Document Type

Thesis

Degree Name

Master of Arts (MA)

Department

Psychology

Faculty/School

Faculty of Science

First Advisor

Robert Gebotys

Advisor Role

Thesis Supervisor

Abstract

Numerous studies suggest that in judging an individual’s category membership, people largely ignore the population frequency membership in the category (base rate information) in favour of the individuating information about that particular individual (more commonly referred to as the “Base Rate Fallcy”: Bar-Hillel, 1980). One particularly interesting strategy that has been shown to increase the use of base rate information is what Ginossar and Trope (1987) refer to as the Prior Activation Assumption. The purpose of this thesis was to extend the boundaries of Ginossar and Trope’s (1987) findings by enhancing participants’ use of base rate information through the manipulation of two training problems that differ in their category membership from the test problem(s). Experiment 1 investigated the effects of the Prior Activation Assumption along with one other strategy through to increase the use of base rate information; that of order in which the information is to be presented. Results of Experiment 1 failed to demonstrate activation of sampling rule across the two training problems to the test problem. A main effect for order of information also failed to be supported. Therefore, in Experiment 2, along with making probability assessments concerning an individual’s category membership, participants also completed Gebotys and Claxton-Oldfield’s (1989) Probability Knowledge Questionnaire. This questionnaire assessed participants’ knowledge of probability concepts that provided the basis to categorize them as either an expert or a novice. Results of Experiment 2 failed to provide direct support for Ginossar and Trope’s Prior Activation Assumption when analyzing the raw probability data, as there was a main effect for the manipulation of BASE RATES. Experts were more inclined to use base rate information and score higher on strategy usage than were novices, regardless of experimental treatment condition. Experts were better calibrated in their responses than novices, better able to use the same or a similar strategy used to solve the training problems when solving the test problem, and better able to transfer that strategy to another problem domain. The applications and limitations for the manipulation of information presented in category membership probability problems are discussed.

Convocation Year

1992

Convocation Season

Fall

Download

Share

COinS