Most murders occur in just 5% of US counties?!

On 4-26-2017, Fox News published a story by Perry Chiaramonte with the headline “US murders concentrated in 5 percent of counties.” This story was based entirely on a report (pdf version) by the Crime Prevention Research Center (CPRC), a hard right source which claims it is dedicated to “conducting academic quality research” but really just publishes anti-gun-control propaganda.† The CPRC report twists the statistics to fit their false narrative.

The CPRC builds its lies on a foundation of truth. Yes, it is true that most murders occur in just 5% of the country’s counties. And the report even tells us that this 5% holds nearly half of the US population, though they don’t make much of that fact. But that’s the exactly problem. They should.

Throughout the report they refer to counties with more absolute numbers of murders as “the worst” and, even more misleadingly, “the most dangerous.” This rhetoric ignores the fact that the murder rate is more important and revealing than the absolute number of murders.

Worse, they do pivot to talking about the murder rate, but they only do so in the context of having ranked these counties by absolute murders first. That’s where the report becomes what can only be described as purposely obfuscatory. They go as far as to make the claim that “Removing the worst 2% or 5% would have reduced the US rate to just 3.06 or 2.56 per 100,000, respectively.” OK, but you’d also be cutting the population in half.

What if we just look at the counties with the worst murder rates to begin with? Well, it turns out that only 1/3 of the CPRC’s “most dangerous” 5% falls in the actually most dangerous TEN percent. Moreover, 3 in 10 of them have a murder rate that is below the national average!

If we want to play the CPRC’s stupid game of removing counties, we would be better off ranking them by murder rate and removing the worst 20%. Firstly, we’d be getting rid of fewer people, reducing the population to 210 million people rather than reducing it to 170 million. Secondly, we’d be making a much greater improvement: instead of the national murder rate dropping to 2.56 per 100k, it would drop all the way to 2.12.‡

But that doesn’t fit the CPRC’s narrative because many of those counties are rural and the lie the CPRC was really trying to tell was that rural areas have less murder despite having higher gun ownership. What they found is that wasn’t the story that would emerge if they ranked counties by murder rate and went from there. But when ranked by absolute number of murders, urban areas necessarily fall near the top. More people means more murders. Only after they achieved the ranking they wanted did they begin examining murder rates.

This intellectual dishonesty was purposeful and strategic.

As an interesting side note, ranking the counties by murder rate and cross-referencing with county level voter data reveals that nearly 3 in 4 of the deadliest 20% of all counties voted Republican in the 2016 Presidential election.


† The only author listed on the report is John R. Lott, Ph.D.  He is a well-known anti-gun-control advocate with some academic credentials and a questionable background. Read up on him here and here.

‡ I used the data made available by the CPRC and the Census Bureau’s 2014 estimated population by county data available here.