Exploratory Study

• Published in: Therapeutic recreation journal 2022-01, Vol.56 (1), p.83-100
• Main Authors: Béland, François, Ouellette, Pierre, Roussel, Hubert, Singleton, Jerome, Hopper, Tristan
• Format: Article
• Language: English
• Subjects: Approximation; Area Studies; Certification; Education; Factor analysis; Individual Characteristics; Intervention; Job Analysis; Job satisfaction; Leadership; Leisure Education; Likert Scales; Mean square errors; Measurement; Professional associations; Professional development; Program Development; Program Evaluation; Psychometrics; Quantitative psychology; Recreation; Rehabilitation Centers; Reliability; Research Design; Research Needs; Research Reports; Responsibilities; Rotation; Scientific Research; Statistical Analysis; Therapeutic Environment; Therapeutic recreation
• ISSN: 0040-5914; 2159-6433
• DOI: 10.18666/TRJ-2021-V56-I1-11070

The aim of this study was to assess the measurement invariance of the Therapeutic Recreation Job Responsibility Scale (TR-JRS). It measures four dimensions of recreational practitioners duties with 27 items: 1) administration, 2) therapy, 3) face-to-face leadership and 4) professional development. Measurement invariance assesses the ability of a single instrument, the TR-JRS, to measure duties among recreational practitioners working in different environments, members of separate professional associations, and to promote different aspects of their work. The sample comprised 417 members of the Canadian Therapeutic Recreation Association and the Federation québécoise du loisir en institution. The initial factorial measurement analysis showed metric invariance as the model of choice, though its overall fit was outside the usual acceptable criteria. A factor analysis with a target rotation resulted in a short scale of 14 items with an appropriate Root Mean Square Error of Approximation and Comparative Fit Index. It included the same four dimensions as the 27-item version. A measurement invariance analysis was also run on the short-form scale. The number of factors, position of items within factors, and factor loadings were equal in both groups. The absolute values of the Root Mean Square Error of Approximation and the Comparative Fit Index associated with the short form were over the acceptable cutoff points by a small margin. As expected, items means differed between the two groups and reliability coefficients were high. Such findings reflect this study's exploratory nature. Hence, further studies on item development could improve fit indices in favor of a stronger invariance status for this scale, which measures an important construct in therapeutic recreation.