This paper provides an empirical test of the rational addiction model, used in economics to model individuals' consumption of addictive substances, versus the utility misprediction model, used in psychology to explain the discrepancy between people's decision and their subsequent experiences. By exploiting a unique data set of disadvantaged Australians, we provide longitudinal evidence that a drop in life satisfaction tends to precede the use of illegal/street drugs. We also find that the abuse of alcohol, the daily use of cannabis and the weekly use of illegal/street drugs in the past 6 months relate to lower current levels of life satisfaction. This provides empirical support for the utility misprediction model. Further, we find that the decrease in life satisfaction following the consumption of illegal/street drugs persists 6 months to a year after use. In contrast, the consumption of cigarettes is unrelated to life satisfaction in the close past or the near future. Our results, though only illustrative, suggest that measures of individual's subjective wellbeing should be examined together with data on revealed preferences when testing models of rational decision-making.