Although universal screening for mental health difficulties is increasingly recognized as a way to identify children who are at risk and provide early intervention, little research exists to inform decisions about screening, such as the choice of informants and the type of information collected. The present study examined the incremental validity of teacher- and parent-rated (primarily mothers) symptoms and impairment in a non-referred sample of early elementary school children (n = 320, 49 % boys, ages 6 to 9) in terms of predicting impairment as rated by a different teacher 1 year later. Teacher-rated symptoms and impairment and parent-rated impairment were each unique predictors of later impairment; however, parent-rated symptoms did not contribute to the prediction of later impairment above and beyond these other indicators. The results indicate that, when screening for mental health difficulties in the school system, impairment ratings collected across settings add useful information, but it may not be necessary to use parent symptom ratings when teacher symptom ratings are available.