6. DISCUSSION
6.3 Research limitations
First and foremost, the greatest limitation of the research is the overall balance in the process model and corresponding findings. The findings on the design process, the responses to a team incentive system, and the degree of success of such a system (to a lesser degree) were mainly derived from the study by Collins (1995). Given the sound methodological quality of the study by Collins (1995), this development did not necessarily make the research fundamentally flawed, but some components of the model may slightly resemble a one-of-a-kind account of a team incentive system implementation. Is it desirable and perhaps even necessary to further develop and reinforce the process model by means of additional process studies and corresponding findings. If additional studies make a meaningful contribution to the model, the components mentioned will be provided with a more nuanced and balanced character.
Moreover, there was also a lack of balance in the specific contribution of each study. The study by Collins (1995) provided by far the most findings and first-order codes, the studies by Edwards and Langley (2007), Ezzamel and Willmott (1998), and Greene et al. (2014, 2015) all made a solid contribution in terms of findings, and the contributions of the studies by Eriksson (2010), Länsisalmi et al. (2000), and Suchan and Hayzak (2001) were barely worth mentioning. This means that the study by Collins (1995) has a disproportionately large share in the process model and that the studies by Eriksson (2010), Länsisalmi et al. (2000), and Suchan and Hayzak (2001) have a disproportionately small share in the model. Although this observation may result in a somewhat less balanced picture of the overall implementation of a team incentive system, it does not necessarily have to be an unfavourable observation. The studies by Eriksson (2010), Länsisalmi et al. (2000), and Suchan and Hayzak (2001) showed a relatively low methodological quality and ended up with poor critical appraisal results for the criteria related to believability and adequate participant representation, investigator impact, and ethics. The study by Collins (1995), on the other hand, achieved practically the best overall appraisal results (only the study by Edwards and Langley (2007) was rated higher). It was not quite unfavourable that the methodological sound study by Collins (1995) managed to make a greater contribution to the process model and that the methodological weaker studies by Eriksson (2010), Länsisalmi et al. (2000), and Suchan and Hayzak (2001) ultimately made a significantly smaller contribution.
Finally, as regards the overall balance in the model, the setting organizations in which the team incentive-related events occurred were mainly active in sectors related to healthcare and (industrial) manufacturing. It was beneficial to the balance in the model that the healthcare organizations were quite different from the industrial and manufacturing companies, but it was a matter of concern that there were almost no findings and first-order codes from other settings. The study by Länsisalmi et al. (2000) was conducted in a different kind of setting organization, a high-technology Fortune 500 consulting firm, but with a total of two findings this study hardly succeeded in making a meaningful contribution. To conclude, the overall balance in the process model constituted a research limitation in that the study by Collins (1995) had a disproportionately large share in the model, some studies hardly made any contribution, and the studies were characterized by a monotonous dichotomy between healthcare settings and manufacturing-related settings.
Logically, there are also other research limitations to be discussed. A second important limitation of the research involves that this qualitative research synthesis did not manage to establish whether there were clear connections between team dynamics and perceptions of a team incentive system, and between staff turnover and the degree of success of a team incentive system. To begin with the latter, it may well be that staff turnover is actually an integral part of the degree of success of a team incentive system, but it may also well be that staff turnover triggers events that ultimately lead to a greater or lesser degree of success. To be precise, a situation may arise in which a high-performing team member and valued colleague voluntarily leaves his/her organization out of frustration over the team incentive system in place. It is entirely possible that the resignation of this team member causes damage to the team dynamics of collaboration and learning, and that this subsequently leads to a lesser degree of success of the team incentive system. In short, the question remains as to what the precise connection between these two components is. In addition, the only thing we know about the connection between team dynamics and perceptions of a team incentive system is that events from the former sometimes lead to the latter perceptions. To be precise, because of the negative
team dynamic of free riding of low-performing peers, some team members perceived the heavy weighting of team performance in a team incentive system as unfair. However, given the findings that were at our disposal, we were unable to draw any further conclusions about the connection between these two components.
A third limitation of the research concerns that this qualitative research synthesis is not extensive or exhaustive. This is due to deliberate ignorance of the literature. Before proceeding, it should be noted that this research limitation has nothing to do with and is unrelated to the aforementioned extensiveness of the data structure. This synthesis pursued the methodology of Gioia et al. (2013) and therefore had the aim to develop a dynamic inductive model on team incentives. According to Gioia et al. (2013, p. 21), "upon consulting the literature, the research process might be viewed as transitioning from 'inductive' to a form of 'abductive' research, in that data and existing theory are now considered in tandem". This synthesis intended to remain as close as possible to the inductive slant and did so by the semi-ignorance of previous research. Obviously, previous research had to be examined to build the theoretical framework (therefore semi-ignorance), but it was attempted to develop the dynamic inductive model without being influenced by previous research. The fact that there was considerable time between constructing the theoretical framework and developing the inductive process model made it easier to hold on to this semi-ignorance. To come to the point, inductive research presumes a level of semi-ignorance and some suspension of belief in the
established wisdom of prior research and according to Gioia et al. (2013, p. 23), the literature reviews and syntheses that follow are therefore never extensive or exhaustive. The inductive approach pursued in this synthesis may not have resulted in an extensive or exhaustive process model, but at least the approach has led to some surprising and promising emergent concepts (e.g., the role of team manager as intermediary and the sense of power and control).
Fourthly, no peer examination was applied in this qualitative research synthesis. The technique of peer examination involves multiple authors each reviewing and coding articles, and makes use of multiple raters to establish inter-rater reliability (Major & Savin-Baden, 2010). It is one of the most common techniques for establishing trustworthiness and validity in qualitative research. By not embracing this technique, the synthesis may have lost some of its trustworthiness or validity. However, to add some nuance, such techniques as peer examination and inter-rater reliability reflect a positivist perspective that may belie and threaten the interpretative nature that is necessary for synthesis (Major & Savin-Baden, 2010). Gioia et al. (2013, p. 22) went even further by citing these techniques as "some sort of back-door positivism sneaking into an interpretive study". This synthesis concerns a master thesis that is generally not conducted by multiple reviewers, but in the case of a scientifically published synthesis a certain trade-off has to be made between embracing the interpretative nature of a synthesis and enhancing trustworthiness and validity through positivist techniques such as peer examination and inter-rater reliability. A final, closely related research limitation concerns that the findings and process model of this qualitative research synthesis have not been confirmed by a 'higher authority'. Such higher authorities may include influential researchers and specialists in the fields of remuneration, team incentive systems, and qualitative syntheses. Obviously, confirmation by a higher authority and constructive feedback/suggestions from such an authority could have led to a more accurate, balanced, and broad-based process model.