Objective To investigate the construct validity ON medication and the reliability both ON and OFF medication of linear encoder muscle power testing in persons with Parkinson's disease (pwPD). Design A study using baseline data from one randomized controlled trial (study 1) and one cohort study (study 2). Setting University exercise lab. Participants Study 1: 35 healthy controls and 70 pwPD. Study 2: 20 pwPD. Intervention Study 1: baseline data. Study 2: 4 chair rise tests (2 ON and 2 OFF medication), in a randomized order, separated by 4 to 16 days. Main measures Linear encoder data were collected from a chair rise test. Known groups validity and convergent validity (i.e., construct validity) were assessed by comparing peak power between pwPD and healthy controls and associations between peak power and functional performance (i.e., 6-Min Walk Test, Timed Up and Go Test, Six-Spot Step Test), respectively. Reliability was assessed as day-to-day variation and by intraclass correlation coefficients. Results Peak power was comparable between pwPD and healthy controls (−7.2%, p = 0.17), but lower in moderately impaired pwPD compared to mildly impaired pwPD (−27%, p < 0.01) and healthy controls (−23%, p < 0.01). Moderate to strong associations were observed between peak power and functional performance (r2 = 0.44–0.51). Day-to-day variation ON and OFF medication were 1.0 and 1.3 W/kg, respectively, while intraclass correlation coefficients were 0.95 (0.87;0.98) and 0.93 (0.82;0.97), respectively. Conclusion Linear encoder muscle power testing shows inconsistent known groups validity, acceptable convergent validity ON medication, and excellent day-to-day reliability ON and OFF medication in pwPD.