Utilizacion de modelos unidimensionales de teoria de respuesta a los items con datos multifactoriales

Abstract

Use of unidimensional item response theory models with multifactorial data. A simulation study is presented in order to investigate the relationship between the factorial complexity and the robustness of parameter estimates from unidimensional Item Response Theory models of one and these parameters. The simulation was carried out in accordance with the Wherry, Naylor, Wherry and Fallis (1965) factorial model. Sixteen data sets were generated varying the number of items, number of factors and its characteristics. The results show that when a dominant factor is present this factor is collected by both models. When several independent factors are present, the these parameter model picks up only one of them, whereas the one parameter model collect a combination of all these factors