UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Using item response tree models for studying response behaviors Park, Minjeong

Abstract

In psychological and educational testing, test-takers’ response behaviors are a critical issue because they have significant impacts on the measurement of the trait and ability. The item response tree (IRTree) model is recently introduced as a promising tool for studying response behaviors. In this dissertation, I focused on the explanatory IRTree model that allows the researchers to include person and item characteristic variables to “explain” response behaviors. Although the explanatory IRTree model provides a useful way to address various queries about response behaviors, it has not gained much attention in the literature. Thus, the goal of this dissertation is to draw researchers’ attention to the potential of the explanatory IRTree model. To do so, I first introduced the IRTree model within an explanatory item response modeling framework. Taking this framework, I explicated how the standard IRTree model (a.k.a., descriptive IRTree model) can be easily extended to the explanatory IRTree model. Following that, I showcased two real-data applications. Study-1 used both the descriptive and explanatory IRTree models to inspect the response styles when answering the Rosenberg’s Self-esteem Scale. As the main findings, this study found the presence of two distinct extreme response styles and the acquiescence response style. Study-2 used the explanatory IRTree model to investigate the effects of person and item characteristics on nonresponse behaviors (not-reached and omitted) when taking the reading test of the Progress in International Reading Literacy Study. The findings showed that item nonresponse behaviors occurred differently depending on gender, test language, item location, and item format. There are three unique contributions of this dissertation. First, it expanded the utility of the IRTree model to be a tool for “explaining” response behaviors. Second, it provided an in-depth understanding of response styles in Likert-type psychological rating scales and nonresponse behaviors in educational testing. Finally, the method and findings of the two studies offered practical implications on the test/scale development and validation.

Item Media

Item Citations and Data

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International