New NCAA research on college academic performance has emerged that may help provide a better barometer in predicting a student-athlete's chances of graduating from college.

The new data were presented for the first time to the Division I Academics/Eligibility/Compliance Cabinet's Subcommittee on Continuing Eligibility during its September 6 meeting in Indianapolis. The data are relevant to ongoing discussions within NCAA governance groups about the accuracy of high-school performance and early college performance as predictors of what college presidents have indicated they would like to see improve: graduation.

The data suggest that there may be "academic checkpoints" during a student-athlete's college career that provide a more reliable indication of the likelihood of that student-athlete's graduation. While high-school grades and test scores have proven to be an accurate indicator of first-year college performance and a useful indicator of the likelihood of graduation, the college-performance data lend credence to the theory that each subsequent year of college brings a more accurate indication of degree completion.

Given that, the AEC Cabinet and the Subcommittee on Continuing Eligibility may start looking more at continuing-eligibility standards to improve graduation rates. The data may even have an effect on the NCAA's ongoing review of initial-eligibility standards.

A group of academic consultants has been charged by the Division I Board of Directors to study the effects of Proposition 16, the Association's current initial-eligibility rules, and determine what, changes, if any, are necessary. That group has deliberated for several months but has not issued any recommendations.

The continuing-eligibility research presented to the subcommittee is based on the NCAA's Academic Performance Census, which was begun in 1994 to assess the impact of Prop 16. The study tracks academic behaviors of student-athletes in Division I colleges or universities between 1994 and the present. The research assesses graduation status after six years for all Division I student-athletes (about 26,000 a year), first-year outcomes and graduation status for all Division I student-athletes who are on athletically related financial aid (about 13,000 a year), and yearly outcomes on a sample of Division I student-athletes receiving athletics aid.

The data show that for student-athletes who eventually do not graduate, a significant drop-off in academic performance appears to occur just before or during the junior year.

Jack McArdle, a statistical consultant on academic issues for the NCAA who helped present the research to the subcommittee, said that identifying those stages at which student-athletes become academically vulnerable can help institutions know when to increase their attention toward those students.

"We're seeing that there is an academic milepost at about the junior year where the eventual nongraduates begin to slip," said McArdle, who also is a professor of psychology at the University of Virginia. "Maybe that's when the special academic programs come back for these kids -- maybe those programs aren't just for the freshman year. If we can at least identify these mileposts, it gives schools a chance to rectify the outcomes."

Tension between access and graduation

The Association's current continuing-eligibility rules stipulate that student-athletes successfully complete 25 percent of the course requirements in a specified degree program before entering their third year in college. The percentage increases to 50 before entering the fourth year and 75 before entering the fifth. Various NCAA groups, including the Student-Athlete Advisory Committees, have suggested that strengthening continuing-eligibility standards might be an alternative way of increasing graduation rates, rather than strengthening initial-eligibility standards to generate a pool of student-athletes that is more likely to graduate.

While those discussions are ongoing, no one has forwarded any concrete proposals.

"But there does seem to be a link between continuing eligibility and initial eligibility," McArdle said. "The question is if you choose to lower initial-eligibility standards in order to increase access, can you manage continuing-eligibility standards to ensure that the people you gain -- who would have been projected to be nongraduates under the old model -- do in fact graduate?"

Past research has strongly suggested that there is an essential tension between providing access to the finite number of athletics scholarships and increasing graduation rates. Increase access and you open the door to prospects who are less likely to graduate; adjust standards solely to improve graduation and you close to door to prospects who can't make the initial bar but might actually have graduated had they had the opportunity and the proper guidance.

"Almost everyone concurs that our continuing-eligibility rules have had a beneficial effect on graduation rates of student-athletes," said David Knight, chair of the AEC Cabinet and faculty athletics representative at the University of North Carolina, Greensboro. "With the new data in the Academic Performance Census that explicitly deals with college courses taken and grades received by currently enrolled scholarship athletes, we might be able to improve our predictions of the likelihood of graduation."

The data have led to discussions of what is being called a "seamless" model for academic standards. The seamless model would attempt to identify an academic milepost during a student-athlete's college career, use initial-eligibility standards to predict that milepost and use continuing eligibility to predict graduation. Under this model, the prediction of both outcomes would be enhanced.

"The seamless model would tie initial and continuing eligibility together," McArdle said. "It's based on academic data indicating that prediction of early academic performance using high-school variables is more accurate than prediction of graduation, yet it maintains graduation as the ultimate outcome.

"It would generate an increase of graduation rates by identifying high-risk students in college rather than identifying high-risk prospects before college."

McArdle said if nothing else, the data do seem to indicate that the predictability of graduation increases the closer you move toward the outcome. It makes sense, then, that high-school variables have a decreasing level of predictability over that same time. In other words, the farther a student-athlete is in his or her college career, the more accurately his or her graduation can be predicted.

That may seem simple, but up until now, the NCAA has not had enough student-level data to examine those kinds of propositions in detail. The Academic Performance Census data recently contributed by all Division I schools has improved the level of details possible.

Now it will be up to the subcommittee and the cabinet to a large degree to determine if some sort of "seamless" model for both initial- and continuing-eligibility standards has merit.

"The data certainly demand discussion of the 'seamless' idea," said Thomas W. Adair, faculty athletics representative at Texas A&M University, College Station, and chair of the Subcommittee on Continuing Eligibility. "It's something the subcommittee certainly will be looking at in the future."

"We've reached the limit of useful predictiveness of high-school variables on college graduation," Knight said. "If we decide there's a better way of predicting graduation than by using high-school parameters, then maybe the high-school parameters can be used to determine whether a prospect can reach that college academic milepost.

"I think that's what we're after."

The subcommittee next meets in February, along with the rest of the AEC Cabinet. By then, the group anticipates the sixth year of data in the Academic Performance Census to have been completed, which is significant because it will have tracked the academic behaviors of at least one full graduating class.

"We hope that will add even more to our deliberations in February," Adair said.