A Metacognitive Stopping Rule for Problem Solving

Although people expect to improve by investing effort in solving a problem, several studies have found negative time-confidence correlations in various problem-solving tasks. The present study employed the metacognitive approach to illuminate why, despite lengthy thinking, people provide solutions in which they have only low confidence. According to the proposed Diminishing Criterion Model (DCM), as people invest longer in a problem, their confidence in their solution increases in a goal-driven manner, in accordance with the common belief. Nevertheless, the process ends up with a negative time-confidence correlation, because people tend to find lower confidence levels as satisfactory as they invest longer in solving a problem, reflecting a compromise in their stopping criterion. The hypotheses derived from the DCM were supported with two problem types. Even when the participants were allowed to submit a “don't know” response, they still provided low confidence solutions after lengthy thinking, suggesting that they found these low confidence solutions to be satisfactory. The study offers reconciliation between beliefs and empirical findings and explains why people end up offering solutions with low confidence rather than continuing attempts to improve or admitting failure (via the “don't know” option).