Applying the Rasch Measurement Model to Measure Changes in College Students’ Mathematics and Statistics Perceptions

• In Event: Measurement Issues in Survey Research

Abstract

In higher education, it is common practice to compare pre- to post- results: typically the raw scores or mean values using the survey data results. This method violates many assumptions of statistical analysis by misusing ordinal data as interval or ratio. Also, the power of analyses in longitudinal studies is decreased due to loosing participants in the process of repeat measurement. Further, the functioning of survey items and rating scale may change from pre- to post implementation. All of these issues threaten the measurement integrity and make it risky to make implementations based on findings. This paper provides an alternative approach to objectively measuring changes across time, and discusses implications for improving survey research in higher education using Rasch measurement

Authors

• Letao Sun, University of Kentucky

• Kelly D. Bradley, University of Kentucky

• Michelle L Smith