Rejoinder to Tomislav V. Kovandzic's Comment on Our Article 'The Effectiveness of Gun Control Laws: Multivariate Statistical Analysis.' (Response to T.V. Kovandzic in This Issue, P. 363)
Kwon, Ik-Whan, Scott, Bradley, Safranski, Scott R., BAe, Muen, The American Journal of Economics and Sociology
We reviewed Tomislav v. Kovandzic's comment article on our article (Kwon et al., 1997) with great interest and at the same time with a great deal of disappointment. We did not benefit from his comments. Kovandzic appears to have missed the main point of our study. The major finding of our study was that "The gun control laws have a very mild effect [italics added] on the number of gun related deaths while socioeconomic variables such as a state's poverty level, unemployment rate, and alcohol consumption, have significant impact on firearm related deaths" (Kwon et al. 1997, p. 41). We further emphasized the importance of socioeconomic factors in our study: "Unless this country directs its efforts toward the socioeconomic ills which appear to bear the strongest relationship to violent deaths by firearms, the fatalities likely will remain high whether this country has gun control laws or not [italics added]" (p. 48-49). Once he missed the major point of our paper, it was relatively easy for Kovandzic to pick out other areas that in our view are inconsequential or unrelated to the main issue.
However, Kovandzic raised an important issue. He maintains that indexing states into two dichotomized groups (states with no restriction on gun purchase and states with some type of regulations) created a serious measurement bias. We not only concur with his concern but also clearly addressed the concern in our study, as we wrote, "The main reason for a difference in findings between this study and the earlier study rest with the characteristics of the laws themselves (between this country and Canada). The Canadian law is a federal regulation governing the use of firearms whereas the laws and regulations used in this study are state laws which vary widely between. Accordingly, the results from this study may not be as clear and strong as the results found by the Mauser and Holmes' study" (Kwon et al. 1997, p. 47). Apparently, Kovandzic missed this statement completely.
Researchers in this field agree that there is a wide variety of research methods involving gun control issues and, therefore, the findings do vary. The study by Kleck and Patterson (1993) that Kovandzic repeatedly cited as his major (and only) source of comments suffers from shortcomings as does any study, including ours. The Kleck-Patterson study used 170 U.S. cities with a 1980 population of at least 100,000. The cities were coded for the presence of nineteen major categories of firearm restrictions. The results indicate that, in general, gun control restrictions have no net effect on the violence rate. However, the same study also mentioned a few exceptions, "of 108 assessments of effects of different gun laws [italics added] on different types of violence, 7 indicated good support (of relationship), and another 11 partial support, for the hypothesis of gun control efficacy" (Kleck and Patterson 1993, p. 249). Is Kovandzic prepared to say that 170 cities have identical gun control regulations?
Even if we accept his hypothetical argument that the 170 cities have a uniform code of regulations, the KP clearly indicates the merit of gun control efficacy in certain areas. For example, the KP study admitted that "... of 108 assessments of effects of different gun laws on different types of violence, 7 indicated good support (of the efficacy of gun control laws), and another 11 partial support for the hypothesis of gun control efficacy" (p. 249). The proceeding statement appears to be in line with our findings.
Kovandzic's blanket statement that there is no relationship between gun control regulation and fatalities not only contradicts the findings of the Kleck-Patterson study but more importantly misleads his readers. The Kleck-Patterson study focuses on city-level data, while our study emphasizes state level data. The results could be different because of the sample design. It is also interesting to point out that a similar study by Kleck and Gertz (1995) that found gun ownership saved lives by preventive use was challenged by Morganstein (1997) who stated that "the [survey] respondent's perceptions of death averted by the defense use of guns may be severely exaggerated [due to flaw of research method]" (p. …