Jean T Shope1, C Raymond Bingham. 1. University of Michigan Transportation Research Institute, Ann Arbor 48109-2150, USA. jshope@umich.edu
Abstract
OBJECTIVE: This study replicated and extended a conceptual model of drinking-driving and its relationship to other problem behavior to determine (1) if the model characterized the problem driving and problem behavior of young adults from the general population and (2) if the model applied to women as well as to men. METHOD: Telephone survey data (measures of drinking-driving, drug-driving, risky driving, problem drinking, drug use and delinquent behavior) collected from young adults (N = 4,230, 53% female) were used in structural equation modeling. Two models were developed--one for problem driving and one for problem behavior--each testing the structural associations among latent variables and testing them as latent indicators of common second-order latent variables. RESULTS: The results supported the hypothesis that drinking-driving (along with problem drinking, drug use and delinquent behavior) is an indicator of problem behavior in the general population sample. Also, drinking-driving, drug-driving and risky driving were demonstrated to be indicators of a common construct: problem driving. The models fit the data equally well for both young men and young women, and a crossvalidation conducted on an independent sample confirmed the results. CONCLUSIONS: The results provide empirical support for the inclusion of drinking-driving and problem driving in an extended model of Problem Behavior Theory. Drinking-driving, drug-driving and risky driving all relate to a common construct: problem driving. These results have important implications for research and for interventions to treat or prevent drinking-driving, especially among young adults.
OBJECTIVE: This study replicated and extended a conceptual model of drinking-driving and its relationship to other problem behavior to determine (1) if the model characterized the problem driving and problem behavior of young adults from the general population and (2) if the model applied to women as well as to men. METHOD: Telephone survey data (measures of drinking-driving, drug-driving, risky driving, problem drinking, drug use and delinquent behavior) collected from young adults (N = 4,230, 53% female) were used in structural equation modeling. Two models were developed--one for problem driving and one for problem behavior--each testing the structural associations among latent variables and testing them as latent indicators of common second-order latent variables. RESULTS: The results supported the hypothesis that drinking-driving (along with problem drinking, drug use and delinquent behavior) is an indicator of problem behavior in the general population sample. Also, drinking-driving, drug-driving and risky driving were demonstrated to be indicators of a common construct: problem driving. The models fit the data equally well for both young men and young women, and a crossvalidation conducted on an independent sample confirmed the results. CONCLUSIONS: The results provide empirical support for the inclusion of drinking-driving and problem driving in an extended model of Problem Behavior Theory. Drinking-driving, drug-driving and risky driving all relate to a common construct: problem driving. These results have important implications for research and for interventions to treat or prevent drinking-driving, especially among young adults.
Authors: Marilyn S Sommers; Michael S Lyons; Jamison D Fargo; Benjamin D Sommers; Catherine C McDonald; Jean T Shope; Michael F Fleming Journal: Alcohol Clin Exp Res Date: 2013-06-26 Impact factor: 3.455