Grading Leniency Is a Removable Contaminant of Student Ratings

It is well established that students' evaluative ratings of instruction correlate positively with expected course grades. The authors identify 4 additional data patterns that, collectively, discriminate among 5 theories of the grades-ratings correlation. The presence of all 4 of these markers i...

Full description

Saved in:
Bibliographic Details
Published in:The American psychologist 1997-11, Vol.52 (11), p.1209-1217
Main Authors: Greenwald, Anthony G, Gillmore, Gerald M
Format: Article
Language:English
Subjects:
Academic grading
College students
Consumer Behavior
Correlation
Discriminant Validity
Educational Program Evaluation
Employee Performance Appraisal - standards
Evaluation
Grade Inflation
Grades (Scholastic)
Grading (Educational)
Higher Education
Human
Humans
Influences
Leniency
Opinion
Professional Competence
Psychology
Quality standards
Rating
Ratings
Ratings & rankings
Reproducibility of Results
Statistical Analysis
Student Evaluation of Teacher Performance
Students
Students - psychology
Teacher Effectiveness
Teacher Effectiveness Evaluation
Teaching
Teaching - standards
Test Validity
U.S.A
Undergraduate students
University of Washington
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is well established that students' evaluative ratings of instruction correlate positively with expected course grades. The authors identify 4 additional data patterns that, collectively, discriminate among 5 theories of the grades-ratings correlation. The presence of all 4 of these markers in student ratings data (obtained at University of Washington) was most consistent with the theory that the grades-ratings correlation is due to an unwanted influence of instructors' grading leniency on ratings. This conclusion justifies use of a statistical correction-illustrated here with actual ratings data-to remove the unwanted inflation of ratings produced by lenient grading. Additional research can profitably seek other inappropriate influences on ratings to identify more opportunities for validity-enhancing adjustments.
ISSN:0003-066X
1935-990X
DOI:10.1037/0003-066X.52.11.1209