Measuring Undergraduate Evolution Learning Using the CANS (Conceptual Assessment of Natural Selection): Assessing Psychometric Properties Using a Rasch Measurement Framework

• In Event: New Developments in Item Response Theory Modeling and Applications

Abstract

Evolution is one of the most difficult biological topics, warranting innovative assessment strategies for enhancing evolutionary instruction. The CANS (Conceptual Assessment of Natural Selection) has advanced the measurement of evolution knowledge. Here, we use Rasch analysis and a large sample of undergraduates (>6000) to expand upon prior validity testing of the CANS. We identified problems related to dimensionality, person-item alignment, item fit, and construct overrepresentation. We also report that, despite pre-test disparities in evolution knowledge, students from all race and gender groups achieved comparable learning gains. This work generates tangible recommendations for instrument improvement and novel insights about evolutionary reasoning.

Authors

• Austin Zuckerman, Cornell University

• Gena Sbeglia, San Diego State University