Using Latent Trait Models for Evaluating Likert and Forced-Choice Formats When Assessing College Major Preference

• In Event: Psychometrics and Educational Measurement Poster Session 1

Abstract

Efficiently and effectively assessing students’ preference over many options, such as college majors, is a challenging task because of the time and cognitive demand imposed on the respondents and the trustworthiness of their self-reports. The College Major Preference Assessment (CMPA) was designed to solve these problems by including Likert and forced-choice formats. This study aims to demonstrate the use of Latent Trait Model (LTM) in Structural Equation Modelling (LTM-SEM) to evaluate the performance of the CMPA for validation purpose. The results suggested the existence of Likert method effect. Forced-choice format was found to be more difficult and discriminating than Likert format. Also, forced-choice format provided a higher level of certainty in estimation the latent trait than did Likert format.

Authors

• Sirui Wu, The University of British Columbia

• Amery D. Wu, University of British Columbia

• Jake E. Stone, Visier Inc.