This study evaluated the ability of the Child Behavior Checklist and the Teacher's Report Form to identify children with autism spectrum disorders (ASD), using a sample of children with ASD (n = 458), referred children without ASD (n = 1109) and children from the general population (n = 999). A ten items ASD scale was constructed using half of the sample and the ability of this scale to discriminate between children with ASD and the other children was tested for the CBCL and the TRF separately and together in the other half of the sample. Using a cut-off score of 8 the combined CBCL/TRF ASD scale demonstrated high predictive values (NPV 95%, PPV 73%) in identifying children with ASD and children in the general population sample. This might be an acceptable percentage of false positives in general screening, considering the chance that these children might have other behavioural, emotional, and developmental problems which also need psychiatric evaluation. In the referred population, using a cut-off of 13, PPV was 49% and NPV was 85%. The high NPV indicates that in a referred population the scale is especially good at identifying children who do not need evaluation with a more ASD-specific instrument.