We evaluated the construct validity of the Child Yale-Brofwn Obsessive-Compulsive Scale (CYBOCS) in a large clinical sample (N=730) using confirmatory factor analysis. Results found inadequate fit for a priori models, though a model accounting for overlapping item content displayed good fit. Parallel obsessions/compulsions items may provide largely redundant information on the CYBOCS. Findings suggest modifying the CYBOCS to reduce burden on researchers, patients, and clinicians, and to more accurately measure pediatric obsessive-compulsive disorder (OCD).