Do Concealed Guns Reduce Crime?

Mar. 1 -- More Guns, Less Crime sounds like an oxymoron, but it is the title of aprovocative book by economist John R. Lott.

The thesis of Lott's book and of the nationwide crusade associated with itis encapsulated in its title: Lott maintains that counties in the UnitedStates that have enacted laws freely allowing for the carrying of concealedhandguns have seen a decrease in confrontational crimes such as murder,assault, robbery, and rape.

Gun as Defensive Tool

Now senior research scholar at Yale's School of Law, Lott has received aninordinate number of kudos and brickbats for his work, which was firstpublished in 1998. He has been called everything from a tool of the gunlobby to a courageous challenger of political correctness.

These are odd ways to refer to someone whose book and papers are full ofarcane statistics and multiple regressions.

The last term is important. A multiple regression is a study of the linear relationship between adependent variable (in this case the crime rate) and a collection ofindependent variables (in this case many factors, including the concealedgun laws, that might affect the crime rate). It attempts to estimate howmuch each of the independent variables affects the dependent variable andhow sure we can be of each of these effects.

The size of the effects is often expressed in terms of so-called regressioncoefficients and our confidence in them involves various other common statistics. Thus many of Lott'scontroversial results (in the book and in his paper on the same subject with DavidMustard) take the dry form of statements about coefficients and confidenceintervals.

If Lott's thesis is correct, regression coefficients relatingconfrontational crime rates to the passage of laws that require officials toissue concealed weapon permits are negative. That is, more guns, less crime. The values ofthese coefficients are also statistically significant, not likely to haveoccurred by chance.