Determining and examining gender and/or experience differences in students' intuitions of chance, probability and uncertainty

In 1997, 156 students enrolled in two pre-statistics courses at two colleges in the southeastern United States answered a questionnaire containing 15 multiple-choice probability problems or questions published previously by several researchers; a subsample of nine of these students was also interviewed as a follow-up. The purposes of the study were to determine if there was any relation between observed differences in responses to these questions and the genders and/or the previous experience of the respondents in statistics or probability; and to determine if there was any difference by gender and/or experience in the way the students reasoned their answers to the questions. Of the respondents, 63% were females, and 64% classified themselves as having no previous experience in statistics or probability. In addition, two of the questions were a recreation of an experiment first conducted by Konold in 1987 which resulted in his theory of a heretofore unreported method of intuitive, probabilistic reasoning, which he called an "outcome requirement" or "outcome" approach.;The methodological bases for of this study were found in the works of Kahneman and Tversky, who, in the early 1970's, set the standard and agenda for the study of human intuition and its role in the day-to-day decisions of people; and within the conceptual framework established by Kant, Poincare and Piaget, in which intuition was defined as a category of cognition in which a person, in conditions of uncertainty, immediately "grasps" the frequency or probability of the uncertain phenomenon, without any need to explicitly explain or justify the phenomenon.;In Part 1 of the study, responses to the questionnaire were tabulated by question and by the gender, experience, and both experience and gender of the respondents. Seven questions were identified with significant differences in responses, to which chi-square tests were made to test for independence between the differences and the genders, experience and/or both experience and genders of the respondents. Only one question tested significantly as having a relation between the differences in responses for that question and both the experience and genders of the respondents. Further testing identified the responses of experienced male respondents as the major factor in those differences. Ironically, most respondents, both experienced and inexperienced, answered this question incorrectly. In Part 2 of the study, the subsample of nine respondents described the reasoning they used to answer the above significant test question and the two questions which were the recreation of the Konold experiment. Comparisons were made of these interview statements by gender, experience and type of reasoning. There appeared to be no difference by gender in the way respondents answered the questions, and how interviewees reasoned their answers. Experience, however, did play an important role in the responses and the statements describing reasoning for the significant test question, but inversely. All experienced interviewees answered the question incorrectly, most using a representativeness heuristic. Two inexperienced interviewees answered the question correctly, using appropriate, normative reasoning. Two other interviewees who answered the question incorrectly, gave reasoning statements that were inconsistent with their responses, and perhaps indicative of reasoning other than normative or heuristic. Responses and reasoning given in this study for the two Konold experiment questions were strikingly similar to those made by 1987 respondents, and confirmed the existence of the outcome approach. (Abstract shortened by UMI.).