Secondary Analysis in Entrepreneurship: An Introduction to Databases and Data Management

Article excerpt

"Low-barrier-to-entry (LBE) research" is a term derived from Northwestern University's Denise Rousseau's efforts (1987) to describe research methods that rely on inexpensive or free-to-the-researcher techniques. Examples of LBE approaches for entrepreneurship research have been outlined before (Katz 1989); but more details on LBE methods are commonly needed because many of the approaches are not easy to use. This article focuses on one of the most daunting of these approaches: secondary analysis. There are two major reasons why secondary analysis can be difficult:

1. There is a knowledge explosion

and growing unfamiliarity with

existing holdings. For example, the

1989 listing of the archives of the

Interuniversity Consortium for Political

and Social Research (ICPSR) is

more than 500 pages long and grows

by about 15 percent per year. Also,

many studies of potential utility to

entrepreneurship researchers are not

readily apparent from the archive

listing description. As the ICPSR

develops better cross-referencing

methods, these problems may lessen,

but they will still remain substantial.

2. Secondary analysis poses problems

that are different from those

experienced by researchers accustomed

to gathering their own data.

For example, the self-employed have

not been looked at as a separate

group in many national studies.

Often, the low error rates associated

with these national surveys are

based on years of detection and

corrections of errors in the data set.

However, since the self-employed are

not studied as often as those who

earn wages or salaries, the error

rates for the data on the self-employed

are usually higher.

Another unique problem is researcher unfamiliarity with the "feel" of the population, survey, or data set. In doing original research, the researcher usually has detailed first-hand knowledge about the people, the questions, and the results of the study. Such information i s invaluable for intuitively identifying errors, anomalies, and paradoxes in the data. To be maximally effective when using secondary data, researchers must develop techniques for improving their "feel" for the study and develop supplementary techniques for error detection and correction of their "big picture" perspective.

This article will introduce the process of secondary analysis, paying particular attention to databases that are potentially useful for the individual-level analysis of entrepreneurship issues.(1) It also will introduce types of data archives, identify 10 relevant data sets available from one of the largest archives, and explain the fundamentals of data management for archival data sets and the typical problems in secondary analyses. An evaluation of the potential of secondary analysis for the field of entrepreneurship will also be offered.

THE NATURE AND AVAILABILITY OF ARCHIVAL HOLDINGS

The underlying theory for this article comes from a number of sources. The philosophy behind LBE research comes from Katz (1989). The specific model for data preparation comes from Geda (1987), while the problem identification and resolution approach comes from Katz (in press). Overall, the underlying approach reflects the methods for secondary analysis pioneered at the Institute for Social Research at the University of Michigan.

There are three major sources for secondary analysis data sets: noncommercial archives, commercial sources, and other researchers. Noncommercial archives are the major source used by academic researchers. Kiecolt and Nathan (1985) list 12 noncommercial archives in the United States, all of which are located on college campuses or in government agencies. Data costs range from free, for ICPSR holdings sought by member schools, to $460 per tape from the National Technical Information Service. …