clock menu more-arrow no yes mobile

Filed under:

HEY WE CAN SEE THE WHOLE WORLD FROM HERE

Yup. Pretty sure we can see the whole world from here.
Yup. Pretty sure we can see the whole world from here.

If you really like to scare the shit out of people, just repeat what our methods professor at Georgia Tech said once about establishing reality: "There are no facts, only correlations." It's rhetorical arson, especially when you're riding with people who don't understand numbers at even the basic cromag level we work at around here. 

Continuing that analogy, the correlative fuel you can siphon out of the CNNSI piece on athlete arrests in D-1 football is weak horse piss compared to the full-powered gasoline you might otherwise use in a more thorough study. This lack of empirical gas has a few root causes, which will be happy to list here.

1. Lack of context. Let's be gentle by starting and saying that the overall data for college students and their general criminality are lacking, and thus the data would be hard to pull. This is the dictionary problem in statistical analysis, i.e. trying to look up the definition to a word, discovering you don't know a word in that definition, and then looking that up and realizing you don't know a word in that definition, and then realizing slowly how much data you don't have all the way down the chain of things you don't know. If you've done this it is the shittiest thing ever and happens all the freaking time.

You have to also consider that the full context may be provided in a longer magazine piece. That said, here come the knives for the soft parts.

If SI wants to throw out a number as being "troubling" or elicit that reaction from the Mark Emmerts of the world, you're going to have to do more than amass a data set and point to it .For instance: right now there are two empty avocado halves sitting to my left. Isn't that level of neglect to basic housekeeping terrible, Mark Emmert?

Mark Emmert: FART snorgle [wakes up from nap] Ahahhhh Cam Neutron is eligibubulll [falls asleep]

Saying the numbers are there is fine, but attempting to pull anything from this whatsoever in terms of defining a reality is horseshit a la mode. (And because it is a well-written piece, there's at least the ice cream covering the poopy numbers and approach beneath on the internet summary of whatever this is going to be. YAYYY ICE CREAM.) There is zero context provided by peer group, by overall general population, or with other college students. Avocado peels everywhere I tell you, and an indication of a creeping problem here.*

*Thus the Frivolous And Wasteful Committee On Avocado Peels is formed. All hail the FAWCOAP and its voluminous, poorly researched reports.

This is anything but definitive, but it's certainly more effort to provide some kind of meaning to these numbers than SI put forward in the piece. Start with 7% of student-athletes being arrested or charged, and then proceed down the rabbit hole with Slow States and watch the math come alive!

Of those seven percent, "nearly 60 percent…were guilty or paid some penalty". If we assume "nearly 60 percent" means 57% (shockingly, the actual numbers and survey methods aren’t given), then 4% of players on top 25 football teams have been actually convicted of, or plead guilty to, a crime.

The number of average college students with the same criminal record? According to this article from Corvallis, Oregon’s Daily Barometer, 3.45%. That’s right: Your typical college football player is one-half of one percent more likely to have a criminal conviction. To put that in perspective, a team of 85 players has half a person more convicted criminals on it than a sample of 85 students drawn randomly. Hide yo kids, hide yo wife.

We don't expect SI to trot out 80 page data tables, but if this were a video game it would be Civilization, and we'd be standing on a desert square swearing the entire world was sand and sorrow. The number cited by Slow States is not hard science, but it's a better attempt than any made by the writers here.

2. LEADERS, LEGENDS, AND LARCENY. Three Big Ten teams in the top ten for total arrests overall, and only one SEC team. THAT DATA SET WAS NOT REVIEWED BY DR. RO-TEL AND HIS STAFF OF CLEAN SHAVEN STATISTICIANS WHO USE BARBASOL HARUMPH HARUMPH. That poking sound should be loud and clear, even from out here in the holler where we're sitting next to a happily ticking still. (This is a troll-ass point we'll soften in part four, but feel free to stab away with it, SEC fans.)

3. Juvenile records are a horrible data set to use. There is a very good reason for this: juvenile records sit behind a wall of legal barbed wire that would make a sapper weep. They vary state-by-state by rule, and are often trumped by individual judge's sealing orders. This is why we don't use them in the Fulmer Cup, and why SI shouldn't have used them for any data set they wanted to call uniform or fair.

For instance, juvenile records in Florida aren't accessible unless the minor involved has committed one felony or three misdemeanors, and even then the judge may seal the case by order. The judge's order trumps the statute in this case, and blocks it from being seen on background checks. This happens frequently, and what you end up with much of the time is a blind "case" that's visible but unclear as to charge, outcome, or identity of participant. In New York all juvenile cases not involving a felony are sealed and therefore invisible. 

So when the nut graf of the piece mentions that only two out of 25 programs conduct background checks on their incoming recruits, there's two instances of serious slippage here. First, programs probably don't do them out of negligence and cost, not because they know that juvenile records searches are sketchy business at best. Second, they assume this means anything when they also write this in the middle of the piece:

Nor did SI and CBS News have access to juvenile arrest records for roughly 80 percent of the players in the study.

The issue of background checks for most recruits in most states is dead before you finish the first page of the article. They should have scrapped this element on day one and focused on what they could establish rather than flapping it out there as something significant.

4. Um, so what can we learn from this? Only that if you feel like trolling hard you can just go back to point #2 above, and that the Fulmer Cup provides a fascinating counterpoint to this since it covers what happens once you get immersed into a program, not what you have coming into it. ( A case where the SEC more than pulls its weight in comparison to what you shoddily sort-of-see in 20% accurate vision with this data.)

Also, if we take as truth the time-worn adage that felons and convicts make for better players, then Dave Wannstedt looks way worse than he already does for getting fired at Pitt despite having a crew of rowdy bastards just dying for a fight.