Abstract Introduction Applying a cross-sectional analysis to a sample of 2,627 African-American and Caucasian adults aged ≥ 45 years from the Johnston County Osteoarthritis Project, we studied the association between educational attainment and prevalence of radiographic knee osteoarthritis and symptomatic knee osteoarthritis. Methods Age- and race-adjusted associations between education and osteoarthritis outcomes were assessed by gender-stratified logistic regression models, with additional models adjusting for body mass index, knee injury, smoking, alcohol use, and occupational factors. Results In an analysis of all participants, low educational attainment (<12 years) was associated with higher prevalence of four knee osteoarthritis outcomes (unilateral and bilateral radiographic and symptomatic osteoarthritis). Women with low educational attainment had 50% higher odds of having radiographic knee osteoarthritis and 65% higher odds of symptomatic knee osteoarthritis compared with those with higher educational attainment (≥ 12 years), by using fully adjusted models. In the subset of postmenopausal women, these associations tended to be weaker but little affected by adjustment for hormone replacement therapy. Men with low educational attainment had 85% higher odds of having symptomatic knee osteoarthritis by using fully adjusted models, but the association with radiographic knee osteoarthritis was explained by age. Conclusions After adjustment for known risk factors, educational attainment, as an indicator of socioeconomic status, is associated with symptomatic knee osteoarthritis in both men and women and with radiographic knee osteoarthritis in women.