Publication at First Faculty of Medicine, Faculty of Physical Education and Sport |
2013
Abstract
The paper deals with an evaluation of dimensionality of the Edinburgh handedness inventory (EHI) using nonparametric item response theory models. The results of Mokken analysis (based on monotone homogeneity and double monotonicity models) revealed that the psychometric properties of the EHI may be enhanced by dropping few items.
These items violated unidimensionality of the scale (box opening, broom) or invariant item ordering (knife) in different subpopulations. Thus we recommend elimination of these items while using the EHI in Czech environment.
The study presents the nonparametric IRT models (e.g. Mokken models) as a useful method for evaluating the structure of questionnaires/inventories used in psychodiagnostics.