Self-reported joint pain, a typical manifestation of osteoarthritis, was examined using 335 twin pairs from the Swedish Adoption/Twin Study of Aging to estimate relative genetic and environmental influences on self-reported joint pain and to examine the relationships between joint pain, health behavior, and psychological variables. Findings suggest that family resemblance for self-reported joint pain represents similar environments more than genetic similarity. Data from the early 1970s, including exercise, physical activity at work, obesity, and neuroticism, were used to predict joint pain in 1993. For men, moderate amounts of exercise decreased the likelihood of joint pain, but strenuous amounts of physical activity in the workplace had the opposite effect. For women, exercise and physical activity were not significant predictors, but past obesity and higher levels of neuroticism increased the likelihood of reporting joint pain in 1993.