Mar. 1, 2001 -- More Guns, Less Crime sounds like an oxymoron, but it is the title of a provocative book by economist John R. Lott.
The thesis of Lott's book and of the nationwide crusade associated with it is encapsulated in its title: Lott maintains that counties in the United States that have enacted laws freely allowing for the carrying of concealed handguns 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 an inordinate number of kudos and brickbats for his work, which was first published in 1998. He has been called everything from a tool of the gun lobby to a courageous challenger of political correctness.
These are odd ways to refer to someone whose book and papers are full of arcane statistics and multiple regressions.
The last term is important. A multiple regression is a study of the linear relationship between a dependent variable (in this case the crime rate) and a collection of independent variables (in this case many factors, including the concealed gun laws, that might affect the crime rate). It attempts to estimate how much each of the independent variables affects the dependent variable and how sure we can be of each of these effects.
The size of the effects is often expressed in terms of so-called regression coefficients and our confidence in them involves various other common statistics. Thus many of Lott's controversial results (in the book and in his paper on the same subject with David Mustard) take the dry form of statements about coefficients and confidence intervals.
If Lott's thesis is correct, regression coefficients relating confrontational crime rates to the passage of laws that require officials to issue concealed weapon permits are negative. That is, more guns, less crime. The values of these coefficients are also statistically significant, not likely to have occurred by chance.
Does Threat Convince Criminals?
I'll spare you the technicalities. Suffice it to say that Lott's formal calculations are not wrong in any blatant way. He has, however, been criticized on many other grounds. Researchers Dan Black and Daniel Nagin have, for example, found that Lott's results are less compelling if small counties under 100,000 people are not considered. The small counties often have no arrests for certain offenses and so the arrest rate for these offenses, one of the many other independent variables in Lott's study, is an unusable value of 0/0.
Others say that Lott's model does not adequately take into account crime trends or unusual situations. For example, the whole state of Florida should have been eliminated from consideration, some argue, because of special problems having to do with Castro's Marial boat lift of 1980 during the term of the study. Critics also maintain that Lott makes no distinction between the crimes of juveniles and adults, that concealed gun laws are not an all-or-non variable, that their effects should vary by state, and that more attention should be paid to the time they've been in effect.
Using data from the 3,054 counties in the U.S., Lott addresses most of these and other issues with some success in my opinion. But more fundamental questions remain.
The most basic is: What is the mechanism for this correlation between more concealed weapons and less confrontational crime? In one word: why?
His basic argument is that the increased cost to would-be perpetrators (i.e., the risk of being caught or shot) convinces them to refrain from murder, rape, robbery, and assault. Guns used defensively, or even the prospect of guns being used defensively by potential victims, is enough to scare some criminals into pursuing less violent careers.
Hot Chocolate Consumption and the Crime Rate
But people with permits for concealed guns are presently those whose work puts them at increased risk or else people who are prudent, but fearful. What would happen if concealed weapons became the norm? Just because a certain relationship (more guns, less crime) is linear and negative over some range of the independent variable doesn't mean it will remain so over a much bigger range.
And even if this negative linear relationship were to hold up with many more people carrying guns, do we really want to decrease the crime rate in this way? There is a considerable psychic cost to being a citizen in a nation of armed people warily navigating their way through their crime-free lives. Some version of the somber anxiety that one experiences going through airport luggage scanners would extend throughout one's whole life.
There remain also the usual problems associated with all correlations and regressions. Is the association, even if real enough over a limited range, a causal one or is it coincidental? Or might there be a factor, not yet identified, leading both to more concealed gun laws and lower crime rates? After all, more consumption of hot chocolate is also associated with less crime and both are brought about by cold weather.
The bottom line is that our understanding of crime rate fluctuations is still very murky. Why the sudden precipitous drop in the murder rate in New York City, for example — the economy, demographics, enforcement? Cultural factors, while hard to quantify, certainly play an important role.
Some countries like Japan have few guns and little violent crime, whereas others have few guns and a lot of crime. Similar variation holds in countries with a lot of guns (or should I say a Lott of guns). In Switzerland and Israel a very high percentage of citizens have guns and the crime rate is low, whereas in this country many people have guns and the crime rate is high (although down considerably from its peak).
Although the iconoclasm, statistical presentation, and simplicity of the book's sales pitch are appealing, I simply don't buy it.
Professor of mathematics at Temple University, John Allen Paulos is the author of several best-selling books, including Innumeracy and A Mathematician Reads the Newspaper. His Who’s Counting? column on ABCNEWS.com appears on the first day of every month.