Despite frequent reliance on teacher and parent ratings of children's behavior for multi-informant assessment, agreement between teachers' and parents' ratings is low. This study examined the predictive utility of teacher and parent ratings for children's self-regulatory outcomes (i.e., executive function, continuous performance task) in four studies. Study 1 included 163 children ranging from 31 to 84 months of age (M = 55.58 months, SD = 8.43). Study 2 included 1,088 children ranging from 48 to 63 months of age (M = 55.15, SD = 3.65). Study 3 included 246 bilingual Spanish-speaking children ranging from 40 to 72 months of age (M = 56.66, SD = 6.06). Study 4 included 280 children ranging from 38 to 75 months of age (M = 55.92, SD = 4.16). Across studies, parents and teachers rated children's externalizing behaviors on the Conners' Rating Scale, the Strength and Weaknesses of ADHD-Symptoms and Normal Behavior Scale, or both; children completed a variety of performance-based self-regulation tasks. The strength of associations of parent and teacher ratings with self-regulatory outcomes was compared, and regression analyses determined the unique and overlapping variance accounted for by different raters. Teachers' ratings had larger associations with self-regulation than did parents' ratings across outcomes and studies--except for two instances in Study 4 where the associations for teachers' and parents' ratings were equal. These findings indicate that teachers supply more useful information than parents, possibly because teachers have better-informed expectations of children's behavior, and they raise questions about the utility of multi-informant assessment, at least with preschool children. [This paper will be published in "Research on Child and Adolescent Psychopathology."]