From Neil deGrasse Tyson

3,400: Americans who died by Terrorism since 2001

3,400: Americans who died by household Firearms since five weeks ago.

Several users have since asked for a source, and checking his Twitter feed he hasn't followed up with a source.

I know that some children die from guns within their household, but I find the statistic very surprising, unless I've misinterpreted "household Firearms [sic]".

Do 3400 Americans die from firearms every five weeks? And are they from guns in their own household?

  • Ya, that term does not make a whole lot of sense. Most firearms have a household as very few guns overall are owned by homeless people. Maybe they mean 3,400 Americans are killed by firearms stored in their own household every five weeks?
    – Jonathon
    Commented Nov 21, 2015 at 3:26
  • 8
    'Household' could reasonably be taken to mean 'personally owned' as opposed to guns owned by the military or similar. Commented Nov 22, 2015 at 4:04

2 Answers 2


Not Quite. But it is large number.

First there is the question of how to interpret "Household firearms". While Technically this leaves open the possibility that the quote is referring to all personal firearms; If we go with this definition we will include pretty much every single death by firearm outside of military and police operations. But personally, based on the words used I believe that they meant people killed by firearms owned and/or stored in their very own homes (which is a very common thing to talk about for the anti-gun crowd).

There are just over 30,000 deaths by firearms per year in the US, or just under 3K per five weeks. So as per the specific claim, I am sure there exist five week stretches that involved 3,400 deaths.

But not all firearm deaths are caused by firearms from the victims own house. The majority of deaths are suicides - around 16-21,000 per year. (I don't have any data on this, but I am taking it as a given that the vast majority of firearm assisted suicides are performed with a firearm they or a member of their family owns.)

Accidents are insignificant, making up just 1.5% of firearm related deaths.

Homicides make up around a third of all firearm related deaths. Some number of these will also be committed with household firearms, but I have been unable to find any specific statistics for this.

Source: Gun Policy.org

  • For full transparency: that source is at best neutral, and seems to be more inclined to be anti-gun based on who runs and funds and supplies info to it.
    – user5341
    Commented Nov 22, 2015 at 19:39
  • Why do you exclude suicides and accidents? The original claim just refers to "died".
    – oefe
    Commented Nov 29, 2015 at 9:20
  • 1
    @oefe I don't. Including all deaths, we get a number very close to the initial statement. But, if you believe that "Household firearms" means something more specific, I talk about excluding the some of the homicide deaths. Either way, it was important to break down the death count. As most people when told that there are 30,000 firearm related deaths probably do not think that most of that was suicides. And people seem far less likely to blame the gun for suicides, so to them the simple generic statistic might be extremely misleading.
    – Jonathon
    Commented Nov 29, 2015 at 13:52

Not if you use meaningful definitions.

First of all, as @Jonathon's excellent answer alredy pares down,

  • Only 33% (~10K) firearm deaths a year in USA is a homicide.

    Given that Particular policies that have restricted access to firearms appear to lead to fewer suicides for particular groups but have relatively little effect on the overall suicide rate (source: "Guns and Suicide: Correlation or Causation?" article by Mark Duggan, 2002)), comparing the causative deaths from terrorism to not-necessarily-causative deaths by gun suicide is not a meaningful comparison, especially given that the comparison occurred in the context of gun control debate which the conclusion above shows to be non-productive to reduce suicide rates.

  • Of those, ~500/year (so, on average, 50 in five weeks) are gun accidents. We'll be charitable and assume 100% of them can be claimed to be "household firearms". (src: Alpers, Philip, Amélie Rossetti and Daniel Salinas. 2015. Guns in the United States: Unintentional Gun Deaths. Sydney School of Public Health, The University of Sydney)

  • Of the rest, we need to estimate how many are by "household firearms". Without having precise definition, it's hard to know what Tyson meant, but let's assume he spoke in the context of gun control debate and as such, he meant guns legally purchased by the owner (who was the crime perpetrator).

    This is a surprisingly (or not, to a criminologist) difficult figure to determine. However, a realistic estimate is between 3% and 10% - which would make this number between 300 and 1000 a year, and between 30 and 100 in five weeks.

    • The best coverage of the topic I was able to find was Politifact's analysis of Joe Scarborough's claim.

      Ultimately, there are holes in the data. But Cook said while 3 percent or 10 percent might not be the exact number of legally purchased firearms used by criminals, the fraction is in that ballpark.

    • Another analysis covered a somewhat different data (ALL crimes vs. homicides only), but the numbers support the picture above:

      The 2013 National Crime Victimize Survey report there were almost exactly 300,000 crimes, including murders, facilitated with a firearm. Of those, not more than 5,000 can be shown to have been facilitated with a firearm legally purchased by the offender.

      This is 1.7% of total crimes that were done with "household firearms".

  • Added up the estimate above to the accidents, we get a ceiling of between 80 and 150 americans who die from household firearms every five weeks - a far cry from 3400 claim, making the claim wildly false.

  • 6
    You are changing definitions. There is absolutely no indication that the OP meant to exclude suicides, and precious little to suggest they meant 'legally purchased by the perpeteastor'. If I steal your gun from you and then kill you with it, you are still killed with a household gun by any definition. Commented Nov 22, 2015 at 22:23
  • 4
    -1 because I disagree with your choice of interpretations. I would read "household firearms" as firearms that were originally purchased for use by a non-military, non-law-enforcement individual, even if they were later stolen/wrested from their grip. Your argument that suicide rates make comparisons misleading is an excellent footnote, but doesn't make the original claim incorrect.
    – Oddthinking
    Commented Nov 22, 2015 at 22:59
  • 2
    Is the claim vague? Yes (not OP's fault - deGrasse Tyson's fault.) Should you rail against the apples & oranges comparison? Yes. Should you specify how you interpret the claim? Yes. I agree/applaud that. But your choice of interpretations seems extreme to me, which turns it into a strawman attack.
    – Oddthinking
    Commented Nov 22, 2015 at 23:31
  • 1
    @user5341 If you don't care what the questioner meant then obviously you aren' t going to give a relevant answer Commented Nov 23, 2015 at 4:23
  • 2
    Then let me try that again: "You are changing definitions. There is absolutely no indication that the claimant meant to exclude suicides, and precious little to suggest they meant 'legally purchased by the perpetrator'" Commented Nov 23, 2015 at 14:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .