This paper examines the distributions of past-year volume of ethanol intake, frequency of drinking 5+ drinks and alcohol dependence in a representative sample of 18,352 U.S. current drinkers aged 18 years or over. Within categories defined by these three partially overlapping domains, it presents rates of self-perceived impaired driving, i.e. driving after having had too much to drink, in the year preceding interview. High volume drinkers, those with an average daily ethanol intake of 1 ounce or more, composed 19.7% of current drinkers and accounted for 66.5% of all reported ethanol consumption, 72.6% of all heavy drinking days, 49.2% of all alcohol dependence and 62.8% of all impaired driving incidents. Frequent heavy drinkers, those who drank 5+ drinks at least once a week, composed 12.3% of current drinkers and accounted for 42.9% of all reported ethanol consumption, 81.9% of all heavy drinking days, 40% of all alcohol dependence and 57% of all impaired driving incidents. Drinkers with DSM-IV alcohol dependence composed 9.9% of current drinkers and accounted for 28.9% of all reported ethanol consumption, 37% of all heavy drinking days and 56.9% of all impaired driving incidents. The overlap of these three high risk groups, each of which had a probability of at least one impaired driving incident per year, was far from complete. Of individuals who met any of these criteria for high risk drinking (i.e. high volume, frequent heavy drinking or dependence), more than half met only one criterion and only one in seven met all three. The group that did meet all three criteria had such a high rate of impaired driving incidents, an average of 5.14/year, that it accounted for 36.4% of all such incidents despite making up only 3.8% of all current drinkers. The results are discussed in terms of their implications for targeting prevention and intervention efforts, e.g. whether targeting one problematic aspect of drinking behavior will reach drinkers with other types of problem behaviors as well.