New Scientist reports:
A survey of 2400 drivers carried out by David Hemenway and his colleagues at the Harvard School of Public Health shows that motorists who carry guns in their cars are far more likely to indulge in road rage – driving aggressively or making obscene gestures – than motorists without guns. Some 23 per cent of gun-toting drivers admitted making rude signs, compared with 16 per cent of those who did not carry guns (Accident Analysis and Prevention, DOI:10.1016/j.aap.2005.12.014).
Not surprisingly, Lott is criticizing the study:
While one regression with a few very basic variables was apparently run (but not shown), no explanation was offered for why such a limited set of control variables were used (e.g., why not trouble with law enforcement, education, income, smoker, race). Trouble with law enforcement (past arrests) would have been obvious
Actually, table 1, which fills an entire page of the paper presents a multivariate regression using all 15 variables including trouble with law enforcement, education, income, smoker and race. And even if he skipped over that page, the results of the multivariate regression are mentioned several times in the text.