Updates to Some Grad School Data

I admit to eagerly anticipating the rankings that appear annually. I regularly find the methodology of the rankings logical. However, I am posting this blog after years of puzzling over what appears to be a conundrum in the statistics of your "Best National Universities" category. To illustrate the peculiarity, below is data from three Big Ten universities for the years 2002, 2003, 2005, 2007, and 2008.

U. of MI-Ann Arbor U of IL-Urbana/Cham. U of WI-Madison

top 10% ACT* top 10% ACT* top 10% ACT*

2002 69 26-30 55 25-30 50 25-30

2003 87 26-30 56 26-30 55 26-30

2005 90 26-30 50 25-31 58 26-30

2007 90 27-31 55 25-30 58 26-30

2008 90 27-31 55 26-31 60 26-30

* 25th-75th percentile

I find it curious that, in 2003, Michigan's top 10% rose miraculously 18% to 87% yet its ACT range remained identical to 2002 and, indeed, matched both WI and IL, both of which edged up only slightly to the mid-50% range. In the ensuing years, MI has gone on to report that 90% of its entering freshmen are in the top 10% of their classes whereas IL and WI have remained in the 55-60% range. The statistics reported by WI and IL are understandable in view of not only their patterns over the past six years but also in comparison to other top tier public and private universities. MI, however, appears to be a very suspicious "outlier". It is difficult to understand how such an outstanding freshman class, 90% of which have graduated in the top 10% of their high school classes, have scored so relatively modestly on their ACT exams. It would be interesting to see how the incoming freshmen at Duke and Stanford, also in the 90-91% range for top 10% of their class, performed on their ACT exams. I would be very surprised if their scores were in the 27-31 range that Michigan is reporting.

One explanation for Wisconsin and Illinois' lower top 10% numbers in comparison to Michigan's, while at the same time maintaining a near identical ACT range, could be that the high schools from which MI recruits are far less competitive than those from which IL and WI recruit, thereby resulting in fewer of their entering freshmen achieving top 10% ranking. But this seems an unlikely explanation.

I would urge you to pay particular attention to what Michigan reports this year. Perhaps they can explain the conundrum. Obviously, integrity of reported data and prevention of "gaming" remains vital to the survey's reputation.

David of IL 3:26PM May 16, 2009