The objective of this study was to evaluate associations between sleep problems and psychiatric symptoms at school. A random sample consisting of 5813 eight- to nine-year-old children was selected from ordinary schools. Both parents' and children's reports of sleep problems were taken into account. The psychiatric symptoms were addressed according to the teachers' reports (the Rutter Scale B). Children with severe sleep problems were more likely to have a psychiatric disturbance according to the Rutter B Scale (OR 2.45, 95% CI 1.85-3.25). Logistic regression models showed that severe sleep problems were highly associated with emotional problems (OR 2.74, 95% CI 1.84-4.13), school attendance problems (OR 2.53, 95% OR 1.45-4.41), behavioural problems (OR 2.44, 95% CI 1.59-3.75) and hyperactivity (OR 2.02, 95% CI 1.30-3.13). Over 95% of severe sleep problems were reported only by the children themselves. In conclusion, children with severe sleep problems have substantially more teacher-reported psychiatric symptoms than those with no or mild sleep complaints. In diagnosing sleep disorders, it is important to include children as informants because relevant information may be overlooked when only parents are questioned.