OK, I WILL Comment on this “Study”.
Ben from Carnaby Fudge sent me a copy of the Social Science & Medicine report entitled “State-level homicide victimization rates in the US in relation to survey measures of household firearm ownership, 2001-2003” authored by Matthew Miller, David Hemenway, and Deborah Azrael of the Harvard School of Public Health. I’ve perused it. It is your typical statistician’s wet-dream with phrases like “multivariate analysis” and “negative binomial regression models.” I am not a statistician. I admit up front that I don’t follow most of the discussion in the nine-page report, so this is my layman’s understanding of the report’s findings and my commentary on it.
Studies like this tend to do a combination of things. One, they state the blindingly obvious, e.g.:
Consistent with previous work, we found that homicide rates were higher in areas with higher rates of urbanization and resource deprivation; like others we also find that homicide rates are higher in the South.
Two, they draw conclusions, but don’t admit to actually drawing those conclusions:
Our study does not establish a causal relationship between guns and homicide. It is possible that a non-causal relationship explains our findings or that the association we observe might have arisen because individuals in states with historically high homicide rates acquired more guns (than did individuals in low-homicide states), as a defensive response to actual high homicide rates in their communities (i.e. “reverse causation”).
Have to cover all the bases, you know. But this won’t be mentioned in any reports in the press. Third, they tend to not mention anything that doesn’t reinforce the message being pushed: “More Guns = More Gun Crime.” In this case, homicide.
I’ve already had a go-around with Dr. Hemenway, but let me see if I can put this report in a little context. The study covers homicide from 2001-2003. Here’s a DOJ chart showing nationwide homicide rates from 1900-2002:
Click on the image for a link to the source data. Note that after 1994, homicide dropped precipitously. The data shows that in 2002 – the middle of the study period – the national homicide rate was 6.1/100,000 population. That is roughly the same rate we had in 1966, 1947, 1940, and 1913. The fact of the matter is homicide rates vary widely with time. However, the number of guns in circulation over time does one thing and one thing only: It increases.
But the argument put forth by this paper is that it is the level of household firearm ownership that is the critical correlation factor, and according to this report “approximately one in three US household contained firearms”. But if household ownership was the critical factor in homicide rates, then why the tremendous swings from 1900 to the present? And why has the homicide rate in the U.S. declined from 1994 until just last year? Surely Dr. Hemenway et al. don’t expect us to believe that the number of households containing firearms nationwide has decreased each and every year in the past decade? The report states:
Case-control studies suggest that the presence of a gun in the home is a risk factor for homicide in the home, that the risk is higher for women than for men, and that when any family member purchases a handgun all members of the household are at increased risk of homicide victimization.
During the past decade we’ve added a minimum of 30 million new firearms in public hands – at least 10 million of which were handguns. Since 1993 we’ve gone from 21 states with “shall-issue” or unrestricted concealed-carry legislation to 39. We’ve had an influx of “assault weapons” and “pocket rockets” – supposed engines of death and destruction far more lethal than the weapons available in the 60’s.
Yet homicides declined. Non-fatal firearm related crime declined.
But we’re supposed to believe that if you or someone in your household buys a gun, it is somehow the overriding risk factor in the probability of your becoming a victim of homicide. The report doesn’t say that, but that is most certainly how it is being reported:
Never mind the fact that if you’re not a young black male living in an inner city, your likelihood of dying by homicide (regardless of weapon) is about equal to that of someone living in Europe.
Nope. A gun in the home is the thing to be feared!
Do you understand why this kind of thing pisses me off?
Edited to add: And do I even have to point out that the “study” makes no distinction between criminal murders and justifiable homicides? As the Albuquerque Tribune recently noted, last year 10% of reported homicides there were of a defensive nature. Three were shootings of home intruders – with guns kept in the home one would assume.