Gre Scores As Predictors Of Minority Students' Success In Graduate Study: An Argument For Change - Statistical Data Included

College Student Journal, June, 2001 by Charles Sampson, Patricia G. Boyer

The query at hand

ETS uses first year (graduate) grade point average (FYA) as its primary criteria. However, ETS cautions admissions officers to consider the test score along with other factors when reviewing an applicant's dossier. However, at many institutions, a preponderance of weight is assigned to the GRE test score. For example, graduate school applicants who earn excellent undergraduate grades and receive only average or above average GRE scores will either be completely eliminated from further consideration or will have a low probability of being selected for admission (Goldberg, 1977; Jacobson, 1993; Williams, 1997). Additionally, numerous studies have suggested that the validity of the GRE in predicting one's ability to actually earn a graduate degree is inadequate (Robertson & Neilson, 1961; Ingram, 1983). Many scholars also questioned the adequacy of first-year grades as a criterion measure of success in graduate school. These concerns have been generated by inconsistent findings across studies as well as problems with construct validity of GRE (Hackman, Wiggins & Bass, 1970: Willingham, 1974).

Thus, the question to be examined in this research is, "How accurate have the GRE and select demographic variables been in predicting the first year graduate work among minority graduate students during the period, 1988 through 1997, at a `Research I' institution?"

Methodology

Study Participants and Measures

The sample for this study consisted of minority students who earned a doctoral, specialist, or master's degree during the period of 1988-1997 from a `Research I' University and who were recipients of a fellowship from the Graduate School.

Data for this study were gathered based on information available in the Graduate School from students' application dossiers and included undergraduate and graduate school transcripts. Undergraduate transcripts were used to compile information on undergraduate grade point averages and undergraduate degrees. The University records were used to collect data concerning students' graduate work.

Since GRE scores are used to predict `success' in the first year and because it is an indicator selected by ETS, first year grade point averages (FYA) were used as the dependent variable for this study. The independent variables in this study were age, gender, undergraduate grade point average (Uggpa), degree conferred (degree), GRE-verbal score (GRV), GRE-quantitative score (GRQ), GRE-analytical score (GRA), majors and undergraduate institution (Uginst). Additionally, the degree conferred is doctorate, specialist, or master degree; majors represent the area in which graduate degrees are being sought; and the undergraduate institution is determined by whether the student attended a predominantly white institution (PWI) or a historically black college and university (HBCU).

Data Analysis

Means and standard deviations were calculated using basic descriptive statistics. The correlation between the independent variables (age, undergraduate grade point average, GRE-verbal, GRE-quantitative, GRA-analytical, gender, graduate school majors, undergraduate institution, degree conferred) and the dependent variable (first year average) were calculated using Pearson product-moment correlation. The variables significant at p [is less than] .05 level were entered into a multiple linear regression model to predict the variables that are significantly associated with first year average.