131 Contribution of MCD to moral craftsmanship the implementation of a complex intervention, such as MCD, can be difficult for various reasons (Schildmann et al., 2019; Svantesson et al., 2014). The hoped for and expected outcome of MCD can also vary depending on the context, as demonstrated by the differences between the professional disciplines. Different outcomes of the subgroup analyses can be logically explained. The post-measurement of office staff re-integration services showed a more positive development on ‘knowing the personal values and norms of their immediate colleagues’. An explanation for this might be that office staff re-integration services usually work individually instead of in a team. Joint reflection during MCD means more contact between this staff, which can help them get to know each other better, and gain more insight into each other’s personal values and norms. Only 34% (54/156) of the intervention group participants indicated participating in 7 or more MCD sessions in the post-measurement. A large proportion of the participants therefore took part in less than 7 MCD sessions. As researchers, we expected that a minimum of 7 to 10 MCD sessions would be required for a clear positive impact on MCS of prison staff. Contrary to expectations, we did not find a greater impact with an increasing number of MCD sessions. However, this may be explained by a lower-thanexpected overall impact of MCD, making it difficult to find an effect in subgroups. Also, the content of some items in the questionnaire were open to multiple interpretations. This should be improved in a new version of the questionnaire. For example, whether a change is positive or negative for MCS depends on the meaning of the item. It is, for example, possible that some participants view an item such as ‘In my work it is allowed to make mistakes’ as a negative statement (for example, because of the focus on safety in prison it is perceived as not being attentive or careful), which may make them more inclined to disagree with this statement. However, one could also interpret this positively: we are open to making mistakes and learn from them. Prison staff who have not yet participated in MCD sessions may respond to items differently than after attending MCD. Staff who participate in MCD can learn a lot from and score higher on elements of MCS. However, it is also possible that, because of what they learn in MCD, they become more aware that some items are more or barely present, or absent in practice, so they score lower in the post-measurement. 5