25 General introduction researchers shared relevant findings from the ongoing research to highlight experiences that could potentially improve the implementation process. Since several locations were involved, an additional national steering committee was formed during the research. This group consisted of the research team, stakeholders of DCIA concerning this part of the craftsmanship program, and the involved coordinator and director of the DCIA Educational Institute. Every three months, this group discussed the progress of the project with MCD and of the research to enhance the collaboration between all involved and gain more understanding of each other’s work. Furthermore, to foster mid-term learning of MCD facilitators and jointly search for possible improvements in their role, peer meetings were organized for MCD facilitators working at DCIA. We organized this together with the DCIA Educational Institute and within the craftsmanship program. These meetings give facilitators the opportunity to continue to develop themselves (Stolper et al., 2014, p. 256) and improve their connection with the participants in this specific context. Most MCD facilitators had no prior experience with working in the prison context. Both the steering committees and these peer meetings created an environment that enabled the responsive approach; researchers shared anonymized research data – based on the interviews with participants, focus groups with facilitators, and data from the evaluation forms – to start a dialogue and collaborative learning process about the results. Adaptations in our design due to the closing of prisons Initially, two institutions were selected as the intervention group: the prisons of Nieuwegein and Zwaag. At these two locations, 11 teams started with MCD and were part of the research. Two comparable locations were selected as a control group: the prisons of Almere and Leeuwarden. This original set-up changed during the research period because in the summer of 2018, one location from the intervention group (Zwaag) and one from the control locations (Almere) were closed by the Ministry of Justice and Security. To collect enough reliable data for our analyses, we started the project at two ‘new’ intervention- and control locations from 2019 onwards. When the data collection of the control location Leeuwarden was completed, they wished to start implementing MCD. Therefore, after the research on the control group Leeuwarden had ended, they started as an intervention group. We selected one new location to be the control group for this phase: the prison of Vught. Eventually, these five prison locations were part of this research project. Figure 2 shows the data collection period for the intervention and control groups. 1