The Role of Inference in Context Effects: Inferring What You Want from What Is Available

It has recently been suggested that a number of experimental findings of context effects in choice settings can be explained by the ability of subjects to draw choice-relevant inferences from the stimuli. We aim to measure the importance of this explanation. To do so, inferences are assessed in an experiment using the basic context-effect design, supplemented by direct measures of inferred locations of available products on the price-quality Hotelling line. We use these measures to estimate a predicted context effect due to inference alone. For our stimuli, we find that the inference effect accounts for two-thirds of the average magnitude of the context effect and for about one-half of the cross-category context-effect variance. Copyright 1997 by the University of Chicago.