The causal pathway from vulnerability to drug use and addiction involves a complex interaction between genetic, environmental, and behavioral factors. An individual can intervene on this causal pathway by two major types of individual decision. There is the inaugural, momentous decision to use a drug for the first time. This decision is influenced by both prior knowledge on the drug and its expected effects, and also by prior self-knowledge on one's own vulnerability. After an individual has used a drug for the first time, there is the decision to repeat drug use. This decision is influenced by the same factors that were involved in the inaugural decision to initiate drug use, except for one crucial difference. The first drug use has now acted on the individual, changing its brain acutely and also potentially persistently in a way that could bias subsequent decision-making in favor of repeated drug use. The goal of this review article is to assess the contributions and limitations of rodent models (i.e., rats, mice) to understand how prior drug use can influence decision-making in a way that favors future drug use. Overall, research on rodents shows that prior drug use can increase impulsive, risky and/or potentially harmful decision-making. However, this does not apparently translate into more drug use when rodents have the choice between a drug and a competing, nondrug option, except when the expected value of the latter is considerably decreased. The delayed drug reward hypothesis is developed to resolve and explain this apparent discrepancy. This novel hypothesis makes several unique predictions, some of them counterintuitive, and suggests that extrapolation of rodent research to humans should not only take into account differences in drug choice situations but also inherent species-specific differences in individual decision-making.