I have had the privilege of working with the NCAA staff and many NCAA committees on academic research about student-athletes.

As a result of this work, I was asked to comment on the academic research in the recent book, "The Game of Life: College Sports and Educational Values," by James L. Shulman and William G. Bowen. This book presents several interesting sets of data and analyses related to collegiate academic issues about student-athletes, and I am pleased to find the NCAA membership is interested in these issues.

The authors describe a unique collection of data, called the College and Beyond (C&B) data, collected by the Mellon Foundation. The C&B data used here permit analyses focused on (a) direct comparisons of student-athletes versus other students in the same colleges, (b) the long-term comparisons of three cohorts (1951 to 1976 to 1989) at the same colleges, and (c) the relation of student characteristics and long-term occupations and economic outcomes.

The Mellon Foundation has shared these C&B data with other researchers (including myself), and anything that can be learned from these data could be useful information for the NCAA membership.

The authors also highlight some key limitations of these data, including (a) the initial selectivity of the schools and (b) the difficulties in separating the selection of students to colleges. While more could be discussed about this statistical methodology, it is clear that the C&B institutions do not represent the 312 Division I institutions.

Unfortunately, this key feature of the data has not been carefully considered when some members of the media and the public have quoted results from this book as if they apply to all NCAA institutions. In contrast, our previous NCAA academic research suggests that many of the findings on academics reported in the "Game of Life" are likely to not generalize the larger family of institutions that make up the full membership of the NCAA. Let me outline a few specific issues that may be relevant to other NCAA members.

* The authors are quite careful to point out that their study, while a large and comprehensive survey of more than 90,000 college students, is based on "highly selective colleges." Their sample of 30 institutions is limited (12 Division I schools) and highly selective by design for three cohorts (1951, 1976, and 1989). No analyses are presented to extend these results to a wider range of NCAA schools.

* The authors report a large "admissions advantage" for student-athletes based on selective information from a single "nonscholarship, highly selective" school that represents only one part of the C&B database. While some admissions preference for high-school athletes over other student applicants does seem likely, the magnitude of this preference is not likely to be indicated by these data. The broad impacts of preferential admission of student-athletes require the study of multiple institutions and a much broader sample of applicants.

* The authors report that the average SAT score of recruited athletes is much lower than other entering students in the same college. While the SAT may be lower on average, the demographic composition of these students is not fully taken into account in these analyses. Previous academic research from the NCAA (and other educational research) shows that demographic analyses are critical in understanding aggregate scores based on all student-athletes. Recent data from the NCAA Initial-Eligibility Clearinghouse indicate that averages of SAT and ACT scores for student-athletes entering NCAA Division I institutions from 1994 to 2000 are above the national averages for all college-bound students.

* The authors of the book did not have the opportunity to examine the newest longitudinal data. In the chart above, a horizontal line connects graduation rates over the past 10 years for a specific set of colleges. The highest line (at about 85 percent) indicates the rates for all students in the C&B schools. The line immediately below (at about 75 percent) shows the graduation rate for NCAA student-athletes in the same C&B colleges. During the past 10 years, student-athletes in these 12 schools do have a lower graduation rate than the student body, and this is in complete agreement with the results reported in "The Game of Life" (for data collected up to 1989).

However, the lower two lines (starting about 50 percent) indicate the overall graduation-rate trends for all Division I schools. Clearly, these are much lower overall rates, and now most of the student-athlete graduation rates are higher than the overall student body rates at the same schools. Obviously, any student-athlete advantage in graduation rate varies greatly over different schools, and further examination of specific groups of students and schools is required for a complete analysis. These kinds of results of interest to the NCAA were not the focus of this book.

* The authors report that the "selective" student-athletes "under-performed" in college compared to their peers even when their prior performance (that is, SAT) was taken into account. However, the magnitude of this under-performance reported in the book was negligible in Division I-A public colleges, small in Division
I-A private colleges and considered large only in the Ivy Group (I-AA) and Division III liberal arts colleges in the C&B database. Recent NCAA research has noted that the entering classes of 1994 to 2000 freshman classes display a clear trend of increases in all aspects of academic profiles. Thus, the under-performance reported in the highly selective schools is not a necessary or even likely outcome when we consider the full family of all Division I schools. NCAA academic data indicate that student-athletes in football and basketball perform better than expected from their incoming characteristics (SAT/ACT and high-school grade-point average) when compared with other student-athletes. It is possible that student-athletes in the broader group of institutions would fare better than the general student body if such a comparison were made.

* The authors report a "clumping" of majors by student-athletes into social sciences and humanities, a postgraduate "over-representation" of student-athletes into specific business-related fields. As a consequence, however, the authors report a decided "earnings advantage" in later life careers for student-athletes compared to the other students in the same colleges. These occupational advantages were unanticipated and were not considered to be of great benefit by the authors of the book. Other readers will find these long-term earnings an issue worthy of further consideration.

In summary, the authors of the "Game of Life" have offered a unique and interesting data collection, as well as some carefully documented results and implications. However, the limitations of these C&B data and analyses also mean that a broad generalization of these results to all NCAA schools is not appropriate.

Without doubt, the NCAA should seriously consider the academic issues raised in "The Game of Life." But we also need to carefully follow the suggestions of its authors and put these new results into an appropriate place in the much broader collection of academic information on all students at NCAA member institutions.

John J. McArdle is a professor of psychology at the University of Virginia and serves as a consultant on NCAA research matters.