We study the impact on criminal activity from a large scale Swedish reform of vocational upper secondary education, extending programs from two to three years and adding more general theoretical content. The reform directly concerns age groups where criminal activity is high and students who are highly overrepresented among criminal offenders. The nature of the reform and the rich administrative data allow us to shed light on several behavioral mechanisms. Our results show that the prolonged and more general education lead to a reduction in property crime, but no significant decrease in violent crime. The effect is mainly concentrated to the third year after enrollment, which suggests that being in school reduces the opportunities and/or inclinations to commit crime.