The current study aimed to determine the extent of association between mental health symptoms and impairment in children using teacher ratings. Strengths and Difficulties Questionnaires (SDQ) data were extracted from an administrative database for consecutively referred children to a school mental health service between 2004 and 2009 (n = 480). Incremental increases in symptoms predicted increasing impairment levels in the sample using both linear and dichotomized impairment outcomes (both p < 0.0001). Unexpectedly, high estimated probabilities of abnormal impairment were observed well within the normal symptom range, particularly for younger elementary school-aged boys. The findings suggest that screening efforts that depend on cut-points on the SDQ total problem scale may be missing many substantially impaired children. Methodological limitations are discussed.