Case selection

The present study represents an embedded multiple-case study (Yin, 1989, pp. 46). The cases were selected by following a theoretical replication logic (see Yin, 1989). The basis for the selection of cases for this study can be found in other studies (see Vesalainen & Murto-Koivisto, 1994; Murto-Koivisto & Vesalainen, 1995). In those previous studies, entrepreneurs representing nine different types of nets were interviewed and the nets were followed up between August 1993 and January 1995, after which the data collection of the nets of the present study continued. The cases for the present study were selected from these nine possible ones by forming a dichotomous scale in regard to two dimensions: the cooperative type and the performance of cooperation. The purpose determining the selection of cases was to choose four nets, two of which would represent joint ventures and two, project groups. The cooperative type is also related to the formality of cooperation, because a project group functions on the basis of a written agreement while a joint venture is a separate company established by partners (more about different cooperative types, see Murto-Koivisto, Routamaa and Vesalainen, 1996). Another criterion was that there would be one 'more successful' joint venture, one 'less successful' joint venture, one 'more successful' project group and one 'less successful' project group.

Success was evaluated by a factor that is usually ignored in performance studies - namely by the cooperative partners' perception of the performance. Subjective measures have been widely neglected in evaluating performance (cf. Hovi, 1995, Mohr & Spekman, 1994; Bucklin & Sengupta, 1993; Geringer & Hebert, 1991), and instead, objective measures such as sales volumes and the duration of cooperation have been favored, e.g., by Harrigan (1988). However, based on personal empirical experience, it is sometimes quite difficult to say which sales are due specifically to cooperation and which are simply due to general economic development. It has also to be remembered that many outcomes are impossible to estimate quantitatively (Bucklin & Sengupta, 1993). Moreover, in practice it became clear that the duration of cooperation did not prove to be a valid measure because so many external factors, e.g. public support, kept the nets 'alive'. The partners carried out subjective measurements by comparing the objectives set for cooperation and the outcomes attained (see also Mohr & Spekman, 1994). Thus, each partner's viewpoint was taken into consideration (cf. Anderson 1990). Although the emphasis was on 'soft' measuring, some objective checking was also used, namely a form of input-output calculation (for more details see Murto-Koivisto & Vesalainen, 1995; cf. Mohr & Spekman, 1994).

Methods of data collection

The present study was carried out using a multi-method approach typical of case studies. The empirical data were mainly collected by means of unstructured and semistructured interviews with open-ended questions, and the interviews were carried out on an ongoing basis during the whole development process. The interviews were held mainly with the entrepreneur/owner-manager of each company, but also with the network broker, who was an outside consultant in some nets. Two rounds of interviews were carried out by visiting entrepreneurs' firms, and the third round by telephone interviews. Parallel with three rounds of interviews, the researcher contacted key informants (see Patton, 1987, pp. 95) in each net monthly to ask them whether something was happening in their nets. The key informants facilitated the gathering of data considerably, because it would have been too time-consuming to keep in contact with every entrepreneur monthly. In addition to interviews, observation and different documents were also used in data gathering. The researcher had access to organizational rules, regulations, memoranda, and any other official or unofficial documents generated by or for the net. Observations were made in two different connections: participation in the meetings of cooperative groups and in connection with the interviews. The methodological triangulation, i.e. the use of multiple methods, was a good method of improving the quality of the study. All in all the nets were followed very intensively for two and a half years. The ages of these nets varied from three to four years at the time when the data collection ended.

The data were analyzed using content analysis. A coding scheme procedure is recommended by Weber (1985) when carrying out this kind of analysis. In this study the categories that were used were a synthesis of the researcher's own and those that emerged from the data. The coding process progressed in such a way that first the transcriptions of the cases were read through very carefully in order to get a holistic view of them. In the second round the sentences and paragraphs which represented certain dimensions were marked in the text. After that there was an attempt to illuminate the development process of each net. Moreover, there was some consideration of what makes a more successful and less successful process different by comparing the cases to each other. The purpose of this kind of procedure is to try to maximize underlying differences between the cases so that the real factors concerning the successful and unsuccessful development path will be found (cf. Hamel, 1991).

Top of Page
Previous Page | Main Menu | Next Page
Return to 1996 Topical Index

1997 Babson College All Rights Reserved
Last Updated 5/1/97 by Jo Ann Mathieu

To sign-up for the Center for Entrepreneurial Studies' publication lists,
please register with the
Entrepreneurship WebTeam.