OBJECTIVE: Enhanced pain facilitation is reportedly an important contributor to the clinical pain experiences of individuals with knee osteoarthritis (OA). Ethnic differences in the prevalence and severity of knee OA in addition to associated pain are also well documented. Temporal summation (TS) of pain is a widely applicable quantitative sensory testing method that invokes neural mechanisms related to pain facilitatory processes. This study tested whether TS of pain, an index of pain facilitation, differentially predicts the clinical pain experiences of African Americans and non-Hispanic whites with symptomatic knee OA. METHODS: A total of 225 study participants underwent assessment of TS of mechanical and heat pain stimuli applied to their most symptomatic knee and their ipsilateral hand (mechanical) or forearm (heat). Using telephone-based surveys, participants subsequently reported their average and worst clinical pain severity across four consecutive weeks after the assessment of TS. RESULTS: In predicting future clinical pain, ethnicity interacted with TS of mechanical pain (but not heat pain), such that TS of mechanical pain at the knee significantly predicted greater clinical ratings of average (b = 0.02, p = .016) and worst (b = 0.02, p = .044) clinical pain for non-Hispanic whites but not African Americans (p values > .30). CONCLUSIONS: These results reveal the importance of considering ethnicity when examining pain facilitation and the clinical pain of individuals with symptomatic knee OA. The results of this study are discussed in terms of ethnic differences in the predictors of clinical pain experiences among African Americans and non-Hispanic whites with knee OA. © 2014 by the American Psychosomatic Society.