Abstract We present a factor analytic study of the Conners Rating Scales for parents and teachers in this article. A comparison is established with the original factor analytic studies (Conners, 1979a, 1979b) and the results obtained by Farré-Riba and Narbona (1997), using a Spanish sample. Five hundred and forty children, ages 4 to 17, were randomly selected in Manizales, Colombia. The shortened Spanish version of the Conners Parents Rating Scale (CPRS; Conners, 1979a) and the Conners Teacher Rating Scale (CTRS; Conners, 1979b) were used. Parsimonious factor analyses for both scales were developed. Three stable factors were found in the CPRS: hyperactivity, somatic symptoms, and inattention. Twenty-two out of 48 items were significantly saturated by these 3 factors. In the CTRS, 4 different factors were found that accounted for 63.5% of the total variance: uncontrolled temperament, inattention, hyperactivity, and difficulties in social relationships. Twenty out of 28 items were significantly saturated by these 3 factors. Factor structure was closer to the Spanish Farré-Riba and Narbona report than to the original findings. We concluded that the CPRS and the CTRS Spanish versions, when used by Colombian children and adolescents, do not seem to evaluate exactly the same underlying behavioral dimensions. We propose selecting only 22 items of the CPRS and 20 items of the CTRS (brief versions) for further epidemiological and clinical use.