Total knee arthroplasty (TKA) is being undertaken in a younger population than before and as a result the functional demands on the knee are likely to be increasing. As a consequence, it is important to define quantitative functional knee tests that can monitor any increase. A valuable functional knee test has to be able to distinguish small differences (selectivity) and has to be independent of pain (content validity). In this study, patient-based questionnaires (WOMAC and Knee Society score) and performance-based tests (sit-to-stand movement, maximal isometric contraction and timed-up-and-go) were used to assess which of these tests are selective and valid to measure knee function. Tests were considered to be selective if they could discriminate between knee patients and healthy control subjects, and to have functional content validity if they were relatively independent of pain. Twenty-eight patients were measured 16 months after surgery and compared to a healthy control group of 31 subjects. The sit-to-stand movement and timed-up-and-go test were both selective and functionally content valid. The timed-up-and-go test can be used for a quick initial assessment of global function and the sit-to-stand movement as a more biomechanical instrument identifying how the knee function of the patient is affected.