Health-related fitness knowledge in adolescence: evaluation of a new test considering different psychometric approaches (CTT and IRT)

Abstract

Fostering health-related fitness knowledge is a common goal across physical education curricula. However, carefully developed knowledge tests that satisfy the psychometric criteria of educational assessment are lacking. Therefore, two studies were conducted to evaluate a newly developed health-related fitness knowledge test within the framework of classical test and item response theory regarding item quality, test reliability, construct validity, and dimensionality. Overall, 794 ninth graders (Mage = 14.3 years, 50.6% girls) took part in Study 1. They differed in the type of physical education classes (minor or major subject) and school (lower or higher educational level) they attended. Study 2 incorporated 834 ninth graders at the same educational level (Mage = 14.2 years, 52.5% girls). Item–test correlation, test reliability, and validity were examined. In addition, item and test quality were investigated using unidimensional two-parameter logistic item response models. In Study 1, pupils at the same educational level with physical education as a major achieved higher knowledge scores than pupils with physical education as a minor (t = −5.99, p < 0.001; d = 0.58), which confirmed the test’s construct validity. In Study 2, the weighted likelihood estimate reliability of the final 27 items was 0.65, and the test–retest reliability reached rtt = 0.70. The items satisfied the assumption of local independence. The final test fulfilled the psychometric criteria of reliability and construct validity to assess health-related fitness knowledge in cross-sectional and interventional studies. This test extends the possibilities of research on health-related fitness knowledge in physical education.