Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, BUSINESS AND MANAGEMENT ( (c) Oxford University Press USA, 2019. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).



date: 20 June 2019

Tracking the Entrepreneurial Process with the Panel Study of Entrepreneurial Dynamics (PSED) Protocol

Summary and Keywords

In the early 1990s business creation was receiving a great deal of attention after it was clear that new firms were a major source of job creation. There was not, however, reliable data on the prevalence of persons participating in firm creation, what they would do to implement new ventures, or the proportion of start-up efforts that became profitable businesses. This hiatus led to the development of longitudinal studies of the entrepreneurial process; 14 projects have now been implemented in 12 countries.

The Panel Study of Entrepreneurial Dynamics (PSED) protocol was designed to provide estimates of the prevalence of individuals involved in business creation and the presence of pre-profit, start-up ventures; data on the major activities undertaken to implement a new firm; and track the proportion that completed the transition from start-up to profitable new firm. A number of challenges were involved in implementing the research program, including the development of efficient procedures for identifying representative samples of nascent entrepreneurs and criteria for determining the dates of entry into the start-up process, the transition to a profitable business, and disengagement from the initiative.

Data collection is a three-stage process. The initial stage is identifying nascent entrepreneurs in a representative sample of adults. The second are detailed interviews on the start-up team and activity related to creating a new venture. The third stage is follow-up interviews completed to determine the outcome of the start-up efforts. A large number of scholars have been involved in development of the interviews and the PSED data sets have considerable information on the perspectives, activities, and strategies of those involved in the start-up process.

Since the initial data sets were made available 15 years ago, there has been considerable research utilizing PSED data sets. One major finding, however, is that the firm creation process is much more diverse and complicated than had been expected. There are substantial research opportunities to be explored. A review of the major features of the PSED protocol and a summary of the existing data sets provides background that will facilitate additional analysis of the firm creation process. Four data sets (Australia, Sweden, and U.S. PSED I & II) are now in the public domain. Critical features of the start-up process have been consolidated and harmonized in a five-cohort, four country data set which is also available.

Keywords: business creation, entrepreneurship, start-up process, firm births, longitudinal, representative cohorts, nascent entrepreneurs, PSED, Panel Study of Entrepreneurial Dynamics

Introduction: Rational and Objectives

Entrepreneurship and small businesses were receiving a great deal of attention in the early 1990s. Those concerned with the role of new and small firms in economic growth were absorbing the dramatic findings from the longitudinal analysis of businesses identified in a commercial credit rating file. Initial analysis suggested that smaller firms were the major source of new job creation; later assessments clarified that new firms were the major source of job gains. There was an increasing interest in serious study of entrepreneurship among a wide range of scholars. The popular media was giving more attention to entrepreneurship and interest among students was growing.

But most research on entrepreneurship or small businesses was based on retrospective accounts of high potential firms, samples of convenience, or data from cohorts of individuals that were difficult to accept as “business creators,” such as college students in entrepreneurial courses, participants in business planning workshops, or small business owners, many who have been firm managers for decades. Research had been completed that indicated it was possible to account for 60% to 90% of the variation in the prevalence of firm creation (new firms per 100 existing firms) across regions in six countries (Reynolds, Storey, & Westhead, 1994); but how individuals reacted to different contexts to implement firm creation was unknown.

There was almost no reliable information on a number of aspects of the business creation process.

  • What proportion of the human population made an attempt to create a new business?

  • What is actually done to implement a new firm?

  • What proportion actually managed to implement a profitable business?

  • What is different between the start-up efforts that reach profitability and those that disengage from the start-up process?

Answers to these questions required a new approach to exploring the entrepreneurial process.

The most useful research programs would have a number of elements. First, data would be based on representative samples that would represent the adult population and the population of nascent ventures; this would facilitate extrapolation to the national population of adults or nascent ventures. Second, information would be collected on a diverse range of activities and features associated with business creation; this would not only provide a comprehensive description, but facilitate testing a wide range of theoretically relevant hypotheses related to the entrepreneurial process. Third, the progress of the start-up efforts would be tracked to determine the proportion that became operational businesses; this may require a long follow-up period.

Over the past two decades a collection of projects, collectively referred to as the Panel Studies of Entrepreneurial Dynamics (PSED) studies have emerged to provide data that will provide answers to these questions. While these desirable features are simple to describe, the actual implementation—and at a cost that can be justified—turned out to be far more complicated than expected. The journey from the pretests in 1992 has been a dramatic challenge and provided major advances in understanding business creation.

The following material is designed as a guide to the assist those interested in utilizing PSED data for analysis. After a review of the conceptualization of business creation and the history of the research program, a number of issues relevant to analysis are reviewed, followed by comments on the impact of the program and future opportunities.

Conceptualization of Business Creation

A central feature of all conceptions of entrepreneurship is the notion that a new venture is created. As new ventures are created by people, the PSED conceptualization presented in Figure 1 emphasizes individuals that take action to start a new firm as they enter the start-up process. As indicated by the circles on the left, these may either be independent individuals acting on their own, nascent entrepreneurs, or those that are taking action as part of their regular job assignment, nascent intrapreneurs.

Tracking the Entrepreneurial Process with the Panel Study of Entrepreneurial Dynamics (PSED) ProtocolClick to view larger

Figure 1. Conceptualization of the Entrepreneurial Process.

Once they enter the start-up process, two outcomes are expected: either they manage to launch a new firm or they disengage. The focus of the research design is to identify those that have entered the process, determine what they do in the start-up process, and track the outcome of their efforts. The entire process occurs in the context of a distinctive social, political, historical, and economic context. There have been a number of conceptual and operational challenges in identifying those active in business creation and developing criteria for the major transitions, such as entry into the process, disengagement, and the transition to a new firm.

Development of the Research Program

The PSED initiative is a research program that consists of a series of closely related projects that share major features. The adult population is screened to locate those active in business creation, they complete a detailed initial interview, and follow-up interviews track progress in business creation and the outcome for the nascent venture. While there continues to be adjustments and enhancements in the operational procedures and interview schedules, the basic design has not changed. Major events in the development of the program are presented in Table 1. There have been three distinct waves of PSED projects.

Table 1. Major Events in the PSED Program Development






Initial Wisconsin Study

Demonstrated feasibility.

Reynolds & White (1993, 1997).


Two proposals submitted to NSF, not funded

Enhanced research design.


Development of Entrepreneurial Research Consortium

34 units, 120+ scholars indicated serious interest in project.

Gartner et al. (2004).


Norwegian Screening

Demonstrated cross national viability.

Reynolds & Curtin (2011), Chap 8.


U.S. PSED I Pilot study

Expanded screening to two items, revised data collection. procedures


U.S. PSED I screening

Confirmed significance.

Reynolds (2000).


Swedish screening

Reynolds & Curtin. (2011), Chap. 9


Netherlands I screening

Reynolds & Curtin. (2011), Chap. 7


U.S. PSED I female oversample


U.S. PSED I minority oversample


Global Entrepreneurship Monitor Implements screening

Expansion of screening to three items.

Reynolds et al. (2005).


Canadian screening

Reynolds & Curtin. (2011), Chap. 3


Netherlands II screening

Reynolds & Curtin. (2011), Chap. 7


U.S. PSED II screening

Reynolds & Curtin. (2009), Appendix A.


Germany screening

Reynolds & Curtin. (2011), Chap. 5


Latvia screening

Reynolds & Curtin. (2011), Chap. 6


Australia screening

Reynolds & Curtin. (2011), Chap. 2


China screening

Reynolds & Curtin. (2011), Chap. 4


U.K. follow-up pretest

Pretest of revised interview schedule.

Reynolds, Hart, & Mickiewicz (2014).


Finland screening


Switzerland screening


Spain screening


U.K. screenings


Five cohort data set created

Facilitates cross national assessments.

Reynolds et al. (2016).

The first published results were from an initial effort in Wisconsin, discussed in a report to a state agency and summarized in a research monograph (Reynolds & White, 1993, 1997). This first project demonstrated the feasibility of the design and led to the development of two detailed proposals were submitted to the National Science Foundation from the University of Michigan Institute for Social Research (ISR). While neither proposal received financial support, this exercise substantially improved the research protocol, increased confidence among entrepreneurial scholars that the project was worthwhile, and subsequently led to the U.S. effort being housed at the University of Michigan ISR.

A project development team (Nancy Carter, Bill Gartner, and Paul Reynolds) then organized an exploratory discussion of a collective effort at a research conference in 1995. The result, after wide distribution of invitations to participate, was the development of the Entrepreneurship Research Consortium (ERC), which consisted of 34 entrepreneurial research units representing over 120 individual scholars. Each participating unit provided financial support with the understanding they would participate in the project design and share access to the resulting data sets.1

There was agreement on the basic design and considerable interest in adding modules to the interview schedules. The effort to provide a comprehensive description of the business creation process and provide a wide range of participants with opportunities for analysis led to a 60 minute phone interview and a 12 page self-completed questionnaire. To identify the unique features of nascent entrepreneurs the design included data collection from a comparison group, adults not active in business creation. Complications associated with developing a cost-effective screening procedure delayed implementation of the project and full-scale national screening was not implemented until 1999.

By this time interest in the effort had increased and a proposal to expand screening to increase the sample of women nascent entrepreneurs was funded; the screening effort was doubled, from 15,000 to 30,000 with only women nascent entrepreneurs retained from the additional contacts. A second supplement to expand the cohort of African-American and Hispanic nascent entrepreneurs was funded which led to a further additional 30,000 screening interviews; in this cohort there was no follow-up on Whites identified as nascent entrepreneurs.

At the same time the U.S. PSED I effort was in development other national projects were implemented. The Norwegian project was the first to complete national screening in 1998. A Swedish project, comparable in scope to the U.S. PSED I effort, completed screening in 1998. The first Netherlands effort completed screening in 1998 and a Canadian effort completed screening in 2000. A second Netherlands effort was implemented in 2001.

These initial sets of projects share several features that were later adjusted. Most significantly, it was widely assumed that the majority of nascent ventures would reach a resolution in 12 or 24 months, so they were designed with a limited number of follow-up interviews. It has been discovered that the median time is much longer, and later studies have included follow-ups of up to 75 months. Further, there have been adjustments in the critical items used in the screening interviews.

Based on the success of cross-national applications of the screening procedures in the first set of PSED projects, this feature was adopted for the cross-national comparisons of entrepreneurial activity emphasized in the Global Entrepreneurship Monitor program (Reynolds et al., 2005). This involved an expansion to three screening items with the application of additional criteria to confirm that pre-profit nascent ventures were being identified.

Many of these adjustments were implemented in the second wave of projects, which included the U.S. PSED II in 2005, a German program utilizing the GEM screening efforts in 2006, Latvia and Australia in 2007, and a Chinese screening in eight major cities in 2009. Much of the U.S. PSED II interview schedules were adopted for the Australian, Chinese, and Latvian efforts, although each project included material unique to their own situation.

The third wave of projects reflected the experience of a U.K. pretest using a cohort identified in 2012 with a single follow up in 2013. Use of a revised and simplified interview schedule, with a much more efficient inventory of start-up activities, was developed for a cross-European effort (Reynolds, Hart, & Mickiewitz, 2014). Screening interviews were initiated in Finland, Spain, Switzerland, and the United Kingdom utilizing major elements of this interview schedule. As of 2016 this data is now being processed and follow-up interviews initiated.

A final development is the creation of a harmonized data set including five cohorts from four countries, completed in 2016. This effort involves harmonizing the presence and timing of start-up activities for the Australian, Chinese, Swedish, and U.S. PSED I and II data sets. This provides standardized measures of the time in process and outcome measures for a total cohort of 3,000 nascent ventures.

The implementation of 14 projects in 12 countries is good news. On the other hand, data from only four efforts are in the public domain. The Australian, Swedish (in English), and U.S.

PSED I and II projects are on public websites and there is no charge for access. Fortunately, these are the four projects with the largest amount of data and three to five follow-up interviews. Further, three of these projects, the Australian, Swedish, and U. S. PSED I efforts, collected data on comparison groups of those not involved in business creation. The five cohort data set is also on a public website. Details regarding access are provided in the appendix.

The PSED research program has been relatively expensive. The three U.S. projects relied on substantial external funding. The initial Wisconsin project, a joint effort by five academic centers coordinated through Marquette University, was sponsored by the Wisconsin Innovation Network, Inc. and the Wisconsin Housing and Economic Development Authority. The first national study, U.S. PSED I, was coordinated through Babson College and initially financed with funds provided by the 34 units of the Entrepreneurship Research Consortium for the screening, initial detailed interviews and 12 month follow-up interviews. Additional support from the Ewing Marion Kauffman Foundation provided for the 24 and 36 month follow-up interviews and two NSF grants financed the oversample for women (NSF SBR-9809841, Nancy Carter, principal investigator) and the oversample for minorities (NSF SBR-9905255, Patricia Green, principal investigator). Sponsorship of the second U.S. national study (PSED II) for the initial screening, initial detailed interviews, and the 12 and 24 month follow-ups was provided by the Kauffman Foundation; the 36, 48, and 60 month follow-ups were supported by the National Science Foundation (NSF SES-0818366 & SES-0919772, Richard Curtin, principal investigator).

Projects in other countries have also required substantial support, usually from major foundations or national agencies. The PSED program has involved millions of dollars in research support and contributions from hundreds of scholars; the dramatic findings based on representative samples of nascent ventures has more than justified this investment.

Screening Procedures

There are two major challenges in locating a representative sample of those active in business creation. The first is what strategy is most efficient to locate those involved in business creation. Because the prevalence of nascent entrepreneurs is generally low, about 5 per 100, 20 individuals must be contacted to locate one nascent entrepreneur. As the fixed cost to identify an adult that represents a population is relatively high, there are major advantages to sharing this expense with other projects. The most cost-effective procedure was to complete the screening with commercial omnibus or multi-client surveys. Many firms interview several representative samples of 1,000 adults each week. As the majority of respondents will answer “no” to the critical items in the short screening module, 90% of the respondents require less than a minute of interview time. The socio-demographic characteristics and case weights are provided for all respondents as part of the service to a multi-client omnibus interview.

The second is the strategy for locating nascent entrepreneurs. Because in all languages there is an imprecise relationship between words to describe business creation and what people are actually doing, the strategy for identifying nascent entrepreneurs involved two stages. The first stage is represented by the screening items. As more projects were completed, these were expanded to capture all individuals that might be involved in business creation. The second stage involved eliminating individuals that did not meet criteria established for nascent entrepreneurs. These criteria were an expectation of ownership, recent activity to implement a new firm, and a lack of profits.

The screening module was adjusted as new information was developed over successive projects. Table 2 represents the development of these items in the projects representing the U.S. program. As can be seen, the initial 1993 effort in Wisconsin involved a single item. Realizing that it may miss those involved in business creation for their employer, a second item was added for the 1999 effort, U.S. PSED I. After this project was completed, the same screening procedure was implemented for a complementary project, the Global Entrepreneurship Monitor. Analysis of data from this multi-national effort indicated that many that considered themselves owning and managing a business had yet to reach initial profits. As such, they would be considered as involved in nascent ventures. This led to the use of a three item module for the U.S. PSED II.2 Further, all items were embellished to clarify that any sales of a good or service to others would be considered a business activity.

Table 2. Screening Items Utilized in U.S. Projects: 1992–2005



Item content



Are you, alone or with others, now trying to start a business?



Are you, alone or with others, now trying to start a business?


Are you, alone or with others, now trying to start a new business or new venture for your employer? An effort that is part of your job assignment?



Are you, alone or with others, currently trying to start a new business, including any self-employment or selling any goods or services to others?


Are you, alone or with others, currently trying to start a new business or a new venture for your employer, an effort that is part of your normal work?


Are you, alone or with others, currently the owner or a business you help manage, including self-employment or selling any goods or services to others?

Source: Wisconsin Entrepreneurial Climate Study (Reynolds & White, 1993). B) U.S. PSED I screening interview schedule. C) U.S. PSED II screening interview schedule.

The second stage of the screening involved applying three additional criteria. Any person that answers “yes” to one or more of the screening items received additional questions. Taken from the U.S. PSED II, these three additional criteria are represented in the items in Table 3. The first (FF2) involved evidence of recent activity to create a business, only those that answered yes were retained. The second (FF3) involves expectation of some ownership of the new firm. Again, only those that answered “yes” were retained. The third criterion is represented by three items (FF4, FF5, and FF6). Only those that answered “no” to at least one of these items were retained. Those that answered “yes” to all three were assumed to be involved in a venture that had become a profitable business.

Table 3. Nascent Entrepreneur Criteria: U.S. PSED II


Item content


Over the past twelve months have your done anything to help start a new business, such as looking for equipment or a location, organizing a start-up team, working or a business plan, beginning to save money, or any other activity that would help launch a business?


Will you personally own all, part, or none of this new business?


Has this business received any money, income, or fees for more than six of the past twelve months?


Has the monthly revenue been more than monthly expenses for more than six of the past twelve months?


Did the monthly expense include salaries or wages for the owners active in managing the business for more than six of the past twelve months?

The basic strategy of the screening procedure is to identify all those that might quality as nascent entrepreneurs and then exclude those that would not meet all the criteria. About three fourths of those initially identified meet the criteria as nascent entrepreneurs.

Initial and Follow-Up Interviews

Developing the modules for the initial and follow-up interviews involved considerable effort. This reflects, in part, the large number of individuals involved in the design of the U. S. PSED I project, which was initially sponsored by 34 different research center representing over 120 scholars. The effort to accommodate a wide range of research issues resulted in a U.S. PSED I 60 minute phone interview schedule with 22 modules and a 12 page self-administered mail-back questionnaire with 14 modules. The U.S. PSED I modules are listed in Table 4. The teams responsible for the different modules have described their rationale in detail (Gartner et al., 2004).

Table 4. U.S. PSED I Phone and Mail Interview Schedule Modules

Module items

Phone Interview


Open ended items on nature of start-up, reasons, expectations


Start-up activities


Firm registration


Nature of start-up effort: Legal form, economic sector, location


Start-up team: composition, background, contributions


Social network: Scope, background, and contributions


Start-up funding: Requirements and expectations


Assessment of customers, markets, competition


Competitive strategy


Knowledge, use of assistance programs


Future expectations sales, jobs


Personal decision-making style


Current labor force activity


Work, career experiences


Residential tenure, migration (R and parents)


Respondent birth order


Parental in-migration


Parents, family business background


Household structure


Household income


Household net worth


Reaction to participation (Hawthorn effect)


Self-administered mail questionnaire


Opportunity recognition, information gathering assessment


Entrepreneurial climate scale items


Start-up problems expectations


Economic sector, community context assessment


Financial management expectations


Work, training background


Item inventory: Reasons for starting a new firm


Assessment of risk preferences


Personal work experience details


Individual problem-solving orientation


Self-assessment inventory I: Work and orientation to start-ups


Self-assessment inventory II: Generalized personal domains


Time-use diary, recent work-day and day-off


Assessment of new business success in the U.S.


Work participation history, previous 11 years

The U.S. PSED II project used a single phone interview schedule, the organization is provided in Table 5, which also indicates the structure of the follow-up interviews. While some of the material included in the U.S. PSED I schedule was not included in the U.S. PSED II schedule, about four-fifths of the U.S. PSED II schedule items were taken from the U.S. PSED I interview schedules.

Table 5. U.S. PSED II Phone Interview Schedule Modules by Wave

Topic Modules


Wave A

Wave B (1,2)

Wave C to F (1,2)

Screening questions


Assessment of criteria for nascent entrepreneur




A.1: Why involved, business opportunity (open ended)


A.2: Confirm same business activity



A.3: Determine status: new firm, quit, continue



B: Type of business, location




C: Legal form




D: Start-up activities




E.1: Start-up finances, entry into firm registries (3)




E.2: Confirm quit, exit interview



F: Orientations toward competition




G: Owners, key non-owners, & helpers inventory




H: Owner demographics




J: Relationships among owners




K: Juristic (legal entity) owners




M: Key non-owner demographics




N: Helper demographics




P: Community resources, support for new firms




Q: Informal start-up financial support




R: Legal entity start-up investments, debts, net worth




S: Competitive strategy and target markets




T: Growth expectations




U.1: Respondent’s motivation


U.2: Employment structure (3)



V.1: Expense structure: summary (3)


V.2: Expense structure: detailed (3)


X: Respondent’s career background




Y: Respondent’s self-descriptions


Z: Respondent & household socio-demographics




Notes: (1) After wave A, modules are provided to all respondents, only those that auit, or those with a new firm (NF), or still active in the start-up process (SU).

(2) After initial interview, modules are repeated to capture changes or new information about the activity or details on the current status.

(3) Based on Kauffman Firm Survey interview schedule (Mathematica Policy Research, 2007).

Both presentations are basically a summary of the content of the interviews, the full details are provided in the specific questionnaires, available on the U.S. PSED website. Because many items are included in multiple interview waves, four for U.S. PSED I and six for U.S. PSED II, the total number of variables is over 5,000 for each data set. The codebooks clarify the response patterns to each item in the different data collection waves.

Each respondent, however, did not respond to all items, as several features varied depending on the situation of the nascent venture. First, the number of start-up activities initiated to implement the firm could vary. Second, those that involved multiple start-up team members involved questions about each member and their interrelationship. Third, the financial structure could vary in complexity, and more complex structures involved more questions about details.

In addition, different modules were provided in the follow-up interviews depending on the status of the nascent venture. This is illustrated in the overview of the U.S. PSED II modules in Table 5. No respondent completed all the sections of the follow-up interviews. The modules to be completed depended on the status of the start-up activity at the start of the interview.

The U.S. PSED projects provide the most comprehensive, multi-faceted description of the business creation process that has yet to be developed. Even so, other topics have been pursued in other PSED projects, suggesting that there is more to explore regarding the entrepreneurial process.

Cohort Development and Follow-Up Success

A summary of the data collection success for U.S. PSED I and II is provided in Table 6. The first rows indicate the timing of the initial screenings. This occurred over 20 months for the three cohorts developed for U.S. PSED I and over four months for the single cohort for U.S. PSED II. Those identified as potential nascent entrepreneurs are listed in the first set of rows.

The low total for U.S. PSED I reflects the exclusion of men in the female oversample and the exclusion of Whites in the minority oversample. The loss between screening identified nascents and those completing the Wave 1 interviews occurs for several reasons. First is the inability to contact some respondents identified in the screening for the Wave 1 interviews. Second, when some of the details regarding the respondents’ situation were verified in the Wave 1 interviews it turned out they did not meet the criteria for a nascent entrepreneur.

Table 6. Cohort Development and Interview Counts: U.S. PSED I & II




Screening Initiated

Jul 98

Oct 05

Screening Completed

Jan 00

Jan 06

Nascent screening sample




Screening identified nascents




Completed Wave 1 interview




Completed Wave 2 interview




Completed Wave 3 interview




Completed Wave 4 interview




Completed Wave 5 interview



Completed Wave 6 interview



One or more follow-up interviews




One or more follow-up interviews, %




Wave 1 to Wave 2 lag: Mths (avg)



Wave 1 to Wave 3 lag: Mths (avg)



Wave 1 to Wave 4 lag: Mths (avg)



Wave 1 to Wave 5 lag: Mths (avg


Wave 1 to Wave 6 lag: Mths (avg)


Comparison group: Population


Comparison group: Minority


As can be seen, there is attrition following the Wave 1 interview. One source is the inability to contact the respondent for the Wave 2 follow-up interview. But after the second wave, another factor is even more important, those that had discontinued work on the nascent venture were dropped from the project. The focus of the PSED projects has been on the outcomes of the business creation process; tracking the career path of discouraged entrepreneurs was left for another effort. The increase in completed interviews for the U.S. PSED I in the third and fourth follow-ups reflects the success at locating and interviewing respondents that were not contacted for a second wave interview in the initial effort.

A primary objective of the projects was to determine the outcome of efforts to start a new business. The proportion with at least one follow-up interview, which averaged 88.3% for the two projects, indicates considerable success with regards to this purpose.

The fifth set of rows provides information on the lag between interviews. While the research plan was to complete follow-up interviews every 12 months, delays in implementing field work and the long field periods with multiple call-backs to get high response rates led to somewhat longer average times. On the other hand, the entire data collection period covers an arbitrary, case-specific window in the start-up process. Adjustments to create case specific timelines minimizes the significance of variation in the follow-up interview time lags.

The bottom rows of Table 6 provide a count of the comparison groups associated with U.S. PSED I.

Issues in Assessment

There are a number of issues associated with utilizing PSED data sets to improve understanding of the firm creation or entrepreneurial process. This includes the lack of consensus on a distinction between business creation and entrepreneurship; the scope and clarity regarding the independent variables that may affect business creation; tracking the presence and timing of start-up activities; operational criteria for the major transitions associated with the business creation process including entry into the start-up process, the transition to a new firm, and disengagement; how to establish case weights for individual nascent entrepreneurs versus nascent ventures; identifying active nascent entrepreneurs in the data set; the unique relationship of the minority cohort in the U.S. PSED I data set; and the availability of a harmonized five cohort data set.

New Firm Creation versus Entrepreneurship

Few popular terms are associated with so much diversity in meaning as “entrepreneurship.” The connotations include a preference for risk, a source of innovation, a source of new businesses or ventures, a focus on growth or market disruption, or anything that is radically different—whether or not it involves business activity (Pozen, 2008). Most conceptualizations related to economic activity include creation of a business venture. The complication that develops is that the vast majority of business creation is not associated with high levels of risk, extreme growth potential, or dramatic new innovations. Indeed, about half of efforts to create new business ventures appear to be one-person operations best described as self-employment.

On the other hand, most of the research that has documented the positive benefits of business creation—and its relationship to job growth, economic adaptation and innovation, and improved sector productivity—has been based on tracking business creation dominated by small-scale initiatives replicating existing business activity.

While the diversity in conceptualizations associated with the term “entrepreneurship” may not be resolved in the near future. The U.S. PSED data sets include items that allow identifying a number of different sub-samples, including nascent ventures associated with a growth orientation, expected to have a market impact, considered as “high technology,” or even initiated by nascent entrepreneurs with a risk orientation. Depending on the focus of the assessment, different sub-samples of pre-profit start-ups can be identified for analysis.

For example, a four-point scale representing technological emphasis is constructed from three dichotomous items. Is research and development spending a major priority? Is the firm using technology not available five years ago? Is the business considered a high technology firm?

As shown in Table 7 about 6% answered “yes” to all three items, reflecting a maximum technological focus, and 46% provided three “no” responses, suggesting no technological focus.

Table 7. Technological Emphasis: U.S. PSED I, II Cohort

High technology focus


Moderate technology focus


Little technology focus


No technology focus



Source: Reynolds and Curtin (2009), p. 197, Table 7.10.

An assessment of expected market impact can be measured with a similar strategy. Three items related to market impact were included in the U.S. PSED II interview. Will all, some, or none of your potential customers consider this product or service new and unfamiliar? Right now, are there many, few, or no other businesses offering the same products or services to your potential customers? Were the technologies or procedures required for this product or service generally available more than a year ago? A four point scale can be constructed and the results are provided in Table 8. Once again, about 5% anticipate a major impact on the market, about 10% a major or moderate impact, and about half no impact on the markets.

Table 8. Expected Market Impact: U.S. PSED II Cohort

Major impact on markets


Moderate impact on markets


Little impact on markets


No impact on markets



Source: Reynolds and Curtin (2009), p. 197, Table 7.10.

The range in growth expectations is reflected in projections of employment and sales after five years, the distribution is provided in Table 9. In this case, the jobs and sales expected in the first and fifth full year of operation are provided. The average annual growth rate is provided in the right column. About 3% plan for 100 or more jobs in five years and 13% expect annual sales in excess of $1 million. Note that the average annual growth is higher for those ventures expecting a lower level of activity in the first year.

Table 9. Expected Growth in Jobs and Annual Sales: US PSED I, II Cohorts


Expected in first year

Expected in fifth year

Average annual growth





1-5 Jobs




6-10 Jobs




12-25 jobs




26-100 jobs




100 or more jobs







Annual sales

Up to $50,000




















$5,000,000 and up







Source: Reynolds and Curtin (2009), p. 198, Table 7.11.

Two things are clear from this assessment. First, only a small proportion of nascent ventures are expected, by those putting them in place, to represent high technology, have a major impact on the market, or become substantial ventures after five years of operation. Second, the U.S. PSED I and II data sets have, in their representative samples, examples of nascent ventures expected to become substantial businesses, emphasizing high technology, and anticipating an impact on the markets.

Independent Variables

Most analysis involves a distinction between causes, or independent variables, and effects, or the dependent variables. The good news about the PSED data sets is that the most popular independent variables such as the nature of the entrepreneurs or start-up team, competitive strategy, perception of the context and the competition, expectations for an access to financial support, use of assistance programs, expectations for the new venture, measures of attitudes and personality traits, work and career backgrounds, family and household context, and the like are measured with traditional, straight forward procedures. There is little ambiguity about the concepts or operational definitions, the teams responsible for the rationale and design of different modules capturing independent variables have provided detailed descriptions (Gartner et al., 2004).

This is not the case with regards to describing the business creation process. There are no widely accepted conceptual or operational definitions of two critical features of the process, when individuals or a team can be considered to have initiated business creation or when the start-up venture has become an established business. Fortunately, the diversity of data assembled in the PSED protocol allows for a variety of criteria for both of these transitions.

Tracking Start-Up Activity

The U.S. PSED protocol provides considerable information on the activities pursued in the start-up process. Those involved in the design of U.S. PSED I developed a list of activities considered significant for business creation to be asked about each nascent venture; this was expanded for U.S. PSED II. For each activity the respondent is asked if work has been initiated and if it has, they provide the month and year this occurred. Many start-up activities are one-time events while others, such as developing a business plan or hiring employees, are ongoing activities where there is no clear ending date; hence, a date for completing most start-up activities is not obtained.

Table 10 provides a listing, ranked by prevalence, of the activities initiated to facilitate new firm creation reported across all waves of interviews for the two U.S. cohorts. The dates for initiation of 29 start-up activities have been assembled for the U.S. PSED I cohort and 52 for the U.S. PSED II cohort. The two right columns provide the unweighted averages for each cohort. They are ranked ordered with the most frequently reported at the top of Table 10; the first column provides the variable names; these variables names have been harmonized in a U.S. PSED consolidated data set (Reynolds & Curtin, 2011) and a five cohort harmonized data set (Reynolds et al., 2016). This unique resource facilitates a description of the sequence and timing of efforts to implement a new firm, which can vary dramatically from one nascent venture to the next. It also facilitated development of criteria for a date of entry in to the business creation process.

Table 10. Start-Up Activities Prevalence: U.S. PSED I and II Cohorts





Total Cases






Serious thought




MBR 1: Invested own money




Began business plan




Developed model, prototype




Purchased materials, supplies, parts




Define markets to enter




Promote products or services




Sales, income, or revenue




Leased, acquired major assets




Talk to customers




Financial projections




MBR 1: Full time start-up work




MBR 1: Saving money to invest in firm




Phone book listing for business




Established bank account for firm




Obtained supplier credit




Began to organize start-up team




First use of physical space




Collect information on competition




Hire employee




Determine regulatory requirements




Asked for formal funding




Cash flow covers expenses, not owners




Federal income taxes




Federal social security payment (U.S.)




Legal form registered




Acquired registration number




Hire accountant




Took class, seminar, workshop




Acquired doing business as name




Patent, copyright, trademark filing




Business phone line established




Joined trade association




Got initial formal financing




Arranged child care, housekeeping




Obtained liability insurance




Filed state unemployment ins (U.S.)




Contact with helping program




Hire lawyer




Business plan finished




Model, prototype fully developed




MBR 1: Signed ownership agreement




Proprietary technology developed




MBR 2: Invested own money




Investment in legal business




Know listed in Dun & Bradstreet (US)




MBR 2: Signed ownership agreement




MBR 2: Full time start-up work




MBR 3: Invested own money




Received patent, copyright, trademark




MBR 3: Signed ownership agreement




MBR 4: Signed ownership agreement




MBR 4: Invested own money




MBR 3: Full time start-up work




MBR 5: Signed ownership agreement




MBR 5: Invested own money




MBR 4: Full time start-up work




MBR 5: Full time start-up work



Source: (1) Based on common names in the consolidated data sets.

(2) NA = not asked in the interview.

Criteria for Entry into the Start-Up Process

Commitments to pursue business creation may emerge over time. This is similar to two people that may gradually realize that they should marry. They may announce a firm date of an engagement, a public commitment to enter the “marriage planning” process. There will be, in many cases, a precise date when a marriage contract is “executed.” But these are, in terms of the development of the relationship, important milestones but may not represent dramatic changes in the relationship. While this is also true of firm creation, precise descriptions of the process are facilitated by a specific date of entry into the start-up process.

Precisely defined, widely accepted criteria for dating entry into the start-up process do not exist. The initial screening or first detailed interview is completed at an arbitrary point in the process. Some may have been involved for weeks and others for years. These are very imprecise indicators of initiating business creation.

The availability of dates for initiating a range of activities in the initial interview has led to a variety of strategies for defining the date of entry into the process. This has included the date of the screening interview, first activity initiated, and date of initial serious thought, or when the business idea developed.3 One assessment utilized the first activity after initial serious thought.4 This criterion would shift date of entry for that 26% of the cases that initiated activity before serious thought to a later date, treat that 32% that report serious thought and an activity in the same month as an entry date, and delay the date for the 41% that report an activity one or more months after serious thought.5

Given that many more potential entrepreneurs think about—and talk about—starting a business than actually pursue firm creation, another strategy is to ignore reports of “initial serious thought” and focus on reports of behavior, represented by the implementation of start-up activities. Assuming that a minimal level of intensity would be an indicator of seriousness of intent, start-up activities are arranged by their occurrence for each nascent venture. The time lags between each successive pair of events are scanned to locate the first two that occur within 12 months. The date of the first of the earliest pair to occur within 12 months can be considered the date of entry into business creation. With this criterion, used in the following assessment, serious thought is reported prior to the date of entry for 53% of the cases, in the same month for 28%, and after date of entry for 19%. For 1% of the cases serious thought occurs 4 or more years after the data of entry.6

While the PSED procedure, screening for active nascent entrepreneurs followed by details about current efforts, does provide a representative sample, it also leads to a potential complication. Those that take a long time to implement a new firm are more likely to be identified as active nascent entrepreneurs and included in the data set. There may be an underrepresentation of those that move quickly to implement a new firm. One solution has been to eliminate cases with an entry date well before the initial screening interview, such as cases that entered the process in the previous nine months7 or two years;8 either strategy leads to a substantial loss of cases. Excluding those cases with a start date 120 months (10 years) before the first interview eliminates the extreme “long in process cases” but retains most of the sample.9

Criteria for New Firm Birth

There are at least five ways that a “firm birth” has been conceptualized and measured: founder’s interpretations, as a new agent engaged in economic transactions, initial listing in any registry, input of work or labor contribution, or initial profits. Different operational criteria can have a major impact on the descriptions of the entrepreneurial process and the post birth survival trajectories (Schoonhoven et al., 2009; Reynolds, 2015).

Letting the firm founder, or founders, define the birth of the firm may be the most common procedure. It is reflected in answers to the question “When was this firm born (founded, created, established, an operating business, etc.)?”10 Widely used by journalists, field observers, and others not concerned with precise analysis of the start-up process, it has the advantage of being easy to implement. The disadvantage, of course, is that every founder has a different conception of “founding, creation, establishment, or an operating business” which leads to substantial differences when firms are compared. In mass screening to identify nascent entrepreneurs and firm owners it was found that many who claimed their firm was a “start-up” had been profitable for some time and many who said they were managing a business had never achieved profits, they were actually in the “pre-profit” stage. Additional questions were required to develop a precise indicator of these ventures’ stage in the firm life course.11

The focus of industrial organization theory, as pursued by economists, is on ventures that participate in markets, either as a purchaser of supplies, facilities, labor, or capital or as one that sells a good or service. From this perspective, a firm—considered a productive entity that will produce a good or service—is “born” when it completes its first economic transaction. Either the first purchase, such as business cards or a dedicated phone line or a website, or the first sale, no matter how small, will define the birth of a new economic actor that can participate in markets. Most of the research, however, focuses on initial entry into markets as a producer. There is a wealth of studies related to the effects of new entries into various markets. New entries are identified by government sponsored data sets on industry participants and the major focus is on the effects on the markets themselves, such as productivity, innovation, and average profitability (Geroski, 1995; Klepper & Miller, 1995). While relatively precise as a conceptual or theoretical concept, the absence of detailed data on new buyers or sellers in the multitude of markets that make up modern economies has precluded much research reflecting this concept of a firm birth.

The most widely used measure of firm birth in the research community may be the initial entry into a business registry or listing. Conceptually, it can be argued that listing in a registry is a public recognition of a new entity. This would be similar to a human birth certificate or a wedding license, both of which may be considered a public listing of a new citizen or a new social unit. There is a wealth of research on new firms that has utilized initial listing as the birth marker, including new listings in phone books (Shapero & Giglierano, 1982), state incorporation records, Dun & Bradstreet credit rating files,12 state unemployment insurance (ES202 files),13 initial federal social security (FICA) filings,14 initial Schedule C federal tax returns, or initial listings in trade association directories.15 This practice reflects the assumption that if a new venture has been added to the registry, they must be an active business.

If a person decides to create a new business, and devotes personal time to this effort, it is possible to assume that this represents the birth of a new business venture. Reflecting this conceptualization, it is reasonable to track reports of a shift in labor force effort from paid employment to self-employment, assuming self-employment involves establishing a new venture. This is the basis for the reports of entrepreneurial activity promoted as the Kauffman Index of Entrepreneurial Activity (KIEA).16 Using the data collected by the monthly U.S.

Current Population Survey, it is possible to identify changes in labor force activity from one month to the next. Those who do not report working on their own business are identified in one month. If they report more than 15 hours of work per week on their own business in the next month and this is more than work reported in a wage or salary job, they are considered an “entrepreneur” for that one month. This measure is clearly using labor input to identify business creation and assumes that business creation is a form of job switching. Job switching can occur in a matter of days, leading the assumption that job creation can occur within one month. The KIEA clearly captures one aspect of the firm creation process. However, as it does not reflect the majority of the firm creation activity it is best considered an indicator, rather than a comprehensive index.17

Those establishing new firms, as individuals or in teams, share one common goal, to achieve profitability. The primary objective of all commercial firm creation is to make money.18 This immediately leads to considering the transition from start-up to new firm as occurring when the firm achieves a period of profitability. Initial profit reflects a period in which income or revenue is greater than all expenses, including salaries for the owner-managers. The items used to measure initial profitability in U.S. PSED I and II are provided in Table 11.

Table 11. Identifying a New Firm Birth: U.S. PSED I & II


Firm Birth Criteria

Interview Items: Dates


Salaries for managers who are also owners included in the expenses.

Q166, Q166A/R624, R624A/S624, S624A/T624, T624a

U.S. PSED II (Primary)

Are the salaries for owners who are also managers included in the computation of monthly expenses?

AE16, AE16A/BE16, BE16A/CE16, CE16A/DE16, CE16A/EE16, EE16A/FE16, FE16A

U.S. PSED II (secondary)

If missing values for primary variables on initial profits.

Were the salaries or wages of the owners who were active in managing the business included in the monthly expenses for more than six of the past twelve months?


This criteria for identifying initial profits has two important roles in the PSED protocols. First, they are used in the follow-up interviews to determine the status of the business. Those in profit complete different interview modules than those considered still in the start-up process.

Second, they are used in the analysis to identify those cases considered to have reached a profitable outcome.

Criteria for Disengagement

Disengagement from the start-up effort is usually associated with the termination of the start-up effort. In a small proportion of team start-ups, the respondent may have disengaged from the team and the other members continue to work on the nascent venture. For this reason, two questions are utilized to determine the current status of the respondent versus the current status of the nascent venture. They are summarized in Table 12.

Table 12. Identifying Disengagement: U.S. PSED I & II


Disengagement Criteria

Interview Items


Are you still involved as an owner of this new firm start-up or business?


No longer worked on by anyone.

R537, R537A/S537, S537A/T537, T537A


Less than 160 hours devoted to venture in past 12 months, expect less than 80 hours of work on the venture in the next 6 months, no longer a major focus of the work career.

BA43A, BA43B/CA43A, CA43B/DA43A, DA43B/EA43A, EA43B/FA43A, FA43B

Others still involved? Current status of initiative?

BE51, BE56/CE51, CE56/DE51, DE56/EE51, EE56/FE51, FE56

In U.S. PSED I only one item is utilized to determine if the respondent is still involved, and it reflects their definition of quitting. They are then asked if others are still working on implementing the nascent venture. In U.S. PSED II multiple criteria were involved to measure personal disengagement: less than 160 hours working on the start-up in the previous 12 months, expected to work less than 80 hours in the next six months, and agreeing that this start-up was part of their current career plans. This tends to minimize the definitional variation that may be introduced by the respondents regarding “quitting.” They are then asked if others are still pursuing the initiative.

In the U.S. projects and most of the other studies those identified as quits in a follow-up interview are not interviewed in subsequent follow-ups.

Weights: Population Sampling Case Weights

A major ongoing issue with all surveys of national populations is the differential tendency of individuals to become involved with and complete interviews. The underrepresentation of young men and overrepresentation of older women in all surveys of adult populations is well documented. In addition, there is a tendency for those at both ends of the socio-economic spectrum not to be involved, the very poor and the very rich are often not included.

The standard procedure is to include measures of age, gender, socio-economic status, and perhaps geographic location for each case in a survey. The sample is then compared with the best available descriptions of the population on these characteristics. This is often a major government effort designed to provide reliable estimates of important population characteristics, such as the unemployment rate. Case weights are then attached to the survey cases such that the survey sample will match the population sample on the critical characteristics. For example, the case weighted sample will have the same proportion of young men and older women as the more precise population sample.

This is of particular concern for studies of activities of young men, and it is well documented that early career men, those under 45 years old, are a major source of nascent ventures. As a result, confidence that young men are accurately represented in any sample of nascent entrepreneurs is higher when case weights reflecting their prevalence in the adult population have been developed.

The most straightforward procedure takes advantage of a well-managed screening of the adult population. In the screening stage a representative sample of the population is identified and asked about participation in business creation. Those that qualify as nascent entrepreneurs become part of the cohort to be followed over time. The case weights developed for the screening sample are provided as part of the screening assessment. These same weights can then be carried over to the sub-sample identified as the nascent cohort. The case weights for the nascent subsample are then re-centered, adjusted so the average value is one. These case weight then provide an adjustment that will compensate for cases that may be over- or under-represented.

Weights from the initial screening to identify nascent entrepreneurs for the U.S. PSED I and II cohorts were attached to the cases in the respective data sets.

Case Weights: Entrepreneurs versus Ventures

A more subtle issue is adjustments to compensate for the size of the start-up team. In both PSED projects, the start-up team is considered to be those that will own part of the new firm.

Any individual that expects to own part of a new firm is considered an acceptable respondent in the screening procedure. As a result, a start-up venture with multiple future owners is more likely to be included in the sample than a one-person start-up. The most straight forward way to adjust for the possibility of oversampling is to reduce the cases weights by dividing by the expected number of owners.19

Team size reflects responses to questions about expected owners of the new firm. Because a small number of cases reporting more than five expected owners, the distribution is constrained to five categories. There are a small proportion of cases where new firm ownership is expected to involve other organizations, such as financial institutions or another business. Because the sampling procedure involves identifying humans involved in firm creation, the team size for the weighting adjustment utilizes only potential human owners.20 The unweighted distributions of the start-up team size for the U.S. PSED I and II cohorts are presented in Table 13.

Table 13. Distribution of Human Start-Up Team Size: U.S. PSED I & II



N cases (unweighted)
















5 or more





(*) Excludes owners representing organizations.

The results of this adjustment for the U.S. PSED II sample are provided in Table 14. As can be seen in this presentation there is a major impact on the weights assigned to start-up ventures with larger teams, but these are a small proportion of the sample. Almost 89% of the U.S. PSED II nascent firms are expected to have one or two owners; only about one in 20 expect to have four or five owners. The case weights for nascent ventures with three, four, or five members are reduced once the size of the start-up team is taken into account.

Table 14. Case Weights Adjusted for Start-Up Team Size: U.S. PSED II

Team size

Proportion of cases: Total sample

Nascent entrepreneurs, Wave 1: Population weights

Nascent entrepreneurs Wave 1; Population weights adjusted for team size

Nascent entrepreneurs with follow- ups: Population weights

Nascent entrepreneurs with follow- ups: Population weights adjusted for team size






































Total cases






Identifying Active Nascent Entrepreneurs

The creation of the longitudinal descriptions of business creation represented by the PSED data sets involves hundreds of different interviewers having conversations with thousands of respondents. There are many opportunities for variation in the use and interpretations of language and response categories as well as problems in recalling past events. To increase confidence that a relatively homogeneous cohort of nascent ventures has been identified, a multi-step procedure for excluding cases that would increase diversity has been developed.

The first adjustments are related to several complications. First is the collection of data over a window of time for a specific nascent venture, 53 months for U.S. PSED I and 61 months for U.S. PSED II. Regardless of the width of the window, these represent arbitrary sections of the start-up process. Second are the ambiguities associated with language, such that asking about the same issues in a different format may produce different responses. This seems to be the case when the presence of initial profits is assessed. Third is related to the phenomena itself. While the initial assumption may be that those involved in devoting time and resources to business creation are serious about reaching initial profitability, it appears that a substantial minority have a modest commitment. Those that report a small number of activities spread over a number of years would appear to be unsure about making a commitment to business creation. The following procedure is designed to identify those active in the pre-profit stage of business creation and exclude nascent entrepreneurs with a very low level of involvement.

There are six steps in the procedures utilized to identify those cases representing active nascent entrepreneurs. These stages and the effect on the size of the two cohorts are presented in Table 15. The criteria are applied sequentially, so some cases could be excluded for failure to meet more than one criterion. The first column provides the values for the variable CASEKEEP, which is included in the consolidated data sets to identify cases that failed to meet different criteria.

Table 15. Case Attrition in Identifying Active Nascent Entrepreneurs: U.S. PSED I & II


Selection Criteria




Total Number of Cases





Profits 3+ months before screening





Less than 3 of 18 start-up acts





No pair of acts in same 12 months





Profits reported before entry date





Entry 10 years before 1st interview





Nascent Entrepreneurs count




Proportion Nascent Entrepreneurs




No follow-up interviews





Nascent entrepreneurs with follow-up interviews




Proportion nascent with follow-up




(*) Values for CASEKEEP variable, found in consolidated data sets.

The first stage involves confirming that the respondent represents a venture in the pre-profit stage. While this is an important part of the screening module, when asked about this issue at the beginning of the first detailed interview a number of ventures were actually in profit at the time of the screening. They should, therefore, be excluded. To compensate for errors or reporting by the respondent or assigning dates a three month cushion is utilized,21 cases with profits up to three months before screening are retained.22

The remaining processing is based on 18 start-up activities associated with a date of initiation (Reynolds et al., 2016). As “serious thought” it is generally reported by 100% of all respondents and does not involve action and it is not included in this assessment. Processing based on more than 18 activities or a different set of 18 may lead to different results.

The second criterion excludes start-up efforts reporting less than three of 18 start-up activities over all waves of the data collection. This can be considered an indicator of a low level of commitment.

The third criteria is related to the intensity of commitment, for there are cases where a number of start-up acts have been initiated, but none within the same 12 month period. This lack of activity may also reflect uncertainty about making a commitment and these cases are also excluded.

For each case, the first pair of start-up activities that occurred within a 12 month period are identified, and the earliest date in this pair is considered the date of entry into the start-up process. Reflecting the diversity in the start-up process for those identified as active nascent entrepreneurs in the screening, for a substantial number of cases the date of entry, or conception, occurs after the Wave 1 interview.

Once the date of entry, or conception, is identified, it is possible to consider the lag between the entry date and the screening interview. While for most this is within the previous two years, there are a small number that entered the process 10 years before the initial interview. These would appear to have a low level of commitment and are excluded.

These criteria identify those cases that may be considered current, active nascent start-ups at the time of the first interview. As shown in Table 15 this is 77% of the initial count of 2,044 cases where there is Wave 1 data.

The survey operation was able to complete one or more follow-up interviews with 91% of these cases, so the total of the two cohorts with at least one follow-up interview to provide outcome data is 1,429.

Hence, there are two samples available for analysis. One consists of 1,567 active as current nascent entrepreneurs in the first detailed interview. The second consists of 1,429 cases with some outcome information. Case weights appropriate for both sub-samples are in the consolidated data sets.

U.S. PSED I: Harmonizing Minority Cohort with Majority Cohorts

The development of the U.S. PSED I data set involved three distinct cohorts, the original representative sample cohort, the female only cohort, and the minority cohort. Funding for the female only cohort was approved in time to allow the screening and follow-up interviews to be completed along with the representative cohort. Funding for the minority cohort was delayed such that the screening and follow-ups were also delayed. The timing of the data collection for the three cohorts and two comparison groups is provided in Table 16.

Table 16. U.S. PSED I: Timing of Cohort Data Collection


Initial Screening

Initial Screening Sample

First detailed sample (Items Q)

Summer 2001 Follow-up (Items R)

Summer 2002 follow-up (Items S)

Summer 2003 follow-up (Items T)

Representative sample

July 1998 to April 1999



1st follow-up

2nd follow-up

3rd follow-up

Female only

Sept 1999 to Dec 1998



1st follow-up

2nd follow-up

3rd follow-up

Minority only

July 1999 to Jan 2000




1st follow-up

2nd follow-up

Control Group: Representative Sample

Nov 1998



Control Group: Minority only

Nov 1999



Source: Based on Gartner et al. (2004), Table A.1, p. 459.

As can be seen in Table 16, the screening to identify the minority cohort was complete while the first follow-up interviews were being completed with the representative and female only cohorts. As a result, the two waves of follow-up data for the minority cases where completed during the last two follow-ups with the representative and female only cohorts. The representative and female only cohorts completed four waves (Q, R, S, and T) and minority cohorts completed three waves (Q, S, and T).

This could lead to some confusion when the data on the minority cohort is compared to the data on the representative and female only cohorts. To facilitate analysis, two data sets and codebooks have been developed for the U.S. PSED I project, the differences and file names provided in Table 17. One provides the minority cohort data in relationship to the questionnaire utilized in the interviews (Q, S, and T), and the other with data for the minority cohort reassigned to the sequence in which the data was collected (Q, R, and S). Data on minorities identified in the representative and female only sample are not affected by this adjustment.

Table 17. U.S. PSED I Alternative Data Sets

Structure Of Presentation for Minority Cases

Codebook file name

Data set file name

Questionnaire Used (Data in items beginning with Q, S, & T)



Panel Sequence (Data in items beginning with Q, R, & S)



Files available online.

Harmonized Projects in Other Countries

A number of national PSED projects have been implemented since the pretest of the protocol in 1993. Five of the projects identified 19 start-up activities that were identical or similar across the five cohorts and similar procedures were utilized to identify outcomes. This made it possible to create a five cohort, harmonized data set representing the results from four countries including Australia (CAUSEE), China, Sweden, and the United States (PSED I and II). The data set includes a total of 3,910 cases of nascent entrepreneurs and outcome data on 2,541. The consolidated data set includes considerable details on the start-up process and outcomes and the case identification numbers from the original sources. As four of the original data sets (Australia, Sweden, and U.S. PSED I and II) are publically available, variables from these original sources can easily be added to the consolidated data set for a wide variety of analyses. This resource facilitates both cross national comparisons as well as more precise assessments of the impact of multiple mechanisms on the outcome.

Both the documentation of the procedures and the consolidated data sets are in the public domain (Reynolds et al., 2016). Reducing the barriers to analysis with this consolidated data set should facilitate utilization of this material to assess a wide range of theories related to business creation.

Basic Patterns in the Start-Up Process

The primary purpose of the PSED projects is to provide a description of the outcomes of the start-up process. This is illustrated in Figure 2, which represents the transitions from start-up to a profitable firm or a discontinued from over the first 72 months after entering the process.

The source of the data are the harmonized five cohort data set, weighted to represent nascent ventures, and reflecting 2,513 cases at entry. Cases are dropped if there is no information on their status so after 72 months 2,059 cases remain. After 72 months about 33% have reported initial profits, 44% have been shut down, and 22% are still active in the start-up process.

Tracking the Entrepreneurial Process with the Panel Study of Entrepreneurial Dynamics (PSED) ProtocolClick to view larger

Figure 2. Outcome Status by Months Since Entry: Five Cohort Data Set.

A description of the activities implemented during the start-up process is presented in Figure 3; it is based on the cases in the five cohort harmonized file. The presentation is restricted to the 19 activities and two outcomes harmonized in the five cohort data set. The light bars represent the proportion of respondents reporting the activity, which is virtually 100% for serious thought and less than 20% for pursuing intellectual property rights or organizing a start-up team. The black bars represent the average months since entry into the process. The start-up activities are ranked, from left to right, in terms of the average time since entry into the process. As it occurs before the date of entry into the process, serious thought is negative at 18 months before entry. The outcomes, initial profits or quits, take the longest to occur and are presented at the far right.

Tracking the Entrepreneurial Process with the Panel Study of Entrepreneurial Dynamics (PSED) ProtocolClick to view larger

Figure 3. Start-Up Activities: Lag from Entry and Prevalence: Five Cohort Data Set.

These two presentations make clear the unique contributions of the PSED protocol. A relatively detailed description of the steps take to implement a new firm, the proportion that can be expected to achieve profitability and the time required to reach an outcome.

PSED Program Impact on Scholarly Community

The PSED research program has been widely used by researchers exploring business creation and the entrepreneurial process. As of July 2015 these data sets are known to have been utilized in 133 peer review journal articles, 12 books and monographs, 71 book chapters, 23 dissertations and theses, five research reports, and 86 peer review conference presentations (Frid, 2015). About four-fifths of this work has used the U.S. PSED data sets, as they are well documented and freely available to all, scholars or the curious. As no registration is required to utilize the data, it has been difficult to track total use and these counts may be conservative. No other family of similar data sets has been so widely used for the study of entrepreneurial phenomena.

Research utilizing the PSED data sets covers a wide range of topics related to entrepreneurship. Because such a wide range of topics have been explored, it is difficult to summarize the impact on scholarly assessments of business creation. A recent assessment of 83 peer review journal articles found that (Davidsson & Gordon, 2012):

  • 27% focused on personal characteristics of the nascent entrepreneur

  • 13% focused on new venture creation processes

  • 8% focused on personal characteristics and venture creation processes

  • 20% focused on the outcomes of the process

  • 12% focused on personal characteristics and outcomes

  • 16% focused on venture creation processes and outcomes

  • 3% focused on personal characteristics, creation processes, and outcomes

The rich sources of longitudinal data provided by the PSED data sets have clearly led to a diverse set of analyses. In this regard they have achieved their primary purpose, a resource for diverse assessments of entrepreneurial phenomena.

A second impact has been the adoption of the PSED screening protocol as the basis for the Global Entrepreneurship Monitor cross-national assessments, where the focus is on comparing the prevalence of business creation across countries (Reynolds et al., 2005). The GEM data sets have been the source of annual reports comparing countries since 1999, hundreds of national reports providing policy relevant assessments of individual countries, as well a growing body of peer review research completed by the hundreds of GEM team members in over a hundred countries (Bergmann, Mueller, & Schrettle, 2014).

Future Opportunities, Challenges

Two major findings from the PSED program have slowed progress. First is that the business creation process takes much longer than anticipated, six years after entering the entrepreneurial process up to one-fifth of the ventures are still in process and have not reached a clear outcome. There are, however, many reasons to expect that the more significant ventures requiring substantial expertise and resources may take longer than less complex businesses such as a restaurant. The second complication has been the complexity of the multiple processes involved. While successful transitions to profitable firms can be measured across many sectors and contexts, there is no one path to profits. And the more processes or factors that affect outcomes, the larger the data set that is required to sort out the relative impact of different mechanisms.

There continues to be a number of substantial opportunities for progress in understanding the entrepreneurial process where the PSED data sets are relevant:

  • Understanding the impact of different national context on processes and outcomes.

  • Understanding how changes in national economic situations affect the process and outcomes.

  • Implementation of the PSED data collection protocol in developing countries.

  • The benefits of larger data sets, such as the five cohort, harmonized data set, allow assessment of the interactions among different mechanisms affecting outcomes.

  • A source of baseline data on the implementation of typical businesses that can be used as a basis for comparison with samples of distinctive start-ups, such as those rare ventures supported by venture capital firms.


Bergman, H., Mueller, S., & Schrettle, T. (2014). The use of global entrepreneurship monitor data in academic research: A critical inventory and future potentials. International Journal of Entrepreneurial Venturing, 6(3), 242–276.Find this resource:

Butani, S. J., Clayton, R. L., Kapan, V., Spletzer, J. R., Talan, D. M., & Werking, G. S., Jr. (2006). Business employment dynamics: Tabulations by employer size. Monthly Labor Review, (February), 3–22.Find this resource:

Caroll, G. R., & Hannan, M. T. (2000). The demography of corporations and industries. Princeton, NJ: Princeton University Press.Find this resource:

Davidsson, P., & Gordon, S. R. (2012). Panel studies of new venture creation: A methods- focused review and suggestion for future research. Small Business Economics, 39, 853–876.Find this resource:

Delmar, F., & Shane, S. (2003). Does business planning facilitate the development of new ventures. Strategic Management Journal, 24, 1,165–1,185.Find this resource:

Fairlie, R. W. (2014). Kauffman Index of Entrepreneurial Activity: 1996–2013. Kansas City, MO: Ewing Marion Kauffman Foundation.Find this resource:

Frid, C. (2015). Publications based on the panel study of entrepreneurial dynamics. Retrieved from

Gartner, W. B., Shaver, K. G., Carter, N. M., & Reynolds, P. D. (Eds.). (2004). Handbook of entrepreneurial dynamics: The process of business creation. Thousand Oaks, CA: SAGE Publications.Find this resource:

Geroski, P. A. (1995). What do we know about entry? International Journal of Industrial Organization, 13, 421–440.Find this resource:

Haltiwanger, J., Jarmin, R. S., & Miranda, J. (2010). Who creates jobs? Small vs. large vs. young (Working Paper CES 10-17). Washington, DC: U.S. Census Bureau.Find this resource:

Hannan, M. T., & Freeman, J. (1989). Organizational ecology. Cambridge, MA: Harvard University Press.Find this resource:

Klepper, S., & Miller, J. H. (1995). Entry, exit, and shakeouts in new manufactured products. International Journal of Industrial Organization, 13, 567–591.Find this resource:

Liao, J., & Gartner, W. B. (2006). The effects of pre-venture plan timing and perceived environmental uncertainty on the persistence of emerging firms. Small Business Economics, 27, 23–40.Find this resource:

Mathematica Policy Research. (2007). Building and sustaining innovative companies: The Kauffman Firm Survey. Retrieved from

Pozen, D. (2008). We are all entrepreneurs now. Wake Forest Law Review, 43, 283–340.Find this resource:

Reynolds, P. D. (2000). National panel study of U.S. business start-ups: Background and methodology. In J. A. Katz (Ed.), Advances in entrepreneurship, firm emergence and growth (Vol. 4, pp. 153–228). Stamford, CT: JAI Press.Find this resource:

Reynolds, P. D. (2009). Screening item effects in estimating the prevalence of nascent entrepreneurs. Small Business Economics, 33(2), 151–163.Find this resource:

Reynolds, P. D. (2015, August). When is a firm born? Alternative criteria and consequences. Peer review paper session at the Academy of Management Annual Meeting, Vancouver, BC, Canada.Find this resource:

Reynolds, P. D., Bosma, N., Autio, E., Hunt, S., De Bono, N., Servais, I., … Chin, N. (2005). Global entrepreneurship monitor: Data collection design and implementation: 1998–2003. Small Business Economics, 24, 205–231.Find this resource:

Reynolds, P. D., & Curtin, R. T. (2009). Business creation in the United States: Entry, startup activities, and the launch of new ventures. In U.S. Small Business Administration (SBA) Office of Advocacy (Ed.), The small business economy: A report to the president 2008 (pp. 165–240). Washington, DC: U.S. Government Printing Office.Find this resource:

Reynolds, P. D., & Curtin, R. T. (Eds.). (2009). New firm creation in the U.S.: Initial explorations with the PSED II data set. New York: Springer.Find this resource:

Reynolds, P. D., & Curtin, R. T. (Eds.). (2010). New firm creation: An international overview. New York: Springer.Find this resource:

Reynolds, P. D., & Curtin, R. T. (2011). PSED I, II harmonized transitions, outcomes data set. Retrieved from

Reynolds, P. D., Hart, M., & Mickiewicz, T. (2014). The U.K. business creation process: The 2013 Panel Study of Entrepreneurial Dynamics pretest. Birmingham, U.K.: Aston Business School, Enterprise Research Centre. doi:10.13140/RG.2.2.27165.51685Find this resource:

Reynolds, P. D., Hechavarria, D., Tain, L., Samuelsson, M., & Davidsson, P. (2016). Panel Study of Entrepreneurial Dynamics: A five cohort outcomes data set. Research Gate. doi:10.123140/RG.2.1.25561.7682Find this resource:

Reynolds, P. D., Storey, D. J., & Westhead, P. (1994). Cross-national comparisons of the variation in new firm formation rates. Regional Studies, 28(4), 443–456.Find this resource:

Reynolds, P. D., & White, S. (1993). Wisconsin’s entrepreneurial climate study. Milwaukee, WI: Marquette University, Center for the Study of Entrepreneurship. Submitted to the Wisconsin Housing and Economic Development Authority, June.Find this resource:

Reynolds, P. D., & White, S. (1997). The entrepreneurial process: Economic growth, women, and minorities. Westport, CT: Greenwood.Find this resource:

Ruef, M. (2010). The entrepreneurial group. Princeton, NJ: Princeton University Press.Find this resource:

Sadeghi, A., Spletzer, J. R., & Talan, D. M. (2009). Business employment dynamics: Annual tabulations. Monthly Labor Review, May, 45–56.Find this resource:

Schoonhoven, C. B., Burton, M. D., & Reynolds, P. D. (2009). Reconceiving the gestation window: The consequences of competing definitions of firm conception and birth. In P. Reynolds & R. Curtin (Eds.), New firm creation in the United States: Initial explorations with the PSED II data set (pp. 219–238). New York: Springer.Find this resource:

Shapero, A., & Giglierano, J. (1982). Exits and entries: A study in Yellow Pages journalism. In K. H. Vesper (Ed.), Frontiers of entrepreneurial research: 1982. Wellesley, MA: Babson College.Find this resource:

Tornikoski, E., & Renko, M. (2014). Timely creation of new organizations—The imprinting effects of entrepreneurs initial founding decision. M@n@gement, 17(3), 193–213.Find this resource:

Walls & Associates. (2012). National Establishment Time-Series (NETS) database: 2012 database description. Denver, CO: Walls & Associates.Find this resource:

Yang, T., & Aldrich, H. E. (2012). Out of sight but not of mind: Why failure to account for left truncation biases research on failure rates. Journal of Business Venturing, 27, 477–492.Find this resource:

Appendix: Access to Data Sets


(1.) A more detailed discussion of the emergence and administration of the ERC is provided in Gartner et al. (2004, pp. xiv–vii).

(2.) An assessment of the effect of the change in two items in the 1999 screening to three items in the 2005 screening indicated that when adjustments were made to control for the change in the number of screening items there was no statistically significant change in the prevalence rate (Reynolds, 2009).

(3.) Summary in Reynolds and Curtin, (2011, p. 19) and Tornikoski and Renko (2014).

(4.) Yang and Aldrich (2012).

(5.) Based on the patterns in 3,212 unweighted cases in four cohorts in three countries in the consolidated file discussed below, omitting Sweden due to absence of dates on serious thought in the first two waves of data collection.

(6.) Based on 2,947 unweighted cases in four cohorts in the consolidated file discussed below, omitting Sweden due to absence of dates on serious thought in the first two waves of data collection (Reynolds et al., 2016).

(7.) Delmar and Shane (2003).

(8.) Liao and Gartner (2006).

(9.) Among 3,639 cases in the five cohort data set, 1.6% have an entry date 10 years prior to the initial Wave 1 detailed interview, 5.0% five to 10 years prior, 8.5% three to five years prior, 73.2% up to 2 years prior, and 1.6% after the Wave 1 interview. These latter cases qualified for entry on the basis of start-up activities initiated after the Wave 1 interview.

(10.) All the initial panel studies, including PSED I, allowed to respondents to determine if the start-up venture was an operating business with a single item in the interview schedules (Reynolds & Curtin, 2011; Table 11.10).

(11.) This involved some revision to the screening items and analysis procedures for both the U.S. PSED II project and the Global Entrepreneurship Monitor harmonized interviews (Reynolds et al., 2005).

(12.) Walls & Associates (2012) have developed comprehensive longitudinal records from the Dun & Bradstreet files, with full contact details for all listed firms.

(13.) The U.S. Bureau of Labor Statistics uses harmonized reports from state unemployment insurance payments (ES202 files) to track new employer firms on a quarterly basis (Butani et al., 2006; Sadegh, Spletzer, & Talan, 2009).

(14.) Substantial data sets of employee organizations have been created by the U.S. Census Center for Economic Studies utilizing listings in the federal social security payment files where new listings are tracked annually, Haltiwanger, Jarmin, and Miranda (2010) is one example of the benefits.

(15.) Most research in the population ecology paradigm pursued by organizational sociologists develops files based on listings in public records or trade associations (Caroll & Hannan, 2000; Hannan & Freeman, 1989).

(16.) The index reflects only reports associated with those 20 to 64 years old. “Older individuals (ages 65 and older) are removed from this sample because retirement in this age group leads to lower entrepreneurial activity rates” (Fairlie, 2014, p. 27).

(17.) As over 80% of those active in firm creation do so while they have a full time job or are managing another business and, as discussed below, the start-up process takes years to complete, this “index” does not capture the vast majority of business creation activity.

(18.) New social ventures, which may not place a primary focus on profits, cannot sustain themselves unless they manage to break even (income equal expenses).

(19.) This complication has been mentioned in numerous commentaries, including Davidsson and Gordon (2012). The procedure used here was proposed by Ruef (2010).

(20.) In the U.S. PSED I and U.S. PSED II about 3% of the owners are either other organizations (financial institutions, other businesses, government agencies) or individuals representing other organizations (Reynolds & Curtin, 2009, Table 5.1).

(21.) Extending this “cushion” to six months had little effect on four cohorts; it would have increased good cases with follow-ups for the SE-PSED cohort by 8%.

(22.) In the CAUSEE and CH-PSED projects, the detailed Wave 1 interview was to be provided immediately after the screening interview. There were, however, some cases where the Wave 1 interview was applied in a second session. For consistency with the other projects, it was assumed that, on average, the screening interview occurred 30 days before the Wave 1 interview.