Possible escape/avoidance functions of stereotypic behavior were investigated in a classroom setting using functional academic tasks. A 6-year-old boy's stereotypic mouthing was assessed during high vs low response activities, familiar vs novel activities and avoidance vs partial-avoidance conditions. Results showed that stereotypy occurred at higher rates during more difficult activities (i.e. those that were novel or required a greater number of responses), and when stereotypy was allowed to effect a delay in instructional demands. Treatment procedures based on these analyses were implemented by the classroom teacher and shown to be effective.