Data from 42 male and 58 female subjects who participated in the Colorado Alcohol Research on Twins and Adoptees (CARTA) project were subjected to model‐fitting analyses. The aim of the present study was to use linear structural equation models to determine whether differences in previously measured psychomotor sensitivity to alcohol predict differences in self‐reported alcohol consumption over a 4‐year period. LISREL model‐fitting results indicate that, for male subjects, only rail walking insensitivity is predictive of alcohol use reported 2 years after their initial CARTA testing. For females, only hand steadiness sensitivity is predictive of alcohol use reported 2 years after their initial CARTA testing. The results for males support a hypothesis that would consider alcohol insensitive individuals at greater risk for alcohol abuse. The female results, however, would argue against such a hypothesis.With only one measure of sensitivity predicting alcohol use at only one out of four time points, in both men and women, the overall results suggest that our three measures of psychomotor sensitivity to alcohol are, in general, poor predictors of alcohol consumption in this sample.