The latest pathetic instance
of anti-gun scholarship, found in the comments of a Tim Blair post
The Effects of Gun Prevalence on Burglary: Deterrence vs Inducement
Philip Cook and Jens Ludwig
The proposition that widespread gun ownership serves as a deterrent to residential burglary is widely touted by advocates, but the evidence is weak, consisting of anecdotes, interviews with burglars, casual comparisons with other countries, and the like. A more systematic exploration requires data on local rates of gun ownership and of residential burglary, and such data have only recently become available. In this paper we exploit a new well-validated proxy for local gun-ownership prevalence -- the proportion of suicides that involve firearms -- together with newly available geo-coded data from the National Crime Victimization Survey, to produce the first systematic estimates of the net effects of gun prevalence on residential burglary patterns.
The proportion of suicides that involves firearms is a proxy for gun ownership? With that sort of attitude toward statistics, you could prove anything you wanted to. As another commenter pointed out,
I remember that study (or one similar to it) in which the suicide proxy was used. Of course, it created widely erratic swings because certain areas where gun ownership is hugely common have such a low incidence of suicide (and crime in general) as to make the study worthless.
Also note that gun suicides are not always reliably reported. They are sometimes hushed up as "accidents."