Research Procedure

The points below describe in detail the research procedure followed in the course of designing and conducting the studies - both at stages 1 and 2. In reference to each stage, the points first describe the sample selection process, accounting for the criteria of the selection of COs, cluster members, and finally, specific individuals selected to participate in the study. Following the sample selection process, the points discuss the data collection, analysis and interpretation methods used in the work. The description of the research procedure is bookended with information on methodological rigor (at stages 1 and 2) applied in the course of the qualitative research.

120 Application of Grounded Theory Research Stage 1

Sample Selection

Qualitative research based on the methodology of grounded theory (stage 1) comprised the most important stage of empirical research. The research was conducted with a view to achieving the general objective discussed in the introduction to this work, i.e. the identification of the levels of development of cooperation in COs (selected for the study). At this stage of research, sample selection was performed with regard to its representativeness, accounting for the principle of theoretical sampling used in the methodology of grounded theory. It should be noted that qualitative research is bound to different criteria of representativeness than quantitative research in which sampling is performed on the basis of the principles of statistical inference in order to describe the entire population and reflect the actual distribution of the phenomenon in question to the highest degree possible. In turn, the representativeness of the sample in qualitative research is rooted in its variation and variety' (Flick, 2018). Research focused on formulating a theory requires the careful selection of cases which would allow for the detailed analysis of the given phenomenon, while accounting for its ambiguities, which emerge from the opinions and experiences of individuals or groups of individuals participating in the study. The size and the variety' of the study must - on the one hand - enable the researchers to make the largest possible number of comparisons between the analyzed cases (Maxwell, 2005) and - on the other - enable them to identify' common features. On this canvas, researchers are able to establish a conceptual framework rooted in the perceived relationships and analogies.

In research drawing upon the procedures of collecting and analyzing data developed within the methodology of grounded theory, the sampling process was of particular significance, as the discovery' of conceptual categories, as well as the relationships among them, is a process which is executed with the use of the constant comparative method (Glaser & Strauss, 1967). The aim of this method is - on the on hand - to minimize, and - on the other - to maximize the differences among the compared groups. Similarities in the collected data are meant to lead to the emergence of specific categories and the identification of their conceptual properties, while the act of maximizing differences allows for even further reaching development of the theory with a view to achieving a higher degree of generality. For this reason, according to Glaser and Strauss, the formulation of a theory should draw on the largest possible number of similarities and differences among data, so that it will be possible to attain the largest possible variation among the derived categories (Glaser & Strauss, 1967). By working in accordance with the principles of the methodology of grounded theory, the authors strove to achieve a sample of acceptable adequacy (Glaser & Strauss, 1967), which is judged primarily on the basis of the range of the generated theory - the variation of the categories derived in the research and the level of their saturation, as well as the relationships perceived to exist among them, which can be expressed with the help of a set of integrated research hypotheses. However, it should be noted that for the generated theory, accurate evidence, or the number of cases under consideration,⁴ are of much lower significance than capturing the right balance between the similarities and the variation of the sample.

In an attempt to select a specific logic of sampling and the resulting research criteria, in the first step, the authors analyzed the population, focusing on entities from the same substantive area - formalized COs (operating on the basis of their legal makeup) operating in Poland in the time preceding the study. However, we should note the lack of academic publications with credible data on the specific number of COs functioning in Poland. The calculation of the size of this population was also made harder by the natural process of the fluctuation of the number of COs (resulting from the breakup and emergence of COs). The point of reference for the above calculations were the results of projects executed by the Polish Agency for Enterprise Development: Inventory of clusters, as well as three series of cluster benchmarking research (Benchmarking klastrow w Polsce, 2010; Holub- Iwan, 2012; Plawgo, 2014). The inventory held in 2015 has shown that at that time, around 135 COs (due to the aforementioned fluctuation, the number should be treated as an approximation) operated in different regions of Poland (Buczyriska et al., 2016). The COs were very diverse in terms of their age (over 70% COs operated on the market for at least 5 years), sector, size (the average size of a CO was 44 cluster members), and legal organizational makeup. The screening process comprised three stages (Fig. 4.2).

As the above figure shows, the first stage of sampling pertained to the selection of COs, while the subsequent two stages consisted of selecting the members of the COs selected for the study (stage II) and the individuals representing the selected institutional members (stage III). At each stage, the sampling was targeted.