Measuring Student Success as a Latent Construct With a Partial-Credit Rasch Model

• In Event: Rasch Models: Considerations and Use

Abstract

Previous measures of student success in undergraduate science courses have relied solely on single cognitive course outcomes to both define and measure student success. In this study, we define student success as a latent construct composed of several cognitive (final course grade, GPA, and learning gain on assessment) and affective (sense of belonging, engagement, and science identity) course outcomes, or proxies of success. Via a partial credit Rasch model, these outcomes were used to create a new measure of student success in introductory biology. Our measure indicates that the identified proxies of success can be considered a unidimensional measure of student success that is both valid and reliable.

Authors

• Hannah Huvard, New Mexico State University

• Courtney Donovan, University of Colorado - Denver

• Robert M. Talbot, University of Colorado - Denver