Student athletes all too often get a bad rap in our society. I bet the experience of others is similar to mine: student athletes are often very good students, especially after controlling for the amount of time they dedicate to their sport. While there are some "bad" students who are good athletes, there are also a lot of "bad" students who are not athletes.

Today, The Chronicle of Higher Education released its estimation of graduation rates for 2005. The news is pretty good, and even the Chronicle has a hard time coming up with a negative spin. However, there is evidently a two year negative trend in African-American male basketball players and all women basketball players. I do not subscriber to the "two observations is a trend" school of thought, but alright.

I grabbed the 2005 data and threw it into STATA. Here's a picture of overall athlete grad rates versus student grad rates:

There is a positive correlation between the graduation rates of the overall student body and for the athletes on a given campus. Simple regression results of the relationship are:

. reg athleterate studentrate


      Source |       SS       df       MS              Number of obs =     315

-------------+------------------------------           F(  1,   313) =  296.84

       Model |  2.94129942     1  2.94129942           Prob > F      =  0.0000

    Residual |  3.10144223   313  .009908761           R-squared     =  0.4867

-------------+------------------------------           Adj R-squared =  0.4851

       Total |  6.04274165   314    .0192444           Root MSE      =  .09954

------------------------------------------------------------------------------

 athleterate |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

 studentrate |   .5320662    .030882    17.23   0.000     .4713036    .5928288

       _cons |   .3115148    .018301    17.02   0.000     .2755062    .3475233

------------------------------------------------------------------------------

If we restrict the constant to zero, we get :

. reg athleterate studentrate,nocons


      Source |       SS       df       MS              Number of obs =     315

-------------+------------------------------           F(  1,   314) = 6199.51

       Model |  117.917099     1  117.917099           Prob > F      =  0.0000

    Residual |  5.97240012   314  .019020383           R-squared     =  0.9518

-------------+------------------------------           Adj R-squared =  0.9516

       Total |    123.8895   315  .393299999           Root MSE      =  .13791

------------------------------------------------------------------------------

 athleterate |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

 studentrate |   1.032438   .0131125    78.74   0.000     1.006638    1.058237

------------------------------------------------------------------------------

While the slope coefficient is statistically different from one, the economic/behavioral difference is not significant.

In other words, as the general student body goes, so goes the student athletes. What? Student athletes are similar to non-student students? I know this flies contrary to general opinion, but in my experience this is more often the case than not. What I haven't figured out, either personally nor from the literature, is whether this is in spite of or because of the tutoring/acadmic advising infrastructure dedicated to student athletes on many campuses around the country.

Perhaps stories about student athletes are sexy and "socially responsible," especially on campus where many look down their nose at sports in general and the expenditure of dollars on athletics specifically. However, in my mind the correlation is very reassuring (and not surprising) - it is likely that the vast majority of student athletes are much more like their non-athlete counterparts than is generally believed.

(STATA data file)