An Examination of Levels of Aggregation: Application to Parents' Perceptions of Coaching Sportspersonship

The effects of scale reduction on the reliability, validity, and predictive power of the Sportsmanship Coaching Behaviors Scale for Parents were examined. Three variations of the SCBS-P were tested. The full-scale version included all original sub-dimensions with four items per factor. The partially reduced version retained the three strongest loading items per factor, while the highly reduced version used a single item from each factor. Three studies were conducted using separate samples of parents from youth sports events. The Confirmatory Factor Analysis results indicated that the full-scale and partially reduced models demonstrated acceptable model fit. Measurement invariance testing confirmed metric invariance between the full-scale and partially reduced models. Regression analyses showed that the predictive power of the SCBS-P decreased as the number of items was reduced, with the Models factor being the most significant predictor of perceived program value.