Problem behaviour in schools may have detrimental effects both on students' well‐being and academic achievement. A large literature has consistently found that school‐wide positive behaviour support (SWPBS) successfully addresses social and behavioural problems. In this paper, we used population‐wide longitudinal register data for all Norwegian primary schools and a difference‐in‐difference (DiD) design to evaluate effects of SWPBS on a number of primary and secondary outcomes, including indicators of externalising behaviour, school well‐being, pull‐out instruction, and academic achievement. Indications of reduced classroom noise were found. No other effects were detected. Analyses revealed important differences in outcomes between the intervention and control schools, independent of the implementation of SWPBS, and that a credible design like DiD is essential to handle such school differences.