We evaluate two variants of a school-based, intensive learning camp for pupils who are assessed 'not ready' for further education after compulsory school, using a stratified cluster randomized trial involving 15,559 pupils in 264 schools in Denmark. Next to training pupils in Danish and mathematics, the main variant targets non-cognitive skills, while the alternative variant instead uses this time for more training in Danish and math. In the short-run, in the academic areas that are targeted in the camp, we find small positive effects in math and weak evidence for positive effects in Danish. Yet, we find no evidence of lasting effects and we do not find short-run effects on non-targeted areas in math and Danish or on non-cognitive skills. Further, we find no evidence that training of non-cognitive skills affects academic outcomes. These results provide a perspective on recent evidence regarding the effects of training non-cognitive skills in schools - by running such an intervention with older pupils and in a comparatively high-resource school system.