Behavioral problems in young children can be assessed by asking their parents or teachers to rate their behaviors. Genetic analyses of parental ratings show relatively large heritabilities for emotional and behavioral problems in young children, but data from teachers for this age group are scarce. Sources of variation in the Teacher's Report Form (TRF) problem scales were examined. The TRF was completed for 211 Dutch 5-year-old twin pairs and 4 single twins. Twins rated by different teachers had higher means and variances than twins rated by the same teacher, in addition twin correlations were lower in this group. In both groups monozygotic (MZ) correlations were generally higher than dizygotic (DZ) correlations. A model for twin resemblance was tested that allowed for these effects. For 5 problem scales (Withdrawn, Social Problems, Aggressive Behavior, Rule Breaking Behavior and Attention Problems) a model with genetic and unique environmental sources of variation fitted best to the data. For 3 problem scales (Anxious/Depressed, Thought Problems and Somatic Complaints) there were familial influences but it was not possible to distinguish between common environmental influences or genetic influences. Heritability was 63% for Attention problems, around 45% for Withdrawn, Social Problems, Aggressive Behavior and Rule Breaking Behavior, and around 30% for Anxious/Depressed, Thought Problems and Somatic Complaints.