United States Homicide Rate vs Gun Ownership by State

There is literally 0 correlation between how many people own firearms, and the homicide rate with firearms, or overall homicide rate as below graphs and correlation coefficients illustrate. This means that any policy that seeks to decrease or increase gun ownership within the population will have absolutely no affect on the murder rate within the same population. But this should be common sense. We should know that violent behavior, or any behavior for that matter is not dictated by inanimate objects. 16 oz cups do not make people fat, pencils do not miss spell words, and guns do not cause violent behavior. Tell me this, if you wanted to decrease shark attacks, would you advocate for banning the sale of ice cream? Well gun bans sound just as ridiculous.

This graph shows homicide rate with firearm vs firearm ownership


This graph shows the overall homicide rate vs firearm ownership. Both graphs show no correlation.


In other news scientists can’t figure out how Canadians watch movies and play videos games, and manage not to shoot each other.


  • It appears your data is comparing 2004 gun death data to 2009 gun ownership data.

    Here is where I got my data:

    It disagrees with your wikipedia data in a very serious way.

    Gun deaths by state: the violence policy center data for 2009
    gun ownership by state: http://usliberals.about.com/od/Election2012Factors/a/Gun-Owners-As-Percentage-Of-Each-States-Population.htm

    The resulting graph of deaths (Y axis) vs gun ownership (X axis) shows a positive correlation.

    y = 0.1533x + 6.1237 R² = 0.1726

    It is not clear to me how I post my graph, but it shows a least squares fit that runs from lower right to upper left, meaning:
    More guns means more deaths.

    It also shows, if you correlate the data by which 2012 Presidential candidate won that state, that, in general, the RED states have higher percentages of gun ownership and higher
    gun deaths per 100,000 of population than do the BLUE states.

    Tell me how to post my graph and you will see exactly what I mean.

  • Try a correlation between suicides and gun ownership, then firearm suicides and gun ownership. I get r2 of 0.44 and 0.70 respectively for the years I examined. I encourage you to try it yourself.

  • Clearly the relationship between gun ownership and the homicide (or suicide) rate is complicated, and cultural factors etc have a major influence. Also any comparison has to be multivariate to account for such obvious risk factors for violence such as age. In short simplistic correlations such as the ones here are pretty much useless.

    For a more statistically valid (or at least methodologically valid) analysis see the the Harvard School of Public Health review: http://www.hsph.harvard.edu/hicrc/firearms-research/guns-and-death/

    • Shouldn’t have said that simple gun/homicide correlations are worthless, they just show there is not a very strong correlation. The Harvard review finds that most studies that include other factors show a positive if not overly strong correlation between gun ownership and homicides (by country, state, etc.) And of course correlation is not causation.

    • More specifically, that is the Harvard Injury Control Research Center review. That section of the Harvard School of Public Health is funded directly by the Joyce Foundation, a very anti-gun interest group.

      A Harvard study *not* funded by the Joyce Foundation found opposite results: http://www.law.harvard.edu/students/orgs/jlpp/Vol30_No2_KatesMauseronline.pdf

      But that’s how academia works. Examine the sources of funding, contributor backgrounds, methodology, and then repeat for each cited source to be sure you’re not getting distorted info.

  • Hi Randroid,

    Perhaps you could use updated numbers,


    Or what about a real study, rather than “about.com”


  • Very nicely run regression, but your conclusion “There is literally 0 correlation between how many people own firearms, and the homicide rate with firearms” is not supported by the data.

    First issue is that the murder rate data you are using from Wikipedia is from the FBI’s 2010 data (which is slightly lower than CDC data), while the gun ownership data you are using is actually from a telephone survey conducted in 2001 by the Behavioral Risk Factor Surveillance System (BRFSS) in North Carolina. http://www.schs.state.nc.us/SCHS/brfss/2001/us/firearm3.html

    So your regression is comparing data that is over 10 years old with newer data. In addition, the gun ownership % may be unreliable–it was a telephone survey of 200,000 people. Even further, % gun ownership does not necessarily reflect the number guns in a given state. In the late 1990s, the Department of Justice did a study of prison inmates to find out where their guns come from and found that many of them were black market guns (from 35-50%). None of these guns are accounted for in the above study, because all they were looking at is household ownership, not total number of guns. If criminals had to get guns from households, and every household surveyed had only one gun, the statistic might have at least a sliver of weight. But, sorry to burst your bubble, this data is not proof of a lack of correlation between gun ownership and gun murder.

  • I agree, the data could possibly be unreliable, because it is collected by surveys. However, if that is what we conclude then we can’t perform any statistical analysis since all gun ownership data is a survey.

    I also think this discussion fully illustrates one specific point, that trying to form government policy based on macro statistics is wrong. In any such cost benefit analysis tries to balance people’s live against each other, and this is something is not possible. In practice, it leads to endlessly battling statistical studies, as our discussion illustrates.

    The issue with guns is the threat of force. However, government cannot treat men as guilty until they have proven themselves to be, for the moment, innocent.

    But this is precisely what gun control laws do. Gun control laws use force against the individual in the absence of any specific evidence that he is about to commit a crime. They say to the rational, responsible gun owner: you may not have or carry a gun because others have used them irrationally or irresponsibly. Thus, preventive law sacrifices the rational and responsible to the irrational and irresponsible. This is unjust and intolerable.

Leave a Reply

Your email address will not be published.