Norms from Single Elements

Norms from Single Elements

The present model of category norms assumes both that a single instance suffices to set up a norm and that a reference to the category label serves to restrict consideration to members of that category. These assumptions entail nonregressive predictions and a tendency to neglect relevant base rates.

To illustrate, imagine that you were shown a single exemplar of an unfamiliar species of insect, which was larger than any other insect you have ever seen. How large would you expect another exemplar of that species to be? Would it occur to you that the single insect you saw is likely to be larger than most of its conspecifics? The canons of inference prescribe that the single insect you saw is likely to be larger than most others in its species. Predictions of the size of the next member of the new species should accordingly regress toward the general norm for insects. Intuitive expectations, in contrast, are firmly centered on the single observed value, which constitutes the norm for that category. In the context of social judgment, the ease with which category norms are established leads to radical generalization from a single observation of behavior to an interpersonal norm for the behavior of other people in the same setting, or to a norm for a person’s behavior on future occasions (Nisbett & Ross, 1980; Read, 1984). . . .

Radical generalization from observed behavior to an intrapersonal norm is manifest in the nonregressiveness of behavioral predictions (Kahneman & Tversky, 1973; Ross & Anderson, 1982). A single observed instance of unusually generous tipping appears sufficient to set up a norm that will be consulted in predictions of a person’s future tips. The interpersonal base rate for the behavior of other people is effectively excluded from consideration when thinking about this person’s tips, just like one’s knowledge of the size of insects in the previous example.

Radical generalizations can be made with high and unwarranted confidence. Experimental results suggest that subjective confidence depends mainly on the consistency of available evidence rather than on its quality or quantity (Einhorn & Hogarth, 1982; Kahneman & Tversky, 1973; Tversky & Kahneman, 1971). Subjective confidence in a prediction is likely to reflect the breadth, or variability, of the category norm on which the judgment is based. Confidence will be high if all elements cluster around a single value of the attribute, independently of the number of these elements. However, it would be incorrect to conclude that judgments are entirely insensitive to the quantity of evidence. A category norm based on one or two elements can support confident judgments, but it can also be altered relatively easily under the impact of new evidence.

Many inferences bridge across situations and across attributes. In the absence of better evidence, people readily predict success in graduate school from an IQ test score, research productivity from performance in a colloquium, or the size of a mother’s graduation gift to her daughter from the size of a tip that she gave to a waiter. Such generalizations are involved both in deliberate predictions of future behaviors and in spontaneous inferences about traits and about unobserved aspects of situations.

Inferences that bridge attributes are no more regressive than direct generalizations (Nisbett & Ross, 1980; Ross & Anderson, 1982). Nonregressive predictions can be generated in two ways. First, the predicted value may be chosen so that its position matches that of the known attribute in their respective interpersonal norms (Kahneman & Tversky, 1973). Thus, a predicted graduation gift can be chosen that is as extreme in the norm for gifts as the observed tip is in the interpersonal norm for tips. Second, nonregressive predictions can also be generated by matching descriptive labels (e.g., if a tip is remembered as very generous, what graduation gift would be considered equally generous?). The intention to predict a behavior elicits a search for relevant incidents or for pertinent descriptive labels. The search is guided by the similarity of potential elements to the target attribute, and it is probably concentric: The nearest incidents or labels that turn up in the search are used, nonregressively, to generate a prediction. In the absence of better data, people are willing to make extreme predictions from evidence that is both flimsy and remote. The process of concentric search yields radical (and overconfident) predictions from observations of dubious relevance. However, the same process also makes it likely that distant labels or incidents will be ignored when evidence that is closer to the target attribute is available. Thus, generous tipping habits will not be given much weight in predicting a mother’s graduation gift to her daughter if pertinent incidents of their interaction can be retrieved.