linked to lower probabilities of both marginal survival and growth. t(1041)= 2.06357 p=.03931 Du Rietz and Henrekson (2000) Multivariate regression Regression coefficient (gender vs. sales)

We may conclude that our multivariate tests show female underperformance in the sales variables but not in any of the other three variables, profitability, employment and orders

We posit that successful entrepreneurs are likely to have entrepreneurial parents

Factorial analysis of variance F-ratio (entrepreneurial parent vs. success/failure)

Lead entrepreneurs in successful firms were more likely to have entrepreneurial parents. F(1,25)= 4.28 p=.04905 Dyke, Fischer, and Reuber (1992) (computer services sample) Multivariate regression Regression coefficient (start-up experience vs. profit)

Start-up experience was positively related to the profitability of food manufacturing and computer service firms r(96)=.46 p=.00000 Dyke, Fischer, and Reuber (1992) (food manufacturing sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. growth number of employees)

For food manufacturing, there was a significant and positive relationship with growth in employees r(55)=.30 p=.02337 Dyke, Fischer, and Reuber (1992) (food retail sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. number of employees)

For two industries (food retail and food wholesale) there was a significant and positive relationship with the number of full-time employees, indicating that owners with

entrepreneurial parents had larger firms in these two industries

r(66)=.24 p=.04869 Dyke, Fischer, and Reuber (1992) (food wholesale sample) Multivariate regression Regression coefficient (start-up experience vs. sales) r(64)=.30 p=.01439

(23)

22 (Continued) Table 10: p-curve disclosure table (General P-curve)

Quoted text from original paper with (statistical) results Results Dyke, Fischer, and Reuber (1992) (furniture manufacturing sample) Multivariate regression Regression coefficient (entrepreneurial parent vs. sales)

However, for furniture manufacturers there was a significant and negative relationship with annual total sales.

r(75)=-.23 p=.04419 Fairlie and Meyer (1996) Multivariate regression Regression coefficient (immigrant vs. profit) t(14847)= -2.57736 p=.00997 Gimeno, Folta, Cooper, and Woo (1997)

High psychic income from entrepreneurship should decrease the entrepreneur's threshold level of

performance.

As a result, higher levels of psychic income from entrepreneurship should be negatively related to the likelihood of exit. Multivariate regression Regression coefficient (entrepreneurial parent vs. survival)

It seems that entrepreneurs who are more intrinsically motivated and have a family history in entrepreneurship are simply more likely to accept a lower level of economic performance to remain in business. t(1514)= -2.143 p=.03227 Haganti, Watts, Chaganti, and Zimmerman-Treichel,(2008)

The presence of ethnic-immigrant member(s) in a new venture's founding team will be positively associated with venture performance Multivariate regression Regression coefficient (ethnic group vs. growth)

Ethnic-immigrants on founding teams, however, did make a significant

difference to new venture performance in founding teams that had relatively young members. t(46)=2.25 p=.02928 Jo and Lee (1996) Factorial analysis of variance F-ratio (start-up experience vs. ROE)

start-up experience has a negative impact on performance F(2,29)= 5.78 p=.00772 Lerner, Brush, and Hisrich (1997) H4d: The influence of related occupational experience will be positively related to business performance. Pearson correlations Regression coefficient (industry experience vs. profit)

The entrepreneur's involvement as a founder was significantly associated with gross revenues, whereas previous experience in the industry was similarly correlated (r = 0.28,p < .01) r(206)=.28 p=.00004 Meyer (1990) Multivariate regression Regression coefficient (education vs. profit)

Other interesting coefficients include a significantly higher return to education in self-employment than in wage and salary jobs, t(1592)= 6.51092 p=.00000 Robinson and Sexton (1994) H4: The relationship between experience and self- employment success will be positive and significant Multivariate regression Regression coefficient (general business experience vs. profit)

Hypothesis four (H4: Experience will have a similar relationship to self-employment as does education with a weaker overall impact) is supported in Tables 3, 4, and 5. t(21343)= 12.21 p=.00000 Storey and Wynarczyk (1996) It is therefore possible to construct a gender variable female. A priori the direction of the impact upon firm survival is not specified

Multivariate regression Regression coefficient (gender vs. survival) t(167)= 2.01149 p=.04588 Taylor (1996) Multivariate regression Regression coefficient (education vs. profit) t(436)= 2.46 p=.01428 Van Praag (2005) Multivariate regression Regression coefficient (gender vs. number of employees) t(248)= -21.22 p=.00000

(24)

23

Results

In this section the p-curves will be presented, the information entered and the individual pp-value calculations can be found in the Appendix (table 11-17)

Education

As can be seen in graph 1, the data reveals that education is significant right-skewed (p=<.0001) and there are no significant results for the evidence being flatter than the 33% power (p=.9592) or being left-skewed (p=.9988). Furthermore, has 57% of the p-curves a value of 0.01.

(25)

24 Experience

Graph 2 shows that the evidence for the effect of experience on the performance of an entrepreneur is right-skewed (p=<.0001). 45% of the data has a p-value of 0.01 and 36% a value of 0.02. There is no evidence for the claim that the data distribution is flatter than the 33% power (p=.9839) or left-skewed (p=1).

(26)

25 Gender

In graph 3 below it is shown that the evidence of being a male as entrepreneur is a success factor, is significant right-skewed (p=.0094). 60% of the evidence has a p-value of 0.01 and 20% a value of 0.05. There is no evidence for the data being flatter than the 33% power (p=.5559) or being left-skewed (p=.8363).

(27)

26 Ethnic group

The p-curve of the ethnic group factor is not significant right-skewed (p=.5833), flatter than the 33% power (p=.2262), and neither significant left-skewed (p=.6095) (see graph 4). The evidence distribution is as follows: 25% p-value of 0.02, 50% of 0.03, and 25% has a p-value of 0.04.

(28)

27 Entrepreneurial experience of relatives and friend

Graph 5 shows the p-curve for this factor. The p-curve is not significant right-skewed (p=.3404) and only just not significant left-skewed (p=.0599). However, the p-curve is significant flatter than the 33% power (p=.0074). Furthermore, 43% of the p-values has a value of 0.05.

(29)

28 Initial financial capital

The p-curve showed in graph 6 does test the evidential value of the claim that initial financial capital has a positive relationship with the performance of an entrepreneur. As can be seen below the p-curve is significant right-skewed (p=.0042), not significant flatter than the 33% power (p=.6839), and neither significant left-skewed (p=.9481). Moreover, 58% of the p-values has a value of 0.01 or 0.02.

(30)

Graph 7 present the general p-curve, it shows that the p-curve is significant right-skewed (p<.0001). Furthermore, is the data distribution not significant flatter than the 33% power (p=.2269) and neither significant left-skewed (p=.8857). 39% of the p-values has a value of 0.01.

Evidential value of the success factors for entrepreneurs

Bachelor thesis