Does Entrepreneurial Education Matter for the Performance of Medium-Sized Venture Entrepreneurs?

Abstract

We explore the question of whether entrepreneurial education matters for medium-sized venture performance. We do so to better understand the conflicting evidence indicating that, while such education appears to have significant positive micro-level effects, it has no significant macro-level effects. The growing investment in entrepreneurial education has increased intentions and start-up rates in treatment effects studies but has yet to make an impact on national start-up or survival rates. To address the contradiction, we apply a different empirical approach—one based on the capabilities view—where we survey venture entrepreneurs about their firm’s performance and their education in order to determine whether their skills-enhancing entrepreneurial training is or is not a key driver of success. We find that while that training significantly increases their confidence, when taken as a whole—as a multidimensional measure—that entrepreneurial education does not significantly influence venture performance. We discuss the implications for research, policy, and education.

1. Introduction

It is difficult to overstate the importance of entrepreneurial activity in modern economies, let alone in the US (Reynolds et al., 2004), where over 650,000 new ventures start up each year (Kauffman Foundation—https://indicators.kauffman.org/ [accessed on 1 June 2024]). The vast majority of start-ups are small- and medium-sized enterprises (SMEs)—globally, these are the ventures that power the engines of employment, product and process innovation, poverty alleviation, market creation, and economic development (e.g., Avdullahi & Ademi, 2020; Malesios et al., 2018; Minai et al., 2018). Given that a ‘lack of education’ is one of the three major barriers to such activity (Rideout & Gray, 2013), entrepreneurship education and training (henceforth denoted as EE) has become a major focus of universities, governmental departments, and researchers (e.g., Arend et al., 2025; Kuratko, 2005; Rauch & Hulsink, 2015; Suguna et al., 2024).

For the purpose of this study, EE involves formal courses and training with content specific to entrepreneurship, delivered through a recognized school, college, or online program. It can involve ‘know what’ (e.g., general knowledge) and ‘know how’ (e.g., skills). Students of EE accrue several benefits—they have greater intensions to start businesses, actually start more businesses, and also grow them more (e.g., Dickson et al., 2008; Henry et al., 2005; Martin et al., 2013; Safranski, 2004). As a result, the US (and many other countries) have invested heavily in EE for decades, and at several levels of schooling, in order to increase the positive effects of SME start-ups (e.g., Brush et al., 2003; Greene et al., 2004; Huber et al., 2014; Katz, 2003; Peterman & Kennedy, 2003; Sánchez, 2013; Walter & Block, 2016). In the US, EE has exploded from 16 programs in 1970 to 5000 courses taught by 9000 faculty in 2008, and even with some colleges requiring all students to take at least one EE course in 2013 (e.g., Finkle et al., 2006; Katz, 2003; Kuratko, 2005; Lyons & Zhang, 2018; Safranski, 2004; Solomon, 2007). Further, the number of start-ups funded by EE has increased by twenty-fold in ten years (AngelList, 2015; CrunchBase, 2013), with about one-third of business incubators housed at universities.

Naturally, with the growth, size and economic importance of those investments in EE have come calls for more rigorous research into its effects (e.g., Davidsson et al., 2001; Haase & Lautenschlager, 2011; Rideout & Gray, 2013) in order to assuage doubts over its efficacy (e.g., Aronsson, 2004; Martin et al., 2013; Rae & Carswell, 2000; Weaver et al., 2006). Essentially, the calls for more research seek to answer the question of whether EE matters. To that stream of inquiry, we contribute with the analysis of the more focused question of whether EE matters to ME performance—a question that has not been answered1.

A primary logic underlying the investment in EE is that such training improves the skills of entrepreneurs-to-be, where they, in turn, produce more and better SME activity to improve the economy (e.g., Bhatia & Levina, 2020; Fayolle & Gailly, 2008; Minai et al., 2018). Our analysis focuses on what appears to be an empirical flaw in that logic—a flaw that appears when one compares the small and big pictures of EE impact over the last few decades in the US. At the small scale, considerable evidence exists that EE training at the individual level is significantly positively correlated with several measures of their entrepreneurial activity (e.g., Martin et al., 2013). However, even though since 2018, the number of people trained in entrepreneurship has doubled the number of MEs started up each year, at the large and longer scale, neither the start-up rate of MEs nor their survival rate appears to have improved. In fact, the US Bureau of Labor Statistics data indicates that ME start-up rates have fallen significantly in the last two decades (and more so for the largest firm sizes—i.e., of 50–499 employees) (e.g., Fairlie, 2022). Survival rates based on Canadian data imply no significant changes in the past decades of most MEs (i.e., those with 5–99 employees) either, hovering around a 10-year rate of 55 percent (https://ised-isde.canada.ca/site/sme-research-statistics/en/key-small-business-statistics/key-small-business-statistics-2023#s1.2—Key Small Business Statistics, 2023, from Innovation, Science and Economic Development Canada [accessed on 1 June 2024]). This is not a temporary trend; it covers a significant time lag between taking EE and starting a business. While the interest in entrepreneurial careers has dramatically increased over the recent decades, supported and stoked by explosive EE investment (Rideout & Gray, 2013), SMEs have consistently continued to suffer high failure rates due to weak performance (Machirori & Fatoki, 2013; Minai et al., 2018). In other words, even though the levels of EE investment and training have increased by magnitudes over the past four decades, the quantity and quality levels of MEs have not actually increased at all. Our focus is on investigating the latter outcome—on identifying whether and how EE, controlling for factors at the founder and firm levels, affects ME performance along several dimensions. The results of our investigation help explain this inconsistency between the small and big pictures.

Our study is different from most if not all, previous research on the effects of EE. We contribute to the entrepreneurship literature in three major but related ways, the first and most important being the application of a capabilities-based perspective to EE. Given our dependent variable is ME performance, we apply the standard approach in the capabilities view of trying to identify which firm factors contribute to their success on the margin (e.g., Barney, 1991; Collis, 1994). This retrospective approach differs from the vast majority of EE research that takes a prospective, treatment-effects approach, where the focus is on whether an EE experience (versus none) differs in its effects on subsequent entrepreneurial behaviors like start-up activity. Such a difference in perspective implies that the appropriate sample for our study is existing MEs (rather than being a set of individuals, some of whom had EE training and some of whom did not)2. Our second major contribution, based on the virtue of our perspective, is how we capture EE—we do so as a multidimensional influence rather than as a one-shot treatment. Because starting an ME entails buy-in at a non-trivial scale that requires entrepreneurs to prove a wide range of skills (Lazear, 2004, 2005) and because most entrepreneurs wait a significant time to start a venture after receiving training (e.g., Lange et al., 2014) and most entrepreneurs who receive EE training take more than one course (Nabi et al., 2017) and those courses can differ greatly by quality and method (Rideout & Gray, 2013) and some of the outcomes of EE training lie beyond the skills taught (e.g., such as provided by networking—Eesley & Lee, 2021), our assessment of EE spans several measures. The third contribution, relating to the first, involves the number of controls we have in our analysis to identify the marginal effects of the EE factors. While previous studies may have controlled for a handful of factors, mostly at the individual level (e.g., gender and age), we control for over ten relevant factors at both the individual and the firm levels (e.g., ME age, size, and its industry’s turbulence). We believe this is a more reasonable and objective way to verify the additional value of EE itself on ME performance.

Our empirical results indicate that the overall EE—as a multidimensional collection of measures—has no significant positive effect on a variety of ME performance outcomes when individual and firm-level factors are controlled for. That being noted, we did find a significant positive effect on confidence-related measures, such as on the entrepreneur’s self-confidence and on her expectation of her ME’s future business success. Our main result is consistent with the big picture issue for EE impact (e.g., a lack of recognizable change of failure rates at the national level), while our analysis provides a story that is consistent with the small picture evidence—a story about EE providing its students with an option value that is often enough exercised to show significance in past treatment-based studies on entrepreneurial activities. In the remainder of the paper, we review the relevant literature, argue our hypotheses, explain our empirical methods and sample, describe our results, and then discuss their implications for EE and its stakeholders.

2. Literature Review

It is no surprise that the question of EE’s effectiveness has been an active subject of research—there were 79 relevant studies published prior to 2013 alone. However, such research has itself been questioned. Many studies have been shown to be flawed due to biases in their treatment effects analyses (e.g., all but six of 79 studies failed to use randomized assignment for their treatment and control groups—Martin et al., 2013) and due to other methodological issues (Rideout & Gray, 2013; Rauch & Hulsink, 2015; Safranski, 2004; Shadish et al., 2001). Those flaws have caused some researchers to conclude that there is little evidence of EE’s effectiveness (Fiet, 2000; Rideout & Gray, 2013; Weaver et al., 2006). Such concerns are based not only on methodological issues but also due to the focus of most studies on short-term, subjective impact measures such as intentions and start-up activities rather than on long-term measures such as firm performance. From reviews of the literature, a call has arisen for future research to address that gap in focus (e.g., Garavan & O’Cinneide, 1994; Henry et al., 2005; Nabi et al., 2017; Pittaway & Cope, 2007). Here, we attempt to answer that call with a focus on the performance of MEs—MEs being the firms of most interest to policymakers because they start off with a positive employment impact and because they grow more than SEs.

Historically, the two fundamental questions linked to EE are Can it be taught? and Is it necessary? (Kozlinska, 2011). To the first question, while disagreement continues, the majority of studies affirm that entrepreneurship skills can be taught (e.g., Mayhew et al., 2012). EE generates entrepreneurial competencies (Minai et al., 2018)—the kind that is supposed to improve firm performance. Further, we expect that the students gaining those skills will likely become entrepreneurs given that Career Theory explains that individuals will enter specific professions when they have the related education (e.g., Rideout & Gray, 2013). Agency Theory also supports EE in its ability to cultivate agentic capabilities in its students, who then are more likely to pursue a wider array of opportunities and become more successful in realizing desired futures (e.g., Bandura, 2006; Schunk & Zimmerman, 1998). Recent EE work reveals the role of education in building entrepreneurial skills and efficacy (Chatterji et al., 2019; Eller et al., 2021; Galvão et al., 2020; Kotha et al., 2019; Lyons & Zhang, 2018; Sarasvathy, 2001). Because such skills have been regarded as a key factor for success in business and innovation, EE programs and incubators have exploded in prevalence over the past decades (Lyons & Zhang, 2018; Rigby & Ramlogan, 2016).

The question of teaching entrepreneurship may not simply stop at whether specific skills can be learned; EE is also often expected to provide the confidence, goal-setting, motivation, and self-efficacy that help ensure that those skills are actually applied (e.g., McClelland, 1961; Rideout & Gray, 2013). The Theory of Planned Behavior (e.g., Ajzen, 1991; Krueger & Carsrud, 1993) hypothesizes that intentions precede behaviors, where the former have been proven to be heightened by EE (Rideout & Gray, 2013). In fact, most models of EE’s impacts include a path to outcomes being paved by improved skills and confidence (e.g., Erikson, 2003; Rae & Carswell, 2000).

To the second question of EE’s necessity for entrepreneurial success, existing evidence points to possible sufficiency as the best case. The review of 159 published articles from 2004 to 2016 indicated that EE has at least a short-term positive impact on select success measures (although most of those articles involve greatly under-described EE pedagogies and severely limited controls for selection effects—by Nabi et al., 2017). An earlier meta-analysis, based on 42 independent samples (N > 16,000), found that EE was significantly and positively related to entrepreneurship outcomes, with a correlation above 0.15 (Martin et al., 2013). Start-up, survival, and performance of ventures are all driven by the skills of entrepreneurs—skills that can come from EE exposure (Minai et al., 2018; Tehseen & Ramayah, 2015). While such evidence supports the argument that EE can be sufficient for positive venture outcomes, there is no support for its necessity. That is because of the existence of substitutes like real experience acting as viable means to gain skills and confidence (e.g., as a protégé to a mentor entrepreneur, from working at a startup alongside an industry-savvy team, or from having to start-up one’s own business due to economic necessity). Experience-as-an-EE-substitute forms the basis for the Theil Scholarship and for many press-friendly stories of tech-billionaire college dropouts.

3. Hypotheses

We now consider arguments about whether EE should matter to ME performance. Our main hypothesis concerns the expected impact of the ME CEO’s EE experience on their venture’s performance. We consider the arguments both for and against a positive impact because there are two schools on the subject—one that affirms that EE can effectively develop relevant, applicable skills and one that denies it (e.g., Mayhew et al., 2016).

3.1. The Affirmative School

Grounded in a capabilities-based approach, the answer is that EE does matter, yes, because better factors are supposed to lead to better organizational performance (e.g., Barney, 1991). Given here that a set of those factors—managerial skills—are made better through EE, then EE should lead to better organizational performance, even when controlling for the differences in other relevant factors at the individual manager and organizational levels. And, that should be even easier to determine when a fuller assessment of the EE exposure (e.g., measuring both its quantity [in the number of courses taken] and quality [such as direct entrepreneurial experience]) is accounted for through a multi-dimensional measure.

Previous research (e.g., Eller et al., 2021; Galvão et al., 2020; Martin et al., 2013; Mayhew et al., 2012) and extensive reviews of the literature (e.g., Gorman et al., 1997; Kuratko, 2005; Pittaway & Cope, 2007; Weaver et al., 2006) support this affirmative school (Mayhew et al., 2016). The logic underlying those findings is rooted in the training literature (Bell & Kozlowski, 2008) and grounded in human capital theory (Becker, 1964; Mincer, 1958)—a theory drawn upon in the entrepreneurship literature to explain aspects of entrepreneurial success (e.g., Martin et al., 2013; Pfeffer, 1994). As a process involving a set of related skills (e.g., in opportunity recognition and resource acquisition and use—Shane, 2003), entrepreneurship theoretically can be cultivated through EE experiences (e.g., Gorman et al., 1997; Katz, 2003; Pittaway & Cope, 2007; Martin et al., 2013; Mayhew et al., 2016). The affirmative school is also supported by translating the past evidence of EE’s positive significant effects on intentions, skills improvements, and self-efficacy—through the Theory of Planned Behavior, Agency Theory and Career Theory—into the expectation that those factors will translate into positive, sustained and successful entrepreneurial activity by students. Further, it is even expected that such training has a reasonable shelf-life so that if a would-be entrepreneur were to wait years to strike at the right opportunity, then EE would still have a positive impact (Kolvereid & Moen, 1997).

3.2. The Dissenting School

The dissenting school argues that EE does not matter to ME performance. Scholars in this school cast doubt on the past evidence of the positive linkage between EE and venture success. They also raise difficult questions about the limits of EE. These scholars believe that the available research into the impact of EE is of insufficient quantity, quality, and consistency to prove its effectiveness in creating more and better entrepreneurs (e.g., Cox et al., 2002; Davidsson, 2015; Martin et al., 2013). Further, healthy skepticism exists over whether one can learn entrepreneurship in a classroom setting (e.g., Bhatia & Levina, 2020; Kuratko, 2004). It is difficult to imagine how one can teach effective ways to address specific internal politics, build trust among a specific set of stakeholders, increase alertness, build a jack-of-all-trades expertise, make decisions under ambiguity, be lean yet not unsafe, fake it until you make it (without being dishonest), or beg, borrow, and steal needed resources (legally). The reality that many of these key lessons for entrepreneurial success can only be taught out of the classroom has been exposed in stories of billionaires—in the Internet boom and beyond—who had little, if any, formal EE training (Mayhew et al., 2016).

3.3. Finding a Synthesis of the Opposing Schools in Real Options Thinking

A synthesis of these opposing schools is offered to reconcile their opposing logic, arguments, and evidence. It is a logic that, on the one hand, recognizes the underlying potential value of EE training while, on the other hand, also recognizes the constraints and conditionals of realizing that value in any specific application to an ME. We find that synthesis in a real options-based perspective (which is a new argument to the literature on EE effectiveness).

Our theoretical contribution lies in extending the capabilities perspective—the approach that drives our empirical investigation—by considering the underlying real option value of EE-based capabilities, specifically in how that value is or is not realized when the option is exercised in a given context. The investment in EE generates a valuable real option (e.g., McGrath, 1999) for the entrepreneur who then ‘owns’ the capability but does not have the obligation to start up and manage a new venture (such as an ME). However, like most real options, the exercised value is often much lower than the holding value. Not only are the exercise costs usually higher-than-expected (e.g., in terms of start-up and foregone opportunity costs), but the time-to-exercise and the upper bound to the market are often lower than expected due to reactions by imitators and defending incumbents (e.g., Copeland & Tufano, 2004; Van Putten & MacMillan, 2004).

Taking on this real options perspective shifts the question debated by the two schools—about whether EE has value connected to ME performance—to the question of whether that value is often enough realized in its application. Because of the capabilities-based logic and the associated evidence from treatment-based effects research, we must agree with the affirmative school that EE could have a positive significant effect on ME performance. Simply put, EE does provide the relevant skills and self-efficacy to its students that are applicable. However, we must also recognize the concerns of the dissenting school regarding the realization of that effect. First, we must recognize that there are constraints on the skills and lessons that EE can provide—constraints that affect the potential positive impact of EE. Second, we must recognize that there are contextual conditions that may significantly limit the positive impacts of those skills in specific applications (i.e., to a single firm, with a single idea, at a single time and place). We now consider those caveats that limit EE’s exercised option value.

EE is constrained by many factors. Consider the following issues involved in teaching entrepreneurship in an institutional setting: Content must be taught widely versus narrowly, given the variety of students’ interests and the level of generality of most relevant theories and frameworks. The intensity and complexity of the personal stakes involved, of the pressures applied, of the politics or legal or financial concerns must all be tamped down for student safety in any educational experience. The prescription of ‘marginal’ business behaviors (e.g., the fake it ‘til you make it attitude, various corner-cutting tactics, and creative accounting) that rivals and others are likely to use is disallowed in accredited institutions. And, no EE program alone provides the training in the underlying venture idea’s skill (e.g., in coding, engineering, science, design, and crafting to generate that ‘ten-thousand-hours-level’ expertise that is often required to generate a novel and commercializable idea—Gladwell, 2008).

These constraints imply that the value of EE as a real option is likely to be low when exercised in any specific ME. This is because those EE capabilities—taught as general, unpressured, ideal, loosely linked skills—are likely to lose much of their impact when applied to a specific idea, in a specific industry, in a specific market segment at a specific time, under specific start-up resource constraints, under real pressures, in a tightly-linked structure at a specific ME. When the application context differs from the learned context there must be a loss of a skill’s value, at a minimum, until adjustments can be made (e.g., McKeachie, 1987). When that application context not only differs in specificity and pressure but involves greater uncertainty and dynamics as well, that degradation will be significant (e.g., Comfort & Wukich, 2013; Moynihan, 2008). Added to these concerns is the likelihood that the EE training was incomplete—in that critical gaps, often concerning tacit skills (e.g., in leadership, sales, hiring, negotiation, pivoting, reading rival moves), would have gone unfilled, rendering any of the learned skills much less impactful (e.g., Jackson, 2012; Jackson & Chapman, 2012).

In sum, the evidence and logic on EE can be supported by a real options-based capabilities story that can predict both an affirmative and a dissenting outcome. On the one hand, the capabilities-based logic and the historical evidence of treatment-based research predicts that EE will increase general skills related to the entrepreneurial process—that EE has a high potential option value to students—and so could positively affect ME performance. On the other hand, there is insufficient evidence that, once that option value is exercised in a specific application, EE must result in better performance (e.g., due to mismatches in capabilities and opportunities—Williamson, 1999). The real options logic strongly suggests that the value of general skills will be severely degraded due to a myriad of factors, including possible gaps in necessary skills (e.g., concerning the underlying technology of the venture and the ability to apply general lessons to the specifics of a new and uncertain opportunity); the competitive reactions of others with similar, better, or more specialized skills; a real-life lack of information for the idealized decision-making processes taught; a lack of resources and time; and the costly internal and market frictions that commonly exist with new endeavors. Thus, our first (competing) hypotheses are as follow:

H1a. 

EE will have a significant positive impact on ME performance.

H1b. 

EE will not have a significant impact on ME performance.

EE not only builds skills but also builds confidence. The training in a skill includes becoming more familiar with its history, value, and general use. Such training often includes the application of that skill but almost always in an artificial setting (e.g., where any possible failure is limited, where information and feedback are timely, constructive, and mostly costless, and where a positive outcome is generally expected). EE builds the confidence needed for start-ups to occur, or as McMullen and Shepherd (2006, p. 141) note, “[…] through learning efforts and/or the encouragement of others, one may be able to overcome doubt and act [on a given opportunity]”. Research shows that EE increases positive perceptions (e.g., of feasibility and desirability) of entrepreneurial activities (Athayde, 2009; Souitaris et al., 2007) and that such increased feasibility is only possible with increased confidence in personal skills.

An undisputed goal of education, let alone EE, is to increase a student’s self-efficacy (e.g., Shinnar et al., 2014; Wilson et al., 2007). That growth in confidence results from the positive experience that the student should experience in using new skills with increasing success in pseudo-direct, indirect, and social contexts. In EE, the pseudo-direct applications are through simulated or constructed, low-risk, often-isolated real-world activities; the indirect applications are through cases, mentorships, guest lectures, limited consulting, and possibly some valuation and low-risk investment exercises, and the social interactions occur through discussions with other students and by partaking in business plan competitions. All are experienced in controlled environments supplemented by constructive, timely feedback.

We expect that the positive impact of EE on student confidence to not only apply to the entrepreneur’s skills and self-image—of her past and present image—but also to project onto her current venture as well (e.g., Cassar, 2010; Von Bergen & Bressler, 2011). Obviously, if she were not confident in her skills relating to the current venture, or worse, if she were not confident in the future of that venture, she would not rationally be working on it in the present. Our second hypothesis follows:

H2. 

EE will have a positive and significant impact on the entrepreneur’s confidence in herself and her venture’s future.

4. Empirical Methods

4.1. Sample and Procedures

In the tradition of the capability view’s empirical testing approach (Barney & Mackey, 2005), we sample from a population of firms of interest, recording from each firm several measures of specific factors and relevant performance outcomes for analysis. The focal factors are regressed, with controls, on performance in order to determine whether significant correlation is present and in the expected direction. Here, the sample of firms are MEs, the focal factors involve the multiple dimensions of EE, the controls are at the individual and venture levels, the performance outcomes span several perceptions of relative success, and the regressions are hierarchical in nature.

Our sample data come from an online survey of compensated participants using Amazon MTurk’s marketplace. Filters were used to access the targeted population of ventures (i.e., sized between 5 and 500 full-time equivalent employees [FTEs] and located in the US) and respondents (i.e., the entrepreneur managing the ME). The survey was conducted online through Qualtrics in mid-2021. That approach resulted in 213 responses from a cross-section of industries and demographics, with only a handful of missing or inconsistent data, which provided 197 complete records. This approach is consistent with the majority of related studies that use cross-sectional data (e.g., in the capabilities view research and in the entrepreneurial activity research, including that involving EE), and especially when surveys are employed (e.g., Nabi et al., 2017; Padilla-Angulo et al., 2021).

Among the 213 entrepreneurs in our sample, 60.1% were male, the average age was 39.5, average work and entrepreneurship experience was 15 years and 9 years, respectively, most with a business education degree. On average, participants had two prior start-up experiences, had taken over 5 entrepreneurship-specific courses, and were involved in a firm with an average of over 130 FTEs and an average age of over 14 years.

We used the extrapolation technique (Armstrong & Overton, 1977) to test for non-response bias, which is commonly adopted in the literature (e.g., Engelen et al., 2015; Simsek et al., 2007). Based on the assumption that late respondents could be similar to non-respondents, we split the data into two groups and used t-tests to compare the mean values of main constructs between early and late respondents. We found no significant differences between the mean values, indicating that non-response bias was not a threat to our study.

To address the problems related to common method bias, we took several recommended precautionary measures. To mitigate the biasing effects, we ensured anonymity to the participants and informed them that their answers would be strictly confidential and that the data would only be published in aggregate form in an academic journal; we encouraged objectivity by indicating that there are no right or wrong answers to the questions; we separated the questions capturing independent and dependent variables to reduce perception of any linkages between the constructs; and we carefully worded survey items and, whenever available, used established validated measures (Johnson et al., 2011; Podsakoff et al., 2012).

We also used the correlational marker-variable technique (based on a variable theoretically unrelated to other substantive variables) to detect common method bias (Lindell & Whitney, 2001; Richardson et al., 2009). There were no significant correlations between the marker variable (i.e., a variable measuring whether seeking out information was the primary way the venture deals with uncertainty—a variable that was not directly related to the research questions at hand) and the other substantive variables in the study that were subject to common method bias. Moreover, there were no significant changes in the partial correlation coefficients and significance levels between the independent and dependent variables when the marker variable was controlled for. Based on these results, we concluded that common method bias was not a significant issue for this study.

We chose the empirical approach used in most capabilities-based research because we were focusing primarily on the potential drivers of venture performance. Our methodological choice was intended to address a growing call for complementary approaches that target practicing entrepreneurs as the unit of analysis (e.g., Mayhew et al., 2016) in a research stream dominated by treatment effects studies. We are not the first to answer that call with a sample limited to entrepreneurs (e.g., Lange et al., 2014; Parker & van Praag, 2006), but we appear to be the first to focus on MEs.

Relating to the previous EE studies, two empirical issues have been raised in their reviews. The first is a lack of controls, especially at the individual level, that would help reduce the selection effects for those more inclined to choose EE (e.g., Martin et al., 2013; Nabi et al., 2017; Rideout & Gray, 2013). The second is a lack of robustness in the EE measure (e.g., by failing to assess issues of quality). Our empirical contribution to this study is to address both of these issues. We control for over twenty-five individual-level and firm-level factors that have been associated with entrepreneurial outcomes in past research in order to better isolate the marginal effects of EE. Moreover, we assess a much richer impact of EE across multiple dimensions, including measures of quantity, width, intensity, length, and even harm. That array of characteristics helps capture the full weight of EE’s effect on the ME entrepreneur, helps ensure a participant’s better recollection of it given the necessarily retrospective survey approach used (i.e., through the use of the multi-point prompting), and is most appropriate for these respondents (many of whom have had several EE experiences over a range of areas).

4.2. Dependent Variables

To test our two hypotheses, we consider seven dependent variables—four measures of ME performance and three measures of confidence: Performance covers general success, growth, COVID-era revenue change, and innovation. BIZSUCC measures the respondent’s agreement with the statement, “My current business has been very successful”. (7-point Likert scale from strongly disagree to strongly agree). LoGrHiGr measures the respondent’s perception of the venture’s growth (7-point Likert scale from low-growth to high-growth). CovidPerf measures the respondent’s answer to the question over, “The amount that your firm’s monthly revenue been affected by the COVID-19 pandemic?” (selection from nine levels increased by 76–100%, 51–75%, 26–50%, 1–25% to stayed the same to decreased by 76–100%, 51–75%, 26–50%, 1–25%; where midpoints are used as the data points). IP5YR measures the respondent’s answer to the question, “Has your organization produced any patents or other valuable and innovative intellectual property in the last 5 years?” (captured as yes = 1, or no = 0).

The measures of confidence included personal and projective assessments: ENTSUCC measures the respondent’s agreement with the statement “Overall, I have been a very successful entrepreneur.” (7-point Likert scale from strongly disagree to strongly agree). SELF-CONF measures the respondent’s answer to the question “Please rate your self-confidence in your entrepreneurial capabilities” (5-point Likert scale from far below average to far above average). Lastly, FUTBSUCC measures the respondent’s agreement with the statement, “My current business will be very successful”. (A 7-point Likert scale from strongly disagree to strongly agree).

Table 1. Variable descriptions.

Variable	Range	Survey Q
E_AGE	9 levels	Your age [under 18; 18–24; 25–34; 35–44; 45–54; 55–64; 65–74; 75–84; 85+]
EDUCLVL	7 levels	Your education [less than HS; HS grad; some college; 2-year degree; 4-year degree; professional or masters; doctorate]
BUS_ED	1, 0	Most recent formal education mainly in the area of business
ENT_IMF	1, 0	Does entrepreneurship run in your family? (1 = yes, in my immediate family)
ENT_EXTF	1, 0	Does entrepreneurship run in your family? (1 = yes, in my extended family)
WRK_EXP	0-50	How many years of work experience do you have?
ENT_EXP	0-40	How many years of entrepreneurial experience do you have?
#STARTUPS	0-10	How many different businesses have you started in your life?
MALE	1, 0	Identify as: (1 = Male)
MINORITY	1, 0	Identify as: (1 = Minority)
EXTROV	7pt scales	Here are a number of personality traits that may or may not apply to you. Please indicate next to each statement the extent to which you agree or disagree with that statement. You should rate the extent to which the pair of traits applies to you, even if one characteristic applies more strongly than the other. AVERAGE of [Extraverted, enthusiastic] and reversed [Reserved, quiet]
OPENESS	7pt scales	…AVERAGE of [Open to new experiences, complex] and reversed [Conventional, uncreative]
RISK_LOV	7pt scale	…Risk-loving, uncertainty-seeking
STRATMS	5pt scale	Consider the following four statements about how you approach important tasks at work, and please indicate how true they are for you: [from Not at all true of me to Very true of me] AVERAGE OF [Throughout a task, I keep thinking about the specific strategies I could use to perform it well; Throughout the task, I keep track of how effective each of my approaches to it is; Whenever I notice that one approach is less effective, I reassess whether I am using the best strategy; As I go about a task, I continuously evaluate what is going right or wrong in the way I am doing things.]
FOUNDERE	1, 0	Where you work, what role do you primarily identify with? (1 = founding entrepreneur)
LN_FTE_	5-500	In the past FIVE years, what has been the largest SIZE of your business in (equivalent) Full Time Employees (FTEs) [natural logatithm(size + 1)]
LN_AGE_	0-50	What is the AGE of your business (in years) [natural logatithm(age + 1)]
FAM	1, 0	Would you consider your business to be a family business (1 = yes)
MINORTY	1,0	Would you consider your business to be a minority business (1 = yes)
WOMEN	1, 0	Would you consider your business to be a women-owned business (1 = yes)
MFG	1, 0	Which one of the following best describes your organization’s primary industry? (1= Manufacturing)
RETAIL	1, 0	Which one of the following best describes your organization’s primary industry? (1 = Retail Trade)
TECHSVC	1, 0	Which one of the following best describes your organization’s primary industry? (1= Professional, Scientific and Technical Services)
LOWCOST	1, 0	Which of the following best describes your firm’s positioning strategy? (1 = we sell our products/services at a competitive price because we have low costs)
G_S	1, 0	Which of the following best describes your firm’s business model? (1= most of the revenues come from selling our goods and services)
FORPROF	1, 0	Business type: (1 = strictly for-profit)
WH_RE	7-pt scale	Business characteristic: from Wholesale to Retail
B_M_INT	7-pt scale	Business characteristic: from Bricks&Morter.only to Internet.only
INDTURB	0-100	Rate the turbulence of your firm’s primary industry, where higher turbulence means higher amounts of entry and exit, and higher amounts of technological and regulatory change
ENTCOURSES	0-20	Approximate number of entrepreneurship-related courses taken through any form of education
EELEVELS	0-3	Which levels of your education/training included content that was specific to entrepreneurship? [middle school and/or high school; undergradaute college/technical school and/or graduate school; online or certificate training from a non-governmental source and/or training from a goevrnmental source (e.g., SBDC)]
ENTSPEC	1, 0	What has been the most important educational course to your entrepreneurial success? (1 = it was entrepreneurship specific (e.g., writing a business plan))
SICOLLEGE	1, 0	For starting or running this venture, which educational sources, if any, have been significantly important or helpful? (1= undergraduate college/technical school OR graduate school)
EE_WIDTH	0-20	For starting or running this venture, please indicate the areas where those educational sources have been significantly important or helpful? [business planning & vision; strategizing & goal-setting; human resource management; accounting; conducting competitive intelligence; business law; supply chain management; negotiations; managing the triple bottom line/social responsibility; effective communication; decision-making; leadership; innovation; marketing; opportunity recognition; sales; resource acquisition; finance; managing information technology; managing other technology]
DIRECT_EXP	1, 0	In what ways have those educational sources been the most helpful? (1 = by providing direct exposure to starting/running a venture)
INDIR_EXP	1, 0	In what ways have those educational sources been the most helpful? (1 = by providing indirect/vicarious exposure to entrepreneurship)
NETWRKG	1, 0	In what ways have those educational sources been the most helpful? (1 = by providing networking and setting up ways to socially interact with other entrepreneurs)
RES_NEEDD	1, 0	In what ways have those educational sources been the most helpful? (1 = by providing exposure to what resources are needed and are available to pursue entrepreneurship)
HARMS	0-3	In what ways, if any, have those educational sources been harmful? [by providing misleading information about the entrepreneurial process; by providing overly optimistic or incomplete information about the entrepreneurial process; other {not including reducing your confidence in your entrepreneurial abilities}]
EE_INTENSITY	7pt scale	Please indicate next to each skill the extent to which you agree or disagree with the statement that “I have relied mainly on my educational sources to learn the skill of ____”.AVERAGE SCORE [Strongly agree to Strongly disagree] FOR [social intelligence; market intelligence; emotional intelligence; financial intelligence; communication; compassion; compromise; personal motivation; identifying opportunities; exploiting opportunities; tough decision-making; taking ownership; managing people; growing a business; dealing with risk; dealing with uncertainty beyond risk]
BIZSUCC	7pt scale	Please indicate next to each statement the extent to which you agree or disagree with that statement [from Strongly agree to Strongly disagree]- My current business has been very successful.
LOG_HIG	7-pt scale	Business characteristic: from low-growth to high-growth
COVIDPERF	9-level	To the best of your knowledge, how has your monthly revenue been affected by the COVID-19 pandemic? (incr 76–100%; incr 51–75%; incr 26–50%; incr 1–25%; stayed the same…) [midpoints used]
IP5YR_	1, 0	Has your organization produced any patents or other valuable and innovative intellectual property in the last 5 years? (1 = yes)
SELF-CONF	5pt scale	Please rate your self-confidence in your entrepreneurial capabilities [far below average to far above average]
ENTSUCC	7pt scale	Please indicate next to each statement the extent to which you agree or disagree with that statement [from Strongly agree to Strongly disagree]- Overall, I have been a very successful entrepreneur.
FUTBSUCC	7pt scale	Please indicate next to each statement the extent to which you agree or disagree with that statement [from Strongly agree to Strongly disagree]- My current business will be very successful.
INFO_SEEK	1, 0	correlational marker-variable (1 = seeking out information is the primary way my venture deals with uncertainty)

summarizes the variables used in our analysis.

4.3. Explanatory Variables

In order to fully capture the multi-dimensional effects of EE, we consider eleven measures over eight characteristics: EE_courses measures the number of EE courses the respondent completed, capturing the quantity characteristic. Direct_Exp, Indir_Exp, Netwrkg, and Res_Needd measure, respectively, whether the respondent’s EE included helpful direct exposure to entrepreneurial activity, helpful indirect/vicarious exposure, helpful networking/social interaction, and helpful identification of resources needed and available for pursuing such activity. These capture the characteristics of the learning modes experienced.

EE_levels measures the number of different educational commonly-sequential sources of EE the respondent partook in, and captures the length characteristic. EE_width measures the respondent’s perception of where EE has been helpful across a significant range of skills and captures the width characteristic. EE_intensity measures the respondent’s reliance on EE as the source of the main skills involved in entrepreneurial activity and captures the depth characteristic. Harms measures the number of possible harm types that the respondent experienced in EE and captures a danger characteristic.

SI_college measures the respondent’s perception that college EE has been the most significantly helpful, capturing a formality/quality characteristic. And ENT_spec measures the respondent’s perception over whether an entrepreneurship-specific course (versus a more general business course) has been the most important to their entrepreneurial success, capturing the specialization characteristic of EE. (Please refer to Table 1 for descriptions of all variables used in this study.)

While some of these characteristics (and their measures) have been used in past EE research, many have not, and no studies have used such a comprehensive array of factors to capture the full potential impact of a respondent’s EE experiences. It is important to note that many of the characteristics and measures—for example, those that assess the doing (direct) experience, the observing (vicarious) experience, and the interacting (social) experience—those missing in many previous studies, have been acknowledged as crucial in understanding EE quality (e.g., Eesley & Lee, 2021; Erikson, 2003; Plaschka & Welsch, 1990; Porter & McKibbin, 1988). Based on past studies of other complex process inputs akin to EE, measures of its length, width, depth, level, and dangers were all considered appropriate here (e.g., Jones, 1991; Kyriakides & Creemers, 2008).

4.4. Control Variables

In order to properly assess the marginal effects of EE on ME performance (and on entrepreneur confidence), we applied twenty-nine controls split over two levels (i.e., the individual level and the venture level). At the individual level, we considered fifteen controls across three categories: In the demographic category, we measured the respondent-entrepreneur’s age (E_age), gender (male), minority status (minority), and education level (educlvl). In the psychographic category, we measured the respondent’s extroversion (extrov), openness (openess), risk-tolerance (risk_lov), and strategic mindset (stratms). In the entrepreneurship-exposure category, we measured the respondent’s work experience (wrk_exp), entrepreneurship experience (ent_exp), number of startups (#startups), whether the respondent’s most recent formal education was business-related (bus_ed), whether the respondent founded the current ME (founder), and whether the respondent had any immediate or extended family who were entrepreneurs (ent_imf and ent_extf). We chose these factors because the same or similar control types and measures have been used in past studies of EE and entrepreneurial success (Charney & Libecap, 2000; Eller et al., 2021; Lange et al., 2014; Mayhew et al., 2012; Padilla-Angulo et al., 2021; Parker & van Praag, 2006; Walter & Block, 2016).

At the venture level, we considered fourteen controls across four categories. In the basics category, we measured venture age (ln_age) and employment size (ln_fte). In the ownership-type category, we recorded whether the venture was a family firm (family), a minority firm (minority), or a women-owned firm (women). In the industry sector category, we recorded whether the venture was primarily in manufacturing (MFG), retail (Retail), or professional/scientific/technical services (Techsvc) and measured the perceived industry turbulence (IndTurb). In the orientation category, we recorded the venture’s generic strategy (lowcost), its main revenue source (goods&svcs), whether it was for-profit (for_prft), whether it was more retail than wholesale (wh_re), and how much it was internet-based (b&m_int). As with the individual-level controls, we included these venture-level factors because the same or similar control types and measures had been used in past studies of entrepreneurial success (e.g., Lange et al., 2014; Parker & van Praag, 2006) or in past capabilities-view-based studies that of firm performance (e.g., Barney & Mackey, 2005; Miller & Shamsie, 1996).

4.5. Descriptive Statistics

Table 2. Descriptive statistics.

	Variable	Mean	StdDev	Min	Max	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22
1	E_AGE	39.467	10.725	21	79.5	1.00
2	EDUCLVL	4.033	0.939	1	6	0.16	1.00
3	BUS_ED	0.559	0.498	0	1	0.14	0.13	1.00
4	ENT_IMF	0.554	0.498	0	1	−0.07	0.05	0.18	1.00
5	ENT_EXTF	0.244	0.431	0	1	0.06	0.05	−0.05	−0.62	1.00
6	WRK_EXP *	15.209	8.652	2	50	0.70	0.04	−0.02	−0.21	0.02	1.00
7	ENT_EXP **	9.175	5.395	1	27	0.53	0.06	0.13	0.00	0.06	0.57	1.00
8	#STARTUPS	2.249	1.639	0	10	0.23	0.08	0.09	0.10	0.02	0.13	0.31	1.00
9	MALE	0.601	0.519	0	3	0.03	−0.01	−0.01	−0.08	−0.01	0.03	−0.08	−0.03	1.00
10	MINORITY	0.437	0.497	0	1	−0.17	−0.06	0.04	0.16	−0.11	−0.13	−0.01	0.19	−0.07	1.00
11	EXTROV	1.042	2.518	−6	6	0.14	0.05	0.17	−0.02	0.05	0.12	0.07	0.02	−0.08	−0.02	1.00
12	OPENESS	2.695	2.545	−5	6	−0.04	0.09	0.06	−0.05	−0.03	0.01	−0.12	−0.09	0.03	−0.21	0.32	1.00
13	RISK_LOV	0.221	1.597	−3	3	−0.14	0.01	0.03	−0.13	0.21	−0.13	−0.09	0.05	0.09	0.11	0.29	−0.02	1.00
14	STRATMS	0.000	0.721	−1.675	1.075	0.13	0.12	0.08	0.02	−0.10	0.07	0.03	−0.04	−0.03	−0.21	0.19	0.44	−0.06	1.00
15	FOUNDERE	0.451	0.499	0	1	0.00	0.02	0.15	0.06	0.05	−0.18	−0.04	0.01	−0.12	−0.11	−0.01	0.00	−0.07	0.05	1.00
16	LN_FTE_	4.178	1.376	1.609	6.215	0.13	0.20	0.20	0.17	0.11	−0.12	0.18	0.32	−0.15	0.04	−0.04	−0.21	0.04	−0.08	0.11	1.00
17	LN_AGE_	2.379	0.789	0	3.912	0.34	0.15	0.16	0.20	0.10	0.15	0.46	0.31	−0.18	0.13	−0.05	−0.32	−0.09	−0.20	0.06	0.57	1.00
18	FAM	0.493	0.501	0	1	−0.08	−0.10	−0.06	0.13	−0.08	−0.06	−0.13	−0.11	0.10	0.02	−0.09	−0.22	0.14	−0.27	−0.03	−0.07	−0.04	1.00
19	MINORTY	0.225	0.419	0	1	−0.04	−0.06	0.04	0.08	−0.02	−0.05	0.00	0.24	0.06	0.30	−0.09	−0.15	0.11	−0.10	−0.10	0.04	0.00	−0.07	1.00
20	WOMEN	0.221	0.416	0	1	0.02	−0.10	−0.07	−0.01	0.09	0.02	−0.02	−0.07	−0.53	0.03	0.08	−0.06	−0.04	−0.06	0.04	0.06	0.00	−0.08	−0.05	1.00
21	MFG	0.113	0.317	0	1	−0.08	0.02	−0.03	0.11	0.00	−0.13	−0.03	0.13	−0.10	0.17	0.02	−0.24	0.00	−0.18	−0.01	0.23	0.17	0.14	0.15	0.02	1.00
22	RETAIL	0.141	0.349	0	1	−0.17	−0.01	0.04	0.14	−0.16	−0.14	−0.14	−0.13	−0.14	0.02	−0.02	0.16	0.02	0.18	−0.02	−0.04	−0.17	0.09	−0.07	0.10	−0.14	1.00
23	TECHSVC	0.122	0.328	0	1	0.20	0.14	−0.15	−0.14	0.07	0.20	0.08	0.06	0.07	−0.04	0.06	−0.09	0.00	0.01	0.05	−0.16	0.06	−0.10	−0.06	0.05	−0.12	−0.15
24	LOWCOST	0.549	0.499	0	1	−0.07	0.02	0.03	0.14	−0.15	−0.10	−0.03	−0.06	0.08	0.11	−0.09	−0.22	0.06	−0.13	0.02	0.02	0.03	0.23	0.09	−0.05	0.05	0.00
25	G_S	0.742	0.439	0	1	−0.07	−0.15	0.05	−0.18	0.00	0.07	−0.04	−0.20	0.04	−0.15	0.12	0.24	0.09	0.13	−0.11	−0.36	−0.39	0.11	−0.01	0.00	−0.13	0.15
26	FORPROF	0.822	0.384	0	1	−0.16	−0.23	−0.18	−0.16	0.00	0.02	−0.14	−0.10	0.18	0.09	0.08	0.00	0.07	0.00	−0.19	−0.30	−0.32	0.14	0.14	0.01	0.07	0.11
27	WH_RE	5.174	1.694	1	7	0.12	0.11	−0.02	0.00	−0.17	0.13	0.07	0.02	−0.05	−0.09	−0.06	−0.02	−0.15	0.20	0.13	−0.02	0.08	0.05	0.00	−0.07	−0.05	0.20
28	B_M_INT	4.385	1.741	1	7	0.01	−0.01	−0.23	0.04	−0.03	−0.06	0.02	0.22	−0.11	−0.01	−0.28	−0.24	−0.08	−0.13	0.08	0.06	0.11	−0.04	0.04	0.07	0.06	0.00
29	INDTURB	50.700	24.009	0	100	0.12	0.24	0.25	0.11	−0.01	−0.01	0.09	0.24	−0.14	0.06	0.11	0.02	0.04	0.07	0.13	0.45	0.37	−0.12	0.04	0.04	0.14	−0.06
30	ENTCOURSES	5.324	4.516	0	20	0.16	−0.03	0.23	0.06	0.13	0.05	0.26	0.53	−0.02	0.26	0.00	−0.17	0.05	−0.16	0.12	0.30	0.36	−0.05	0.13	0.07	0.12	−0.05
31	EELEVELS	1.333	0.520	1	3	0.08	0.03	0.01	0.04	0.01	0.17	0.13	0.09	0.04	−0.11	0.02	0.15	−0.04	0.10	−0.04	−0.08	0.06	−0.03	0.01	−0.07	−0.10	−0.16
32	ENTSPEC	0.413	0.494	0	1	−0.03	−0.11	0.04	0.05	0.02	−0.02	0.08	0.18	−0.05	0.26	0.05	−0.05	0.18	−0.05	−0.06	0.06	0.12	−0.01	0.17	0.06	0.09	−0.01
33	SICOLLEGE	0.653	0.477	0	1	−0.16	0.10	−0.01	−0.16	0.15	−0.07	−0.20	−0.14	0.18	−0.07	0.09	0.07	0.17	−0.10	−0.07	−0.12	−0.21	0.21	0.01	−0.06	0.10	0.04
34	EE_WIDTH	5.469	3.938	0	20	−0.03	0.06	0.04	0.08	−0.02	−0.11	−0.08	−0.09	−0.12	−0.14	0.14	0.31	0.11	0.30	0.06	−0.02	−0.18	−0.05	−0.10	0.19	−0.05	0.23
35	DIRECT_EXP	0.380	0.487	0	1	−0.07	0.08	0.14	0.14	−0.19	−0.02	0.05	0.10	−0.04	0.13	0.15	0.09	0.20	0.01	−0.13	−0.01	−0.05	0.01	0.11	0.02	0.10	0.08
36	INDIR_EXP	0.296	0.457	0	1	−0.05	−0.01	0.06	0.07	−0.01	−0.08	−0.06	0.06	−0.07	0.14	−0.03	−0.13	0.13	−0.10	0.05	0.05	−0.02	0.09	0.36	0.08	0.11	0.08
37	NETWRKG	0.385	0.488	0	1	0.02	0.24	0.12	0.11	0.03	−0.15	−0.15	0.16	−0.07	−0.02	0.00	0.11	0.03	0.08	0.15	0.21	0.06	−0.12	−0.02	0.00	0.02	−0.05
38	RES_NEEDD	0.441	0.498	0	1	−0.07	−0.06	−0.03	0.09	0.03	−0.14	−0.11	−0.06	−0.08	−0.05	0.10	0.16	0.05	0.19	0.07	0.03	−0.14	0.10	−0.04	0.13	0.02	0.20
39	HARMS	0.568	0.645	0	3	0.07	0.04	0.05	0.18	0.00	−0.09	0.03	0.25	−0.13	0.14	−0.04	−0.09	0.00	−0.06	0.17	0.29	0.25	0.00	0.15	0.00	0.08	−0.01
40	EE_INTENSITY	1.040	1.076	−3	3	−0.01	0.11	0.25	0.05	0.02	−0.15	0.00	0.07	−0.11	−0.02	0.13	0.21	0.27	0.24	0.14	0.24	0.04	−0.03	−0.01	−0.04	−0.05	0.15
41	BIZSUCC	1.516	1.127	−2	3	−0.04	0.08	0.18	−0.02	0.06	−0.09	−0.12	−0.13	−0.10	−0.10	0.21	0.32	0.17	0.35	0.08	0.14	−0.02	−0.11	0.01	0.05	−0.05	0.08
42	LoGrHiGr	4.930	1.407	1	7	0.03	0.20	0.19	0.14	−0.04	−0.19	0.07	0.24	−0.06	−0.07	−0.01	0.04	0.02	0.23	0.23	0.44	0.20	−0.14	0.07	−0.04	0.11	0.02
43	COVIDPERF	4.765	37.503	−88	88	−0.03	0.06	0.00	0.16	−0.11	−0.04	−0.06	0.25	−0.13	0.20	−0.17	−0.24	−0.05	−0.17	0.17	0.23	0.24	0.07	0.15	0.04	0.14	0.01
44	IP5YR_NO	0.488	0.501	0	1	0.17	0.19	0.18	0.27	0.04	−0.14	0.14	0.34	−0.10	0.16	−0.11	−0.24	0.01	−0.16	0.11	0.58	0.51	−0.15	0.18	0.04	0.29	−0.11
45	FAILS_NO	0.455	0.499	0	1	−0.04	−0.15	0.10	0.02	0.04	−0.07	0.01	0.35	−0.01	0.14	0.03	0.00	0.12	−0.10	−0.01	0.08	0.08	−0.03	0.02	−0.02	0.12	0.00
46	SELFCONF	0.531	0.919	−2	2	0.02	−0.02	0.14	−0.03	−0.01	0.05	−0.07	−0.18	−0.03	−0.19	0.17	0.29	−0.05	0.34	0.03	0.00	−0.07	−0.20	−0.04	−0.02	−0.19	0.04
47	ENTSUCC	1.408	1.148	−3	3	0.08	0.17	0.19	0.04	0.00	−0.04	−0.04	−0.07	−0.08	−0.13	0.23	0.29	0.09	0.42	0.06	0.22	0.03	−0.04	−0.07	−0.01	−0.05	0.07
48	FUTBSUCC	1.695	1.188	−2	3	−0.01	0.07	0.18	−0.08	0.06	−0.05	−0.13	−0.10	−0.06	−0.11	0.26	0.42	0.14	0.44	0.06	0.11	−0.06	−0.08	−0.09	0.03	−0.14	0.11
49	INFO_SEEK	0.207	0.406	0	1	0.01	−0.01	−0.03	0.11	0.03	−0.01	0.07	−0.02	0.00	0.02	0.03	−0.04	−0.05	−0.05	−0.13	0.06	0.03	−0.05	−0.04	0.05	−0.09	0.01
		23	24	25	26	27	28	29	30	31	32	33	34	35	36	37	38	39	40	41	42	43	44	45	46	47	48
23	TECHSVC	1.00
24	LOWCOST	−0.07	1.00
25	G_S	−0.05	0.07	1.00
26	FORPROF	0.12	0.18	0.31	1.00
27	WH_RE	0.03	0.08	−0.11	0.00	1.00
28	B_M_INT	0.18	0.10	−0.15	−0.06	0.17	1.00
29	INDTURB	−0.08	−0.11	−0.27	−0.29	0.09	−0.02	1.00
30	ENTCOURSES	0.00	0.06	−0.13	−0.04	0.04	0.18	0.31	1.00
31	EELEVELS	0.02	−0.12	0.02	−0.17	0.11	−0.05	0.00	0.01	1.00
32	ENTSPEC	0.04	0.05	0.08	0.07	−0.03	0.01	0.07	0.24	−0.02	1.00
33	SICOLLEGE	0.04	0.09	0.18	0.23	−0.02	−0.11	−0.22	−0.02	−0.05	−0.08	1.00
34	EE_WIDTH	−0.09	−0.13	0.19	−0.12	−0.02	−0.09	0.08	−0.14	0.18	0.12	−0.01	1.00
35	DIRECT_EXP	−0.12	0.13	0.06	0.03	0.05	−0.01	0.08	0.08	0.01	0.19	0.13	0.20	1.00
36	INDIR_EXP	−0.05	0.17	0.02	0.05	0.05	0.03	0.07	0.18	−0.11	0.29	0.04	0.13	0.14	1.00
37	NETWRKG	−0.10	−0.10	−0.08	−0.20	0.01	0.00	0.19	0.13	0.18	−0.06	−0.02	0.29	−0.04	0.07	1.00
38	RES_NEEDD	−0.13	0.11	0.14	0.04	−0.09	−0.13	−0.05	0.00	−0.09	0.04	0.04	0.37	−0.05	0.12	0.04	1.00
39	HARMS	0.01	−0.04	−0.28	−0.18	0.09	0.16	0.21	0.21	−0.05	−0.04	−0.04	0.08	0.09	0.18	0.17	−0.06	1.00
40	EE_INTENSITY	−0.19	−0.07	−0.04	−0.12	−0.08	−0.10	0.31	0.15	0.03	0.11	−0.06	0.33	0.21	0.08	0.20	0.08	0.10	1.00
41	BIZSUCC	−0.17	−0.02	0.01	−0.14	0.04	−0.20	0.16	−0.04	0.08	−0.02	0.03	0.26	0.07	0.02	0.16	0.20	−0.03	0.42	1.00
42	LoGrHiGr	−0.12	−0.10	−0.33	−0.22	0.28	0.19	0.36	0.22	0.04	0.03	−0.08	0.07	0.15	0.02	0.26	0.00	0.13	0.24	0.22	1.00
43	COVIDPERF	0.02	0.04	−0.30	−0.15	0.09	0.33	0.16	0.15	−0.10	0.03	−0.16	−0.22	0.06	0.08	0.11	−0.09	0.18	−0.09	0.04	0.25	1.00
44	IP5YR_NO	0.01	−0.04	−0.47	−0.39	−0.05	0.24	0.42	0.38	−0.02	0.11	−0.18	0.01	0.11	0.08	0.24	−0.06	0.30	0.22	0.12	0.50	0.29	1.00
45	FAILS_NO	0.01	−0.05	0.02	0.14	−0.03	0.20	0.03	0.23	0.01	0.23	0.10	0.01	0.04	−0.12	0.05	−0.02	0.17	0.03	−0.07	0.03	−0.03	0.16	1.00
46	SELFCONF	0.02	−0.20	0.02	−0.16	0.03	−0.17	0.18	−0.20	0.17	0.04	−0.14	0.30	−0.10	0.00	0.17	0.08	−0.22	0.18	0.29	0.14	−0.18	−0.04	−0.22	1.00
47	ENTSUCC	−0.11	−0.09	0.05	−0.09	0.00	−0.19	0.19	−0.03	0.04	0.01	−0.06	0.20	0.08	0.00	0.09	0.18	−0.07	0.47	0.67	0.23	−0.03	0.08	−0.10	0.31	1.00
48	FUTBSUCC	−0.07	−0.14	0.11	−0.03	0.04	−0.22	0.19	−0.08	0.08	0.07	0.02	0.29	0.06	−0.06	0.16	0.19	−0.15	0.48	0.69	0.21	−0.08	−0.02	−0.01	0.36	0.71	1.00
49	INFO_SEEK	−0.06	0.12	0.00	−0.01	−0.14	−0.12	−0.06	0.09	0.02	0.04	0.05	−0.04	0.03	−0.06	−0.02	0.01	−0.06	0.00	−0.01	−0.10	−0.10	−0.02	−0.01	−0.04	−0.04	−0.01

Notes: N = 213, except * = 201, ** = 206; correlations greater than 0.17 are p < 0.01 significant.

depicts the descriptive statistics for this study. The majority of the variables entail responses across their full possible ranges. There are reasonable means and variances, although there is a positive bias in perceived success measures. There are several high correlations between the personal and venture perceptions of success. And expectedly, there are several high positive correlations of individual EE measures (e.g., for intensity, width, and direct experience) with measures of venture success and entrepreneur confidence.

4.6. Regression Analysis

We test our three hypotheses using hierarchical regression. For all dependent variables except IP5YR—because it is dichotomous—we use OLS regression; for IP5YR, we apply a Probit analysis. This analysis method is standard in the capabilities view literature, in addition to being applied (although somewhat sparsely) in the related EE research (e.g., Walter & Block, 2016). All regressions were checked for multi-collinearity with VIF levels, and no issues were identified (i.e., all levels were well within standard tolerances).

5. Results

Table 3. Incremental Improvement in R² of EE variables ^a.

	Dependent Variable
	BIZSUCC			LoGrHiGr			CovidPerf			IP5YR b
Control variables	R2	ΔR2	ΔF/df	R2	ΔR2	ΔF/df	R2	ΔR2	ΔF/df	R2	ΔR2	ΔF/df
Entrepreneur characteristics		0.25 ***	4.12 (15, 181)		0.27 ***	4.45 (15, 181)		0.23 ***	3.57 (15, 181)		0.33 ***	5.92 (15, 181)
Venture characteristics	0.34	0.09 †	1.58 (14, 167)	0.50	0.23 ***	5.45 (14, 167)	0.33	0.10 *	1.81 (14, 167)	0.63	0.30 ***	5.20 (14, 167)
Independent Variables
Entrepreneurship education and training	0.41	0.07 †	1.68 (11, 156)	0.54	0.04	1.38 (11, 156)	0.39	0.06	1.39 (11, 156)	0.68	0.05	0.65 (11, 156)
	ENTSUCC			SELFCONF			FUTBSUCC
Control variables	R2	ΔR2	ΔF/df	R2	ΔR2	ΔF/df	R2	ΔR2	ΔF/df
Entrepreneur characteristics		0.26 ***	4.29 (15, 181)		0.21 ***	3.19 (15, 181)		0.33 ***	6.02 (15, 181)
Venture characteristics	0.35	0.09 †	1.67 (14, 167)	0.32	0.11 *	1.89 (14, 167)	0.41	0.08 †	1.61 (14, 167)
Independent Variables
Entrepreneurship education and training	0.45	0.10 **	2.54 (11, 156)	0.45	0.13 ***	3.49 (11, 156)	0.53	0.12 ***	3.49 (11, 156)

^a n = 197. BIZSUCC = degree to which business has been successful; LoGrHiGr = venture’s growth; CovidPerf = amount that monthly revenue has been affected by the COVID-19 pandemic; IP5YR = any patents or valuable intellectual property produced in the last 5 years; SELFCONF = self-confidence in your entrepreneurial capabilities. ENTSUCC = perception of entrepreneurial success; FUTBSUCC = expectation of future success. ^b Probit analysis reporting McFadden pseudo-R². ^† p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.01.

and

Table 4. Standardized regression coefficients estimating venture and entrepreneur success ^a.

	Dependent Variable
	BIZSUCC	LoGrHiGr	CovidPerf	IP5YR b	ENTSUCC	SELFCONF	FUTBSUCC
EE variables
EE_courses	0.05	0.08	−0.08	0.02	0.02	−0.11	−0.08
EE_levels	0.02	0.02	−0.05	−0.04	0.02	0.03	−0.01
ENT_spec	−0.06	0.04	0.01	0.33	−0.02	0.05	0.04
SI_college	0.03	0.04	−0.11	−0.06	−0.07	−0.04	0.04
Direct_Exp	−0.01	0.16 *	0.16*	0.52	0.05	−0.09	0.02
Indir_Exp	−0.01	−0.04	0.01	−0.22	0.05	0.08	−0.03
Netwrkg	0.03	0.13 †	0.12	0.65	−0.06	0.13	0.05
Res_Needd	0.09	0.02	0.04	0.34	0.08	−0.02	0.08
EE_width	0.03	−0.05	−0.21 *	−0.04	−0.07	0.22 *	0.02
Harms	−0.07	−0.12 †	−0.02	0.05	−0.11	−0.28 **	−0.19 **
EE_intensity	0.28 **	−0.04	−0.12	0.37	0.35 ***	0.03	0.35 ***
R2	0.41	0.54	0.39	0.68	0.45	0.45	0.53
F (40, 156)	2.73 ***	4.64 ***	2.49 ***		3.21 ***	3.22 ***	4.36 ***
Chi2 [40]				184.98 ***

^a n = 197. Standardized regression coefficients are shown. Entrepreneur- and venture-level controls are included in the models. BIZSUCC = degree to which business has been successful; LoGrHiGr = venture’s growth; CovidPerf = amount that monthly revenue has been affected by the COVID-19 pandemic; IP5YR = any patents or valuable intellectual property produced in the last 5 years; SELFCONF = self-confidence in your entrepreneurial capabilities; ENTSUCC = perception of entrepreneurial success; FUTBSUCC = expectation of future success. ^b Probit analysis reporting McFadden pseudo-R². ^† p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001.

depict the main results from our hierarchical OLS analyses regressing EE on ME performance and entrepreneurial confidence. There is overall support for H1b over H1a, but there are individual elements of EE that provide spotty support for H1a. The marginal impact of EE—as a set of measures collectively capturing its multiple dimensions of influence—is not statistically significant (p < 0.5) on any of our four ME performance outcomes. It is a non-significant but positive (i.e., in terms of the sign of the sum of the component standardized coefficients) influence on currently perceived ME business success [BIZSUCC], with one component (EE_intensity) being individually highly positively significant (ß = 0.28, p < 0.01). It is a non-significant but positive influence on perceived ME business growth level [LoGrHiGr], with three components (Direct_Exp, Netwrkg, and Harms) being individually significant in the expected directions (ß = 0.16, p < 0.05; ß = 0.13, p < 0.1; ß = −0.13, p < 0.1). It is a non-significant and negative influence on COVID-era sales revenue change [CovidPerf], with two components (Direct_Exp and EE_Width) being individually significant in the different directions (ß = 0.16, p < 0.05; ß = −0.21, p < 0.05). As such, we surmise, consistent with our real-options perspective on EE, that when the application context is much more specific or unusual than what was taught, the value of such expertise can degrade to the point of getting in the way of what should be performed (as in some ‘failure to drop the tools’ in an emergency—Weick, 1996). And finally, EE is a non-significant but positive influence on intellectual property generation [IP5YR], with no significant components.

These results are consistent with our options-based narrative. The comprehensive EE impact is non-significant on ME performance, but the most application-related factors—the ones considering the depth and personal involvement (EE_intensity and Direct_Exp) of the training—are significant. So, while general skills provide high option value, only the hands-on, immersive competencies translate into added performance. We also note that EE’s influence on ME performance is significant when regressing on the measures without the individual and firm-level controls, which may help explain some results of those past studies that have involved fewer controls and fewer dimensions of the EE construct.

Table 3 and Table 4 also reveal consistent support for H2. The marginal impact of EE is (statistically) significant on all (three) entrepreneurial confidence measures. It is a significant (p < 0.01) and positive influence on self-perceived entrepreneurial success [ENTSUCC], with one component (EE_intensity) being individually highly positively significant (ß = 0.35, p < 0.001). It is a significant (p < 0.001) and positive influence on perceived self-confidence in one’s own entrepreneurial abilities [SELF-CONF], with two components (EE_width and Harms) being individually significant in the expected directions (ß = 0.22, p < 0.05; ß = −0.28, p < 0.01). And it is a significant (p < 0.001) and positive influence on perceived confidence in the ME’s future success [FUTBSUCC], with two components (EE_intensity and Harms) being individually significant in the expected directions (ß = 0.35, p < 0.001; ß = −0.19, p < 0.01). We interpret these results, especially on SELF-CONF, as confirming that EE has a strong and continued impact on self-efficacy (consistent with many past studies)3.

6. Discussion

To answer the research question of whether EE matters to ME performance, our analysis provides results consistent with our expectations that, although EE increases confidence, the skills taught tend not to lead to significantly increased venture success. That said, in almost all of the analyses, the collective impact of EE was positive. These results support our narrative that EE creates a high option value embodied in the entrepreneur’s general skills and self-efficacy, that such options are often exercised in the form of new venture creation, but that exercising such an option by creating a new venture entails a realized value that is often significantly reduced by the specificities and frictions found in any particular start-up attempt by the entrepreneur focused on a given core idea, resource set, and market context. Our narrative contains an explanation for that reduction in exercised option value—one that is based on the limitations of EE itself as being constrained in the specificity and depth of the skills that can be taught and on the reality that EE substitutes readily exist (e.g., in first-hand experiences) that are often more specific and that many rival managers may have obtained.

One important reason for investigating this research question was to better understand the glaring contrast between the input and the output of the ‘EE-related industrial complex’ (i.e., the massive investment in programs, faculty, centers, chairs, and incubators) in the US and in similar economies (e.g., Canada). While the input side has enjoyed explosive investment and growth in the past few decades, the output side—at the macro-level (e.g., in start-up and survival rates)—has been too stable (and, over some measures, has worsened). That contrast is all the more frustrating because we know that entrepreneurial activity matters to economies, to politicians and policy-makers, to places of higher learning, and to the existence of billion-dollar institutes like the Kauffman Foundation, which all desire to see the innovation and job-creation EE promises. What is confusing is that this EE industrial complex has mostly avoided confronting its macro-level ineffectiveness. Instead, most of its studies are treatment-based and carry a positive and micro-level bias towards EE’s impacts on the short-term and on perhaps more trivial outcomes, such as the increased intentions of students to engage in entrepreneurship. Cynically, one could see why that has occurred—it is more difficult to assess macro-level impacts (especially when generated from audited numbers from several government agencies that each partially capture entrepreneurial activity) than micro-level impacts based on numbers from self-interested, proprietary, program-specific, primary data sources. Regardless, the lack of evidence that EE makes a significantly positive macro-level impact—which should most logically result from the accumulation of the well-documented micro-level impacts—is disappointing. It is all the more so for two reasons: the money and expertise to study that paradox readily exist, and the fact that EE is at such a scale now that every ME entrepreneurial team could be individually trained each year, which, given the evidence that such training should have a short-term positive impact, logically would provide the desired macro-level change.

We find cold comfort that our main result is consistent with the disappointing lack of impact of EE, and specifically where we would like to see it most (i.e., on the performance of the ventures that are most important to the economy—MEs). Regardless, our study’s findings also reflect the outcomes of several related empirical investigations: EE does not disproportionately create very successful firms (Fairlie et al., 2015), EE does not have a significant statistical impact on the sales growth of SMEs (Avdullahi & Ademi, 2020), EE has no lasting effect on venture performance (Fairlie et al., 2015), and EE has little to no impact on subsequent operating performance of new ventures (Lange et al., 2014). But, while our research produced similar results, it brought added insight: Ours is the first study to consider the real option value of EE, specifically in how that is translated into short- and longer-term outcomes. In the short run, EE’s provision of general skills is likely to increase start-up intentions and even rates, especially when fueled by increased confidence. In the long run, while the influence of EE on confidence may linger, the exercise value of the skills-based real option in starting up and managing an ME is likely to decrease for several reasons, including the degradation during the delay, the missed updates; the specificity required for successful application; the uncertainties involved in exploiting a new opportunity (that, by definition, cannot be pre-taught); and the impact of rivals who can imitate observed skills or access substitutes for EE training (e.g., from mentors, investors, relatives, business partners, and experienced others).

While our real options-based narrative fits the past evidence, at this point, it is worthwhile to consider some alternative explanations. First, the macro-level issues may be being driven by forces outside of the influence of EE: There may be structural and cultural forces at play at the macro-level (in the US). It may be that the capacity of the economy to absorb and sustain new offerings from new ventures (especially those resulting from EE experience) is limited (e.g., given the time it takes to legitimize a new market—Aldrich & Fiol, 1994). It may be that more and more R&D races are being won by deeper-pocketed incumbents (Schroth & Szalay, 2010), making capitalization rather than education the more important factor for success. Or it may be that the adaptability and reactivity of incumbents have improved so much as to offset the actions of entrepreneurial startups (e.g., there may be Red Queen effects involved—see Barnett & Hansen, 1996). Second, it may be that the real option value of EE involves weaknesses that are more likely brought out at a macro rather than at a micro-scale. This may be based on EE being incomplete—in that it teaches many skills but not all, and so it is neither necessary nor sufficient for success (Edelman et al., 2008; Lange et al., 2014). This is because there are specific gaps in most programs: in teaching social and emotional intelligence (Murray et al., 2018), in teaching technical skills (Czuchry et al., 2004), in teaching hiring and selling skills (Gendron, 2004), and so on. Further, EE is constrained by its institutional contexts—places where it is simply not possible to teach students how to act in chaos and ambiguity (Aronsson, 2004; Gendron, 2004), even though that may be what is required in order to succeed. A related alternative explanation is that EE has quality problems—such that teachable content is simply not being absorbed by students due to issues regarding instructors, delivery modes, and so on. Third, it might be that we are using the wrong macro-level measures here to assess success. Perhaps the focus should be on whether the student brings entrepreneurial insight to any firm she works at, regardless of whether it is a SE, ME, incumbent, governmental institution, NGO, or other. While we are not advocating moving the goal-posts, it may be prudent, given both the large investments that have been made in EE and the inconsistency between micro- and macro-level impacts, to identify intermediate goals and measures in order to better explain the value generated by EE and where to find it. Linking EE to the wrong measures may work to decrease its effectiveness and detract from its societal benefits (Zawadzki, 2019)4.

6.1. Implications

Our analysis implies a real options story for EE, while our discussion implies important issues that help explain EE’s short-run, micro-level positive impacts in the face of disappointing, non-significant long-run, macro-level ones. In turn, that implies that determining the proper goals, measures, and associated pedagogies would help address those issues. Taking further actions on those implications at this point, however, is confounding, at least for those parties currently benefiting from the EE industrial complex. There seems to be little incentive to rock a boat when it has been building out so well for decades and with few leaks. There is a clear motivation to keep publishing short-term, positive-impact treatment studies, to keep feeding the media and alumni success stories, and to keep growing the programs as long as interest remains.

To the entrepreneurship students, the implication from our analysis is to use EE as an option, one that needs to be supplemented with both more specific (e.g., in technological expertise and in skills application experience) and more general (e.g., in a jack-of-all-trades span) training to be most impactful. Students need to be patient, both in obtaining that extra training and in identifying the right opportunity where they can use their more-generalized skills to add the most value.

To policy-makers, two big questions remain, however—about whether EE has been worth such a massive investment and what can be used to justify continuing it at present or higher levels (Walter & Block, 2016). Some scholars suggest that EE may not be the most cost-effective method of addressing innovation and job creation, let alone of addressing deeper issues over human capital, discrimination, or inequity (Fairlie et al., 2015). We have raised the former point directly, given the scale of EE today—where, hypothetically, every US ME start-up team could be personally trained through the courses offered each year but is not. So, we would urge policy-makers to reconsider how those resources should be best deployed (after identifying what goals are best and proper for EE to pursue).

To educators, the implications of our analyses support a re-examination and framing of best pedagogical approaches. Perhaps working with recent graduates who have gone on the find market success could yield innovative thinking that helps shift the focus in the ways we teach to more practically relevant methods and topics. And, perhaps it is time to practice more of what we preach, moving out of the traditional lecture/theory-oriented classroom and into one where students are placed in risky, uncertain situations that require agile thinking. This may be a classroom where students get comfortable with failure, learn from it, and assess the failures of others to learn more. This may involve more skills in being lean, in learning from fast and cheap experimentation (e.g., based on minimum viable products—see Stevenson et al., 2024). This is likely to be more of a place where experience and practice are prioritized and one where empathy and need recognition drive creative and design thinking (e.g., Bender-Salazar, 2023) as a process that produces greater value through better testing and market offerings. As such, we encourage educators to continue to reimagine and experiment with what EE can be (e.g., in terms of curricula that help students translate the general to the specific, that tailor some lessons to different venture types, and that work on building personal resilience and adaptability, including teaching students how to analyze and build on relevant feedback and available data).

6.2. Limitations and Areas for Future Work

As with every survey-based, restricted sample-sized, cross-sectional, national empirical study, our research product involves several limitations. Our survey involved recollection, which raises valid concerns about historical accuracy. Our sample was restricted in size while applying many controls, raising some concerns over statistical power (although significant results and reasonable R²s were obtained). Our sample was collected in the second year of the pandemic and restricted to US-based MEs. The usual warnings about over-generalizing beyond such a context may apply (especially geographically), and so we call for further such studies to be done in other countries. Using MTurk—which is an accepted source—also carries possible concerns about the kind of respondent who applies to this mode of survey call (and that involves a nominal monetary payment that other surveys often do not). Sampling across industries, while providing a fuller picture of the impact of EE across the economy, may hide some possible industry-specific effects. And, relying on the respondents for information about both the focal factors and the focal outcomes, rather than applying some triangulation, again while standard, raises possible response-related biases. Further, we acknowledge that EE (with its various impacts) represents a complex phenomenon (given its many dimensions of quality, delivery, timing, practicality, and intensity) that is difficult to explore in any one study.

Acknowledging all of these limitations, we remain confident in our main results—given that bias-related testing (e.g., over non-response and single response) did not reveal issues and that (many) more than the usual controls and outcome variants were used in the analysis.

Future work could be valuable in several areas. First, follow-on research would be useful to address the empirical limitations described above (e.g., using triangulated and longitudinal data to increase validity and better explore causation and possible co-evolution). Second, it would be useful to have future studies replicate and extend our findings (e.g., to new countries, to specific industry sectors, to specific venture types, to new measures of EE, to different measures of success, with more controls). Third, further insights may be gleaned with data analysis involving specific substitutes to EE, and different forms of EE that could be provided in separate educational settings (at different costs and through different delivery modes).

6.3. Conclusions

Even if the promise and the reality of EE differ, it remains a well-intentioned activity that doubtlessly builds significant real option value for its students. We explored an important research question about the impact of EE to help explain the conflicting evidence of its positive effects in the short-run at the micro-level versus its non-effects in the long-run at the macro-level; we did so by surveying US-based entrepreneurs about their ME firm’s performance and their own educational experiences. We found that while EE positively impacted their confidence, it was non-significant in improving venture performance when considered as a multi-dimensional driver. We argue that EE may be constrained in offering sufficiently specific and comprehensive application-targeted skills, and so, as an option, loses value when exercised, especially in contexts where substitutes and uncertainties exist. We expect that further investigation into the impact and limits of EE—as a decades-long investment that has reached a scale where it should have a macro-level influence—will reveal more answers that can provide new directions and objectives to help improve the outcomes we desire in future entrepreneurial activity.