The revised 28-item Conners' Teacher Behavior Rating Scale (TBRS) is subjected to factor analysis replication, with a referral sample of 354 children. The factor structure is more clearly defined than in the original normative study, and a primary factor of hyperactivity emerged, accounting for 39% of the variance. Six factors that emerged, all with eigenvalues of 1.00 or more, accounted for 69% of the variance. Hyperactivity and conduct factors, which are consistent over multiple studies, emerged. Items consistent with inattentive, unsociable, and passive behavior also separated into discrete factors. This suggests that using a mixed clinical population produces a clearer delineation of factors, which may provide the basis for subtyping a clinical population. Use of the TBRS as the diagnostic measure in medication studies, and its relationship with the abbreviated Teacher Rating Scale (ATRS), are explored.