Dueling Stats: CDC vs. FBI

When you have two government agencies generating conflicting data, which do you believe disbelieve less?

I ran across this story today:

CDC: Firearm-Related Homicides Surged During Obama’s Last 2 Years in Office
According to the CDC, “the number of firearm-related homicides was relatively stable during 2010–2014,” then firearm-related homicides sharply increased. For example, there were 11,008 firearm-related homicides in 2014 and that number rose to 14,415 in 2016. That is a 31 percent increase.

Sure enough, the CDC’s WISQARS is now reporting 14,415 homicides by firearm. Which struck me as odd, because I’ve been using the figure 11,004 for 2016.

11,004 is exactly what the FBI’s UCR is reporting for 2016; that’s 3,411 fewer firearms-related murders than the CDC retroactively claims.

In fact, the UCR figure for total murders — any weapon — is 15,070, only 655 more than the CDC’s firearm-only murders claim. The CDC claims 19,362 total murders (+4,292 difference).

Oh, and the 2014 numbers for firearms-related homicides?

CDC 2014: 11,008

UCR 2014: 8,124 (A difference of 2,884 murders.)

Odd.

The agencies use different sources for their data. The UCR is based on police incident reports. WISQARS comes from incident coding in medical reports. One or the other — or both — could be wrong. But clearly both need some quality assurance checking to see if either come close to reality.

I’d love to do a little research into the discrepancy myself, but as an unpaid volunteer, it’s beyond my means. If you’d like to see me dig into this, consider contributing a little to the cause.

ETA, 8//2018: Gee, how ’bout that?

CDC Finally Corrects Bad Data On Number of Firearm Deaths
However, in 2017, John Lott found a discrepancy. He noted that due to an error in Tennessee’s data, the total count was about 100 higher than it should have been. He notified the CDC and they acknowledged the mistake.

It just took them a year to finally fix it.


Carl is an unpaid TZP volunteer. If you found this post useful, please consider dropping something in his tip jar. He could use the money, what with truck repairs and recurring bills. Click here to donate via PayPal.

Facebooktwittergoogle_plusredditpinteresttumblrmail

5 thoughts on “Dueling Stats: CDC vs. FBI”

  1. The firearms homicide numbers in the FBI report are only those cases in which the local/state agency reported the instrument of homicide. I don’t think Florida reports any supplemental data.

    1. Still doesn’t up. If you include the UCR “Other weapons or weapons not stated” — 903 — in the firearm count, you’re still thousands short of the CDC claim. 11,004 + 903 = 11,907 vs. CDC 14,415.

      The CDC firearm count alone is almost as high as UCR total murders (15,070).

      1. because those cases – the other weapons or other firearms – are still cases in which the local agency reported the data to the FBI. That 15,070 doesn’t include any murders in the state of Florida and only 3 in the entire state of Alabama. If you look at table 1 in the FBI report, it lists 18,208 total murders/non negligent manslaughter. They didn’t receive supplemental data on all of those, which is why there are “only” 15k murders listed in the weapons tables. The FBI report is really just an aggregate of state reports produced by state and local agencies. Not every agency reports their data to the FBI.

  2. I believe it’s been well publicized previously that the UCR doesn’t record every incident, only the ones that are reported to them. Some jurisdictions do not make it a requirement (as posted in another comment) even entire states may not participate. I am not claiming to know which ones…but the FBI itself has said the data they have is incomplete due to reporting inconsistencies, errors, etc. (https://ucr.fbi.gov/crime-in-the-u.s/2013/crime-in-the-u.s.-2013/resource-pages/nations-two-crime-measures/nations_two_crime_measures) .

    If you look at the UCR numbers, from 2014 to 2016 gun homicides went up 35%…4% more increase than the CDC reports…with a much lower total.

    That being said, CDC not only has the same probability of omissions and errors, but an impetus to over-report allowing them to classify gun homicides as a public health issue, increasing their authority and oversight of the issue, as well as their budget for the added responsibility. The CDC has certainly been looked at by some political factions as a possible source for an end-run around Congress on gun control.

    Personally, I think the real numbers will fall somewhere between the two…depending on who’s doing the counting.

Leave a Reply

Your email address will not be published. Required fields are marked *