Background/Purpose: Even though self-efficacy (SE) and social supports (SS) are crucial correlates with physical activity level among able-bodied individuals and persons with physical disabilities, few studies have been conducted in persons with intellectual disabilities (PID) due to the lack of validity evidence in existing scales. New scales to measure SE and SS designed for PID have recently been developed by Peterson and colleagues (2009). This study purpose, therefore, was to investigate the psychometric properties of this SE and SS scales of PID, utilizing a Rasch model.
Method: Responses from a total of 146 persons with mild or moderate intellectual disabilities (78 male & 68 female; 18-60 yrs old; including Down syndrome, cerebral palsy, and developmental brain injury) were used for the data analysis, in which the data was collected in previous study for the development of the scale. SE scale is composed of 6 items. SS scales included three sub-categories: family (7 items), staff (6 items), and roommates/peer (5 items). All items were calibrated using many-faceted Rasch model (age, gender, diagnosis, secondary condition, and impartment) employing the FACETS program and logit scores were calculated. Infit and outfit statistics (≤.7- ≥1.3)were employed, respectively.
Analysis/Results: As a result of the fit statistics, item 6 from the family SS scale (payment by family member) did not fit the model. Therefore, a total of six SE items and 17 SS items were used for the final analysis and fit well with all facets employed (infit = 0.8-1.3 logits). The item difficulty was between -.78 (logit) and .68 in the SE scale and -1.34 and 1.71 in the SS scale, respectively. Among the six SE items, "participation in PA almost every day" was the most difficult item (.68) and "participation in PA when sad or depressed" was the easiest item (-0.78). Overall, staff support was more important (0.65±0.79) than family or peer supports. "Transportation support by staff" showed the highest score (1.17) and "peer demonstration of PA" was the lowest perceived SS (-1.34) among the 17 items.
Conclusions: The SE and SS scales developed for PID provided psychometrical quality. More validation evidence is needed to apply the scales in PA interventions among PID.