In spite of a multitude of scoring approaches, the ability of the Rey-Osterrieth Complex Figure (ROCF) to measure executive functions versus grapho-motor skill is still open to question. To clarify this issue, we examined the performance of an ethnically and socioeconomically diverse sample of preadolescent girls (ADHD-Combined, n = 93; ADHD-Inattentive, n = 47; and comparison girls, n = 88) on the ROCF, scoring both immediate copy and delayed recall performance. Girls with ADHD performed the task following a stimulant medication washout. Dependent measures were the established Developmental Scoring System variables of Organization, Accuracy, and Style, plus an operationalized and extensively detailed scoring of errors, featuring an error proportion score (EPS). The major finding is that only EPS differentiated girls with ADHD from comparison girls (a) on both immediate and delayed performance and (b) with stringent statistical control of Performance IQ, fine motor speed, and performance on the Porteus Mazes (as well as comorbidities), with effect sizes in the medium range. Perseverative errors contributed significantly to EPS. Overall, error scores on the ROCF appear to tap planning, a key executive function, and are quite sensitive to deficits in this domain for girls with ADHD.