Optimistic Predictions as Self-Fulfilling Prophecies

Optimism May Lead to (Almost) Self-Fulfilling Prophecies

When people have some control over an outcome, an optimistic prediction may encourage actions that help bring it about. That improvement can offset some of the costs of an inaccurate forecast. Studies that have manipulated people’s expectations about how they would perform on particular tasks have found that the induction of positive expectations can lead to significant improvements in performance (Armor & Taylor, 2000; Buehler & Griffin, 1996; Campbell & Fairey, 1985; Peake & Cervone, 1989; Sherman, Skov, Hervitz, & Stock, 1981). This appears to be true even for predictions that can, at the time they are made, be considered to be overly optimistic. In an early and influential set of studies (Sherman, 1980), people were found to predict that they would behave in more socially desirable ways than base-rate data indicated (thereby demonstrating unrealistic optimism); however, people who had initially made these “unrealistic” predictions were subsequently more likely to behave in a socially desirable manner than were people who had not been first asked to make predictions about their behavior. In these studies, then, optimistic errors in prediction could be considered “self-erasing.” Additional evidence of self-erasing errors has been obtained in studies of voting behavior, in which merely asking people if they intend to vote has been found to increase the likelihood that they will vote (Greenwald, Carnot, Beach, & Young, 1987) and of consumer decision making, in which a mere inquiry of purchasing intent increased the likelihood of purchasing even big-ticket items such as automobiles (Morwitz, Johnson, & Schmittlein, 1993).

An important caveat, however, is that optimistic predictions are rarely completely fulfilled: Although people who make optimistic predictions tend to attain more than they would have had they not made those predictions, they do not necessarily achieve the standards they initially set for themselves. A study by Buehler and Griffin (1996) illustrates this point. In this study, expectations about how long it would take to complete a particular assignment were manipulated by anchoring participants’ predictions around relatively optimistic or pessimistic estimates (cf. Tversky & Kahneman, 1974). Although neither group completed their assignments within the time they said they would, the group that had been induced to predict relatively early completion times did complete the assignment significantly earlier than did the group induced to predict relatively late completion times. These results suggest that even overly optimistic expectations tend to get people further toward their goals than they would have otherwise.

The authors describe several possible mechanisms by which optimistic expectations may enhance performance. One explanation is outcome based, and derives largely from social learning theories (e.g., Bandura, 1977, 1997), expectancy-value theories of motivation (e.g., Atkinson, 1964) and, more recently, from cybernetic models of self-regulation (e.g., Carver & Scheier, 1981, 1998; Scheier & Carver, 1988). According to these models, goal-directed behaviors are strongly influenced by people’s expectations about what the outcomes of their behaviors might be. If people anticipate success, behavior is initiated (and ongoing behaviors maintained); if success appears to be less likely, people disengage from their pursuit (or refrain from initiating that pursuit) and abandon either the task or the currently selected means for completing that task. These models make a straightforward prediction: At least in situations where tenacity is rewarded (cf. Aspinwall & Richter, 1999), people who are optimistic – even people who are unrealistically optimistic – should achieve more than people with more modest expectations.

Another possible explanation for how even unrealistically optimistic expectations may influence behavior is more process-oriented, and draws on information about how these expectations are derived. According to one of the leading explanations for why people exhibit optimistic biases, people tend to infer the likelihood of different outcomes on the basis of case-specific plans or scenarios about how the future will unfold, and several lines of evidence suggest that the very processes of constructing and considering these scenarios tends to render people prone to bias (Buehler et al., 1994; Griffin & Buehler, 1999; Kahneman & Tversky, 1979, 1982; Newby-Clark, Ross, Buehler, Koehler, & Griffin, 1999). To the extent that the scenarios people generate in the context of making predictions provide a mental script for how to behave, however, these scenarios may facilitate effective performance (Anderson, 1983; Pham & Taylor, 1999). According to this model, optimistic forecasts may be associated with the attainment of desired outcomes (or the avoidance of undesired ones) only when the person making the forecast has a clear idea of how the considered outcome could reasonably be attained (or avoided). If the scenarios that people generate in an effort to divine the future provide a sort of mental road map for how that future might be attained, then it may be that one of the causes of optimistic bias will be a source for its cure.

Several other explanations for the prediction–performance link have been offered, although the unique aspects of these explanations have not yet been subjected to empirical test. One hypothesis is that the statement of prediction implies some amount of self-definition (e.g., “I am the type of person who can accomplish X”), and that people strive to meet their predictions in order to maintain consistency with the implied self-image (e.g., Sherman, 1980; cf. Aronson & Carlsmith, 1962). Another possibility is that the statement of optimistic expectations may be associated with an enhanced motivational state (here optimistic bias could be seen either as a cause of this motivational state or as a consequence of it), and it may be this state that either directly or indirectly facilitates the attainment of desired outcomes. The authors conclude that the evidence available to them does not establish which explanation is best.