Background: There is a growing interest in the use of cognitive approaches for the treatment of psychological difficulties with people who have intellectual disability. The aim of the present study was to investigate the properties of an anger inventory, the relationship between self-reported and staff-reported levels of anger, and the relationship between different staff measures. Method: Two studies are described that use self-report and staff-report measures. The first study employs a cross sectional design and the second, a longitudinal design. Results: No relationship was found between self-reported and staff-reported levels of anger on the first measurement occasion, but on the second measurement occasion after some intervention a significant relationship was found. If staff had accompanied individuals during an intervention a larger correlation was found. Conclusion: These results suggest that staff may require specific training or exposure to a therapeutic environment before they can predict the response of people with intellectual disability on the anger inventory. This suggests that involvement of care staff in interventions may be useful therapeutically. (Contains 3 tables.)