Abstract
Background: Young drivers represent a disproportionate number of the individuals involved in alcohol-impaired driving. Although there is a known association between drinking and alcohol-impaired driving in young drivers, the link between early onset drinking and early onset alcohol-impaired driving has not been explored. Objectives: The present study aimed to assess this link along with potentially confounding factors. Methods: The assessment used a proportional hazards model with data collected from the Buffalo Longitudinal Study of Young Men, a population-based sample of 625 males at aged 16–19. Results: Controlling for the effects of potentially relevant confounds, the early onset of drinking was the most influential factor in predicting the early onset of alcohol-impaired driving. Race and the early onset of other forms of delinquency also played a significant role in the early onset of alcohol-impaired driving. Conclusion: Preventing an early start of drinking among adolescents may be the most critical factor to address in preventing an early start of alcohol-impaired driving.
Acknowledgements
This research was supported by Grant R01AA016161 from the National Institute on Alcohol Abuse and Alcoholism.
Notes
1The study was approved by the Institutional Review Board from the Research Foundation, the State University of New York for the protection of human subjects.
2The sample attrition was 4.6% for the second wave and 7.6% for the third wave. The small attrition did not affect the nature and characteristics of the sample as we assessed (Citation33). Also, the attrition was not likely to have a significant effect on our current analysis because the overall measures of age onset of alcohol-impaired driving, age onset of drinking, and age onset of other forms of delinquency counted any respondents who reported or did not report any age onsets of the behaviors in any of the waves.
3A total of 840 eligible households were recruited. Of the 840, 625 were interviewed and 215 were randomized out based on screening, availability, and agreement. As a result, the response rate was 74% for both young males and their parents, since the survey interviewed 625 complete pairs. Although the sample might not represent the general population of adolescents in the area because of oversampling of those who had higher risk backgrounds, it provided sufficient data in delinquency for analysis.
4The 33 delinquency items were adopted from National Youth Survey developed and conducted by Elliott and his colleagues (1985) (Citation35).
5The results from our correlation analysis indicated the reliability and stability of the survey items across waves which provide a foundation for the overall measures across waves (see Appendix 2).
6The Cox Regression procedure is the proportional hazards model which estimates the probability of the event to occur. The negative sign of the coefficient in our Cox Regression analysis is in the predicted direction of the relationship between early onset of drinking and early onset of alcohol-impaired driving. Early onset of drinking was measured in age. The negative relationship indicated that respondents who started drinking at lower age were likely to have a higher probability of an early start of alcohol- impaired driving.
7There was minimal multicollinearity among the independent variables as examined. The highest correlation was 0.35 between age onset of drinking and age onset of delinquency.
8Also, New York State does not have a use/lose law for drivers under age 21. The state has a junior license for those under 18. During the day from 05:00–21:00 h, the junior license holder must be accompanied by an adult driver, and at night they must be accompanied by a parent.