Abstract
Conservative political orientation is a strong predictor of gun ownership in the United States. We explore the extent to which this relationship is mediated by two related belief systems: Christian nationalism and Right-Wing conspiratorial thinking. Drawing on nationally representative data from the sixth wave of the Baylor Religion Survey (N = 1,248), we use logistic regression and the Karlson-Holm-Breen method to analyze gun ownership, disaggregated by type of gun and reason for ownership. Christian nationalism and conspiratorial thinking underlie political effects on automatic and semi-automatic guns, handguns, and guns owned for protection, especially among non-Hispanic white respondents. Christian nationalism and conspiratorial thinking are less salient to driving political differences in long gun or recreational gun ownership. Findings elucidate the belief-based mechanisms underlying the societally important link between conservative politics and gun ownership, demonstrating how beliefs pattern who owns guns (and why) in the United States.
Notes
1 It is important to note that Christian nationalism is conceptually and empirically distinct from both Evangelical religious affiliation and having high religiosity (see Whitehead & Perry, Citation2020b).
2 It should be noted that recent research (Gorski & Perry, Citation2022) shows variation in the effects of Christian nationalism by race, arguing that “white Christian nationalism” is a better term for the strand of this belief system characterized by racial threat. To be sensitive to this distinction, we control for race and perceived racial threat in all models and conduct supplemental analyses including only non-Hispanic white respondents (Appendix C).
3 See Lizotte et al. (Citation1994) for more on the distinct drivers of gun ownership for protective and recreational reasons among adolescents.
4 Primary analyses were replicated using listwise deletion instead of multiple imputation (N = 917 for political affiliation models; N = 920 for political ideology models), with substantively similar results (available upon request).
5 Republicans and Democrats could describe themselves as “strong,” “moderate,” or “leaning.”
6 Conservatives and Liberals could also describe themselves as “leaning” or “extremely.”
7 The racial threat scale is highly skewed, with about 75% of respondents perceiving all these groups to be “not a threat.” Supplemental analysis that recoded this measure as binary (1 = some threat perceived from any of the groups) produced substantively identical results, available upon request.
8 See Kohler et al. (Citation2011) for details on the Stata program used to implement this method in our analysis.
9 However, see Appendices D and E for estimates from models that included the potential mediators separately. Implications of these sensitivity analyses are discussed following presentation of the primary results.