Objective: Exploratory application of the Rasch Measurement (RM) Model for evidence for reproducibility, conceptual/content validity, and structural validity of the Moorong Self-Efficacy Scale (MSES).
Study design: Secondary RM analysis of data collected in a randomized controlled trial comparing two exercise interventions for persons living with spinal cord injury (SCI).
Setting: Community-dwelling persons living with SCI enrolled in an exercise study.
Participants: Adults (n = 79) enrolled in the parent study had a traumatic SCI > 3 months prior, injury level C5 to T12.
Interventions: Not applicable.
Outcome measure: The original MSES is a 16-item measure of self-efficacy with a 7-level response scale for un/certainty which was developed for use with persons living with SCI.
Results: We addressed item misfit, infrequent category endorsement, and category step disorder by removing two items and reorganizing the rating scale. Rating scale changes removed category 4 (Neutral), combined categories 1-3 (Very Uncertain, Somewhat Uncertain, and Uncertain) for all items, and further combined certainty categories for two items. Principal components analysis of the residuals indicated a possible second dimension with a first-contrast Eigenvalue of 2.4. However, the contrasted item groups had explained variance <10% and a dis-attenuated correlation = 0.92 indicating they measure the same underlying trait. The small sample size precluded examination of differential item functioning.
Conclusions: Exploratory RM analysis of MSES produced a 14-item Rasch version which identified structural and content validity evidence concerns inherent in the original MSES. However, results could be biased by a small sample size and further study should examine the item content and rating scale structure with larger, more diverse samples of persons living with SCI.
Keywords: Psychometrics; Questionnaire design; Rasch analysis; Rehabilitation; Self-efficacy; Spinal cord injuries.