GMAT Validity

Validity refers to how well the GMAT exam does what it is intended to do—predict students’ academic performance in the core courses of a graduate management or business program.
The GMAT is the first and only test designed to predict student performance in business school, and it does so more reliably than any other known measure, including the Graduate Record Examination (GRE) and other standardized assessments.

For 50 years, GMAC has maintained the reputation of the GMAT exam as the best assessment for business school admissions by continually studying the performance of test scores as a predictor of academic performance to ensure the test’s validity.

• For the GMAT exam, validity is measured as the relationship between scores on the test and the outcome it is intended to predict—first-year or midprogram grades in a graduate management program.
• The GMAT exam is the best individual predictor of graduate management grades, better even then undergraduate GPA, but the best prediction results from combining GMAT scores with undergraduate grades. On a scale ranging from 0.00 to 1.00, the validity of the GMAT as a predictor of midprogram grades is 0.53.
• A validity range of 0.30 to 0.40 is generally considered good for standardized admissions tests, making the GMAT an outstanding predictor.
• The validity of the GMAT exam averaged over hundreds of validity studies is 0.48.
• Although either GMAT Verbal or GMAT Quantitative scores may be more effective predictors for a single program, on average, scores from both these sections of the GMAT predict equally well.

Validity Studies

Information from validity studies has been collected as long as the test has been in use.

• The first set of validity studies was reported in 1957, using scores from 10 schools that admitted students on the basis of the first administrations of the GMAT exam in 1954.
• Researchers at various schools who have conducted validity studies for different graduate management programs have found results similar to those of studies conducted through the VSS at GMAC. A meta-analyses of more than 400 publicly available studies of the validity of the GMAT estimated the validity of GMAT Total scores alone to be 0.47 (Kuncel, Crede, and Thomas, 2004).

Validity for Groups

Validity studies show that the test is not biased for any population.
• Consistently, different studies examining gender have shown absolutely no difference in the validity of the GMAT exam for males and females.
• Multiple studies examining race have also shown that GMAT scores predict grades equally well for the U.S. subgroups.
• Other studies have shown no differences in validity across groups based on age, native language, undergraduate major, and citizenship regions.

Other Predictors

Few predictors are able to add much to what is explained by GMAT scores and undergraduate grades.

• Although work experience is frequently required for admission to a graduate management program, individual studies rarely see much improvement to validity when the amount of work experience is used as a predictor.
• Amount of work experience, time since undergraduate graduation, and age have sometimes been found to be negatively related to graduate school performance. Students who are no longer used to studying frequently or intensely may receive lower grades, which appears as a negative relationship between age (or another time-based variable) and academic performance.
• Some admission factors, such as Test of English as a Foreign Language (TOEFL) scores and interview ratings, may not be available for all applicants, making it difficult to accurately assess the impact on prediction.

Other Outcomes

Although the GMAT exam was designed to predict performance in the core courses of a graduate management program, studies using other outcomes have also shown some predictive validity.

• The GMAT exam has been shown to have predictive validity for measuring second-year grade point average (GPA) and graduating GPA, as well as first-year GPA.
• Some studies have shown that GMAT scores have some effectiveness in predicting job success as measured by salary.
• The GMAT exam has been specifically designed to predict early graduate management academic performance. Use of the GMAT exam to predict any other outcome should be thoroughly studied in a specific situation before any decisions based on the scores are considered.