Assessment of the construct validity of the German motor test 6-18 for 9 - To 10-year-olds

Utesch T., Strauß B., Tietjens M., Büsch D., Ghanbari M., Seidel I.

Abstract

To date a one-or a two-dimensional model (coordination/condition) for the eight items of the German Motor Test 6-18 (Bös et al., 2009) has not been confirmed. Probabilistic (mixed-Rasch modeling) and classic (confirmatory factor analysis) models are used to test whether the recommended norm-categorizations or alternative categories are empirically valid. Using a cross-validation-design, two samples of 9-to 10-year-olds (N1 = 1495, N2 = 1357; 49.2 % female) were analyzed regarding the norm-quintiles, z-normalized norm-performance-classes, and t-transformed standard scores of the empirical data (eT-categories). This study could not show the construct validity for the two norm-categorizations published. Analyses of the eT-categories yielded a one-class-solution (mixed-Rasch modeling) if two items were excluded: "stand and reach" owing to theoretical assumptions and "balancing backward" because of local model violations. Confirmatory factor analyses confirmed the factorial validity. Thus homogeneity in a one-dimensional motor scale was shown for 9-to 10-year-olds. The six homogeneous items comprise a motor-/fitness-score, which can be interpreted validly in assessments of motor ability.