Or: Academic rankings are more interesting when football is half the grade
Americans love rankings like their freedom. They also love the freedom to read too much into such rankings despite their arbitrary metrics and questionable significance. And so, once again the coming of fall marks the reemergence of a quintessential ranking of dubious methodology and incommensurate importance: the preseason Coaches Poll the USNews & World Report college rankings.
You’re probably now asking yourself what this could possibly have to do with football and I am here to reassure you: it really has nothing to do with football. But I as a red-blooded American will read way too much into this great begetter of high school parent angst and bring some college football into the equation. For despite their questionable merits, these rankings will allow me today to pinpoint the cradles of American scholastic-athleticism. And that, in effect, probably explains why rankings are so appealing: bereft of this list I would have to write for you an expository tome describing the essence of the scholar-athlete and what it means in America today. Instead, rankings in hand, I will crunch a few numbers and call it a day. (Note: while it would be fun to include APR scores to look at how the actual teams do academically, here I’m recognizing schools that have smart kids and good football teams… not necessarily smart football players).
There is something to be said for the USNews rankings which combine some useful statistics like acceptance rate and test scores of incoming students and give schools a single number they can sugarcoat to their hearts’ content (while simultaneously saying that no ranking can truly capture the college experience). Schools then set out to improve their ranking by attracting better students to learn from better professors (much how they hire better coaches to teach better players—hey not so different after all!).
More generally speaking, the effect of athletics on academics is an important topic of debate as big-time football becomes a larger part of university budgets. In that vein there have been studies that examined whether sports success yielded higher academic rankings (synopsis: schools that win national titles in football or basketball get a boost in the USNews rankings due to a bump in applications leading to higher selectivity, but schools with merely good teams don’t receive any benefit).
It takes only a glimpse at the top 25 in any given week to believe that academics and football prowess are largely uncoupled, but here I’ll use the USNews rankings and the Sagarin rankings to see which schools are most worthy of the "scholar-athlete" designation. For the plots and tables shown here I’ve considered the USNews rankings to be "preseason rankings" in effect by matching them up with Sagarin’s final Predictor rankings for that season. How these should be matched up is certainly debatable, but for this context largely irrelevant because the academic rankings are so static (if schools really wanted to generate some excitement in this area perhaps they should restrict professors to four years of eligibility) and so I’ve matched the rankings to be as close together in time as possible (i.e. September to January). While USNews confusingly names their rankings like car models, meaning the latest are technically the "2012 Rankings", I’ll hereafter refer to them by the year they are released (and the football season they correspond to).
I’ve also removed teams that don’t appear in both sets of rankings so that, for example, a I-AA team doesn’t displace a I-A team in the rankings and vice-versa (and so if I used the latest USNews rankings that means that Stanford would be considered first and Duke second). Finally I’ve done a simple conversion by turning the ranking into a score by reversing them, much as how a vote for #1 in the AP poll gives a team 25 points; the USNews ranking thereby becomes a "Brains" score and the Sagarin rankings become a "Brawn" score.
It may be somewhat surprising, but if you consider all of the schools in Division I-A that are included in the "national university" segment of the USNews rankings (this includes 107 of the 120 I-A teams, with notable exceptions like the service academies) there actually is a small positive correlation between football and academic ranking success. Figure 1 shows a scatter plot of the "Brains vs. Brawn" for the 2010 season/2010 USNews rankings. As the trendline shows we have a (very weak) positive correlation, with a correlation coefficient (R) of 0.29. But I don’t think it’s just noise (unfortunately I only have the comprehensive USNews rankings from 2010 to look at all the teams in I-A), as it can very easily be understood as the flagship state university effect—the primary state school is (usually) the most research-driven and has the biggest sports program and best football team. Toledo might argue with the latter point.
Furthermore, this effect completely disappears when only BCS conference teams (plus Notre Dame) are included, as shown in Figure 2. Figure 2 shows exactly what we might expect—no real relationship between football and academic ranking success (while a small negative correlation exists for 2010, it varies from very weakly positive to very weakly negative on an annual basis). So there you have it: as I hinted in the first paragraphs, there is no profound and apparent connection between the football and academic spheres, and we’re just here to have fun with rankings today.
Figure 1: USNews versus final Sagarin rankings for 2010 for the 113 I-A schools considered "National Universities" by USNews.
Figure 2: USNews versus final Sagarin rankings for 2010 for the 66 BCS conference schools (including Notre Dame, excluding Utah). The weak positive correlation between the rankings disappears.
The table below (Table 1) shows the good stuff: which schools best combine scholastic and football pursuits. I also threw in the number of Academic All-Americans to recognize examples of individual (rather than university-wide) academic/athletic excellence. You may be surprised by the team decisively in the #1 slot, but remember, USC is so exclusive it grants 40% fewer scholarships than other teams. There are two simple ways to combine the Brains and Brawn scores that yield similar results with a few exceptions. You can take a simple average or sum of the scores, or the method I prefer, the product of the two scores. The latter method prefers more well-rounded schools (sorry Duke); a college with scores of 60 and 2 will have the same average as a college with scores of 31 and 31 but the latter will have a much higher product. For the most part the resultant rankings are very similar, but I’ve included the mean results for comparison. Table 2 shows the bottom ten teams (worth mentioning that Duke just missed this list at 56th, but using the mean they would rank 28th). The full rankings can be found here. Note: I have complete data for BCS conference teams in the USNews rankings going back to 2004, so all data shown is the average for the 7 year period 2004-2010 (Michigan probably isn't too happy about me using this time frame).
Table 1: The top 10 schools.
|Rank||School||Brawn||Brains||Product||Mean (Rank)||# Academic All-Americans (2004-10)|
|5||Ohio State||58.4||41.7||2437||50.1 (5)||5|
|6||Notre Dame||36.3||61.6||2234||48.9 (6)||5|
|7||Penn State||46.0||47.7||2195||46.9 (8)||17|
|10||Boston College||38.3||52.6||2013||45.4 (10)||0|
Table 2: ...and the Bottom 10.
|Rank||School||Brawn||Brains||Product||Mean (Rank)||# Academic All-Americans (2004-10)|
|65||Mississippi State||21.7||9.3||202||15.5 (65)||0|
|63||Washington State||14.7||15.7||231||15.2 (66)||3|
|62||South Florida||27.4||8.6||235||18.0 (61)||0|
|61||Kansas State||21.6||11.0||237||16.3 (62)||2|
It’s nice that both the top and bottom 10 lists have broad conference representation. I would also like to note the impressive number of Academic All-Americans from Baylor and Kentucky in recent years. To some degree, it may be easier for football players to have academic success at lower-ranked academic universities (where there is reduced competition), but being an Academic All-American anywhere is no mean feat (note: I counted first and second team selections). All but the Big East are represented in the Top 10 (but hey, they get a team if you count basketball affiliation!) and all but the ACC are represented among the Bottom 10 teams. The rankings for the conference follow in a similar fashion with the ACC on top and Big East at the bottom (here the Brain and Brawn are the averages of all football schools in the respective conference); I’ve used the conference affiliations during the 2010 season (not sure, but I don’t think using the 2011 league compositions would’ve swapped the PAC12/Big 10 order) (Table 3).
Table 3: Conference Rankings.
|2||Big 10||29.6||44.4||1314||37.0 (2)|
|5||Big 12||32.0||24.3||776||28.1 (5)|
|6||Big East||29.2||24.3||711||26.8 (6)|
And finally, despite the relative constancy of the academic side of the equation, I’ll leave you with a few animations of how the individual teams have fared over the past seven years, by conference, with an extra plot for the top ten teams. Like I said at the beginning, the USNews rankings may be meaningless, but in no way should that preclude putting way too stock in them. So go ahead, use these rankings for your next tailgating taunts. The numbers don’t lie. Much.
The Top 10 teams from the ranking.