In recent years, various governments and education agencies have developed stricter policies to reduce truancy levels, mainly based on the argument that truancy is associated with risk behaviour, crime and substance abuse. In this article, we use a large, 28-nation comparative survey among 14 year olds to detect general patterns in consequences and causes of truancy by using multilevel analysis differentiating between the individual, school and country levels. The analysis shows that schools can have a major impact on truancy levels by promoting school involvement from parents and by providing a supportive and authoritative environment. The analysis also demonstrates that high truancy levels at school reduce test scores of pupils, even after including a battery of socio-economic control variables. In the conclusion, we argue that efforts to reduce school absenteeism cannot be limited to repressive "truancy sweeps", but should also take into account school climate and the interaction between pupils, parents and schools. Reducing truancy should not be seen just as a law and order problem since truancy has a significant effect on educational achievement, especially among children with lower status family backgrounds. (Contains 6 tables and 1 figure.)