Valid and reliable teacher ratings serve as the foundation for screening and assessment of youth with behavioral disorders and twin studies offer an opportunity to study those ratings. We conducted a meta-analysis of 15 empirical investigations of aggressive and rule-breaking behavior using teacher ratings in the context of a twin research design. We retrieved n = 53 correlations from n = 7,885 pairs of monozygotic (MZ) twins and n = 67 correlations from n = 11,696 pairs of dizygotic (DZ) twins with two goals: (a) to test the best-fitting models for predicting similarity (correlations) in teacher ratings of MZ and DZ twin pairs and (b) to identify significant predictors in the respective models. We found that, for both MZ and DZ twins, the best-fitting model included all predictors and interactions. In the MZ model, we found a significant positive interaction between the percentage of same teachers who completed the ratings and male twins; in the DZ model, we found a significant negative interaction between the percentage of same teachers who completed the ratings and opposite sex twins. Teacher ratings converged and diverged in expected ways, advancing research in the context of the attributions-bias-context (ABC) model of informant ratings.