Bad grades, not criminal behavior predict dangerous play in NCAA football

Are there any factors that predict whether a college football team will commit violent, and potentially harmful behaviors on the field?

Earlier this year, I watched a game involving the Minnesota Golden Gophers, who were in the midst of a disturbing off-the-field issue. During the broadcast, the announcers also mentioned that Minnesota was among the nation’s highest offenders for personal fouls.

This immediately made intuitive sense to me: teams with disciplinary issues outside of football would likely have disciplinary issues on the field. The reckless, sometimes violent nature of a lot of personal fouls made comparisons to off-the-field misbehavior seem very natural.

Let’s take a look at the on-field penalty /off-field crime link. I compiled data from three college football seasons (FBS teams, 2010-2012), recording total arrests (ArrestNation) and personal fouls (from Jake Levy; see here) for each team.

arrests-and-pf

(Horizontal axis): Team arrests per year ||| (Vertical axis): Team personal fouls per game

Take a look at the above graph. While it may appear to have a small positive trend, there is, surprisingly, no significant relationship between team arrests and team personal fouls.

Academics

After ruling out criminal activity as a predictor of personal fouls, let’s try some different factors. Perhaps academics play a role. For instance, maybe “smarter” teams are less likely to commit personal fouls.

academics-and-pf

(Horizontal axis): Team academic score from NCAA; high score = better grades ||| (Vertical axis): Team personal fouls per game

The above graph uses data from the same time frame as before (2010-2012), and uses academic performance from the NCAA (Academic Progress Rate; higher scores indicate higher academic performance).

Results show that teams with better grades teams commit significantly fewer personal fouls. Statistically, there is a great deal of support for the hypothesis (26% of the total variance be explained by academics; < .001).

Why academics?

Why, exactly, would academics play a role in predicting personal fouls? There are no clear answers, but here are a few possibilities:

  1. Intelligence: Book-smart players can make more intelligent decisions on the field, such as avoiding needless personal fouls (e.g., unsportsmanlike penalties).
  2. Top-down discipline: Coaches that exert more control on the field also exert more control in the classroom, causing fewer penalties and better grades.
  3. Bottom-up discipline: Similar to 2), but swapping players for coaches. Specifically, players who are disciplined on the field are more disciplined in the classroom, leading to fewer penalties, and better grades.

Addressing alternative explanations

So, what about alternative explanations? Perhaps the academics-personal fouls relationship is an illusion, and it’s explained by alternative factors. One may be team quality. Perhaps better teams commit fewer personal fouls than worse teams. And, perhaps these teams teams just happen to be smarter, leading to the appearance of a strong academics-personal fouls relationship, where there is actually none.

multiple-regression

The graph above looks at a number of potential alternative factors: overall penalties (courtesy of TeamRankings), arrests, year, and team ranking (from Massey Ratings). The analysis utilizes a statistical technique known as multiple regression, which measures the relationship between a single factor and outcome, while controlling for the influence of other factors.

The results suggest that team academics continue to significantly predict personal fouls, when controlling for other, relevant factors. Importantly, as indicated by the last row of the table (“All factors”) adding extra factors does little to improve predictions of which teams commit more personal fouls (.29 for all factors versus .26 for academics only).

Rise in annual grades, rise in good behavior?

Interestingly, classroom grades are on the rise in a number of Division 1 sports. Will it contribute to fewer bad behaviors on the field? It seems doubtful. Teams may not be getting smarter or more disciplined, but rather, better at learning how to achieve the academic criteria for NCAA’s progress reports. But, it’s something to keep an eye out for.

Advertisements

2 thoughts on “Bad grades, not criminal behavior predict dangerous play in NCAA football”

  1. I’m confused with multiple regression results. You show a Beta of .26 and then refer to it as a correlation. Generally standard notation would use Beta for coefficients of the regression line — and coefficients are not correlations.

    Liked by 1 person

    1. Hi David,

      Changed! Thanks for the response. The values in the multiple regression are standardized beta coefficients, which I inappropriately described as “correlations”.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s