Correction of Dispositional Inferences

Yet Is It Correction?

The studies described so far suggest that people initially take action as an index of the actor’s dispositions and then, under some circumstances, correct those inferences. Although the data are congenial with a correction account, high level accounts of what happens inside people’s heads are sometimes useful but always fanciful stories, of which inventive theorists can usually spin any number (see Gilbert, 1999). For example, whereas a correction account suggests that people draw dispositional inferences and then correct them when they can, a selection account suggests that people draw dispositional inferences when they are busy, but not when they are non-busy. According to the selection account, non-busy perceivers never “pass through” one stage on their way to another; that is, there is no moment in time during which the non-busy observer has drawn anything like a dispositional inference. This account is a perfectly reasonable post hoc description of the effects of busyness on inference that cannot be refuted by the kinds of evidence described so far. However, it can be refuted by another kind of evidence.

We all know that beliefs can be changed in a moment (“Yikes – a monster! Oh, wait, its just Uncle Tim in his Halloween costume”), but that emotions subside rather more slowly (“Gosh, my palms are still sticky”). This suggests that if nonbusy observers do draw dispositional inferences that last for just a moment before they are corrected, then the emotional reactions that those inferences generate should linger even after the inferences have been undone. In a follow-up to the study just described (Gilbert, Giesler, & Morris, 1995), participants took a test of their schizophrenia detection ability and then watched a model perform very well or very poorly, knowing all the while that the model’s performance was not a meaningful standard for comparison with their own. The results of the dispositional inference measures were very much like those shown in Table 9.6, suggesting that only busy participants took the model’s performance as an index of the model’s ability (and hence a positive or negative reflection on their own abilities). But in addition to measuring participants’ inferences, we also measured their emotional reactions at various points in time, and the results were striking. Although busy participants did, and non-busy participants did not, report drawing dispositional inferences, both groups experienced identical changes in their emotional states. Those who saw the model perform poorly felt a brief jolt of positive affect whereas those who saw the model perform well felt a brief jolt of negative affect – and this happened just as strongly and reliably for non-busy as for busy participants. In other words, non-busy participants experienced precisely the emotions one would expect them to have experienced had they briefly drawn a dispositional inference about the model. These findings provide compelling support for the correction account, and are simply unexplainable by the selection account. This does not mean that the account will go away, of course; it merely means that it should.