Aims: Does mastering abilities have an independent effect on disability pensioning? Are some of the large divides across education explained by mastering factors when adjusted for health? Methods: Data from the National Insurance Administration were linked to a health survey of 10,720 Oslo inhabitants aged 40, 45, and 60 years, 46% of the total populations in these age groups who were invited to a Health Survey in 2000-01. Measures of general health, chronic somatic diseases, psychiatric symptoms, and musculoskeletal pain were self-reported. The mastering test Generalized Self Efficacy was used. Results: 10.5% of our eligible sample had disability pension at the time of the survey. The risk was more than five times higher for those with primary school than for those with university education and 50% higher for women than for men. Lowest score on the Generalized Self Efficacy test (poor mastering) had an age-, gender-, and health-adjusted OR of 2.4 compared with the highest level of mastering. Adjusting for mastering lowered the educational divide but not that of gender, when health indicators were taken into consideration Those reporting poor general health had a seven times higher risk than those with good health, and those with a chronic somatic disease, musculoskeletal pain, or poor psychiatric health had a somewhat lower increase in risk of disability pension. Health measures did reduce the impact of education, but not of gender, when adjusted for mastering. Conclusion: Poor mastering was associated with disability pensioning, and reduced the differences across educational level and health but not across gender.