How Problematic Is Self-Selection Bias in College Student Research? An Examination of Experiential Propensity

• In Event: Roundtable Session 29
  In Roundtable Session: Validity and Measurement Issues in Student Learning and Development Research

Abstract

A substantial body of research has explored the relationships between college experiences and outcomes. Many of these studies may face potentially serious issues with students’ self-selection into their experiences, which prevents strong conclusions from being drawn about any causal effects. The present study explored the extent to which self-selection is a concern by directly modeling student reports of their propensity to engage in certain experiences. Within a large multi-institutional sample, the findings indicated that substantial self-selection bias may be present across various outcomes even when including a rigorous set of precollege, college, and institutional control variables. This study suggests that “college impact” research may often contain misleading results.

Authors

• Nicholas A. Bowman, University of Iowa

• Cindy Ann Kilgo, University of Iowa

• Teniell L. Trolian, University of Iowa