Data Analysis of USAToday’s NFL Arrest database: 15 Surprising Insights

Arrests frequency is not increasing

A simple plot of the number of arrests per year gives the appearance that arrests are slightly increasing over time at 1.36 arrests per year (but down from a three-year peak from 2006-2008). However, from a statistical point of view, the arrest count per year is relatively constant. The 2014-15 season should have about the same number of arrests as the last two seasons:

he blue lines show arrests by year and a projected likelihood "cone" of arrests. The red lines a machine prediction for next year. The net: Arrests are not increasing. (Click for full size)
The blue lines show arrests by year and a projected likelihood “cone” of arrests. The red lines a machine prediction for next year. The net: Arrests are not increasing. (Click for full size)

However, the new scrutiny on this issue may introduce an increase in arrests. Only time will tell

NFL players are arrested less often than everyday US residents

At first reaction, 730 arrests across 544 players “sounds” like a large number. It turns out it is not. It represents a 2.7% annual arrest rate for NFL players. This is one-third lower than 4%, the arrest rate for the US population at-large.

Nevertheless, by most socio-economic theories, we would expect NFL players to arrested less often. When comparing the median NFL player against the median US resident:

  • NFL players make 15.2x as much income
  • Virtually all of them attend “some college” (vs. 84% for the median similarly aged US resident). This is not a surprise as college is essentially the “farm league” for NFL recruitment
  • Their 6-year college graduation is also nearly triple the US populations. Even accounting for a year of “red-shirting”, NFL players graduate at a much higher success rate
Comparison of education, arrest rate and salary between the median NFL player and median US resident. The net: NFL players are arrested less often than most. (Click for full size image.)
Comparison of education, arrest rate and salary between the median NFL player and median US resident. The net: NFL players are arrested less often than most. (Click for full size image.)

That, however, makes the first arrest pattern even more interesting.

However, lightning often strikes the same place twice (it can even strike the same place nine times)

For many of these players, getting arrested is not a “once in a lifetime” occurrence. Of the 544 players arrested between 2000 and today:

  • 124 were arrested more than once
  • 39 were arrested three times or more
  • 13 were arrested four times or more

One player, Adam Jones, was arrested 9 times (this equated to one arrest every 11 months over a 8.2-year period):

Number of players who were arrest at least this many times. The net: NFL players have a 34% net recidivism rate. (Click for full size.)
Number of players who were arrest at least this many times. The net: NFL players have a 34% net recidivism rate. (Click for full size.)

Furthermore, multiple arrests for the same player were not concentrated within a single team contract. Thirty-nine players were arrested multiple times, across multiple teams. Two players (David Boston and John Abraham) were each arrested across separate contracts with three different teams.

As striking as these numbers are, they are a statistical understatement. Many player arrests (69 of them) were arrests for multiple counts, across multiple charges. For example, the recent arrest of Jonathan Dwyer counted as (only) one arrest but involved six counts across five criminal charges.

Next: Guilt vs. Acquittal and Team Response