Investigation of Response Changes in the GRE Revised General Test

Area of Science:

• Educational Measurement
• Psychometrics
• Cognitive Psychology

Background:

• Extensive research over 80 years indicates that most examinees achieve score gains by changing their initial answers on multiple-choice tests.
• Prior studies consistently show a net benefit from answer modification, suggesting a valuable metacognitive strategy for test-takers.

Purpose of the Study:

• To investigate the impact of response changes on test scores within a high-stakes admissions context, specifically the Graduate Record Examinations (GRE).
• To examine how examinee ability level influences the benefit derived from changing answers on the Verbal Reasoning and Quantitative Reasoning sections of the GRE.

Main Methods:

• Analysis of response data from a large sample of examinees (N=8,538 for Quantitative, N=9,140 for Verbal) who took the GRE revised General Test across 12 countries.
• Statistical examination of score changes resulting from answer revisions, correlated with examinee ability levels.

Main Results:

• Findings align with previous research, confirming that most examinees benefit from changing answers.
• A significant positive correlation was found between examinee ability and the magnitude of score improvement gained through response changing.
• The benefit of revising answers was evident in both Verbal Reasoning and Quantitative Reasoning sections.

Conclusions:

• The study validates the advantage of allowing answer review and revision, a feature often absent in computer-adaptive testing formats.
• Test agencies should consider the implications for test design, particularly regarding the utility of revision opportunities.
• Test takers can be encouraged to utilize their ability to review and revise answers, as it demonstrably improves performance.