Limits of the Durability Bias

What Are the Limits of the Durability Bias?

In our studies we found the durability bias wherever we looked; but, of course, we looked where we thought we would find it. The durability bias may well be a pervasive phenomenon, but surely it is just as important to know its limits as it is to know its reach. One possible limit has to do with the valence of the event about which the affective forecast is made. In half of our studies, we examined forecasts and experiences of positive affect, and across those studies we found no significant evidence of a positive durability bias. Yet we all know that people occasionally overestimate the duration of their positive experiences, and we have found reliable evidence for such a bias in some of our own studies (Wilson et al., 1998). Nonetheless, the relative ease and difficulty with which these two biases are produced suggests that one may be more robust than another. Why should that be the case?

One possibility was noted at the outset: The psychological immune system ameliorates negative affect but does not augment positive affect, and therefore immune neglect produces only a negative durability bias. Thus, in any complex situation in which numerous mechanisms are simultaneously at work, more mechanisms are conspiring to produce a negative than a positive durability bias. Another possible explanation for the weakness of the positive durability bias has to do with the way in which affective forecasts guide behavior. People naturally avoid those events that they believe will produce negative affective consequences (“No, I’d prefer not to eat snails, thank you”), and hence may fail to learn that such beliefs are sometimes mistaken (e.g., Herrnstein, 1969; Seligman, 1975). Conversely, people may seek those events that they believe will produce positive affective consequences (“But yes, a few more vodka tonics, please”), and thus they may have ample opportunity to learn that such beliefs are sometimes mistaken. If people consistently act on their forecasts, they will inevitably experience fewer disconfirmations of their overly pessimistic predictions than of their overly optimistic predictions, and thus experience may cure the positive durability bias more quickly than the negative durability bias. Indeed, old age may be characterized by a loss of idealism in part because people may learn that the things they once thought would make them permanently happy did not actually do so; but, because they avoided the things that they believed would make them permanently unhappy, they may have failed to learn that those beliefs were equally untrue.

If the valence of an affective forecast describes one limit on the durability bias, then the direction of misprediction describes another. Simply put, people may underpredict as well as overpredict the duration of their affective reactions. For example, people may be surprised to find that the death of a great uncle pains them for much longer than they would have thought possible, that a new sports car gives them greater daily pleasure than they could have imagined, or that a decision to forego a job offer or marriage proposal led to years of unanticipated regret (Gilovich & Medvec, 1995). Although instances such as these surely do occur, our analysis suggests two reasons why overprediction of affective duration is probably more common than underprediction. First, some of the mechanisms we have identified (such as misconstrual, incorrect theories, motivational distortion, and focalism) can produce both underestimation and overestimation, but others (such as undercorrection and immune neglect) can produce overestimation only. We know of no mechanism that would produce underestimation only. Second, underpredictions may be more likely to be remedied by experience than are overpredictions. For example, Rachman and colleagues showed that people frequently overpredict their fear and anxiety, and that these overpredictions are slowly reduced over many experimental trials in which the person makes explicit predictions and then, moments later, makes an explicit experiential report that contradicts that prediction (see Rachman, 1994). However, when people occasionally underpredict their fear or anxiety, this mistake is usually eliminated in just one trial, and it is not difficult to see why: One may touch a stove gingerly several times before coming to believe that it is indeed cooler than anticipated, but it requires just one good scorching to remedy the opposite misapprehension. If underpredictions of negative affect are met with unexpected punishment, whereas overpredictions yield pleasant surprises (when they are noted at all), then we might well expect that over time the overpredictions will become more common than underpredictions. Although we see little evidence of underestimation in either our lives or our laboratories, it is certainly possible that such evidence is simply waiting to be found. For now, we place a public bet on the predominance of the durability bias, fully confident that should our faith prove misplaced, we will not be embarrassed for long.

* Portions of this chapter are reprinted from: Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blumberg, S. J., & Wheatley, T. P. (1998). Immune neglect: A source of durability bias in affective forecasting. Journal of Personality and Social Psychology, 75, 617–638. Copyright © 1998 by the American Psychological Association. Adapted with permission.