Psychometric properties and construct validity of PLAYself: a self-reported measure of physical literacy for children and youth

Abstract

PLAYself is a tool designed for self-description of physical literacy in children and youth. We examined the tool using both the Rasch model and Classical Test Theory to explore its psychometric properties. A random selection of 300 children aged 8–14 years (47.3% female) from a dataset of 8513 Canadian children were involved in the Rasch analysis. The 3 subscales of the measure demonstrated good fit to the Rasch model, satisfying requirements of unidimensionality, having good fit statistics (item and person fit residuals = –0.17–1.47) and internal reliability (Person Separation Index = 0.70–0.82), and a lack of item bias and problematic local dependency. In a separate comparable sample, 297 children also aged 8–14 years (53.9% female) completed the PLAYfun, Physical Self-Description Questionnaire (PSDQ), Physical Activities Measure-Revised (MPAM-R), a physical activity inventory (PLAYinventory), and repeated the PLAYself 7 days later. The tests with this sample confirmed test–retest reliability (intraclass correlation coefficient = 0.81–0.84), and convergent and construct validity consistent with contemporary physical literacy definitions. Overall, the PLAYself demonstrated robust psychometric properties, and is recommended for researchers and practitioners who are interested in assessing self-reported physical literacy. Novelty: The PLAYself is a self-reported measure of physical literacy This study validates the measure using the Rasch model and classical test theory The PLAYself was found to have strong psychometric properties