The authors first address the methodological flaws in previous studies of this issue. They note that the current study's use of panel data provides advantages over more commonly used time-series or cross-sectional data, with the most important being the ability to enter proxy variables for omitted variables that cause crime rates to vary across time and space. Of the 189 cities with populations greater than 100,000 in 1990, 77 were in States that had enacted SI laws between 1980 and 2000. No evidence was found to support the thesis that the longer SI laws are in effect the greater their deterrent impact on violent crime; however, the impact of the SI laws on violent crime rates varied significantly across States. The dangers of estimating a single aggregated effect are particularly strong because of differences in permit fees and training requirements for a concealed handgun permit and where concealed handguns can be taken; publicity and news coverage associated with passage of the laws; and the number of persons in the adult population with concealed handgun permits. There are more States where the passage of SI laws has led to statistically significant increases in violent crime than there are States with statistically significant decreases. Neither the current study nor previous evaluations of SI laws have explicitly measured total rates of civilian gun carrying; consequently, conclusions about the net effect of civilian gun carrying on violent crime rates based on this body of research are not warranted. 2 tables, 11 notes, and 81 references