Searching the internet for explanations: aligning standards improves the accuracy of metacognitive confidence judgments

Two experiments investigated a novel mechanistic explanation for inflated metacognitive confidence during information problem solving with the internet (IPS-I), proposing that overestimation is attributable in part to how individuals initially define explanatory tasks prior to engaging with online information. To test this, we introduced explicit standards in the form of a rubric specifying criteria for explanatory depth, conceptual complexity, and knowledge integration, with the goal of aligning learners' idiosyncratic standards with externally specified criteria and thereby reducing the previously reported overestimation bias. Participants answered open-ended explanatory knowledge questions with or without internet access and provided predictive and postdictive metacognitive confidence judgments. Experiment 1 ( n = 82) followed a 2 × 2 within-subjects design (Internet vs. No-Internet; Standards vs. No-Standards). Experiment 2 ( n = 143) replicated this design and additionally varied the presentation order of standards between participants. Across experiments, internet use improved explanation quality and increased metacognitive confidence, whereas external standards systematically reduced overestimation, with effect sizes ranging from small to large. Contrary to prior findings, the Overestimation-with-Internet-Bias was attenuated and not consistently replicated to the same extent. Together, the results suggest that misaligned task definitions contribute to overestimation in IPS-I and that clarifying evaluative criteria can improve metacognitive calibration in complex, open-ended explanatory tasks commonly encountered in educational and professional contexts.