Metacognition

Metacognitive knowledge is stored, stable knowledge about your own mental processes and those of others — knowing that you find maths harder than languages, that a task requiring divided attention is harder than one requiring focused attention, or that retrieval practice works better than re-reading. Metacognitive regulation is the active, moment-to-moment use of that knowledge during a task: planning which strategy to use before you begin, monitoring your comprehension as you read, and evaluating whether your performance met your goal. Knowledge without regulation (knowing retrieval practice is effective but never doing it) produces no benefit; regulation without knowledge (actively checking progress against the wrong criterion) misdirects effort.

The Dunning-Kruger effect (1999) is the empirical finding that people with low competence in a domain systematically overestimate their own performance, while highly competent people tend to underestimate theirs slightly. Kruger and Dunning argued this occurs because competence in a domain and the metacognitive ability to evaluate competence in that domain rely on overlapping skills — novices lack both, creating a double burden. They tested this across logical reasoning, grammar, and humour; in each case, bottom-quartile participants placed themselves above the 60th percentile. Highly competent performers underestimate because they assume the task is comparably easy for others.

Retrieval practice (the testing effect) works because the act of retrieving a memory strengthens its storage and consolidates it for future retrieval. Re-reading produces a sense of fluency (the material looks familiar) that is often mistaken for knowing — an illusion of knowing. When you attempt to recall, the effort required (desirable difficulty) strengthens the memory trace. Roediger and Karpicke (2006) had participants study a prose passage, then either re-study it or take a free recall test. After one week, the test group retained 50% more than the re-study group. Spaced retrieval (spreading tests across time) is more effective still, as it requires reconstruction from a faded trace.

Evidence-based approaches include: (1) Self-explanation — prompting students to explain their reasoning aloud or in writing while problem-solving, which surfaces gaps in understanding. (2) Calibration training — having students predict their test scores before and after tests, then reflecting on over- or under-confidence. (3) Think-aloud protocols — verbalising mental processes during tasks, which externalises metacognitive monitoring. (4) Explicitly teaching retrieval practice and spaced study schedules, with explanation of why they work. Meta-analyses consistently find that metacognitive strategy instruction improves academic achievement, particularly for low-achieving students, with effect sizes of approximately 0.6–0.7 standard deviations (Hattie, 2009).