The critical role of psychological features in talent identification research is well established. What remains open is how they should be captured, as self-report as well as coach-rating approaches can be subject to bias when used for talent selection purposes (e.g., social desirability and halo effect). The purpose of this study was to examine which of three different approaches (self-report, coach-rating, and a combination of both) to assess constructs related to achievement motivation offers the highest validity for predicting performance levels of youth athletes in individual and team sports over a two-year period using binary logistic regression models and the DeLong test. Descriptively, the combination model displayed the highest predictive validity, followed by coach-ratings and self-reports. However, only the comparison of the combination and self-report models displayed meaningful differences. The findings do not allow a definitive conclusion, but coach-ratings might be the least biased and most cost and time effective approach.
© Copyright 2023 Journal of Sports Sciences. Taylor & Francis. All rights reserved.