Hester Paanakker

7.2 Limitations of the Study and an Agenda for Future Research As with any research, the research in this thesis has some limitations that affect the reliability and generalizability of the findings and need to be taken into account to determine the scope and positioning of the research (R.K. Yin, 1989). The reflection on the limitations also provides suggestions for future research on value (divergence) and street-level craft. This will be discussed below. First, a common limitation of case study research is its scope in terms of the number of cases and respondents taken into account (R.K. Yin, 1989). The data in this thesis was collected only in the prison sector in one country, although in two different facilities to account for interorganizational differences. The results generated value patterns that were internally very consistent. This seems to confirm that the findings reveal value approaches and dynamics that are widespread and endemic to penal service delivery, and therefore generalizable to the Dutch prison sector at large. The analysis did not generate significant differences between the two cases, which seems to confirm that there is a shared sense-making and problematizing of value (divergence) between penal facilities, despite and beyond their differences in geographical setting, size, and phase of detention, and confirm that, for the purpose of this explorative study, the set-up of the study was broad enough in its scope. The same can be argued for the number of respondents. An inherent limitation of the case study design in this thesis is that the respondent categories vary in number of respondents. For instance, each facility only has a team of three or four managing directors in total, so this group is by necessity smaller than the group of prison officers. When the groups reflect on each other’s views, this inevitably also generates slightly different numbers per respondent group, which could distort the findings. Although no indication of skewed balance was found, it is acknowledged that a larger number of respondents, especially from policy level and management levels, could contribute to a more representative picture. However, it should be noted that, for each of the two cases, the entire staff of directors was interviewed, and a very large share of its middle managers, which adds to the credibility of the findings and increases intra-case generalizability. Nevertheless, additional quantitative empirical research can help to process the qualitative assessment of value approaches into the development of penal surveys that can incorporate a larger number of cases, particularly with a larger number of respondents 169 Conclusions and Discussion