Kutlu, ÖmerKartal, Seval Kula2022-08-082022-08-082018http://hdl.handle.net/20.500.12575/82952In this study, it is aimed to investigate dimensionality of Academic Motivation Scale items by depending on graded response model, generalized graded unfolding model, bifactor model and DIMTEST. Within the scope of this aim, the motivation scale was implemented on 1858 students who are studying at Ankara University. The fit of data matrix to dimensionality assumption is examined based on the item parameter estimations, general model data fit statistics, item and person fit statistics which are produced by the models. As result of the comparisons made among the models, it is found out that bifactor model provides the most consistent results with the theoretical foundation underlying the items and produces highest parameters for the items. It is revealed that generalized graded unfolding model and bifactor model enable better results than graded response model with regard to general model data fit. In terms of item fit statistics, the models that provides the best fit are bifactor model, generalized graded unfolding model and graded response model respectively. The index values obtained based on bifactor model also bring out the existence of a strong general dimension on which scale items can be ordered. However, multidimensional item parameters show that items have high discrimination parameters not only on the general factor but also on the group factors. The results of DIMTEST analysis also support that scale items are multidimensional. Based on the results of the analysis, it is concluded that bifactor model statistically and theoretically provides the best fit to the scale items.trÇok boyutu madde tepki kuramıİki faktör modeliGenelleştirilmiş aşamalı monoton olmayan modeldoctoralThesis