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

Relation of Age, Size, Speed and Strength to NFL Arrests

Several in the media have opined that the push for NFL players to get stronger and stronger could be a driver leading towards increased aggressiveness (leading to higher likelihood of arrest. This was the primary driver that led me to mash-up the NFL Combine and NFL.com height, weight and age data with the USA Today NFL Arrest database.

The following diagram depicts the relative size and strength of the typical (i.e., median) arrested NFL player in comparison the typical adult male, adult female, female teen, and male child. The adult ages match the median age of the arrested NFL player (for comparison in crimes ranging from arguments at bars to domestic violence). The female teen age of 17 was picked as this was the most commonly occurring age mentioned in crimes against teenage females. The male child matches the age of the child affected in the recent Adrian Peterson arrest.

To-scale image comparison of size (height and weight) of typical (median) arrested NFL players to others. The net: NFL players are big and very strong. (Click for full size.)
To-scale image comparison of size (height and weight) of typical (median) arrested NFL players to others. The net: NFL players are big and very strong. (Click for full size.)

The reason this diagram does not show size, speed and strength by criminal charge is discussed below.

Age was not a factor in arrest rate (or legal outcome)

The median age at time of arrest across all arrests was 25 years, 6 months. In addition, 88% of all arrests were of players under 30.

At first blush, arrest looks skewed towards the young. However, the distribution of age at time of arrest virtually matches that of the distribution of age across all NFL players:

Distribution of Age of Player at Time-of-Arrest vs. Age of all Players. The net: Players were young at the time of arrest, but no younger than others. (Click for full size.)
Distribution of Age of Player at Time-of-Arrest vs. Age of all Players. The net: Players were young at the time of arrest, but no younger than others. (Click for full size.)

Some interesting observations perusing the data:

  • The youngest player (at time of arrest) was Terrell Suggs (age 20.5). He was arrested (and eventually Acquitted) for Assault
  • The oldest player was Bruce Smith (age 39.9). He was arrested for a DUI (BAC 0.07) and later Acquitted

Height and weight were not factors as well

It would have been interesting to see if larger players (or smaller) players had a higher propensity to be arrested—in general or for a specific set of crimes. However, that was not the case:

  • Neither height nor weight was any statistical indicator of likelihood or arrest (or legal outcome), in general or any crime
  • Furthermore, there was less then 5% variation in height or weight for the median, upper quartile or lower quartile for players when comparing arrests by criminal charge (or arrests by legal outcome)

Some interesting observations perusing the data:

  • The tallest player arrested was King Dunlap (6’11”). He was arrested for Disorderly Conduct (results unknown).
  • The shortest was Garrett Wolf (5’7”). He was arrested for theft. The charges were dropped.
  • The heaviest player arrested was Aaron Gibson (410 lbs.). He was arrested for Assault and pleaded Guilty.
  • The lightest player arrested was Alex Ardley (173 lbs.). He was arrested for a DUI and pleaded Guilty.

Speed was not a factor

Speed (as measured by 40-yard Dash time on the NFL Combine) was not a factor in arrest likelihood, criminal charge or outcome. Even the speed of arrested players had little variance (less than 0.2 seconds variance on a 4.68-second mean 40-yard Dash time).

Some interesting observations perusing the data:

  • The fastest players arrested were Randy Moss, Darrius Heyward-Bey and Geoffrey Pope (4.25-second dash). Moss was arrest for Reckless Driving and Drugs (he pleaded Guilty). Heyward-Bey for a DUI (pleaded Guilty). Pope for Assault (the case was dropped).
  • The slowest was Trai Essex (5.61-second dash). He was arrested for Disorderly Conduct (the case was dropped)

Strength was not a statistical factor…but it does show a surprising pattern regarding arrests for Sexual Assault

The NFL Combine includes a test to see how many times (i.e., reps) each player can bench press 225 lbs. (about 100 kg.) in one set. Justin Ernest holds the record at an amazing 51 reps. (To be clear, Ernest was NOT arrested for anything, he simply holds the NFL Combine strength record)

Strength was pretty diverse across the NFL Arrest Database:

  • The strongest player arrested was Albert Haynesworth (39 reps). He was arrested four times (Assault, Reckless Driving, Misdemeanor Sexual Abuse). He pleaded Guilty on two cases, other two were dropped
  • The least-strong players were Willie Andrews, AJ Jefferson, and Cortland Finnegan (a three-way tie at 7 reps). Andrews was arrested twice (Domestic Violence, Drugs, Gun charges); results are only known on the Gun charges (he pleaded guilty). Jefferson was arrested for Domestic Violence (results are Unknown). Finnegan was arrested for (and pleaded Guilty to) a DUI charge.

Several in the media have opined that the push for NFL players to get stronger and stronger could be a driver leading towards increased aggressiveness (leading to higher likelihood of arrest). This was the primary driver that led me to mash-up the NFL Combine and NFL.com height, weight and age data with the USA Today NFL Arrest database.

It turns out that strength was not a statistical factor in likelihood of arrest or even general category of crime. Even in cases when players charged with violent crimes where not statistically stronger (or weaker) than the typical NFL player.

However, strength does reveal a surprising pattern–but one completely opposite of media conjecture. The distribution of strength scores for those arrested for Sexual Assault was lower than the distribution for any other criminal charge. It was not just a “little bit lower”, it was a lot lower: the strongest people arrested for Sexual Assault had lower NFL Combine bench press scores than the “weakest” (i.e., 25% quartile) players arrested for virtually every other type of crime:

Distribution of Strength (25th - 75th percentile) by criminal charge at arrest. One of these (those arrested for Sexual Assault) stands out from the others (these players have much lower strength scores). This is a potential indicator but still lacks sufficient data for statistical validity. (Click for full size).
Distribution of Strength (25th – 75th percentile) by criminal charge at arrest. One of these (those arrested for Sexual Assault) stands out from the others (these players have much lower strength scores). This is a potential indicator but still lacks sufficient data for statistical validity. (Click for full size).

I do need to point that that the sample is not statistically large enough to make a decisive statement. However, there exposes a potential pattern that is worth further analysis, especially looking at social dynamics and crime.

Next: Wrap up and methodology