Exploring Students' Misinterpretations of Linear Regression Models

Authors

• Heli Kauppinen Finnish Institute for Educational Research, University of Jyväskylä https://orcid.org/0000-0003-1339-7644
• Sara Taskinen University of Jyväskylä https://orcid.org/0000-0001-9470-7258
• Markus Hähkiöniemi University of Jyväskylä https://orcid.org/0000-0002-6337-5366

Abstract

This study investigated introductory statistics course students’ misinterpretations of linear regression models by examining students’ final exam answers using mixed method research approach. In total 259 university students participated in the study in 2022-2024. The study examined qualitatively the misinterpretations of regression coefficients and the statistical significance of regression coefficients. The analysis revealed several misinterpretation categories. The most common errors were related to the interpretation of regression coefficients. Statistical significance was incorrectly inferred using non-statistical reasoning such as common sense. After the qualitative step, the dependency structure of misinterpretation categories and its relationship to the student background variables was examined quantitatively.