The Conjunction Fallacy

Abstract

The present investigation adopted a debiasing approach to the judgmental error known as the conjunction fallacy (Tversky & Kahneman, 1982). Such an approach was used to determine the extent to which the conjunction fallacy reflects task specific misunderstanding of particular judgment problems. The results suggest that (a) subjects' misunderstanding of conjunction problems is indeed somewhat task specific, and (b) a debiasing approach can effectively lower but not eliminate the conjunctive error rate for problems that do not strongly implicate representativeness thinking. Educational strategies based on statistical and probabilistic knowledge are discussed as an approach to debiasing inferential errors like the conjunction fallacy.

On the nature of the conjunction fallacy

“…Nevertheless, I think this technique may still provide a good initial approximation of how people interpret one of the key options. Morier and Borgida (1984) also found the same result. They added "Linda is a bank teller who is not a feminist" in the list of options.…”

“…Morier and Borgida (1984) provide another instance of application of this technique. They use a different set of options to avoid the misunderstanding.…”

School students' reasoning about conjunction and conditional events

“…Some studies followed the suggestion of Tversky and Kahneman to focus research interest on the form of expression for responses, such as ranking alternatives, qualitative measurement, and degrees of quantitative measurement. Morier and Borgida [18] found students performed better when asked to estimate probabilities (80% conjunction errors) than when asked to rank order alternatives (95% conjunction errors), and also when other probabilities such as P(A or B) were requested (48±69% conjunction errors), making students pay attention to the logic of the conjunction for assigning P(A and B). Yates and Carlson [19] found fewer conjunction errors (30%) for a problem involving speci®ed numbers of marbles in an urn than for`naturalistic problems' involving more subjective estimates of probability (55±71% conjunction errors).…”

“…As the items on conjunction and conditional probability in this study were part of a larger survey, it was not possible to include as many formats and settings for questions as covered by earlier researchers. Based on the experiences of Morier and Borgida [18] and Pollatsek et al [15], it was decided to use the open-ended format allowing students to produce their own estimates of likelihood. Aware of the opinion of Shaughnessy [13], it was decided that task complexity should be minimized to focus on the concepts of research interest.…”

“…Rather than using in-school contexts such as drawing objects from a bag with and without replacement, however, out-of-school contexts similar to those of Fielder [20] and Pollatsek et al [15] were used in an e ort to explore student reasoning in everyday contexts and to re¯ect the application goals of the school curriculum. The use of probability and frequency forms of questions was intended to extend the results of Morier and Borgida [18] and Pollatsek et al [15] to school students. Five research questions were of interest.…”