Psychometric properties of the General Self-Efficacy Scale in Parkinson's disease.

OBJECTIVE This study aimed to investigate the psychometric properties of the General Self-Efficacy Scale (GSE) in people with Parkinson's disease (PD). More specifically, we investigated data completeness, scaling assumptions, targeting, reliability, and construct validity. MATERIALS AND METHODS This study involves data available from two different projects that included people diagnosed with PD for at least 1 year, yielding two samples (1 and 2). The combined total sample (N=346; 60% men) had a mean (SD) age and PD duration of 71 (8.9) and 9 years (6.3), respectively. Both samples received a self-administered survey by mail, which was administered twice in sample 2. Additional data (e.g., clinical assessments) were available for Sample 1. RESULTS Total GSE scores were computable for 336 participants (97%). Corrected item-total correlations exceeded 0.4. Principal component analyses identified one component (the eigenvalue of the first component extracted was 6.9), explaining 69% of the total variance. Floor and ceiling effects were <6%. Internal consistency (coefficient alpha) was 0.95. Analyses of test-retest reliability yielded (ICC) values from 0.69 to 0.80. The highest value refers to those (n=47) with identical self-ratings of mobility (in the on condition) at both tests; the standard error of measurement was 3.1 points. Construct validity was further supported by correlations in accordance with a priori expectations. CONCLUSIONS This study provides support for the validity and reliability of GSE scores in people with PD; the GSE can thus serve as a valuable outcome measurement in clinical practice and research.