Liquor-by-the-drink and alcohol-related traffic crashes: a natural experiment using time-series analysis.

A quasi-experimental study was conducted to estimate the impact of liquor-by-the-drink (LBD) on alcohol-related traffic accidents in North Carolina counties. Time-series analysis for the period from January 1973 through December 1982 found LBD was associated with statistically significant increases of 16 to 24% in both the number of police-reported alcohol-related accidents and in single vehicle nighttime accidents among male drivers 21 years of age and older in counties implementing LBD. No change in alcohol-related accidents was found for non-LBD counties. Single vehicle nighttime accidents involving male drivers under 21 did not change for either the experimental or comparison groups.