Does Irrelevant Information Influence Judgment? : A dissertation submitted in partial fulfillment of the requirements for the degree of Ph.D. in Cognitive Science

Abstract

When people judge how tall a person is it is rational to assume that they ignore all irrelevant information about the color of his/her hair and eyes, the fact that he/she wears glasses, or the profession of that person, and focus on the relevant dimension “height” only.
This dissertation is based on a model of judgment, JUDGEMAP-1, that proposes certain mechanisms that underlie judgment. This model assumes that in order for judgment to happen people activate a number of exemplars from the same category in Working Memory and this active set is called the comparison set. The target stimulus is included in this comparison set and altogether they are mapped on the scale using the same mechanisms that underlie analogical mapping. Thus, key for the rating, which the target stimulus will obtain, is what the other members of the comparison set are (if smaller exemplars happened to be involved in the comparison, the target will most probably receive a higher rating as it will stand out in the comparison set). The basic mechanism responsible for retrieving exemplars in WM in JUDGEMAP-1 is spreading activation and thus any kind of shared features will make the retrieval of a particular exemplar more probable. This means that the color of the hair and eyes, the profession of the person may actually play a role, since the comparison set will consist of people with similar characteristics. This theoretical prediction was backed up by a particular simulation experiment with JUDGEMAP-1 and thus formed the main prediction to be tested within this dissertation.
A series of experiments has been performed that test this prediction under various circumstances. Even though the first experiment confirmed the prediction, , the size of the effect was very small (although significant), so there were serious doubts whether this result can be replicated. Thus, the series of experiments replicates the initial experiment varying some parameters in the hope to enlarge the size of the effect. The experiments varied the dimensionality of the stimuli, the presentation manner and presentation time of the stimuli, and the type of stimuli – from simple lines to rabbits to abstract numbers. Even though we could not increase the size of the effect, the effect was replicated quite robustly in almost all experiments.