By Lawrence M. Rudner

Let’s look at the evidence regarding the validity of the GMAT and GRE exams as predictors of academic success in business school. Although at a cursory glance the tests may look similar, there’s a world of difference in their evidence of validity.

GMAC has long recognized that evidence of validity is critical to determine quality and has encouraged schools to conduct validity studies. As a consequence, there is a great deal of evidence that the GMAT works for business schools, with consistently high validity coefficients. If you look through the peer-reviewed literature, you will find at least 10 quality validity studies published between 2005 and 2008 (also available on gmac.com), including one Educational and Psychological Measurement meta-analysis of 273 validity studies involving 41,338 business students taking the GMAT in recent years. We worked with admissions personnel to simplify the validity study process and make our reports directly useful to schools. Our free Validity Study Service uses your data to determine GMAT validity for a specific program, and the confidential report we provide you shows how GMAT scores work with other factors in your admissions process. The validity of the GMAT has been demonstrated for MBA, EMBA, other business master’s degrees, and doctoral programs. Validity has also been documented for numerous population groups, including by gender, citizenship, language, undergraduate major, US subgroup and age. We have conducted more than 300 validity studies in the past decade, and not a single school has dropped the GMAT after completing a study. In fact, schools that conduct validity studies often either add programs or repeat the study in subsequent years.

Validity evidence for the GRE, found at ETS’s website GRE Research on Validity, highlights a single study, Nathan Kuncel’s 2001 Psychological Bulletin meta-analysis, in which a few business schools were lumped in with social science programs. That study was conducted in 1998, and many of the studies in the analysis were based on the old paper version of the GRE. Further portions of the GRE examination were changed in 2002.

With a little searching, you can find two more studies—a study written in 2004 and published in 2008, and a study written in 1999 and published in 2005. These studies offer evidence of GRE validity for biology, chemistry, education, English, and psychology students. Aside from not addressing business students directly, both studies use GRE data sets from the mid-1990s. That’s about it. Because the test’s content has evolved, it’s not possible to generalize from these older studies to today’s GRE.

Nonetheless, ETS claims that the GRE and the GMAT measure the same thing. Its advertisements present a handful of seemingly similar questions from the two examinations. A closer look at the sample questions reveals very real differences. The GRE examples test basic skills. The GMAT examples use similar content, but assess much higher-order skills of logic and reasoning ability

Lacking validity studies concerning the use of the GRE in business admissions, ETS has produced a comparison tool to predict GMAT scores from GRE scores. The tool uses averages, which can be useful for research. Admissions, however, is about individuals. The comparison tool’s mapping might produce correct averages, but for the vast majority of individuals, the comparison will be wrong, particularly given the tool’s large error of prediction. Statistical games do not make the tests and scores equivalent.
 
While the broad subject area titles for the GMAT and GRE are the same (Quantitative, Verbal, Writing), the skills and the mix of skills within each area are quite different. GRE is a broad-based test originally designed for all graduate level performance. It is often supplemented by one of eight GRE subject tests. The GRE general test can’t be as good as a specialized test such as the GMAT, LSAT, or MCAT.

The GMAT is developed specifically for prospective graduate business school students. Questions are written for potential business students and pilot-tested with them. GMAT test takers and GRE test takers are not random samples from equivalent populations. These differences affect the scales, bias analysis, and item level statistics – almost everything psychometric about the test.

There is little doubt that the GRE can probably help identify the potential super-student as well as those who are not at all ready academically. Those are easy calls. The real challenge is whether the test identifies meaningful differences among your applicants, and there is little evidence that it does. GMAT, on the other hand, has been shown repeatedly to work well for all business school applicants, including those in the middle of the ability distribution, where quality, objective, standardized information is most important.

Lawrence M. Rudner, PhD, MBA, is Vice President for Research and Development at GMAC.