Topic: Gun Homicides vs. Gun Ownership

This is a one-post topic examining the relationship between gun ownership and crimes committed with guns in America.  Strictly speaking, this is not a concern of environmental science.  But it is an issue that can be studied with the same sort of methods and analysis used by environmental science to tackle environmental problems, and therefore an appropriate topic for this blog.

Consider the following question: Is there a correlation between high rates of gun ownership and murders committed with firearms in the 50 states? To help answer it, I found three sets of data:

  • FBI 2010 crime statistics from their Unified Crime Reports (UCR).
  • U.S. Census Bureau 2010 statistics on population.
  • Washington Post, quoting a Center for Disease Control (CDC) 2001 survey on gun ownership.

The last source is problematic as it was done in 2001 rather than 2010. I’m assuming that gun ownership percentages did not change significantly between 2001 and 2010, but I could be wrong1. It is also not an exact count but an estimation using samples. There is no official record of who is a gun owner in America. If you recall, the gun lobby is vehemently opposed to a national gun registry, and so we can only estimate the number of gun owners using surveys. This particular survey was conducted in all 50 states by the Centers for Disease Control’s Behavior Risk Factor Surveillance System. Respondents were chosen at random and were asked the following question: Are any firearms now kept in or around your home? Include those kept in a garage, outdoor storage area, car, truck, or other motor vehicle.

Therefore, when the survey states that 41.7% of respondents in Missouri claimed to have a firearm in the household in 2001, it doesn’t necessarily mean that 41.7% of all households in Missouri have firearms. There is a confidence interval or margin of error surrounding the sample statistic within which the true percentage is likely to fall. For example, suppose the 95% confidence level is plus or minus 3%. Then if 41.7% of the Missouri sample answered yes to the question, I can claim with 95% confidence (that is, I expect to be right 19 times out of 20) that the true percentage of Missouri households possessing firearms is somewhere between 38.7% and 44.7%. I still use 41.7% in my calculations because that is my best estimate, but readers must understand that the results of those calculations are going to be fuzzy.

Actually, I tend to think that gun owners are underrepresented in these surveys. I suspect many gun owners are chary about admitting they own a gun, and may refuse to participate or may lie (think about it, if a stranger called you up and asked you if you owned a gun, would you readily admit it?). But unfortunately this survey is the best we can do for now, so we will use it with the understanding that its results may not be totally precise, but with the belief that they are not far from the truth.

I blended the three sets of data into one spreadsheet. Please click here now to view the spreadsheet (a spreadsheet program such as Microsoft Excel™ or OpenOffice Calc is required to view the spreadsheet).

As you can see, the states (and the District of Columbia) are sorted by percentage of households possessing a firearm in ascending order. Some states such as Hawaii have low ownership percentages and low firearm murder rates, but many other states such as Idaho, Montana, South Dakota, and Montana have very high firearms ownership percentages but also have very low firearm murder rates. Louisiana has a high ownership percentage and the highest firearm murder rate in the country, but Maryland, Delaware, and New Jersey have comparatively low ownership percentages but have disproportionately high firearm murder rates.

That suggests that levels of gun ownership alone cannot predict the firearm murder rate and I think there is a great deal of truth to that. However, if we view a scatter plot of the data, the picture proves a little more complicated. I couldn’t get my graph software to draw the plot I wanted, so I drew it myself. Please forgive its crude appearance. When I get the satisfactory software, I’ll redo it. Here is the graph:Murders vs Gun Ownership2

The graph shows states with relatively low firearm ownership levels with both high and low levels of murder by firearm, and states with relatively high firearm ownership levels with both high and low levels of murder by firearm. But there also appears to be some structure. There appears to be an arc of states, going from left to right, Hawaii, Massachusetts, Rhode Island, Connecticut, New York, Illinois, California, Florida, Delaware, Arizona, Pennsylvania, Michigan, South Carolina, and Mississippi. This arc, which is made up mainly of heavily populated eastern states (and California), suggests a mathematical relationship between level of firearm ownership and murders committed per 100,000 people, logarithmic, I think.

Above this logarithmic curve are New Jersey, Maryland, Missouri, and Louisiana. Beneath this curve in the medium-to-heavy firearm level range are three tiers of states.

Right beneath the curve are New Mexico, Georgia, Virginia, North Carolina, Tennessee, Ohio, Texas, Oklahoma, Kentucky, Alabama, Arkansas, and Alaska. Most of these states are southern with considerable economic activity.

Below that in the second tier are Indiana, Kansas, Washington State, Nebraska, Wisconsin, and West Virginia. Mostly midwestern states.

On the bottom tier are Colorado, Maine, Minnesota, Oregon, Utah, North Dakota, South Dakota, Idaho, Montana, Wyoming, Iowa, New Hampshire, and Vermont. Mostly a mixture of midwestern and western states, many of them sparsely populated.

I have an hypothesis why this is so. When I divide numbers of murders committed by population, I’m assuming that all other factors being equal, number of murders committed varies linearly with population. Double the population, and you double the number of murders. But perhaps that is not so. Supposing, all other factors being equal, murders varied with the square of population, that is, if you double the population, the number of murders rises four-fold. That might explain why sparsely populated states have much lower rates of murder by firearm.

There may be other factors as well which should be investigated. Perhaps states with lower-than-expected murder rates:

  • Have lower gang or organized crime activity. The presence of gangs or other forms of organized crime can push up the murder rate significantly.
  • Have a culture of respect for guns and have in place strong social inhibitions regarding the abuse of guns. It is understood that guns as weapons are to be used only in self defense, never to settle an argument.
  • Have less residents who are members of dysfunctional subcultures, such as the subculture of many inner cities, that glorify or at least tolerate violence.

Of course, my analysis is wholly inadequate to base public policy on. States are large, heterogeneous places. I really should take a look at a few states and analyze them county by county. Does the Texas panhandle have the same characteristics as Houston or Dallas? Does central Florida have the same characteristics as Miami or Orlando? Are all the counties in Iowa the same or do they have different murder rates?

Nevertheless, this simple analysis does lead me to some conclusions. First, that restricting gun ownership may not be necessary to address the problem of gun violence. Also, that it may not be fair to demand one Federal gun policy for all 50 states. Why lay restrictions on Wyoming with its low murder rate? Wyoming doesn’t have a gun problem (at least with regards to crime). Neither does New Hampshire or Iowa. New Jersey has a gun problem. So does Maryland, Missouri, and Mississippi. Louisiana and the District of Columbia have a really big gun problems, and that’s where we should be focusing our energies.

One thing we should demand from all 50 states is their cooperation in preventing the purchase of firearms in their states with the intention of committing crimes in other states. But other than that, I would leave low-gun-crime states like Wyoming alone.

Where we need much tougher gun policy is in states and areas that have higher rates of murder by firearm. I would want to first concentrate on those states that form the logarithmic curve that I mentioned above. I would then want to spend special attention on those four states that have even higher rates: New Jersey, Maryland, Missouri, and Louisiana. Less urgent are those states in the first two tiers below the curve.

As I mentioned in a previous post2, to effectively target gun violence we need to target the most important sources of gun violence: gangs, robbery, and especially arguments, the cause of 40% of all gun homicides in the U.S.3 We need to gain the upper hand on gangs, particularly juvenile gangs. We should concentrate on the most likely weapons to be used in crimes: handguns, and not shotguns, rifles, and the so-called assault weapons which are used in the most sensational crimes but are involved in only a small minority of homicides3. By directing our efforts wisely, we should be able to reduce the level of gun violence in this country.


  1. New York Times, “Share of Homes With Guns Shows 4-Decade Decline” by Sabrina Tavernise and Robert Gebeloff, March 9, 2013 (click here to read). The article references two surveys done on gun ownership in the U.S. The first, the General Social Survey (GSS) has been conducted by the National Opinion Research Center (NORC) every year since 1973. It shows a general decline in gun ownership as a percentage of American households (click here to read). The second survey is by the Gallup organization. The New York Times article claims that Gallup found less of a decline in gun ownership, but when I went to the Gallup site, I found that their report claimed that gun ownership is at an all-time high (click here to read). Even according to the GSS, the decline in gun ownership shows little change from 2000 to 2006, (34.3% to 34.5%) so my figures should not be that far off and my conclusions should hold.
  2. “Background Check Bill Goes Down in Defeat”. Click here to read.
  3. United States Census Bureau website, The 2012 Statistical Abstract, Law Enforcement, Courts, & Prisons: Crimes and Crime Rates, specifically “Murder Victims–Circumstances and Weapons Used or Cause of Death”. To view, click here, then go to option 310. You can choose having the data displayed as a PDF file or on an Excel spreadsheet (although OpenOffice Calc works just as well).

2 responses to “Topic: Gun Homicides vs. Gun Ownership

  1. I don’t know of any major figure arguing that there is a correlation between population size and violence, or rate of ownership. Michael Moore already demonstrated that Canada has a higher per capita rate of gun ownership than the US with a significantly lower homicide rate.

    There are simply too many factors, such as poverty rates and urban crime. Proximity to Mexico is probably also a factor. That is why there should be a basic federal policy with final discretion left to individual states. The biggest mistake was allowing Clinton’s assault weapons ban to expire in 2004, for which both Bush and Obama are to blame.

  2. Pingback: Welcome to the Environmental Analyst | The Environmental Analyst

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s