Further Tests of the Metacognitive Advantage Model: Counterfactuals, Confidence and Affect

Abstract

This study tested whether people have an accurate sense of how good their reasoning is, as measured by their confidence in their responses, and how good they feel after they give those responses. First, incorrect responders were unjustifiably confident in their responses. However, correct responders were even more confident, and this confidence boost was found to come from their awareness of alternative solutions that are intuitive but incorrect. An affect measure revealed the same pattern: correct responders felt better, and incorrect responders felt worse, after they solved reasoning problems, but this was only the case when post-reasoning affect was measured after participants were instructed to think of alternative solutions. Implications are discussed for the possibility of implicit error monitoring, the role of counterfactual thinking in meta-reasoning, and the use of affective measures in meta-reasoning research.

Keywords

• meta-reasoning
• confidence
• affect
• counterfactual thinking
• conflict detection