BACKGROUND: Data from monkeys with deafferented forelimbs and humans after stroke indicate that tests of the motor capacity of impaired extremities can overestimate their spontaneous use. Before the Motor Activity Log (MAL) was developed, no instruments assessed spontaneous use of a hemiparetic arm outside the treatment setting. OBJECTIVE: To study the MAL's reliability and validity for assessing real-world quality of movement (QOM scale) and amount of use (AOU scale) of the hemiparetic arm in stroke survivors. METHODS: Participants in a multisite clinical trial completed a 30-item MAL before and after treatment (n = 106) or an equivalent no-treatment period (n = 116). Participants also completed the Stroke Impact Scale (SIS) and wore accelerometers that monitored arm movement for three consecutive days outside the laboratory. All were 3 to 12 months post-stroke and had mild to moderate paresis of an upper extremity. RESULTS:After an item analysis, two MAL tasks were eliminated. Revised participant MAL QOM scores were reliable (r =0.82). Validity was also supported. During the first observation period, the correlation between QOM and SIS Hand Function scale scores was 0.72. The corresponding correlation for QOM and accelerometry values was 0.52. Participant QOM and AOU scores were highly correlated (r = 0.92). CONCLUSIONS: The participant Motor Activity Log is reliable and valid in individuals with subacute stroke. It might be employed to assess the real-world effects of upper extremity neurorehabilitation and detect deficits in spontaneous use of the hemiparetic arm in daily life. ©2006AAN Enterprises, Inc.