Doing the same thing but expecting different results: The role categorization effect of “startup” versus “large corporation” on perceived entrepreneurial self-efficacy

Ming D. Leung University of California, Berkeley Haas School of Business Berkeley, CA 94720-1900 [email protected]

Nathanael J. Fast University of Southern California Marshall School of Business Los Angeles, CA 90089-0808 [email protected]

Working Draft August 31, 2016

0

Doing the same thing but expecting different results: The role categorization effect of “startup” versus “large corporation” on perceived entrepreneurial self-efficacy

Abstract Employees of startup firms are much more likely to found entrepreneurial ventures than employees of large corporations. Drawing on categorical learning and self-efficacy theories, we propose and test a novel theory of role categorization to explain this phenomenon. While working for a small startup firm, the skills an employee gains are categorized as pertaining to the ‘startup’ domain. Because one’s experiences have been imbued as being more relevant to entrepreneurship, one will be more confident in one’s ability to succeed in such an undertaking. Individuals therefore learn to develop their entrepreneurial self-efficacy through the experiences they gain by working with a startup versus a large corporation. It is this entrepreneurial self-efficacy that leads to a greater likelihood of them considering an entrepreneurial career path. To alleviate selection issues, we performed two experiments on undergraduates and virtual freelance workers to test our theory. We find that subjects who performed tasks labeled for a “startup,” were more likely to indicate a greater likelihood of taking more entrepreneurship classes as well as be more likely to claim to want to start a business, than subjects who performed identical tasks for a “large corporation.” Entrepreneurial self-efficacy mediated the relationship between our conditions and startup intentions.

1

From causes which appear similar we expect similar effects. This is the sum of all our experimental conclusions [Hume 1748]

Organizational theorists who study the founding of entrepreneurial ventures have focused on how the decision to enter or pursue such ventures is contingent on the contextual factors that individuals face (Thornton 1999, Aldrich and Ruef 2006). In particular, a stream of work has proposed and demonstrated how an employee’s prior relationship with an established organization affects their likelihood of embarking on an entrepreneurial path (Burton et al 2002, Audia and Rider 2006, Sorensen and Fassiotto 2011). Findings suggest that organizational size and age exert particularly important impacts on rates of entrepreneurship (Shane and Venkataraman 2000, Dobrev and Barnett 2005). Gompers, Lerner, and Scharfstein (2005) demonstrate that venture-backed firms are more likely to be started by former employees of younger and smaller firms. More concretely, Sorensen (2007) calculated that the founding rates of entrepreneurial ventures by employees of large corporations is almost three times lower than the rate of founding exhibited by employers of small firms. Theories used to explain this puzzle generally rely either on selection arguments or the belief that what an employee does and experiences at large established firms differ markedly from what they do and experience at small startups. For example, to the extent that the skills an employee acquires will play a central role in their decision to embark on an entrepreneurial venture, then smaller and younger firms are likely to provide a different learning opportunity than larger and older organizations. Larger organizations may have greater role differentiation, therefore providing fewer broad experiences valuable to an entrepreneur (Dobrev and Barnett 2005) or perhaps because larger firms insulate employees from entrepreneurial opportunities more than smaller firms (Saxenian 1994). While these extant contextual theories rely on the differences between what an employee actually learns at younger and smaller firms versus older and larger ones, the literature has not grappled with how that learning may categorized differently by employees at small firms versus those employed by large

2

firms. Nor has it addressed how such differences in categorization might, in turn, have downstream consequences on their likelihood of embarking on entrepreneurial ventures. In the present paper, we propose that individuals learn to develop their entrepreneurial self-efficacy through the experiences they gain by working with an entrepreneurial venture, and that it is this entrepreneurial self-efficacy that leads to a greater likelihood of them considering an entrepreneurial career. Drawing on categorical learning and self-efficacy theories (Medin et al 1993, Bandura 1991), we propose a theory of role-categorization and entrepreneurial learning. While working for a small startup firm, the tasks an employee learns are categorized as pertaining to a ‘startup’ or ‘entrepreneurial’ domain. In these instances, because one’s experiences have been imbued as being more relevant to entrepreneurship, one will be more confident in one’s ability to succeed in such an undertaking. Therefore, individuals that work for startups will be more likely to choose such an employment path because they feel more confident in their abilities in those domain, what has been referred to as entrepreneurial efficacy. Conversely, individuals who work at large corporations, will have no such confidence as their experiences are not categorized as such. This perspective we offer advances a more subtle line of thinking that theorizes on the potential differences in how individuals learn to adopt different values working at smaller versus larger organizations, resulting in employees being more or less amenable to entrepreneurial forays. What Sorensen and Fassiotto (2011) refer to as ‘fonts of beliefs and values.” For example, Nanda and Sorensen (2010) argue that the reason work colleagues who were former entrepreneurs increases the rate of entrepreneurship in a firm are due to informal training processes that shed light on the entrepreneurial process to others. Similarly, Stuart and Ding (2006) believe that the normative stigma attached to entrepreneurship among university scientists influence their reluctance to enter the private sector. However, our theory departs from this line of thinking because it operates independent of differences among fellow employees or historic norms. In this sense, our theory is more parsimonious as it requires almost no assumptions regarding differences between small and large firm contexts (or the employees who work in them) beyond how the actual work contexts are categorized.

3

We support our theory with evidence from two experimental studies. In our first study, we were interested in seeing whether our theory would alter the human capital investment intentions in the classes that of a population of undergraduates indicated they would take. We found that when subjects were told that the task they were performing was pertinent to a “small startup” they were more likely to indicate they were planning on taking more entrepreneurial classes in the future than a set of subjects who performed the same task but were told it pertained to a “large company.” In our second study, we tested whether our theory would alter the entrepreneurial intentions of a set of online freelance workers on MTURK. We found that subjects who were told the task they were performing pertained to a “startup” were more likely to indicate their likelihood of starting a business in the future. Furthermore, in support of our theory, entrepreneurial self-efficacy mediated the relationship between the treatment and the outcome. A non-trivial confounding factor in all observational studies of entrepreneurship is the fact that selection into those positions is non-random (Sorensen 2007). Individuals that are more disposed to starting ventures may also be more likely to sort into startups and also more likely to hold higher entrepreneurial self-efficacy (Chen et al 1998, Zhao et al 2005). Our experimental approach guards against this. By randomly assigning participants to complete identical work categorized either as a “startup” or a “large corporation” we are able to isolate the causal variable and avoid differences in selection. In addition, the experimental evidence we provide also allows us to move beyond mere speculation as to the mechanisms behind the observed differences as is the case of the observational studies to which we are contributing.

Decisions to embark on an entrepreneurial career The past experiences that individuals have affect the decisions they make as to which career path to choose in the future. For example, in Halaby’s (2003) examination of the Wisconsin Longitudinal Survey of 10,317 men and women, he demonstrated that even experiences as far back as ones family or school

4

backgrounds exert consequential effects on the likelihood that an individual will value an entrepreneurial versus a more bureaucratic career path. More recent past experiences individuals have also direct their career intentions. Shane and Khurana (2003) demonstrate how an individual’s past career experiences affect whether potential entrepreneurs decide to choose to found a new venture. They suggest that past career experiences affect individual career decisions by mitigating people’s perceptions regarding the risks their new venture will face, thereby leading those individuals who have careers in which they have observed a greater number of successful firm foundings to be more likely to found a new firm as well. Ultimately, firm foundings are individual level decisions that are subject to biases as are any other decisions people make (Baron 1998). As Moore and his colleagues (2007) demonstrated, individuals look inward, to themselves, to develop their understanding as to the possible success they will have in founding a venture while ignoring the realities of the external competitive marketplace. In doing so, they often misinterpret how competent they may be in pursuing such a risky venture. Extant research suggests that entrepreneurs may be even more biased in their self-assessments than managers in large organizations (Busenitz and Barney (1997). Perhaps this is why Cooper and his co-authors (1988) found that one third of the 2,994 entrepreneurs he surveyed rated their odds of success at 10 out of 10. The reality is bleaker. For example, almost half of all new manufacturers fail within their first four years (Mata and Portugal 1994). Despite the grim statistics faced by startups, one reason why individuals choose to embark on risky career paths is because they believe they are better equipped to do so relative to others. Indeed, higher levels of perceived self-efficacy in terms of competence significantly affects one’s aspiration levels, commitment to goals, persistence in tasks, and attitudes (Heath and Tversky 1991). The more positively one feels about one’s abilities and the more competent one believe oneself to be as a decision maker, the more opportunities one will see in risky choices and the more likely one will be to take those risks (Ajzen 2002). For example, when individuals are told they performed particularly well in decision making tasks they are more likely to take riskier gambles in the future (Krueger and Dickson 1994). 5

Entrepreneurial self-efficacy The fact that “people’s beliefs about their capabilities to exercise control over their own level of functioning and over events that affect their lives” (Bandura 1991: 257) affects the decisions they make has not escaped theorists of entrepreneurship (Krueger and Brazeal 1994, Scherer et al 1989). Termed entrepreneurial self-efficacy (ESE), this type of self-efficacy is defined as the strength of an individual’s belief that he or she is capable of successfully performing the roles and tasks of an entrepreneur. A stream of literature has identified a strong correlation between being an entrepreneur and those individuals’ perceptions of their self-efficacy in particular domains. For example, Chen et al (1998) find that individuals with higher measures of ESE are also more likely to be entrepreneurs than those with lower levels of ESE. While the extant literature has established a strong link between entrepreneurial self-efficacy and the likelihood of being or considering an entrepreneurial career path, the causes of such a correlation are unknown. On one hand, entrepreneurial experience could lead to increase entrepreneurial self-efficacy. The extant work on self-efficacy would posit that entrepreneurs, because they are actually performing tasks that are entrepreneurial, may be developing a recognition of their efficacy in such a domain. In this case, acting and performing as an entrepreneur would cause them to have greater levels of ESE because they are already in such a position. On the other hand, because individuals may differ in the levels of selfefficacy that pertain to entrepreneurship, then having a high level of entrepreneurial self-efficacy could lead individuals to choose an entrepreneurial career. For example, Chen et al (1998) conducted a survey that demonstrated that entrepreneurship students had higher ESE scores than management and psychology majors and were also more likely to indicate their intentions to start a business. Extending this stream of work, Zhao and his colleagues (2005) demonstrated that increased levels of entrepreneurial self-efficacy is correlated with one’s intentions to start a business in a pre-post design. In this study, they surveyed MBA students at the beginning of their program and then again at the end of their program. In the first survey they asked questions regarding a student’s past experiences in 6

entrepreneurship. In the second wave, they asked students their perceptions of formal entrepreneurial learning, their perceived level of entrepreneurial self-efficacy, and their entrepreneurial intentions. They found that both past entrepreneurial experiences (gathered in the first wave) and perceived entrepreneurial learning (gathered in the second wave) were positively correlated with entrepreneurial intentions. Furthermore, they found that entrepreneurial intentions (surveyed in the second wave) fully mediated the relationships. While this study clearly demonstrates the value at examining the concept of entrepreneurial selfefficacy as an individual cause of entrepreneurship, it does not establish a causal link nor does it provide insight into the antecedents of entrepreneurial self-efficacy. This is because they did not measure entrepreneurial self-efficacy in the first wave. This leaves open the question as to whether students who entered with high self-efficacy (perhaps as a result of their past entrepreneurial experiences) were those that were more likely to choose entrepreneurship classes as well as score high on entrepreneurial selfefficacy (as demonstrated in the second wave survey). In short, the direction of the relationship between higher ESE and entrepreneurial intentions is unclear.

Role categorization of tasks The first step in realizing one’s self-efficacy is learning about ones abilities. The development of one’s self-efficacy arises from the acquisition of complex cognitive, social, and linguistic skills through experience (Bandura 1977, Gist 1987). The old adage, learning by doing is descriptive of the process by which individuals come to recognize and increase their own self efficacy. This is because people attribute their experiences to particular skills they may be developing or possess (Gist and Mitchell 1992, Stajkovic and Luthans 1998) and in turn, one’s self-efficacy is formed through an assessment of their success in that task.

7

However, learning is highly contextual, so attribution of one’s self-efficacy from the performance of particular tasks to the successful completion of another will be subject to a myriad of factors (Gist and Mitchell 1992). One perspective, since at least Hume (1748), is that in recalling how one’s experiences may be applicable to the challenge at hand, individuals not only draw from their most recently past experiences, but also those that are similar to the situation they face. Cognitive theorists consider this as the contingent aspect of learning from one’s experiences (Hertwig and Erev 2009) or instance based learning (Gonzalez et al 2003) and economists label it as case-based decision making (Gilboa and Schmeidler 1995). For example, when firefighters attempt to predict how a fire will behave, they will tend to draw from their memory the past situations that are most similar to the current one (Hertwig and Erev 2009). One heuristic that individuals use to recognize similarities between experiences is to consider how they are categorized. Categorical theories suggest that social items, if they are identically grouped, will be considered more similar to one another than those social items which are not identically labeled (Murphy 2002). For example, the labeling for movies into genres induces film goers to consider identically labeled movies to be more similar to one another than movies that are labels as a different genre (Hsu 2006). A movie that is categorized (or otherwise, labeled) as being a “romance” will be expected to share many features that are similar to other movies that are also labeled as a ‘romance’ movie. Conversely, the movie will not be expected to be similar to a movie that is labeled as a ‘horror.’ In this way, labels serve to direct our attention whereby similarly labeled items are going to be considered more similar to one another than items that are not similarly labeled. In learning, the experiences an individual faces are stored in memory along with a list of features that were related to the task. These stored examples (Medin and Schaffer 1978) are clustered by these similar features and categorized accordingly. When an individual faces a new experience, particularly in ill-structured environments, they compare these novel experiences to those already stored in memory in order to identify similar past experiences with which to compare the novel experience (Medin et al 1993). 8

The more features of the new experience that matches with stored ones, the more likely the new experience will be categorized as such. This process leads people to draw on past experiences which are categorized similarly to inform them of how they may behave given decision at hand (Gonzales et al 2003). These findings lead us our theory of role-categorization and learning. To begin, we posit that the experiences one has while working at a startup will be stored in memory as pertaining to a startup. That is the learning inherent to performing tasks while working at a startup will be categorized in one’s memory as being associated to a startup because the category of ‘startup’ will be a salient component to the tasks one does. Conversely, because small startups represent such a markedly different organizational form than large corporations, individuals who work for a large company will categorize the tasks performed there in their memory as experiences which are associated with a large company or at minimum, not associate the experiences with a startup. These differences occur both explicitly (e.g., in the vocabularies used in company materials, websites, and meetings), as well as more subtly (e.g., when describing one’s work to a friend). While our theory does not necessarily require that the two types of organizational forms be diametrically opposed, the distinction between startups and large companies is one that is highly salient and has been summarized in more detail by others (Sorensen 2006). Given this, the experiences that an individual will have while working at a startup will be seen as being related to future decisions and challenges associated with other situations that pertain to startups. For example, if an individual were seeking a new job and were asked what startup experience they have, they would likely simply draw from their recent past experiences of working in a startup to answer such a question. On the other hand, those individuals who have been working for a large corporation would be unable to adequately answer such a question by drawing on their experiences at their employer. Because individuals look inward to develop a conception as to how likely they are to be successful as an entrepreneur, then in deciding whether to pursue such a career path, those individuals who have worked for a startup in the past will be more likely to draw on those experiences are providing 9

evidence of their entrepreneurial self-efficacy. Because those past experiences have been categorized as related to a startup, then they will be easier to draw from when one considers their ability to succeed in an entrepreneurial endeavor. On the other hand, those individuals who have worked for a large company in their past would be unable or unlikely to draw on those experiences to inform the question as to whether they have sufficient entrepreneurial self-efficacy in order to succeed in an entrepreneurial venture. Because their categorized experiences are not associated with startups, they do not believe themselves to possess the appropriate abilities. We therefore summarize our theory with the following three interrelated hypotheses. Hypothesis 1: Individuals performing tasks categorized for a ‘startup’ will be more likely to consider entrepreneurial careers than individuals performing the same tasks categorized for a ‘large corporation.’

Hypothesis 2: Individuals performing tasks categorized for a ‘startup’ will have greater levels of entrepreneurial self-efficacy than individuals performing the same tasks categorized for a ‘large corporation.’

Hypothesis 3: The relationship between performing tasks categorized for a ‘startup’ and the consideration of an entrepreneurial career is mediated by an individual’s entrepreneurial self-efficacy.

Note that our theory does not require any assumptions regarding differences between the actual tasks performed at a startup versus a large company. For example, tasks performed in a payroll function likely do not differ dramatically between startups and large companies. In this case, our theory would still posit that an individual working in a startup in the payroll function would believe their skills to be more applicable to a startup context. They therefore should have a greater likelihood of working for a startup again or embarking on their own compared to an employee working in the payroll function of a large company.

10

To the extent that the tasks actually differ, then we expect this effect would be even stronger, as the beliefs one holds of their entrepreneurial self-efficacy will be further bolstered by the actual observable differences in skills or abilities gained in performing those tasks. For example, the skills one acquires pitching to venture capitalists would likely be more applicable to a startup situation than to presenting ideas to one’s superiors in a large company. If one does pitch to Venture capitalists, not only would one believe themselves to possess experiences related to startups, one might actually hold experiences that are valuable to such a context – then a higher level of entrepreneurial self-efficacy may be actually warranted. Working for a large corporation is likely to be more of a default condition than working for a startup, perhaps because many more individuals work for large companies than do for small startups. Because of this, we would not be surprised to find that the effect of working for a large corporation and the categorized learning effect of such experiences would be as strong for individuals to believe in their large company self-efficacy. Therefore, we do not hypothesize on the opposite effects here and discuss this more in detail in below.

Overview of the present research Because the confounding effects of both selection into working for startups versus a large company are so great, and also because there may be actual differences between working for the two different types of organizations, we tested our theory using experimental methods. Doing so allows us to not only control for these confounding factors but also provide a conservative test of our theory. In Experiment 1, we examined whether completing work for a ‘startup’ would produce stronger entrepreneurial intentions than completing the exact same work for a ‘large company.’ Entrepreneurial intentions are the indication that an individual provides that they plan to start a venture in the future (Boyd and Vozikis 1994, Kolvereid 1996). We captured such intentions in Experiment 1 by assessing the 11

degree to which undergraduates majoring in business indicated a desire to pursue entrepreneurship by investing human capital resources (time and tuition money) toward taking entrepreneurship classes. In Experiment 2, we assessed whether the proposed mechanism—entrepreneurial efficacy—drives the relationship between working for a startup and adopting entrepreneurial intentions. We used a population of virtual workers, obtained through Amazon’s Mechanical Turk, to complete identical tasks that were either labeled as for a startup or for a large company. We then measured entrepreneurial efficacy and entrepreneurial intentions.

Experiment 1 In our first study, we were interested in whether the experience of working for a startup would influence an individual’s likelihood to report that they would make human capital investments towards an entrepreneurial career path (Hypothesis 1). Participants were asked to complete a set of data gathering tasks that would be used to decide on an office relocation. In the treatment condition, the participants were told this task was for a “small startup” to identify a potential city to relocate to. In the control condition, the information was being provided to a “large corporation” for a relocation decision. We then obtained a behavioral measure of entrepreneurial intentions—desire to take entrepreneurship classes while in college.

Participants and design Participants consisted of 153 students (81 men, 66 women, 6 unidentified) from a large West Coast university, ranging in age from 18 to 29 years (M = 19.82, SD = 1.63), participated in this study. These students participated in the university business school’s subject pool, and received class credit for participation in our experiment. We randomly assigned participants to one of two conditions: the startup condition (n = 77) or the large company condition (n = 76). Subsequent to the manipulations and tasks,

12

we asked participants to indicate their intentions to take entrepreneurship classes. Participants then answered several demographic questions.

Manipulations and measures To set up our manipulation, we had participants conduct online work that was ostensibly for an existing company. The students were informed that they had been recruited to help provide insight to an actual company. Specifically, they read the following:

Thank you for your interest in this project. Rather than conducting a research study, we are providing you the opportunity to earn 1 credit by helping us provide valuable insights to a company we're consulting with. The company is interested in getting opinions and assistance from undergraduate business students like you, as they would like to attract college graduates as employees. The project will require approximately 45 minutes of your time and you must complete it within the next four hours. To proceed, please click on the button below.

Our goal was to make the actual work and description of the task identical with the one exception being the type of company for which the work was conducted. Participants were instructed as follows:

[Startup condition]: We are providing consultation assistance to a new company. The company is a small start-up in the high-tech software industry and is looking for the ideal city in which to locate its main office. We are in the process of compiling background information about various U.S. cities and getting opinions about the livability of the cities from people like you.

[Non-startup condition]: We are providing consultation assistance to a large corporation. The company is a large corporation in the high-tech software industry and is looking for the ideal city in which to expand with a new office. We are in the process of compiling background information about various U.S. cities and getting opinions about the livability of the cities from people like you.

Participants were given the names of 5 cities (Pittsburgh, Atlanta, Phoenix, Denver, and Charlotte) and then provided answers for a variety of questions based on research conducted on www.wikipedia.org. Questions included topics such as population, colleges and universities, businesses, entertainment 13

opportunities, strengths and weaknesses, and additional notes. They were then asked to rank order the cities in order of most to least desirable. Finally, they were asked to identify the one city that would be best for a startup (in the startup condition) or the best one for expansion (for the control condition). In order to hide the true purpose of the study, we asked participants to answer a number of demographic variables, including age, gender, ethnicity, along with items assessing the number of entrepreneurship classes they had taken in college as well as the number they planned on taking while in school. The latter served as our dependent measure of entrepreneurial intentions. Answers ranged from 0 to 10, M = 2.38 SD = 1.56.

Results and discussion We performed a Chi-Square test to assess our prediction that doing the same work while categorizing it differently leads to differing levels of entrepreneurial intentions. Specifically, we examined whether condition (startup versus control) influenced students’ desire to take at least one entrepreneurship class. As predicted, those in the startup were more likely to indicate desire to take one or more entrepreneurship classes (96%) than those in the control condition (86%), x²(1) = 4.24, p = .037. Follow-up analysis using the square root of total classes, to eliminate skewness, demonstrated a significant difference between startup (M = 1.52, SD = .52) and control (M = 1.33, SD = .65), F(1,146) = 3.99, p = .048, condition for average intended number of entrepreneurship classes. Experiment 1 demonstrated that the ways in which people categorize the work they do influences their subsequent professional intentions. In particular, those participants who believed they were conducting work for a startup company were more likely to demonstrate interest in pursuing entrepreneurship (by taking more entrepreneurship courses) than individuals who believed they were performing identical work for a large company. To test our proposed mechanism for this effect, we conducted Experiment 2.

14

Experiment 2 In this experiment, we sought to examine the notion that working for a startup increases entrepreneurial efficacy (Hypothesis 1) as well as the idea that entrepreneurial efficacy mediates the relationship between categorizing one’s work as startup-related and entrepreneurial intentions (Hypothesis 2). We tested this theory by studying a population of virtual workers on Amazon’s Mechanical Turk. We hired workers and subjected them to the same task and treatments. We predicted that individuals in the ‘startup’ condition would reported a greater likelihood of wishing to start a business and, more importantly, that this effect would be mediated by the individual’s level of ‘entrepreneurial efficacy.

Participants and design Participants consisted of 99 workers (49 men, 50 women) that we hired from Amazon’s Mechanical Turk (Mturk), ranging in age from 18 to 55 years (M = 29.21, SD = 8.38). Workers agreed to complete the tasks in exchange for $12. We randomly assigned participants to one of two conditions: the startup condition (n = 51) or the large company condition (n = 48). Subsequent to the manipulations and tasks, we asked participants to indicate their entrepreneurial efficacy as well as their intentions to start their own business. Participants then answered several demographic questions.

Manipulations and measures To set up our manipulation, we hired participants to conduct online work. Similar to Experiment 1, workers were informed that they had been recruited to help provide insight to an actual company. Specifically, they read the following:

[Startup condition]: Our company is a small start-up in the high-tech software industry and we are looking for the ideal city in which to locate our new office. We are in the process of compiling background information about various U.S. cities and getting opinions about the livability of the cities from people like you. 15

We will be providing you with the names of 5 cities. Your task is to generate an electronic report for each city so that we can compile everyone's findings into the same file. It should take you around 1 hour to complete the work assignment. You will be given a 4-hour window in which to complete the reports. Once you have completed the reports, fill in your code at the end of the survey and submit it.

[Non-startup condition]: Our company is a large corporation in the high-tech software industry and we are looking for the ideal city in which to expand with a new office. We are in the process of compiling background information about various U.S. cities and getting opinions about the livability of the cities from people like you. We will be providing you with the names of 5 cities. Your task is to generate an electronic report for each city so that we can compile everyone's findings into the same file. It should take you around 1 hour to complete the work assignment. You will be given a 4-hour window in which to complete the reports. Once you have completed the reports, fill in your code at the end of the survey and submit it.

As in Experiment 1, workers were given the names of 5 cities (Pittsburgh, Atlanta, Phoenix, Denver, and Charlotte) and asked to provide answers for a variety of questions based on research conducted on www.wikipedia.org. Questions included topics such as population, colleges and universities, businesses, entertainment opportunities, strengths and weaknesses, and additional notes. They were then asked to rank order the cities in order of most to least desirable. Finally, they were asked to identify the one city that would be best for a startup (in the startup condition) or the best one for expansion (for the control condition). To assess intentions to be an entrepreneur, we used two measures. First, we used a two-item measure of intentions to pursue a startup: “I plan to I plan to take actions toward the creation of a new business within the next year,” and “I would like to pursue starting a new business.” The items were anchored by 1 (“Not at all”) and 7 (“Very much”), (M = 3.37, SD = 1.75, α = .85). As a second measure of entrepreneurial intentions, we assessed the degree to which participants viewed themselves as entrepreneurs. Items were assessed using the above 7-point scales, and included 4 items: “Being an 16

entrepreneur is an important reflection of who I am,” “Overall, being an entrepreneur has very little to do with how I feel about myself”(reverse-scored), “In general, being an entrepreneur is an important part of my self-image,” and “Being an entrepreneur is unimportant to my sense of what kind of person I am” (reverse-scored), (M = 3.34, SD = 1.71, α = .90). We assessed entrepreneurial efficacy using a 5-item scale. The items were anchored by 1 (“Not at all”) and 7 (“Very much”). The items included the following: “When facing entrepreneurship-related tasks, I am certain that I can accomplish them, “I believe I can succeed at most any start-up-related endeavor to which I set my mind,” “I will be able to successfully start a new company if I decide to do so,” “I am confident that I can perform effectively in a start-up organization,” and “I am more able than the average person to effectively create new companies” (M = 4.50, SD = 1.36, α = .93). We next asked participants to answer a number of demographic variables, including age, gender, and number of business they had started in the past (answers ranged from 0 to 2, M = .34, SD = .56).

Results and discussion Three participants took longer than the allotted time (i.e., longer than 8 hours) and we excluded them from our analyses, reducing our total number of participants to 96. We controlled for number of previous start-ups and gender because they were correlated with the dependent variable and the mediator. We used linear regression to assess the relationship between condition and intentions to start a business in the future. As predicted, and consistent with Experiment 1, those in the startup condition were more likely than those in the non-startup condition to report intentions to start a future business, B = .20, t(95) = 2.00, p = .048. Similarly, those in the startup condition were more likely than those in the nonstartup condition to identify as entrepreneurs, B = .19, t(95) = 2.12, p = .037. Thus, we obtained additional support for Hypothesis 1. Next we assessed the effects of our manipulation on entrepreneurial efficacy. According to Hypothesis 2, engaging in work that is categorized as entrepreneurship-related will result in greater entrepreneurial efficacy than engaging in the same work if the work is categorized differently. Consistent 17

with this notion, those in the startup condition were more likely than those in the non-startup condition to report having high entrepreneurial efficacy, B = .24, t(95) = 2.51, p = .014. Thus, we obtained support for Hypothesis 2. Finally, we tested the notion that entrepreneurial efficacy mediates the link between working for a startup and adopting entrepreneurial intentions. As shown in Figures 1 and 2, the mediation was significant for both dependent variables, offering strong support for Hypothesis 3.

Figure 1: Entrepreneurial efficacy as a mediator of the effects of engaging in startup-related work on entrepreneurial intentions. Numbers represent standardized regression coefficients; numbers in parentheses represent simultaneous standardized regression coefficients. As hypothesized, entrepreneurial efficacy fully mediated the effect of startup work on intentions to start a business; the 95% bias-corrected bootstrapped confidence interval (CI) did not include zero, demonstrating significant mediation (CI = .14 to .95).

*

.24

Entrepreneurial Efficacy

***

0.56

***

(0.54 )

.20* (.06) Completed Startup Related Work *

**

p<.05, p<.01,

Intentions to Start a Business

***

p<.001

Figure 2: Entrepreneurial efficacy as a mediator of the effects of engaging in startup-related work on identifying as an entrepreneur. Numbers represent standardized regression coefficients; numbers in parentheses represent simultaneous standardized regression coefficients. As hypothesized, entrepreneurial efficacy fully mediated the effect of startup work on identification as an entrepreneur; the 95% biascorrected bootstrapped confidence interval (CI) did not include zero, demonstrating significant mediation (CI = .10 to .71).

18

*

.24

Entrepreneurial Efficacy

***

0.39

***

(0.37 )

.19* (.10) Completed Startup Related Work *

**

p<.05, p<.01,

Identification as an Entrepreneur

***

p<.001

Experiment 2 provides support for our predictions. Not only did we find additional support for the notion that how people categorize their work influences their subsequent professional intentions, but we found evidence revealing the psychology behind this effect. In particular, participants who believed they were conducting work for a startup company experienced an increase in entrepreneurial efficacy which led, in turn, to greater intentions to start a business.

Discussion and Conclusion We proposed and tested a role-categorization theory of entrepreneurship. We posited that employees who perform tasks categorized as being related to a ‘startup’ will possess more entrepreneurial self-efficacy than employees who perform tasks categorized as being relevant to a ‘large company.’ Possessing greater entrepreneurial self-efficacy leads people to be more likely to indicate that they intend to embark on an entrepreneurial career path. We demonstrate our theory with two experiments. Individuals hired to do a tasks, ostensibly labeled as being relevant for a ‘startup’ were more likely to indicate that they planned to invest their human capital on taking more entrepreneurial courses than those individuals who performed an identical task, but labeled as pertaining to a ‘large corporation.” Freelance contract workers, who were hired to perform tasks labeled as pertaining to a ‘startup’ were more likely to indicate future entrepreneurial intentions than those performing identical tasks categorized as being relevant to a ‘large corporation.” This relationship was mediated by their level of entrepreneurial self-efficacy. We demonstrate a novel and more parsimonious theory to explain the observed phenomenon that employees of small organizations are disproportionately more likely to embark on entrepreneurial 19

ventures than employees of large corporations. While extant theories rely on the belief that the tasks and responsibilities individuals learn, the norms of the organizations, or social ties individuals develop differ between small startups and large organizations – our theory does not require these assumptions. At the extreme, our theory would operate even if all these factors were held constant, and only the type of firm one worked for was labeled differently. The use of experiments in our identification strategy provides two benefits. First, it allows us to eliminate the alternative explanation of selection. Those individuals who plan to take an entrepreneurial career path are also likely to have higher levels of entrepreneurial self-efficacy. So any results we observe that do not account for selection cannot necessarily be attributed to role-categorization. Second, experimental evidence also allows us to demonstrate the particular effect of role-categorization on developing entrepreneurial self-efficacy. However, one criticism could be that our results may not be replicated in the real world. We disagree. We are able to realize effects after having our subjects work for merely an hour on our tasks. In a real entrepreneurial setting, employees are constantly reminded of the fact they are working for a startup, which can be due to things as subtle as the free drinks and food provided to employees to more obvious signals. Given that employees are constantly being reminded of working for a startup, we would expect that this effect may even be stronger outside of the lab. Our findings inform the literature on intrapreneurship – or the development of entrepreneurial ventures within large companies. This stream of work examines how large companies may encourage the entrepreneurial ideas of their employees by developing structures within their firms to create new ventures from internal employees and ideas. Our theory would suggest that one mechanism by which these employees may be able to develop their entrepreneurial self-efficacy by encouraging them to label the experiences and tasks they perform as being ‘entrepreneurial.’

20

References Aldrich, H. E., and M. Ruef. 2006 Organizations Evolving, 2d ed. Thousand Oaks, CA: Sage. Ajzen, I. (2002). Perceived behavioural control, self-efficacy, locus of control, and the theory of planned behavior. Journal of Applied Social Psychology, 32(4), 665-683. Audia, P. G., and C. I. Rider. 2006 ‘‘Entrepreneurs as organizational products revisited.’’ In R. Baum, M. Frese, and R. Baron (eds.), The Psychology of Entrepreneurship: 113–130. Mahwah, NJ: Lawrence Erlbaum Associates. Bandura, A. 1977. Self-efficacy: Toward a unifying theory of behavioral change. Psych. Rev. 84 191– 215. Bandura, A. 1982. Self-efficacy mechanism in human agency. American Psychologist 37:122–147. Baron, R. 1998. Cognitive Mechanisms in Entrepreneurship: Why and when entrepreneurs think differently than other people. Journal of Business Venturing 13, 275–294. Boyd, N.G., and Vozikis, G.S. 1994. The influence of self-efficacy on the development of entrepreneurial intentions and actions. Entrepreneurship Theory and Practice 18:63–90. Burton, M. D., J. B. Sørensen, C. M. Beckman. 2002. Coming from good stock: Career histories and new venture formation. M. Lounsbury, ed. Research in the Sociology of Organizations, Vol. 19. Emerald Publishing, Bingley, UK, 229–262. Busenitz, L., J. Barney. 1997. Differences between entrepreneurs and managers in large organizations; biases and heuristics in strategic decision-making. J. Bus. Venturing 12 9–31. Chen, Chao, Patricia Gene Greene, and Ann Crick. 1998. Does Entrepreneurial Self-Efficacy Distinguish Entrepreneurs from Managers? Journal of Business Venturing 13, 295–316. Cooper, A. C., C. Y. Woo, W. C. Dunkelberg. 1988. Entrepreneurs’ perceived chances for success. J. Bus. Venturing 3 97–109. Dobrev, S. D., and W. P. Barnett 2005 "Organizational roles and the transition to entrepreneurship." Academy of Management Journal, 48: 433-449. Gilboa, I. and Schmeidler, D. (1995) Case-based decision theory. Q. J. Econ. 30, 605–639 Gist, M. E. (1987). Self-efficacy: Implications for organizational behavior and human resource management. Academy of Management Review, 12, 472–485. Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical analysis of its determinants and malleability. Academy of Management Review, 17, 183–211. Gompers, P., J. Lerner, and D. S. Scharf stein 2005 "Entrepreneurial spawning: Public corporations and the genesis of new ventures, 1986 to 1999." Journal of Finance, 60: 577-614. Gonzalez, C. et al. (2003) Instance-based learning in real-time dynamic decision making. Cogn. Sci. 27, 591–635 Halaby, C. N. 2003 "Where job values come from: Family and schooling background, cognitive ability, and gender." American Socio logical Review, 68: 251-278. Heath, Chip and Amos Tversky. 1991. Preference and Belief: Ambiguity and Competence in Choice under Uncertainty. Journal of Risk and Uncertainty, 4:5-28. Hertwig R and Erev I 2009. The description-experience gap in risky choice. Trends Cogn Sci 13(12):517– 523.

21

Hsu, Greta. 2006. “Jacks of all trades and masters of none: Audiences’ reactions to spanning genres in feature film production.” Administrative Science Quarterly 51:420-450. Kolvereid, L. 1996. Prediction of employment status choice intentions. Entrepreneurship Theory and Practice, 21(1), 47-57. Krueger, N., and Dickson, P. R. (1994). How Believing in Ourselves Increases Risk Taking: Perceived Self-Efficacy and Opportunity Recognition. Decision Sciences, 25(3), 385-400. Mata, J., P. Portugal. 1994. Life duration of new firms. J. Indust. Econom. 42 227–246. Medin, D. L., Goldstone, R. L., and Genter,D. (1993). Respects for similarity. Psychological Review, 100(2), 254–278. Medin, D. L., and Schaffer, M. M. (1978). Context theory of classification learning. Psychological Review, 85(3), 207–238. Moore, Don A., John M. Oesch, and Charlene Zietsma. 2007. What Competition? Myopic Self-Focus in Market-Entry Decisions. Organization Science. Vol. 18, No. 3, pp. 440–454. Murphy, G. L. 2002. The big book of concepts. Cambridge, MA: MIT Press Nanda, R., J. B. Sørensen. 2010. Workplace peers and entrepreneurship. Management Sci. 56(7) 1116– 1126. Sørensen, Jesper B. and Magali A. Fassiotto. 2011. Organizations as Fonts of Entrepreneurship. Organization Science, Vol. 22, No. 5, pp. 1322–1331 Shane, S., and S. Venkataraman 2000 "The promise of entrepreneurship as a field of research." Academy of Management Review, 25: 217-226. Shane, S. and R. Khurana. 2003. Bringing individuals back in: The effects of career experience on new firm founding. Industrial and Corporate Change, Vol. 12, 3: 519-543. Sørensen, J. B. 2007. Bureaucracy and entrepreneurship: Workplace effects on entrepreneurial entry. Admin. Sci. Quart. 52(3) 387–412. Saxenian, A. 1994. Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Harvard University Press, Cambridge, MA. Stajkovic, A. D., and Luthans, F. 1998. Self-efficacy and work-related performance: A meta-analysis. Psychological Bulletin, 124(2), 240-261. Stuart, T. E., and W. W. Ding 2006 "When do scientists become entrepreneurs? The social structural antecedents of commercial activity in the life sciences." American Journal of Sociology, 112: 97144. Thornton, P. H. 1999 "The sociology of entrepreneurship." Annual Review of Sociology, 25: 19-46. Zhao, H., Seibert, S. E., & Hills, G. E. (2005). The Mediating Role of Self-Efficacy in the Development of Entrepreneurial Intentions. Journal of Applied Psychology, 90(6), 1265-1272.

22