Do Uncommon Names Turn Kids Into Criminals?

INSERT DESCRIPTIONPhoto: tomeppy

The answer to that question is almost certainly “no,” but a new study that is getting lots of media attention does claim that there is a correlation between having an uncommon name and being more likely to show up in the juvenile justice system. The study finds this relationship to be true both for blacks and whites.

As the authors clearly state in the abstract of their paper, “Uncommon names are likely not the cause of crime but correlated with factors that increase the tendency toward juvenile delinquency.”

To me, that makes the result a lot less interesting. It is kind of like saying that we know that people who regularly wear orange jumpsuits are more likely to be criminals, when orange jumpsuits just happen to be the required uniform in the state prison. The wearing of orange jumpsuits isn’t the cause of the criminal activity, but it is highly correlated with past criminal involvement.

And to be honest, I think there is a good chance that even the claimed correlation between unusual names and criminality is wrong. I only quickly glanced through the academic paper, but it seems to me like the authors have made a mistake that will bias their results.

The authors first compute criminality for each name by taking the ratio of the number of juvenile delinquents with that name and dividing it by the number of children total with that name. The higher that ratio, the more criminal the name. But then the authors take the log of that ratio. The problem is that the log of zero is equal to negative infinity, so any name for which that ratio is equal to zero gets dropped from the analysis.

The kinds of names that will have a ratio of zero are uncommon names for which no one with that name is a juvenile delinquent.

If I understand correctly what they are doing, if exactly one person has a particular name, the only way that the observation for that name will be included in their sample is if that person is a juvenile delinquent! This leads to a powerful bias toward mistakenly concluding that people with uncommon names are more likely to be criminals.

TAGS: ,

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.

 

COMMENTS: 61

View All Comments »
  1. Elana says:

    Gee I hope not…
    I might be in trouble otherwise…

    Thumb up 0 Thumb down 0
  2. Francis says:

    If the authors did indeed make the mistake that Levitt accuses them of having made, I agree that the result will probably be inaccurate.

    However, if uncommon names were indeed correlated iwth higher criminality, this would be a far more interesting observation than the one about orange jumpsuits and crime. The difference is that in Levitt’s example, crime causes orange jumpsuits. No one is seriously suggesting that committing a crime causes you to go get a funny name. This would be a more interesting clue to follow up: if committing criminal acts isn’t causing you to have an uncommon name, why are the two variables correlated? I agree that it’s not as exciting to know things correlated with crime as it is to know things that cause crime, but you have to take what you can get.

    Thumb up 0 Thumb down 0
  3. Josh Penslar says:

    Good catch! If I taught Stats 100 this would definitely be an extra-credit question on my exams. If nothing else, it would pick out the Freakonomics readers in the class!

    Thumb up 0 Thumb down 0
  4. William says:

    Clever observation, Steven!

    This could be exacerbated further, since they did NOT change the names of people with unusual spellings for common names (Patric is one example they gave). A birth certificate misspelling could result in a one-of-a-kind name and make the study more vulnerable to the effect you describe.

    This could have been made manifestly apparent if they listed the ratio of delinquents to total people with the name of Michael, vs. the ratio of delinquents to total people with unique names.

    Thumb up 0 Thumb down 0
  5. Rob says:

    @ Francis
    Seeing as parents choose the name, one could assume that it’s their influence which is the correlation: if you are a parent who is more likely to bestow an unusual name, then…

    Thumb up 0 Thumb down 0
  6. Robin says:

    Yesterday, I saw a report about this which disturbed me. It suggested that this study’s results could be used by police to catch criminals. I don’t know about everyone else, but that sounds a bit too much like Minority Report to me, and a serious misapplication of this study even if it is sound.

    Thumb up 0 Thumb down 0
  7. Paula Hall says:

    It’s been a while since I read Freakonomics, but didn’t the authors find that uneducated and/or single parents were more likely to give their children uncommon names? If that is the case, then that this study found a correlation would not be surprising.

    Thumb up 0 Thumb down 0
  8. Ben says:

    In reality, though, how many TRULY unique names (that is, only 1 person in the entire country) are there? My suspicion is that it is incredibly small, and does not/would not have very much of an effect when compared to the total number of ‘unique’ names (as defined by the study).

    But it is still something worth considering.

    Thumb up 0 Thumb down 0