The major focus of research on alcohol is not on the majority who drink without problems, but on the small minority who have extreme problems. Difficulty in conceiving, measuring, and analyzing non-problem drinking lies in the exclusively problem-drinking orientation of most drinking measures. Drawing on conventionally used scales (e.g. Short Michigan Alcoholism Screening Test) and other established concepts in the alcohol literature (e.g. craving, hangover), a set of 24 items was selected to classify all persons in a sample from Tecumseh, Michigan, as to their alcohol-related behaviors (N = 1266). A Sensible-Problem Drinking Classification (SPDC) was developed with five categories: very sensible, sensible, borderline, problem, and impaired. A variety of known alcohol and psychosocial variables were related monotonically across these categories in expected directions. Ethanol ounces per week was only modestly related to SPDC groups: R2 = 0.09 for women, R2 = 0.21 for men. The positive relationship of problem and non-problem SPDC groups to high and low blood pressure was P = 0.07, while ethanol (oz/week) was uncorrelated to blood pressure (mm Hg) in this subsample (N = 453). The development of SPDC requires additional items measuring self and group regulatory alcohol behavior. However, this initial analysis of no-problem subgroups has direct import for public health regulation of alcohol use by providing a model of a sensible view of alcohol use.