Objective: A low level of response (LR) to alcohol characterizes groups at high risk for alcoholism and predicts future heavier drinking and alcohol-related problems. Because LR might change with increasing age and years of drinking, there is a need to measure this phenotype as early as possible in the drinking career. Method: Data were generated from 1,106 12- to 13-year-old subjects in the Avon Longitudinal Study of Parents and Children (ALSPAC). Information about alcohol and other substance use and problems was obtained using a structured interview, and LR from the Self-Rating of the Effects of Alcohol (SRE) questionnaire. Results: Drinking was reported by 80 (7.3% of all subjects) boys and girls. Boys comprised 62.5% of the sample; the mean (SD) age of subjects was 12.9 (0.16) years. Among those 80 drinkers, alcohol had been consumed on an average of 5.4 occasions over the prior 6 months, the average maximum drinks ever imbibed was about 3.3, and 23.8% had any of 26 possible alcohol-related problems. In this group, 36.3% had smoked cigarettes, and 12.5% had used marijuana. The First 5 SRE score among drinkers correlated with the maximum number of drinks consumed at one time at .61 (p < .001), the number of alcohol problems at .25 (p < .01) and the frequency of drinking at .29 (p < .01). When considered along with smoking and marijuana histories, sex and weight, the SRE score contributed significantly to the prediction of maximum drinks per occasion (β = .60, p < .001) and drinking frequency (β = .24, p < .05), with a trend (β = .17, p = .08) for alcohol problems. Conclusion: These results support the conclusion that a low LR to alcohol correlates well with the maximum number of drinks consumed even early in the drinking career among individuals for whom acquired tolerance or other factors were unlikely to have explained the relationship.