Hester Paanakker

effects and responses this creates at the frontline, may well be generalizable to other public sector organizations. The motivation to select the Dutch prison sector as a case study was partly due to thematic and cross cutting issues that have a strong potential to be applicable to many other (types of) frontline organizations in public service delivery. For instance, the prison sector was selected because the role of values in its public service delivery is markedly complex, with multiple inherently conflicting public service values, a large distance between policy formulation level and policy implementation level, and street-level workers operating in demanding and unpredictable contexts in terms of the nature of the work and the beneficiaries they attend to, as well as in terms of the political volatility that shapes its institutional context (Gofen, 2013; Stewart, 2006; Stewart & Kringas, 2003). In addition, the prison sector provides a context of public sector change and reform, cutbacks and neoliberal strategies and performance rhetoric that inherently affects values and their attainment, also in relation to a frontline public craft that comes increasingly under pressure as a result (Connell et al., 2009; Maroulis & Wilensky, 2014; Pollitt, 2008; Pollitt & Bouckaert, 2017; Trommel, 2018; Tummers et al., 2015; Van de Walle, 2006; Wynen et al., 2019). These characteristics describe the settings of many public service organizations nowadays. Hence, frontline organizations throughout the public sector may very well experience a comparable high level of value divergence on value enactment and comparable dynamics of toxic stereotyping throughout policy sectors. In addition, frontline workers from other organizations may employ the same types of strategies to cope with value divergence, indicative of their adaptivity and resilience on the one hand, or of their alienating attitudes otherwise. This constitutes an interesting avenue for future research. Another important limitation of the case study method employed is that the findings do not always allow for hard causal inference (R.K. Yin, 1989). This does also not fit the explorative aims and set-up of the study, but does provide important leads for future studies to build on. Specifically, future research should assess whether, and how, the degree of value divergence impacts on the level and nature of implementation problems and moral dilemmas. Larger scale quantitative data collection and analysis can help to provide insight into the exact nature and strength of these relationships. It could also provide more insight into the differences in causation of experienced and real value divergence, an issue this thesis provides strong indication of – with the former having graver impact than the latter – but could not be settled in 171 Conclusions and Discussion