This study provides a test of the presumption that police recruits with a diverse background, undertaking comparatively long academic training, will refrain from non-legalistic practices. This is tested by longitudinal survey data, covering two cohorts of Swedish police recruits. The results show stable support for the legalistic perspective during academy training. However, during on-the-job training, the recruits become more positive towards non-legalistic practices. This reorientation takes place quite irrespective of the type of duty to which they are assigned. Additionally, neither the recruits' nor their parents' level of education seems to matter. There is some effect of age and gender; young male recruits are somewhat more prone to adopt Dirty Harry-inspired measures-that is, achieving essential ends by tarnished means.

Frontline workers play a crucial role in implementing activation policies. Nevertheless, research on what competencies are required for activation work is limited. We explored activation competency based on a survey of 1,735 frontline workers in the Norwegian labour and welfare administration. Factor analysis revealed two distinct underlying dimensions in activation competency: market competency and user‐oriented competency. We found that the social workers in the study viewed themselves as having significantly less market competency and slightly more user‐oriented competency than non‐social workers have, but the differences were small. The results also indicate that these effects are partially mediated by attitudes towards conditionality. The results give reason to treat activation competencies as twofold and raise the question of whether social work education improves frontline workers' competency in activation work compared with frontline workers with other educational backgrounds.

a b s t r a c tBackground: The Stockholm Prevents Alcohol and Drug Problems (STAD) programme has been regarded as one of the most successful programmes to date, in reducing alcohol-related violence. This multicomponent Responsible Beverage Service (RBS) programme was implemented in Stockholm, Sweden, and has been documented to be extremely effective in reducing alcohol-related nightlife violence. The SALUTT programme in Oslo, Norway was carefully modelled on the STAD project. Aim: We investigate whether the results from STAD were replicated in the SALUTT intervention. Design: Using geocoded data, the level of violence in the intervention area was compared with different control areas before and after the intervention. Statistics: Autoregressive moving average models (ARIMA). Findings: The SALUTT programme had no statistically significant effect on violence. However, the level of violence in the different potential control areas of Oslo fluctuated without a clear common trend. Hence, it was difficult to establish proper control areas. Conclusions:The results from the Swedish STAD-intervention were not replicated in Oslo. Successful interventions are not necessarily replicated in other contexts, and the current literature does not shed sufficient light on the conditions under which such interventions actually work. Moreover, more attention should be devoted to the identification of adequate control areas in future research.

No abstract