Making judgements when information is missing: Inferences, biases, and framing effects

A model of multi-attribute evaluations based on incomplete information is investigated which proposes that subjects make inferences about missing attribute-information and incorporate them into the overall judgment of choice alternatives, in terms of classic studies of persuasion.

Abstract

A model of multi-attribute evaluations based on incomplete information is investigated which proposes that subjects make inferences about missing attribute-information and incorporate them into the overall judgment of choice alternatives. Judgments are a function of: (a) inferences based upon known information, and (b) an Incomplete Information Bias which is consistent across levels of the presented attribute-information. The independence of the two effects is demonstrated in an experiment about the evaluation of two-attribute consumer items—varieties of ground beef. First, the effect of inferring missing information about product quality is manipulated, without affecting the Incomplete Information Bias, by reminding subjects of the generally perceived relationship between price and quality. Second, two factors are found to affect the bias that do not affect the inference process; the (favorable or unfavorable) descriptive label, or frame, of the attribute for which information is missing and the subjects' prior knowledge of the class of alternatives. The role of prior knowledge is discussed in terms of classic studies of persuasion which compared the effectiveness of one-sided and two-sided messages.