real-world social situations. Many studies suggest that adolescents, as well as adults, may make less than optimal decisions when personal goals, beliefs, prior experience, values, social expectations, and emotions are added to the decision making equation (Jacobs and Klaczynski, 2002). This outcome is especially true for social decisions (like choosing whether to drink or how much to drink). This is so for a variety of reasons.

First, outcomes of decisions in social situations are probabilistic, meaning that negative consequences of bad decisions may not occur and may not even be highly likely, although they are devastating if they do occur. For example, while the probability of having a car crash after drinking is much higher than after not drinking, drinking and driving does not always end in a crash or a ticket. Because outcomes are probabilistic, adolescents may interpret the fact that they previously drank too much and drove home without a crash as evidence that they can drink and drive safely (Jacobs and Ganzel, 1994). In one study, older adolescents who had a lot of experience drinking and driving, but had not experienced a negative outcome, such as a traffic citation or crash, believed that they were in little danger of having an accident after drinking (Finken et al., 1998), this result suggests that engaging in risky behaviors without consequence may have caused them to lower their perceptions of the risks of drinking and driving. Other correlational studies have shown that greater involvement in risk-taking behaviors was related to lower perceptions of personal risk (e.g., Halpern-Felsher and Cauffman, 2001; Goldberg et al., 2002). Second, the norms for social decisions are not typically known. Instead, individuals are often forced to make judgments on the basis of their own estimates of the norms of social behaviors or attitudes. This general dilemma, faced by people of all ages, is even more difficult for adolescents because they must make decisions based on a limited amount of experience and little feedback from earlier decisions (Jacobs et al., 1995). Several studies indicate that most adolescents overestimate the number of others who drink alcohol (e.g., Basch et al., 1989; Jaccard and Turrisi, 1987). Not surprisingly, the overestimation is greatest for those individuals who drink. This same pattern has been found for other risk-taking and deviant behaviors (e.g., Benthin et al., 1993; Nucci et al., 1991), and it may be related to the fact that those who drink have friends who drink and so they begin to believe that everyone is drinking. In one longitudinal study, adolescents who spent time with peers who encouraged drinking later reported more positive views of drinkers (Blanton et al., 1997).

In addition, studies indicate that adolescents make more biased estimates when they are reasoning about populations with greater variability and when they are reasoning about unfamiliar others (Jacobs, 2004) Underage drinking and other forms of risk taking are likely to occur in social situations and when adolescents find themselves with large groups of unfa-