Despite the achievements described above, there are some limitations that concern the research study. They primarily result from the adopted methodological approach and the flaws of the techniques used to collect and analyze the data.
First of all, the main perspective that determined both the theoretical considerations and the research activity undertaken for the purpose of this work was the context of a CO, and more precisely, the potential of a CO to create new market opportunities for the entities that create it (hence the reason for taking into account the role and opinions of the coordinators of the COs included in the research). The research focused primarily on one category of cluster entities, i.e. enterprises, which is the core component of any CO (the other entities associated in COs such as R&D institutions, business environment institutions, and educational institutions were included in the first stage of the research only to provide a better understanding of the context in which cooperative relationships among the enterprises developed). Therefore, the cluster enterprises were analyzed only through the prism of their membership in a CO. The issue of their potential to take advantage of the opportunities created by the CO was omitted.
Second, according to the symbolic-interpretive paradigm, the results of the conducted research refer only to a limited part of the reality. The research was carried out in seven deliberately selected COs representing three sectors of the economy - in four COs from the metal and ICT industry (the first stage of the research) and in three COs from the energy industry (second stage). Therefore, the selected groups did not meet the criteria of representativeness, which limits the possibility of generalizing conclusions. Nevertheless, it should be emphasized that the conducted research was not focused on generalization of the results to the entire population of COs, instead, the aim was to explore the area of cooperation among the cluster enterprises. However, assumptions regarding the selection of the research sample made in the first stage of the research, during which the theoretical concept was created (the logic of extreme case sampling in order to ensure maximum variability and diversity), allow to put forward a thesis about a wider universality of the findings, which could basically be applied to all COs. This is evidenced by the results obtained in the second stage of the research, which relate to the application of the developed concept in the selected COs and do not differ from those obtained in the first stage. However, to confirm the supposition of the universality of the developed concept, it would be necessary to conduct new comparative studies on a larger, more representative sample.
The third limitation stems from the criteria of selecting the research sample. The adopted criteria (the age and size of the CO and the scope of activities undertaken by it) did not guarantee that the CO would achieve the expected organizational maturity. This maturity could only be verified after completing the study. The selection of the COs in the second stage was based on previous experiences of cooperation with selected COs, which assessed their maturity much earlier.
The fourth limitation is related to the authors’ knowledge and experience in the field of clustering. On one hand, the accumulated knowledge about the concept of cluster and CO constituted an essential theoretical foundation, providing a broad context for the research studies. Additionally, the immersion in the particular environment allowed the authors to extend the knowledge in the analyzed area. On the other hand, knowledge and experience are a serious burden while carrying out research in accordance with the adopted inductive and abductive approach (especially in the context of using grounded theory). Although the concept of CO is poorly developed in the literature (which partly alleviated the discussed limitation), the authors adopted an idea of departing from the existing knowledge as much as possible, which was applied in the design and sampling stages of the research project.
The fifth limitation results from subjectivity inscribed in the specific nature of the research carried out in social sciences. Stage 1 was vitiated by the greatest subjectivity, mainly due to the adopted strategy as well as the data collection and analysis techniques. The methodology of grounded theory (which involves individual interpretation of the received data) and the qualitative interviews (unstructured and non- standardized or partially standardized, enabling the respondents to freely express their opinions) left ample room for mutual subjectivity (of respondents and researchers). However, it was limited by strict rules of conduct imposed by the methodology of grounded theory.
The sixth limitation relates to the data statics. The developed theoretical concept reflects the dynamics of COs - cooperative relationships among cluster enterprises develop according to a specific trajectory (defined on the basis of the conducted research). Meanwhile, the first stage of the research was carried out in each of the four COs at a specific point in time, and therefore presented the particular CO at its current stage of development. It should be emphasized that the study focused primarily on analyzing the relations among the variables and not their change over time. The dynamic nature of the cluster cooperation was effectively captured in the first stage of the research due to the assumptions made during the sample selection stage (the COs included in the sample are at various levels of development) and the data collection techniques (interviews provide information from the past). In the second stage, the dynamics of cluster cooperation was better reflected as a result of the case study strategy and two, or even three (when taking into account the group interview), measurement moments.