Abstract

According to the diversity principle of philosophy of science, hypotheses are better supported by heterogeneous rather than homogeneous sets of data. The purpose of this dissertation was to determine whether people conform to this diversity principle in argument testing. Argument testing is defined here as hypothesis testing in the context of categorical arguments. More specifically, suppose you want to test a tentative conclusion of an argument for which you already have some evidence in the form of a true premise. Would you look for an additional premise that is similar or dissimilar to the premise you already have? According to the diversity principle, you should look for a dissimilar instead of a similar additional premise to test the conclusion. The results of three experiments reported here suggest that people do indeed conform to the diversity principle in argument testing. Subjects test the conclusion of an argument by choosing a dissimilar instead of a similar additional premise, and they also do so across several different arguments, and when choosing among several additional premises