Open Access
American Research Journal of Humanities and Social Sciences
ISSN (Online): 2378-7031
DOI: 10.46568/arjhss
Gun Ownership in the United States: Development and Validation of a Predictive Model
Abstract
Background
Objective: This study aims to develop and validate a predictivelogistic regression model of household ownership of firearm in the United States.
Methods: Data from the Behavioral Risk Factor Surveillance System2017 was used. The list of potential predictor variables is largely based on existing literature on factors associated with gun ownership.Stepwise logistic regression analysis was employed to build the model. The model was then tested using Kolmogorov-Smirnov (KS) statistic and Area under the ROC Curve (AUC) as metrics to measure if the model is a good fit.
Results: About 48.8% participants reported having any guns in their families. From stepwise logistic regression analysis, 12 variables out of 14 are selected in the final prediction model. Factors that affect the likelihood of gun ownership include income level, education, veteran status, marital status etc.The resulting model is promising, with 72% percent of accuracy according to the ROC and a KS of 0.35. Conclusion: A predictive model of gun ownership among U.S. households was developed and validated.
Objective: This study aims to develop and validate a predictivelogistic regression model of household ownership of firearm in the United States.
Methods: Data from the Behavioral Risk Factor Surveillance System2017 was used. The list of potential predictor variables is largely based on existing literature on factors associated with gun ownership.Stepwise logistic regression analysis was employed to build the model. The model was then tested using Kolmogorov-Smirnov (KS) statistic and Area under the ROC Curve (AUC) as metrics to measure if the model is a good fit.
Results: About 48.8% participants reported having any guns in their families. From stepwise logistic regression analysis, 12 variables out of 14 are selected in the final prediction model. Factors that affect the likelihood of gun ownership include income level, education, veteran status, marital status etc.The resulting model is promising, with 72% percent of accuracy according to the ROC and a KS of 0.35. Conclusion: A predictive model of gun ownership among U.S. households was developed and validated.
