AAMC Reporter, September 2000: Study Looks at MCAT's Predictive Power

VOLUME 9, NUMBER 12

JORDAN J. COHEN, M.D., PRESIDENT

SEPTEMBER 2000

Study Looks at MCAT's Predictive Power

Do high MCAT scores go hand in hand with success in medical school? The answer from a recently completed 10-year AAMC study of the MCAT's predictive validity is yes - and no.

"The predictive power of the MCAT varies across schools," says Ellen Julian, Ph.D., assistant vice president of the AAMC's Section for the MCAT. "In some medical schools, if you know a person's undergraduate GPA and MCAT scores, you can predict with some accuracy what their grades will be and what their licensure examination scores will be. In other schools, there is essentially no relationship." The study examined longitudinal data on two years' worth of entering student populations from 14 representative schools.

So what makes the difference? Why is the MCAT predictive of success in some schools and not in others? According to Dr. Julian, it may have to do with a school's focus and mission. "In our study, the school with the least relationship between MCAT scores and success is a state school whose mission is to prepare practitioners to serve in the underserved part of their state," she explains.

This school, she notes, is willing to forgive MCAT scores for students who are exceptional in other ways. "They use the MCAT as a diagnostic tool in predicting who's at risk, and invest large amounts of resources in helping these students be successful."

Schools that experience a stronger relationship between MCAT scores and medical school success, on the other hand, may be more research-intensive schools, where academic accomplishment is very important. "They're going to invest their resources in their superstars," says Dr. Julian. "If you're struggling when you come in, you're going to struggle all the way through, but if you have the potential to be a superstar, they'll groom you, nurture you, and help you blossom."

Ultimately, Dr. Julian hopes that analyses underway this summer will help clarify the characteristics of the schools in which MCAT scores predict well or poorly. These might be used to create a system to help match students with schools. "Both kinds of schools are doing it right for their mission," she says. "If I'm a student who's not well prepared, I should go to a school with remedial help that will focus on me. If I'm a superstar, I'd rather go to a competitive, sink-or-swim place that will give me extra opportunities." A report on the MCAT Predictive Validity study's results has been released as a Contemporary Issues in Medical Education briefing and as a CAMCAM Fact Sheet. The full study will be released as an MCAT monograph.

Information: Ellen Julian, Ph.D., (202) 828-0692

16 March 2004