Chapter Listing | Return to 1997 Topical Index


METHOD

 The Need to Repeat Studies in Different Samples

 A major weakness in our knowledge on the growth of small firms, and entrepreneurship in general, is that it is highly fragmented. The same or similar phenomena are often studied in isolated research projects using different concepts, models, measures and methods. It is rare that researchers build on theories and models developed by others. This problem is illustrated in the reviews of contemporary research carried out by e.g. Low & MacMillan (1988) and Storey (1994). The following quote illustrates this point:
This section reviews a number of empirical studies which relate elements within the entrepreneur, firm and strategy components to the growth of the firm...There are a number of problems in using this approach, because the vast bulk of studies have been conducted independently of each other. Frequently they address issues of specific interest to the researcher, but do so in a way which makes compatibility with other studies difficult. (Storey, 1994, p 125)

To avoid the problem of fragmentation and to build knowledge from the results of multiple studies, three separate studies addressing the same issues with the same measurement instrument in similar samples are jointly analysed in this paper.

Over a ten year period three independent telephone interviews were conducted. The two initial studies in 1986 and 1994 were stratified over the Swedish equivalent of ISIC codes. Independent firms from specific manufacturing, service and retail industries were selected. The samples were also stratified over the standard Swedish size brackets 1-4, 5-9, 10-19, and 20-49 employees. In addition to this, the 1996 sample was stratified over the firms’ growth-rate so that the share of high-growth firms was overrepresented in the sample for all size brackets and industries. The match of the samples concerning industries and size brackets is not perfect. However, for the purpose of this paper, the differences are relatively unproblematic. The samples had 440, 400, and 630 respondents respectively, totalling 1470 respondents, with corresponding response rates of 83%, 55%, and 75%. This gives us a unique sample.

 At the time of the interview, 40 firms had outgrown the largest size bracket and were omitted. The smallest size bracket was left out of the analyses since only one of the studies included firms of this size and since many firms in the size bracket may be part-time businesses where growth ambitions are likely to be very different. This reduced the number of cases by an additional 159 to a total of 1271 cases. The actual number of cases used in the analyses is somewhat lower due to internal non-responses.

Variables and Measures

The dependent variable. To measure growth willingness, i.e. our proxy for behavioural response, respondents were asked two questions; whether a 25% (moderate growth) and 100% (substantial growth) increase in number of employees in five years time would be mainly negative or positive. A seven-point scale ranging from very negative to very positive was used to measure these variables. We are not explicitly measuring an intention or goal to engage in a behaviour that may or may not lead to the expansion of the business, but this over-all measure of growth willingness could be seen as an indication of an intention or goal.

The independent variables. The eight expectancy variables, i.e. our indicators of cognitive response, derived from the literature review and pilot study mentioned earlier, are displayed in Table 1. To measure these variables, respondents were asked how a doubling of the number of employees, regardless of this is desirable or possible, would be likely to affect each of the eight areas. A five-point scale ranging from much more negative to much more positive was used for measurement.

Control variables. Firm size, industry and firm age have frequently been used as predictors of firm growth, i.e. predictors of actual small firm behaviour (cf. Evans, 1987 or Storey, 1994, Ch. 5 for a review of predictors of small firm growth). To the extent that actual firm growth depends on growth willingness, it may very well be the case that these factors
influence the growth willingness of the entrepreneur. Firm size was measured as number of full time equivalents. To determine the industry, respondents were asked in the firm’s main line of business was manufacturing, service or retail. Finally, respondents were asked if they knew what year the firm was founded, which was used to calculate the age of the firm.

To make comparison across the three studies more valid, the wording and relative position of the questions were exactly the same in all three studies.
 

Top of page | Chapter Listing | Return to 1997 Topical Index


1997 Babson College All Rights Reserved
Last Updated 06/01/98