||A multitrait-multimethod (MtMm) procedure was used to investigate the magnitude of association between disability and depression as well as the convergent and discriminant validity of the Health Assessment Questionnaire (HAQ) Disability Index and two measure of depression, the Beck Depression Inventory (BDI) and the Hamilton Interview Ratting Scale for Depression (HRS-D). Three measures of disability on the HAQ (patient self-report, physical therapist assessment and spouse report) were compared with the two measures of depression in 107 subjects with rheumatoid arthritis (RA). Pearson correlations among the three disability measures provide strong evidence in support of the convergent validity of the HAQ Disability Index ( r's- . 8 9 t o . 8 5 , D'S < . 0 0 0 ). Similarly, the correlation between the BDI and the HRS-D ( rs . 6 9 , p. < . 0 0 1 ) provides evidence of convergent validity between the construct of depression and the construct of disability was greater when disease-related somatic difficulties were not included in the measure of depression. The high correlations among the disability measures suggest that subjects report their physical disability on a health status questionnaire with a high level of accuracy. In addition, the small discrepancies that were observed between self-reported and physical therapist assessed disability were clearly not associated with depression. This observation was true even though items that may have been endorsed due to disease-related difficulty rather than affective disturbance were included in the BDI score. However, a measure of helplessness and perceived loss of control, the Rheumatology Attitude Index (RAI), was associated with discrepancy between self-reported and physical therapist assessed disability ( r = . 2 0 , p_<.05), indicating a sense of helplessness and perceived loss of control may have contributed to subject over-report of disability. Finally, although disability and depression were positively correlated in the MtMm matrix, the level of this association was related to the methodological issue of whether or not items reflecting somatic difficulties were included in assessing depression. Due to high somatic and disease-related content on many of the BDI items and the fact that the HRS-D was not scored positively on similar items when the complaints reflected disease-related difficulty, the correlations between the HRS-D and the measures of disability (r's = . 17 to . 2 5 , pjs < - 0 1 ) likely reflect a more realistic level of association between disability and depression in RA than do the correlations between the BDI and the measures of disability ( r 's = . 3 1 to . 5 0 , p's < . 0 0 1 ) .