Differences Across Majors in the Desire to Obtain a License to Carry a Concealed Handgun on Campus: Implications for Criminal Justice Education

Abstract

The current study examines whether students' course of study is related to their desire to carry a concealed handgun on campus. This analysis is motivated by a growing number of studies that have explored differences between criminal justice (CJ) and other majors in the areas of personality, attitudes, values, and career expectations, and also by discussions of the issue of concealed handguns on campuses following recent shootings on college campuses. Using data from over 3,100 students who completed an online survey instrument, results revealed that CJ majors were in fact more interested in carrying a legally concealed handgun on campus, if it were allowed by university policies. These results held even when controlling for other significant predictors of the desire to carry a concealed handgun. Implications for CJ education and the wider debate about concealed weapons on university campuses are discussed.

Notes

1. The online survey was configured to permit only one response per individual invitation. Thus, the variable distributions should not be skewed disproportionately by some individuals appearing multiple times in the sample dataset.

2. During the survey questionnaire administration, no additional explanation of the context or parameters of this issue were provided to participants. For example, students completing the survey were not briefed on specifics regarding CHL requirements or given any statistics describing current permitting trends. The authors view this approach as the most authentic way to gauge support in the student population without introducing new potential sources of bias.

3. Results reported here involved only one multivariate modeling technique, but other models were attempted as well. OLS regression models were estimated using the original dependent variables, although model fit was extremely poor given the non-normal distributions. Several alternative recoding schemes were employed to classify the dependent measures as dichotomous (e.g. 0% = 0, others = 1). Ultimately this approach was abandoned because it resulted in the loss of variability with too few categories. Also, multinomial logit models were estimated using the recoded dependents. In all cases, substantive conclusions regarding the main effects were similar, thus we consider multinomial probit results to be representative of the robustness of the findings regardless of which analytical approach was selected.

aTwo thousand five hundred and fifty-two cases represents the total number of cases for which all variables included in these models had non-missing values.

4. A post hoc statistical weighting procedure was employed with each university’s dataset in order to examine whether any of the multivariate results presented thus far differed when samples were adjusted so that they more closely represented the population characteristics (i.e. percent of CJ majors) from which they were drawn. These weighted-sample multivariate models revealed only minor changes in the parameter estimates, with no substantive differences in the effect of being a CJ major on the desire to obtain a CHL and carry a concealed handgun on campus. For example, the significant coefficient for CJ majors in the unweighted model (in Texas) predicting being in the 1–49% likelihood range for obtain a CHL and carry was 0.71. In the same model, when appropriate weighting was applied, this coefficient remained significant and increased to 0.73. In addition, the models computed with a control for campus location (WA vs. TX) revealed the same substantive conclusions, with the additional result that TX students report high levels of interest in obtaining a CHL (at both the 50-99 and 100% levels). Given that none of these alternative modeling strategies changed the general conclusions presented here, only the original results are reported.