? 2003 for Handbook of Entrepreneurial Dynamics:

    Appendix A - Data Collection

    1Paul D. Reynolds

    Babson College and London Business School

    and

    Richard T. Curtin

    University of Michigan

    Prepared for the

Handbook of Entrepreneurial Dynamics: The Process of Organizational

    Creation

     September 2003

     1 This relies heavily on an earlier publication that described the initial procedures and history of the project prepared

    by the first author, although all figures have been updated. Reynolds, Paul D. 2000. ―National Panel Study of

    U.S. Business Start-ups: Background and Methodology.‖ In Jerome A. Katz (Editor), Advances in Entrepreneurship,

    Firm Emergence and Growth, Vol. 4. Stamford, CT: JAI Press, Pp. 153-228.

    Last printed 7/9/2010 8:02:00 AM Page 1 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    Table of Contents

Conceptualization of the Phenomena:

Design History:

PSED Design:

Screening respondents: Basic Procedure

     Completion of Nascent Screening: Third Criteria

Initial Detailed Interview Schedules

    Operational Outcomes: Respondent Cooperation and Response Rates

Follow-Up Data Collection

Follow-up Interview Schedules

Follow-up Data Collection: Operational Results

    Operational Outcomes: Call-backs and prevalence rates

Commentary

References

Last printed 7/9/2010 8:02:00 AM Page 2 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    List of Exhibits

Exhibit 1 Conceptualization of the Entrepreneurial Process

    Exhibit 2 Research Design Overview

    List of Tables

Table 1 Sample Selection: Cohort Size, Screening Dates, and Follow-up Samples

    Table 2 Initial Interview Topics: Nascent Entrepreneur and Comparison Groups

    Table 3 Indicators of Respondent Cooperation

    Table 4 Follow-up Data Collection: Timing and Responsibility

    Table 5 Interview Topics: Nascent Entrepreneur Follow-ups

    Table 6 First Follow-up Outcome Summary

    Table 7 First and Second Follow-up Outcome Summary

    Table 8 Calls to Complete and Nascent Entrepreneur Prevalence Rates

Last printed 7/9/2010 8:02:00 AM Page 3 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    Where do new businesses come from?

The origin of new businesses is the primary focus of the Panel Study of Entrepreneurial Dynamics.

    The design of the research is intended to document the underlying processes and factors that lead

    individuals to pursue the creation of a new business firm. The optimal strategy seemed quite

    straightforward: locate individuals involved in the business start-up process and determine as much

    as possible about their initial situation and their plans and actions to create a new business. Periodic

    follow-ups would provide information on the outcome of their efforts. Implementation of such a

    complex focus has proven difficult but feasible.

Conceptualization of the Phenomena:

     To facilitate discussion of the start-up process, it has been convenient to conceptualize the

    entrepreneurial process as occurring within a political and economic context as indicated in Exhibit

    1.

     Exhibit 1 – Conceptualization of the Entrepreneurial Process

    Social,Political,Economic Context

    Growth

    Adult

    Population

    NE

    FirmPersist?aStart-up Processes Birth

    NIBusiness QuitFirm?b

    Population

    ?c

Last printed 7/9/2010 8:02:00 AM Page 4 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

     The process is conceptualized as having four stages with three transitions. The first transition occurs when an individual, either on their own or responding to an initiative of an existing firm, elects to pursue a firm creation. These persons are classified as nascent entrepreneurs [NE] if they represent an independent start-up effort, or, if they are sponsored by an existing business, the persons are considered nascent intrapreneurs [NI]--a type of corporate entrepreneurship. Conception may be considered the beginning of this, the firm gestation process. The second transition in the life course of the firm occurs when the gestation process is complete, firm birth has occurred, and an infant firm is in place as an operating business. For many in the start-up phase, however, the next transition is to abandon the effort. The firm is, in a sense, stillborn. The third transition for infant firms is a passage into firm adolescence, a stage where survival is considered more certain and not a constant challenge. A secondary feature of the third transition would be the nature of the growth trajectory of surviving new firms; a new firm pursing a high growth trajectory may be considered to have a different character compared to those that are designed to persist as small firms.

     Three of the critical features of the entrepreneurial process about which very little is known are represented by the question marks in Exhibit 1: ―?a‖ represents the proportion of business start-

    ups that complete the process with the implementation of an infant firm. ―?b‖ represents the proportion of start-ups that never complete the process, although when they are abandoned is also a major issue on which little is known. Finally, ―?c‖ represents the features of the gestation process

    itself, both the length of time required to develop an infant firm and the activities that take place.

     One primary objective of the research program was to provide systematic, reliable data on the basic features of the entrepreneurial or start-up process. This would include information on the nature and extent of variation in the critical aspects of the start-up process: the proportion of the adult population involved in firm conceptions, the activities that compose the start-up process, the proportion of start-ups that complete the second transition to become infant firms, and the survival and growth trajectories of the new firms.

     A second objective was to provide reliable data on those factors or variables that could account for variations in these transitional events. This includes, for example, variations in the number of start-up efforts that become operating new firms or the length of time spent in the gestation phase before the efforts are completed or abandoned. Potential influences would include, but are not limited to, the following (many are discussed in detail in the chapters of this handbook): Last printed 7/9/2010 8:02:00 AM Page 5 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    ? Economic context, including national conditions, local conditions, and conditions in the

    economic sector or market of the new firm

    ? Personal context, including work and family situations and responsibilities, as well as the

    presence of supportive social networks for members of the start-up team ? Personal background and relevant experience, based on early family life, educational or

    occupational training, or in specific economic or market sectors for members of the start-up team ? Personal pre-dispositions, either in terms of decision-making style, risk preferences, desire for

    autonomy, and personal aspirations

    ? Socio-demographic background such as age, gender, race, and other characteristics ? Nature and sequence of start-up activities pursued in the firm gestation process ? Nature of the markets, competition for the new firm, and the strategic focus. ? Access to and use of resources, financial and otherwise

    ? Access to and use of programs designed to assist start-ups and new businesses While assessing whether or not any one of these factors might affect the start-up process was one

    type of research objective, most past research in entrepreneurship has emphasized one or two of the

    factors from this list. A second objective was to estimate the relative impact of each of these factors holding constant the impact of all other factors. Such a comparative analysis requires that data be

    assembled on all factors for each start-up initiative. The major challenge in the design of the project was to collect data on all of these topics from the same set of respondents involved in starting new

    businesses. Such a comprehensive data set was required for the research program to provide a

    complete description and explanation of the critical elements of the entrepreneurial process.

     Based on these objectives, and the constituencies to be served by the results, the final

    research design should have a number of features. First, a technique must be developed to provide

    systematic descriptions of the start-up process from its very beginning—the beginning is assumed to

    occur when somebody decides to implement a new firm, before any behavior occurs. Second, the

    data should be based on a representative sample of business start-ups, so that estimates for the entire population—of people or firms—would be possible. Third, as the creation of a new firm appears to

    be very complex and multi-faceted—with many small factors having direct and interactive

    influences—it is important that all major perspectives be represented in the data collection. This is

    tantamount to trying to measure a wide range of potential independent variables from a variety of

    Last printed 7/9/2010 8:02:00 AM Page 6 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    perspectives. Fourth, as the focus is on the process of implementing a new business, the project would need to be designed as a longitudinal or panel study.

Design History

     The distinctive methodological challenge was the creation of a procedure that would capture

    the very beginning of the business creation process. It was necessary to identify a start-up initiative before it had been included in business registries (such as the Dun & Bradstreet credit rating files) or lists of business activity (such as the phone book Yellow Pages). The objective was to create a procedure that could be used to systematically develop a sample that would allow estimates of the total amount of activity in the population.

     Although later abandoned, the initial effort to locate a random sample of nascent entrepreneurs involved an application of multiplicative sampling. Developed with Charles Palit, of the University of Wisconsin Survey Research Laboratory, this application required a systematic procedure to identify the social network of a representative sample of adults (Sudman, Sirken, and Cowan, 1988; Palit and Reynolds, 1993). This involved establishing a list of each respondent‘s parents, siblings, adult children, co-workers, and significant other (usually a spouse) during the interview. The respondent is then asked if any of these people are currently involved in a start-up effort. If they are, the respondent is asked for details so they may be contacted for a more complete interview.

     While this may appear similar to snowball sampling--where a respondent identifies other individuals with unique characteristics, it differs in one important respect. When a person nominated by the respondent as a nascent entrepreneur is interviewed, they are asked questions that will make it possible to determine the probability they would be nominated by more than one person in the initial sample. This allows the use of the information from the procedure to adjust for potential multiple nominations and compute the probability with which nascent entrepreneurs occur in the total population.

     While the implementation of the procedure involved the creation of relatively complete and very useful descriptions of the respondent‘s social networks, it was abandoned for several reasons. Most important, a person nominated as a nascent entrepreneur was actually interviewed in less than 40% of the cases. There were two main problems. In 20% of the cases the original respondent was not able to provide the full name and phone number for the nominated nascent entrepreneur--they Last printed 7/9/2010 8:02:00 AM Page 7 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    did not have the information. In 40% of the cases, the original respondent refused to provide the full name and location information for the nominated nascent entrepreneurs. They were, after all, talking to a stranger on the phone and may not have fully believed that the interview was related to research. As a result, in only two of five cases was a network nominee called to be interviewed. A pretest effort completed in 1997 using one household member to nominate others in the same household for a nascent entrepreneur interview was also associated with very poor results; interviews were completed with very few, less than 10%, of the within-household nominees.

     Several factors led to the change in procedures--from multiplicative sampling to

    straightforward population screening. First, as discussed above, the multiplicative sampling procedure was not working well--very few of those nominated were interviewed. Second, prevalence rates of nascent entrepreneurs were four to eight times higher than the one to two percent initially expected. Third, it was possible to locate a commercial marketing research firm that could include two appropriate screening questions in a national survey so that the costs and benefits of identifying each nascent entrepreneur was attractive compared with other methods. The procedure amounted to asking all respondents if they were starting a business on their own or for their employer. Other questions were then used to eliminate those not considered active in a start-up, not potential owners, or not in the start-up phase itself.

     The first full application of the research procedure was with the adult population of Wisconsin in 1992 and 1993 (Reynolds and White, 1993, 1997). A second application involved the purchase of time in the monthly Survey of Consumers completed by the University of Michigan Institute for Social Research in October and November, 1993 (Curtin, 1982; Reynolds, 1997). These two studies provided similar results regarding prevalence rates (about 4% were identified as nascent entrepreneurs), demonstrated the technical feasibility of the research protocol, and indicated that costs would be high but affordable. In the first case, the initial screening was part of a special purpose survey. In the second project the screening was one part of a multi-purpose project, where many costs were shared with other research projects.

PSED Design

     The research design for the US Panel Study of Entrepreneurial Dynamics [PSED] is

    presented in Exhibit 2. It has three components and reflects data collection from two types of Last printed 7/9/2010 8:02:00 AM Page 8 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    respondents. The first stage involves large-scale screening to create two samples representative of the U.S. population of adults, those 18 years old and older, excluding residents of Alaska and Hawaii. The first sample included those involved in attempting to start a new business. These respondents are either autonomous start-ups, referred to as nascent entrepreneurs [NEs], or sponsored by an existing firm, referred to as nascent intrapreneurs [NIs]. Second, a representative sample of typical adults to be used as the comparison group [CG] was identified. Both types had to meet certain criteria and be willing to participate in subsequent interviews. Once the screening procedures were completed, the second stage of data collection involved detailed phone interviews followed by a self-administered questionnaires mailed to the respondents. The third stage was the follow-up phone and mail interviews completed to determine the outcome of their efforts to implement a new firm. The initial design included plans for 12 and 24-month follow-up interviews. Additional funding has allowed for 36-month follow-ups to be completed in 2003.

    Last printed 7/9/2010 8:02:00 AM Page 9 of 34 C:\convert\temp\63324695.doc

    ? 2003 for Handbook of Entrepreneurial Dynamics:

    Exhibit 2 – PSED Research Design Overview

    [NE=Nascent Entrepreneurs; NI=Nascent Intrapreneurs; CG=Comparison Group; NF=New Firm]

     12 mth24 mth Initial Screening: 0 months follow-upfollow-up

     35 minCriteria 60 min 35 minNE phone•Active phone phone

    •Owner 10 pg•Not NF NIAdult12 pg 10 pgmail•Willingmailmail Population 200 Mil

     25 minphone CGCriteria •Willing 10 pg mail

    This design involved optimizing a number of desirable, though often incompatible, features.

    Among these issues were choices between the sample size versus the amount of information

    assembled from each respondent, the scope of information to be included in the interviews versus the

    desire to keep respondents involved over multiple data collection activities, which items were best

    suited for the phone or self-administered mail questionnaire, and the simplicity of the interview items

    versus the complexity of the research concepts. There is, of course, no single best solution to this

    optimization problem. The design of this research program was the result of both technical issues

    and the need to provide data for use by over 100 scholars and researchers in the Entrepreneurial

    Research Consortium. A political process led to an acceptable solution for this group of scholars.

    Another research team might have developed another, equally appropriate, solution.

    Last printed 7/9/2010 8:02:00 AM Page 10 of 34 C:\convert\temp\63324695.doc