Graduate school entrance exam results don't create inequalities; they reflect them.
By Danny LaChance
Clear your desk, sharpen your #2 pencil, and choose the BEST answer. The results of graduate school entrance exams…
(a) effectively predict grades earned in graduate school.
(b) often predict the likelihood of success in graduate school better than undergraduate grades do.
(c) effectively predict markers of success other than grades, like the chances of passing a professional licensing exam or publishing research that is cited by other researchers.
(d) all of the above.
It's (d), says assistant professor of psychology Nathan Kuncel. In collaboration with Sarah Hezlett, senior research scientist at Personnel Decisions Research Institutes, Kuncel analyzed the data provided by over 3,000 studies of standardized tests used by the admissions committees of graduate programs—everything from the GRE to the LSAT to the GMAT. In what is one of the largest meta-analyses ever undertaken, synthesizing data collected from nearly a million students over multiple decades, the two found that good scores on these exams correlated with success—in many forms—in graduate school and beyond.
Their findings, published in Science last February, come after nearly 80 years of debate about whether standardized tests are biased against women and minorities, whose scores in some areas lag behind those of their male and white counterparts. Breaking down the data, Kuncel and Hezlett determined that the tests are as accurate at predicting success for minorities and women as they are for test-takers as a whole.
“The tests aren't at fault. It would be great, actually, if it was as simple as bad tests," Kuncel says, for then the skills gap indicated by the tests could be dismissed as a distortion of reality rather than a symptom of it.
Instead, entrance exams seem to reflect inequalities created long before students begin filling in bubbles. “The problem seems to be more societal, more ingrained. These tests are quantifying basic content,
verbal, and quantitative skills, which people don't always have equal opportunities to develop. School quality, treatment in the classroom very early on, and other social issues seem to be what's causing differences in performance," Kuncel says.
Kuncel hopes that his synthesis of the studies will help to steer discussions of educational inequality away from standardized tests and toward root causes. “It's ultimately one of those ‘Let's not shoot the messenger' situations," he says. “Let's spend our energy solving the problem."
