Have Studies Been Done On Criminal Background Check Accuracy?

October 24, 2007 5:53 AM

How reliable, statistically speaking, are criminal background checks?

Criminal background checks from a large national agencies (ADT, ChoicePoint, etc) obviously aren't infallible. But has anyone done actual studies on this topic?

I'm asking from an employer's perspective, not as some scofflaw looking to see what my chances of "slipping through the cracks" are. I need to report on how thorough and complete these checks really are, so I'm looking for hard data.

I found this MSNBC article in which an expert off-handedly mentions "we've done studies and...," and I saw this from the CS Monitor which uses a small sample sized criminology study.

Problem is, I'm not sure these data points represent enough evidence to counter the gut-feeling assumptions of my audience of the infallibility of the background check process.

The ideal study would be of a large sample and would focus on felons slipping through the cracks due to missing data (i.e., credit check data is useless, and details about the many people with data that IS on their report in error isn't as helpful as I'd like).

Anything like that out there?

posted by peacecorn to work & money (8 answers total)

I don't know about studies, but I have worked with criminal background checks, both state and federal, and they did sometimes miss things, especially arrests, so it also pays to check with the court records in the localities where the person lives or has lived if you want to be thorough. Hiring a reputable PI (oxymoron?) will likely get more accurate results than some online service. Even then you are at the mercy of overworked or bored clerks who sometimes just fail to do their jobs, or who put information in inaccurately so that it does not show up in a search, hardly surprising. Worst is when they fail to remove a record, such as an outstanding warrant for arrest, even after a person has been acquitted or even already served their time. Then some poor schmuck gets arrested, not infrequently with a new charge of resisting arrest due to arguments with the police over the validity of the warrant.
posted by caddis at 6:37 AM on October 24, 2007

it also pays to check with the court records in the localities where the person lives or has lived if you want to be thorough.

That ain't thorough enough.

I too have worked in background investigations, of which criminal checks were just one of my many duties in each file. The thing about background checks in general is that they are primarily based on existing records in the areas in which a person has lived. (that's why they ask you to list all of your residential addresses in the last 5 years, etc.)

If, say, for instance, there is an outstanding warrant for my arrest in say, West Virginia, based on something I may or may not have done there once upon a time when I was driving through in the middle of the night, there is slim to no chance that a background investigation is going to turn that up if I have never actually lived in West Virginia, savvy?

Or say I worked in NYC for 5 years but lived just across the river in New Jersey. If I got in a drunken brawl on the subway and had some record of that in Manhattan, since I've never lived in NY, most criminal reports aren't going to catch that. As long as I keep my nose clean when I go home to NJ, I'm set.

Its quite possible for a person to have a number of different blemishes in a number of different places that they've never technically resided at, that most criminal reports will never turn up.

To get to that level of thoroughness, you need to be interviewing friends and families, asking pointed questions about places they may be in trouble with that they have not lived in, etc.. In my experience, 95% of background checks are just trying to catch the repeat offenders that haven't been able to keep a clean record in the areas they reside in - which is only the *most likely* place to catch them with their hand in the cookie jar.
posted by allkindsoftime at 7:19 AM on October 24, 2007

Here's a study based on Navy enlistees. Money quote: "An additional 8% of the sample had self-disclosed law violations that were not discovered by the PRC or NACLC investigations."
posted by backupjesus at 7:24 AM on October 24, 2007

I think that a lot of smaller towns across the US were a little slower to adapt to our information age, and thus a lot of stuff that happened 5+ years ago could easily slip through the cracks.
posted by ian1977 at 7:30 AM on October 24, 2007

...they did sometimes miss things, especially arrests...

They don't "miss" those, they deliberately omit them. There was a big lawsuit about inclusion on criminal records of arrests which didn't lead to prosecution. The plaintiffs claimed that the Constitutional principle of presumption-of-innocence meant that a simple arrest that didn't lead to prosecution shouldn't be on the record, and the courts agreed. (And so do I.)
posted by Steven C. Den Beste at 7:39 AM on October 24, 2007

The ideal study would be of a large sample and would focus on felons slipping through the cracks due to missing data.

Do you want to find a study with this specific result, or do you want the truth?

I ask this because in Massachusetts we have the opposite problem. The criminal background check system issues reports that make people look like criminals even though they have never been convicted of a crime. There is a lot of movement to get this corrected. See Google results for Cori Reform.

That said, the Massachusetts system only detects crimes and issues that occured in Massachusetts. It doesn't show crimes that occured in other states. So it wouldn't require a study to tell you that some felons could come through the Mass Cori system looking clean even though they are not.
posted by alms at 7:40 AM on October 24, 2007

They don't "miss" those, they deliberately omit them.

Pre-employment checks perhaps, but the government databases, such as NCIC, have arrests.
posted by caddis at 7:59 AM on October 24, 2007

Do you want to find a study with this specific result, or do you want the truth?

alms, I realize the truth. On some background checks, erroneous crimes and convictions are detailed.

It is far, far more helpful for my presentation to talk solely about the stuff that is missing than the stuff that is there in error.
posted by peacecorn at 9:53 AM on October 24, 2007