Objective: To evaluate the reliability and validity of accelerometry for measuring upper-extremity rehabilitation outcome. Design: Validation study. Setting: Data recorded in the community. Participants: Consecutive Constraint-Induced Movement therapy (CIMT) patients (n=10) and volunteer community residents with stroke (n=10). All participants were more than 1 year poststroke and had mild to moderate motor impairment of the more affected arm. Intervention: All study participants were asked to wear accelerometers outside the laboratory for 3 days immediately before and after treatment, or for an approximately equivalent no-treatment period (controls). Main Outcome Measures: Participants wore an accelerometer on each arm, the chest, and the more affected leg and completed the Motor Activity Log (MAL), which is a semistructured interview of real-world arm use. Results: Test-retest reliability of transformed accelerometer recordings was greater than .86. There was also a large increase in the ratio of transformed more- to less-impaired arm recordings in CIMT therapy patients (d′=0.9, P<.05), while there was no change for controls. The correlation between this parameter and the MAL was .74 (P<.001). Conclusions: Accelerometry provides an objective, real-world index of upper-extremity rehabilitation outcome and has good psychometric properties. © 2005 by American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation.