Faced with limited resources, schools require tools that increase the accuracy and efficiency of functional behavioral assessment. Yarbrough and Carr (2000) provided evidence that informant confidence ratings of the likelihood of problem behavior in specific situations offered a promising tool for predicting the accuracy of function-based hypotheses developed from staff interviews. The current study evaluated conditions in which a similar rating of informant confidence was effective in predicting the accuracy of functional assessment hypothesis statements. Nine students with problem behavior were identified, and functional behavioral assessment interviews with confidence scores were completed with 58 staff members. Between five and eight adults were interviewed about each student. The adults were selected based on their range of contact with the student (0 to 10+ hours per week) and their self-assessed knowledge about behavioral theory (no knowledge to extensive knowledge). Functional analyses were conducted to assess agreement with functional assessment hypotheses and the predictive value of confidence ratings. Results suggested limitations to the general use of confidence ratings in distinguishing accurate from inaccurate functional hypotheses across school staff with a broad range of contact with the target student. The study did find that informants who were both highly confident and who identified accurate functional assessment hypotheses had significantly higher levels of contact with the student in the target routine than those informants who had low confidence ratings and/or identified an incorrect function for the problem behavior. (Contains 4 tables.)