Reading focus
Syllabus: How Do We Reason and Solve Problems?
Use the crossed validity/believability design to separate an answer pattern from the proposed explanation of it. Compare the evidence from choices and verbal reports.
Why you're reading it. Apply the distinction between logical standards, prior knowledge, and response processes to an experiment.
A question to consider. Would accepting many believable conclusions uniquely support failure to search for counterexamples, or could a response preference produce a similar pattern?
Why this reading
Believing a conclusion and judging whether it follows are different tasks. Crossing believability with logical validity makes that difference experimentally visible. The verbal reports add a second question: how confidently can a person's explanation reveal the process that produced an answer?
Before you read
A syllogism links statements about categories. Logical validity concerns the relation between premises and conclusion, not whether the conclusion happens to be true in the world. An invalid argument can have a believable conclusion. Premise conversion means reversing a relation; some reversals preserve meaning and others do not. The participants are asked to assume the supplied premises are true.
Read for
- Reconstruct the four combinations of valid/invalid and believable/unbelievable conclusions. Look for effects of each factor rather than treating every error as a belief effect.
- Identify what the successive experiments try to control, including premise conversion and response preferences.
- Compare answer patterns with the protocol classifications. Consider how asking people to justify an answer might differ from observing the process that generated it.
Consider afterwards
- A participant accepts more believable conclusions, whether valid or invalid. What additional comparison would help distinguish a greater willingness to say yes from a change in the ability to discriminate valid arguments?
Editorial context
The experiments helped establish belief bias as an explanatory problem; they do not show that belief always defeats logic. Later measurement work separates willingness to endorse a conclusion from discrimination of validity. Trippas and colleagues' 2018 meta-analysis found no unconditional believability effect on discrimination across its corpus, while identifying qualifications involving individual differences. The classic interaction and verbal reports therefore cannot uniquely establish one reasoning mechanism.