The study examined how well GRE verbal, GRE Quantitative, and GRE Analytic scores predicted graduate grade-point averages and scores on an objective comprehensive examination among graduate students in a
master's degree program in counseling and school psychology. The initial sample consisted of 261 graduate students, which was reduced to 183 students (102 with a concentration in counseling, 81 with a school psychology concentration) after deletion of subjects with incomplete data. GRE Verbal scores, GRE Quantitative scores, GRE Analytic scores, graduate grade-point averages, and scores on an objective portion of a comprehensive examination were collected and data were analyzed via correlational techniques. GRE Verbal and Quantitative scores were found to be moderately strong predictors of both graduate grade-point averages and scores on comprehensive examinations. GRE Analytic scores emerged as weak predictors of graduate school performance. Results are related to previous findings.
As a common criterion used to select graduate students, the Graduate Record Examinations (GRE; Educational Testing Service,19851996) often have been the subject of debate, and concern surrounding the use of the GRE in making admissions decisions has been evident for some time. At issue has been the extent to which GRE scores predict performance in graduate school. More than one author has suggested that even with considerable validation evidence, decisive conclusions are not warranted. As Ingram (1983) stated, a review of the literature at that time indicated at least one remarkable finding: "the apparent inability of GRE scores to consistently predict measures of graduate student success" (p. 711). More recently, Nilsson (1995, p. 638) noted the lack of homogeneity that characterizes studies concerning the validity of standardized tests such as the GRE, and Sternberg and Williams (1997) suggest that their findings "underscore the need for serious validation studies of the GRE, not to mention other admissions indexes, against measure of consequential performance" (pp. 638-639).
A fair amount of research has addressed this question. The use of GRE scores and their relationship to aspects of graduate education such as program completion (Hackman, Wiggins, & Bass, 1970) and course grades (Hirschberg & Itkin, 1978; House, Johnson, & Tolone, 1987) have been investigated in a number of studies over the last 30 years or more. Much of the research on program completion has focused on doctoral rather than master's level programs. In one such study, for example, Rawls, Rawls, and Harrison (1969) noted that when students who completed their Ph.D.'s were compared to those who did not, those who obtained the doctorate had entered the program with significantly more undergraduate hours in psychology, biological sciences, and math, and fewer hours in the humanities. The average GRE scores for those who graduated did not differ from those who did not graduate. More than 20 years ago, Willingham (1974) reviewed the many studies that had appeared in the literature up until that time concerning the predictive validity of the GRE. In this review, he summarized the findings of 43 validity studies comprising 138 data sets, all involving the GRE, that occurred between 1952 and 1972. Willingham noted that "the most common predictors used in studies of success in graduate school are undergraduate average and GRE scores" and that "[w]ith respect to criteria of success, . . . [v]ery few studies reported validity data with departmental examinations as the criterion" (p. 275). He "concluded that the GRE is usually more valid than undergraduate grade-point average (GPA), and for various fields of study, the GRE Advanced test was the most generally valid predictor of graduate student success" (Dollinger, 1989, p. 56). More recently, however, Goldberg and Alliger (1992) and Morrison and Morrison (1995) reached considerably less positive conclusions based upon meta-analyses of studies examining the use of the GRE in predicting graduate school success. …