Stage II of the Sampling Process: Selection of Institutional Members
At stage II of the sampling process, the authors selected specific institutional members from each of the four studied COs. The process of selecting the entities was held in accordance with the principle of theoretical sampling, understood as a process of gathering data which is controlled by the generated theory (Glaser & Strauss, 1967). In accordance with the methodology of grounded theory, the election of the first entities to study should be rooted in particular research assumptions while additional entities should be selected in effect of discovering subsequent codes (Glaser & Holton, 2010). In this way, subsequent sources of data collection are determined, as the emerging categories (and their properties) and the identified gaps in the generated theory point to the need of gathering additional data and turning to new groups for access thereto, selected - this time - in accordance with theoretical criteria (such as the theoretical objective and relevance) (Glaser & Holton, 2010).
In keeping with the principle of control over similarities and differences, in the first instance, the authors made attempts to minimize differences, focusing primarily on one type of members - cluster companies. The group of studied entities was extended to include cluster coordinators (organizational and legal entities fulfilling coordinating duties in COs), in the belief that their knowledge of the studied COs and experience may provide additional insights, which will turn out to be a valuable addition to the points of view held by the cluster members. Furthermore, the authors decided to take a flexible approach when deciding on including additional entities associated in COs in the sample (such as R&D institutions, educational organizations and business environment institutions). Should these entities fulfill the theoretical criteria (resulting from the generated concept), their inclusion in the sample would have the goal of better understanding the context of the development of cooperative relationships in the studied COs. The authors also introduced specific limitations with regard to the inclusion of entities in the study (which broadened the degree of similarities) and pertaining to the time of membership and the commitment in the activities of the CO - the above limitations were possible to implement thanks to the information received from the coordinators of the studied COs. It was assumed that cluster members with a longer membership record, who actively participate in activities undertaken in the CO will have the most complete information and hence, the most interesting insights into the functioning of their CO. In turn, in order to maximize differences, it was assumed that different types of companies will be able to participate in the study regardless of their age, size, operational profile, and competitive position.
In each of the COs, the studies were first held with the entities indicated by the coordinator (falling within the above-mentioned research limitations). In the course of the study (in the course of analyzing data and deriving conceptual categories and their properties therefrom), the authors identified gaps in the generated theoretical concepts and decided upon the need to supplement the research material with new data, which required the inclusion of additional entities in the sample (which - potentially - could deliver the sought-after data). To this end, the authors drew upon the knowledge of the coordinator on the specific cluster members (requesting the coordinator to indicate which of the cluster members potentially fulfill the theoretical criteria resulting from the requirements of the generated concepts). During the interview, some of the respondents would naturally refer to other cluster members in connection with the discussed topics. When these issues turned out to be crucial (from the perspective of the generated concept), additional entities were added to the sample on the basis of such referrals.6
In the end, having reached the intended state of theoretical saturation, the authors stopped after analyzing 30 entities (including the coordinators) representing four COs: 9 from Interizon (coordinator, 7 companies, 1 R&D institution), 6 from the MC ICT (coordinator, 4 companies, 1 R&D institution), 9 from the MCLP (MCLP) (coordinator, 4 companies, 1 R&D institution, 2 educational organizations, 1 support institution), and 6 from the MWEC (coordinator, 5 companies). The disparities in the proportions between COs (with respect to the number of studied organizations) primarily resulted from the availability of the cluster members and their willingness to cooperate in the research process.
Stage III of the Sampling Process: Selection of Individuals
At this stage, the authors made the assumption that the respondents selected for the study should have the most complete knowledge on the functioning of a given CO (and be authorized to disclose such knowledge in the course of the study) and the role of their home organizations within the CO. The coordinators were represented in the study by employees delegated to oversee the CO. As far as cluster members are concerned, the respondents were primarily selected among the owners of companies or individuals from the highest echelons of management (i.e. CEOs, general managers, members of the Board), and in singular cases, individuals designated to represent their company in the CO. The representatives of the remaining cluster members participating in the study were individuals who fulfilled various functions in their home organizations, engaging on their behalf in activities undertaken within the CO. In the study, each of the selected entities was represented by a single person (and in four cases - two people).