Measuring Mathematics Testing Confidence and Anxiety: A Scale Analysis Using Rasch Modeling

• In Event: 14.037 - Applications of the Rasch Model in Multiple Disciplines

Abstract

The purpose of this research was to develop separate scales for mathematics testing anxiety and confidence. Two theoretical frameworks were applied - one conceptual and one statistical. Leveled construct maps were developed to illustrate the conceptual theory as a hierarchical continuum under the statistical framework of Rasch measurement. Items were designed as scenarios with a third-person perspective and comparative response options. Unidimensional Rasch analysis was employed to assess the extent to which the data fit the one-parameter item response theory model. Specifically, variable maps, threshold estimates, category characteristic curves, and misfit statistics were analyzed. From this research, we conclude that mathematics testing anxiety and confidence are unidimensional, hierarchical constructs for which a student can progress up or down the scales.

Authors

• Caroline Vuilleumier, Boston College

• Kelsey Klein, Boston College