OBJECTIVES:To evaluate the construct validity of two dexterity measures, the 9-Hole Peg Test (9HPT) and Purdue Pegboard Test (PPT) in people with Parkinson's disease (PD). DESIGN:Cross-sectional observational study. SETTING:Testing was conducted at the university or in participants' homes. PARTICIPANTS:Thirty community dwelling people with mild to moderately severe PD and no major upper limb comorbidities or cognitive impairments. INTERVENTIONS:Pegboard tests were administered in the 'on' and 'end-of-dose' phases of participants' PD medication cycles. Participants rated hand function with two self-report questionnaires - the Manual Ability Measure-36 (MAM-36) and a subset of upper limb items from the MDS-UPDRS. To explore construct validity, we compared 'on' phase pegboard scores with normative values for unimpaired men and women and investigated relationships between pegboard scores and hand function questionnaires. RESULTS:In the 'on' phase, pegboard scores were poorer than normative values. Differences in individual subtest scores ranged between 10 and 41%. Correlations between self-reported hand function and pegboard scores were weak to moderately strong in the 'on' phase (r=0.21-0.51), and weak at 'end-of-dose' (r=0.13-0.22). Higher correlation coefficients were observed between hand function and PPT subtest scores than with hand function and 9HPT scores. Most participants reported difficulty with daily hand tasks. CONCLUSIONS:We found evidence for construct validity supporting the use of the 9HPT and PPT to evaluate people with mild to moderately severe PD when 'on', but not at the 'end-of-dose'. Results also suggest that the PPT may be more sensitive to PD-related changes in dexterity than the 9HPT.