skip to main content skip to footer

The Relationship Between College Grades and Aptitude Test Scores for Different Socioeconomic Groups

Author(s):
Schultz, Douglas G.
Publication Year:
1953
Report Number:
RB-53-18
Source:
ETS Research Bulletin
Document Type:
Report
Page Count:
21
Subject/Key Words:
Aptitude Tests, Background, College Qualification Tests, Grade Prediction, Performance Factors, Selective Service College Qualification Test, Socioeconomic Influences

Abstract

The purpose of this research was to compare the relationship between college grades and aptitude test scores for students from families of high, middle, and low socioeconomic status. The subjects were approximately 1700 male students in seven colleges who took the Selective Service College Qualification Test (SSCQT) in May or June of 1951. A questionnaire, including items on income and occupation of family head, was sent to the subjects soon after they had taken the SSCQT. On the basis of responses to each of these items, three socioeconomic groups were formed. The regression of grades on SSCQT scores was determined for each of the three groups separately and the three regressions compared using an analysis of covariance technique. The same procedure was applied to the regressions of grades on scores from an aptitude test (SAT or ACE) administered before or soon after admission to college. The differences among the regressions for the status groups with regard to standard errors of estimate, slopes, and intercepts were found not to be statistically significant (with one minor exception) and to be small and inconsistent. This was true even though there was some tendency for aptitude test scores alone to increase from low to high status levels. It is concluded that Scholastic Aptitude Test scores predict grades equally well and neither overpredict nor underpredict for all socioeconomic classes among college students and this equivalence is not altered after a period of college attendance. These conclusions must be interpreted in light of the fact that among the subjects there was a very inadequate representation of the lowest socioeconomic levels.

Read More