OBJECTIVE: A quantitative measure of wrist position sense was developed to advance clinical measurement of proprioceptive limb sensibility after stroke. Test-retest reliability, normative standards, and ability to discriminate impaired and unimpaired performance were investigated. DESIGN: Retest reliability was assessed over three sessions, and a matched-pairs study compared stroke and unimpaired subjects. Both wrists were tested, in counterbalanced order. SETTING: Patients were tested in hospital-based rehabilitation units. PATIENTS AND OTHER PARTICIPANTS: Reliability was investigated on a consecutive sample of 35 adult stroke patients with a range of proprioceptive discrimination abilities and no evidence of neglect. A consecutive sample of 50 stroke patients and convenience sample of 50 healthy volunteers, matched for age, sex, and hand dominance, were tested in the normative-discriminative study. Age and sex were representative of the adult stroke population. MAIN OUTCOME MEASURES: The test required matching of imposed wrist positions using a pointer aligned with the axis of movement and a protractor scale. RESULTS: The test was reliable (r = .88 and .92) and observed changes of 8 degrees can be interpreted, with 95% confidence, as genuine. Scores of healthy volunteers ranged from 3.1 degrees to 10.9 degrees average error. The criterion of impairment was conservatively defined as 11 degrees (+/-4.8 degrees) average error. Impaired and unimpaired performance were well differentiated. CONCLUSIONS: Clinicians can confidently and quantitatively sample one aspect of proprioceptive sensibility in stroke patients using the wrist position sense test. Development of tests on other joints using the present approach is supported by our findings.