A Comparative Evaluation of Methods of Adjusting GPA for Differences in Grade Assignment Practices

Pui Wa Lei, Dina Bassiri, E. Matthew Schulz

Research output: Contribution to journalArticle

3 Scopus citations

Abstract

Numerous methods have been proposed for constructing an adjusted grade point average (adjusted-GPA) that controls for differences in grading standards across college courses and departments. Compared to the raw GPA, adjusted-GPA measures are generally more predictable from preadmissions variables, such as standardized tests and high school achievement. Relative rankings of students on adjusted-GPA measures are also more consistent with their relative standings within courses. This study compared the performance of 4 polytomous IRT and 3 linear models for constructing adjusted-GPA measures. Unlike previous studies, the regression weights of predictor variables and the course parameter estimates used to compute adjusted-GPA were cross-validated. Adjusted-GPA retained noticeable advantages over raw GPA on cross-validation. The largest advantages were seen in the multiple correlation of adjusted-GPA with preadmission variables, when adjusted-GPA was constructed with the rating scale and partial credit IRT models. The cross-validity of adjusted-GPA was the weakest with the graded response model.

Original languageEnglish (US)
Pages (from-to)70-86
Number of pages17
JournalJournal of applied measurement
Volume4
Issue number1
StatePublished - Dec 30 2003

All Science Journal Classification (ASJC) codes

  • Medicine(all)

Fingerprint Dive into the research topics of 'A Comparative Evaluation of Methods of Adjusting GPA for Differences in Grade Assignment Practices'. Together they form a unique fingerprint.

  • Cite this