Objectives: The Arthritis Self-Efficacy Scale (ASES) is a widely used self-report measure of beliefs reflecting confidence in one's capacity to function despite pain and control pain or other symptoms of arthritis. Despite evidence linking higher ASES scores to lower levels of impairment, pain and emotional distress, numerous modest, non-significant associations have also been observed. In this meta-analysis, we evaluated overall associations between ASES scores and adjustment in RA and OA samples as well as potential moderators that may explain the heterogeneity in these associations. Method: Data from 48 samples that met all 10 inclusion criteria (N = 9222 patients) were subject to analyses. Results: ASES scores had significant medium average effect sizes with functional impairment, pain severity and emotional distress but substantial heterogeneity was evident for each association. ASES-impairment associations were moderated by the diagnosis, ASES version and ASES subscale content: significantly larger effect sizes were found for studies that included RA patients, used the original 20-item ASES and assessed subscale content reflecting the pursuit of daily activities despite pain (i.e. functional self-efficacy) than for studies based exclusively on OA patients, the eight-item ASES and ASES pain control and other symptom subscales. Relations of ASES scores with pain severity and emotional distress were moderated by ASES version and subscale content, respectively. Conclusion: The ASES has significant overall associations with key areas of functioning. Moderator analyses of the measure provide empirically grounded suggestions for optimal use of the ASES within OA and RA patient samples.