Belief bias in informal reasoning

Abstract

In two experiments we tested the hypothesis that the mechanisms that produce belief bias generalise across reasoning tasks. In formal reasoning (i.e., syllogisms) judgements of validity are influenced by actual validity, believability of the conclusions, and an interaction between the two. Although apparently analogous effects of belief and argument strength have been observed in informal reasoning, the design of those studies does not permit an analysis of the interaction effect. In the present studies we redesigned two informal reasoning tasks: the Argument Evaluation Task (AET) and a Law of Large Numbers (LLN) task in order to test the similarity of the phenomena concerned. Our findings provide little support for the idea that belief bias on formal and informal reasoning is a unitary phenomenon. First, there was no correlation across individuals in the extent of belief bias shown on the three tasks. Second, evidence for belief by strength interaction was observed only on AET and under conditions not required for the comparable finding on syllogistic reasoning. Finally, we found that while conclusion believability strongly influenced assessments of arguments strength, it had a relatively weak influence on the verbal justifications offered on the two informal reasoning tasks.

Keywords:

• Belief bias
• Informal reasoning
• Argument evaluation task
• Law of large numbers
• Syllogistic reasoning

Notes

¹Our pre-test differed somewhat to Stanovich and Wests's in that our rating scale included a “don't know” option.

²Scores on the three rating scales were highly correlated; to avoid redundancy, we computed and analysed the mean of the three ratings.

Additional information

Notes on contributors

Valerie Thompson

The authors contributed equally to the preparation of this manuscript. The research was funded by a Discovery Grant from the Natural Sciences and Engineering and Research Council of Canada to Valerie Thompson. We would like to thank Shira Elqayam for her comments on an earlier draft of this manuscript.