Study: One of the most popular theories for America's huge crime drop is wrong

One of the more exotic explanations for America's incredible decades-long drop in crime is the lead theory. As the theory goes, we know that lead exposure is really harmful — it can cause learning disabilities, lower IQs, impulsivity, and aggressive behavior. So when America began taking lead out of its gasoline, it reduced all these bad outcomes — and crime, statistics show, fell right alongside rates of lead exposure.

A new study published in the Journal of Quantitative Criminology suggests, however, that this correlation may not exist at all, calling the supposed drop in both lead and crime a statistical "artifact" based on faulty crime data. And that means one of the most popular theories about the crime drop could be wrong.

What did the study find?

Journal of Quantitative Criminology

First, the study acknowledges that it's possible to find a correlation between lead and crime with some data — specifically, the FBI's Uniform Crime Report (UCR). According to the study, serious crime levels, measured by UCR, and lead exposure, measured by gasoline lead consumption per person, have a "very strong" relationship. So based on the UCR data, one would expect crime to go down as lead exposure goes down.

But the study argues that UCR, which uses data from police departments across the country, is unreliable. It compares UCR to two different measures: the National Crime Victimization Survey (NCVS) and homicides, which are typically considered the most reliable proxy for measuring serious crime levels. While NCVS figures closely correlate with the homicide rate, UCR does not, suggesting that UCR is much less reliable.

The problem: UCR seems to undercount crime levels, particularly "rape, aggravated assault, or a summary measure of serious violence." So as other data suggested that crime had plateaued and dipped, UCR data still suggested that crime was rising. This made UCR data correlate with a rise in lead exposure that other data sources didn't strongly correlate with.

But based on NCVS's data, there was literally zero correlation between lead and crime — although there was a very small correlation between lead and the homicide rate.

In other words, the only way to find a strong national correlation between lead and crime is to use faulty UCR data. When looking at the more reliable data, there's little to no relation between lead and crime.

Is this it for the lead theory?

Scott Olson/Getty Images

Over at Mother Jones, Kevin Drum, one of the biggest proponents of the idea that a drop in lead exposure led to a drop in crime, pushed back at the Journal of Quantitative Criminology study. He argued that the study looked at too narrow of a time period (from the 1970s to 2000s, instead of 1950s to 2000s), and missed what's going on at the local, state, and international levels:

The lead-crime correlation only works for the entire period from approximately 1950 to 2000. If you pick out a subsection of that period, it's not surprising that you might lose the signal. That's especially true when you choose the middle of the period, a time when lots of cohorts all had extremely elevated lead levels. With levels that high among a broad swath of the population, it's entirely possible that other factors were temporarily more important. (Just as they are now, when lead levels are low across the board.) …

On a statistical level, there are state-level correlations, city-level correlations, and, most importantly, international correlations. All of those together paint a very convincing picture. On an outcome level, there are prospective studies that track children with elevated lead levels. All of them show that they get arrested at higher rates the higher their childhood lead exposure. On a medical level, there are now brain scan studies that provide a plausible pathway for lead exposure to cause permanent neurological damage to parts of the prefrontal cortex associated with impulse control.

The authors of the study acknowledge part of this view, noting that their findings only dispute the correlation between crime and lead on a national scale, not a local or personal level.

Still, the national correlation is pretty important to explaining the national trend. It is possible that the study is looking at too narrow of a time period, and maybe the correlation between lead and crime at the local, state, and international levels is more suggestive. But when trying to explain why crime dropped in the US as a whole, the lead theory is a lot less convincing if there aren't strong national numbers to back it up.

We have a request

Vox's journalism is free because we believe that everyone deserves to understand the world that they live in. That kind of knowledge helps create better citizens, neighbors, friends, parents, consumers and stewards of this planet. In short, understanding benefits everyone. You can join in on this mission by making a financial gift to Vox today. Reader support helps keep our work free, for everyone. Will you join us?

One-Time Monthly Annual

$95/year

$120/year

$250/year

$350/year

Other

$

Yes, I'll give $120/year

Yes, I'll give $120/year

We accept credit card, Apple Pay, and Google Pay. You can also contribute via