What Predicts Variation in Reliability and Validity of Peer Assessment? A Large-Scale Cross-Context Study

• In Event: Hilton Roundtable Session 1
  In Roundtable Session: Development, Validation, and Psychometric Testing of Assessments

Abstract

In the current study, we address the research question on what factors predicts variation in peer assessment reliability and validity by analyzing the data collected from a web-based peer assessment platform used in 374 courses across 88 different institutions, involving a total of 19,722 students. Different levels of factors are examined in the study, including factors at the institution-level, course-level, assignment-level, and rubric-level. We used structural equation modeling approach to examine the relations between those factors and reliability/validity of peer assessment.

Authors

• Yao Xiong, Pearson

• Christian D. Schunn, University of Pittsburgh