No doubt you've come across this awful story of a little girl who took a handgun out of her grandmother's purse during a shopping trip and shot herself. Our friend Kathy thought I might want to take this occasion to discuss the proliferation of state laws allowing people to carry concealed handguns. When I decided to take her up on it, I didn't know what I was getting into.

This is certainly not the only anecdote about a legally carried concealed weapon that ended up doing something it shouldn't oughteradone. Kathy even sent in a few. Seattle's mayor wants to ban guns on city property after a shooting at a folklife festival, but under Washington law, he may not be able to. Here's a guy who shot his friend in the back. And so on. I believe I told you earlier about the guy whose gun fell out of his shirt pocket in the Planet Hollywood in Indianapolis and put a bullet through two American Public Health Association conference goers. The Chief of Police went on TV to say there would be no charges filed because "he had a puhmit fo the weapon." On the other hand, we don't seem to hear many stories about people fending off muggers and rapists with their concealed handguns. That's not to say it never happens but, the question is, what is the net effect on public safety of letting people go around with concealed hand guns?

The NRA aggressively pushes so-called "shall issue" laws, which require authorities to give a permit to carry a concealed firearm to anybody who applies, who isn't disqualified by a criminal record or mental illness. From 1979 to 1998, 23 states adopted such laws. But the effect these laws have had turns out to be a major point of controversy, which is difficult for me to untangle for you hear.

In 1997, JR Lott and DB Mustard published a study entitled "Crime, deterrence and right-to-carry concealed handguns." (Mr. Mustard, in the library, with a pistol?) They compared crime statistics before and after states passed such laws, and ran various regression models See my explanation here) to try to isolate the effect of the laws. Two of their 21 models found significant effects. They did find reductions in rape, aggravated assault, and homicide overall, but they found increases in larceny and auto theft, and they did not find significant effects on rates of robberies or the percentage of homicides that involved strangers.

Daniel Webster and colleagues published a detailed critique of these findings in American Journal of Public Health later in that year. Unfortunately, the arguments for the most part are technically complex and almost certain to be opaque to policy makers and to you, dear reader, so we have been left with a scientific controversy, rather like those which long surrounded tobacco, lead paint, and global warming, in which people's prior beliefs and commitments generally determine which expert they believe. In 2005, Rosengart et al did a new analysis using similar data sources and came to essentially the opposite conclusion: "When a 'shall issue' law was present, the rate of firearm homicides was greater." They didn't look at rape or assault. However, this may be a good time to point out that one of the reasons Lott and Mustard's findings seem odd is that contrary to popular misconceptions, rapes committed by strangers accosting people in public places are quite rare. There is no logical reason to expect concealed carry laws to have any effect on the typical kinds of sexual assaults, which involve acquaintances or family members.

There are various technical reasons why Lott and Mustard's analyses are flawed, all of which are corrected in the analysis by Rosengart, et al. Therefore, I am personally convinced that these laws are harmful and that only law enforcement officers should be carrying handguns around in public. However, ideological commitments are so strong that people who support such laws will not believe me. If I take the time to try to explain the reasons for this conclusion, as better qualified people than I have done, it will just add to the he said/she said din and do no good whatsoever in advancing the cause of truth. (I will do it anyway if there is popular demand, just for the heck of it, but it will be boring.)

This is extraordinarily frustrating, to say the least. When scientific findings get caught up in public policy controversies, views held by only a small minority of specialists get equal weight, so long as there are powerful interests to promote them. That does not mean, of course, that minority views in scientific controversies are necessarily wrong, but ceteris paribus, policymakers should normally treat them as unlikely. That may sound like a weak statement, but it's the strongest one I can make.