Metacognition: knowing what you know#

Picture the night before an exam. You’ve read the chapter three times, your notes are a rainbow of highlighter, and a warm, settled feeling tells you: I’ve got this. Then the question sheet lands, and the warm feeling evaporates. The material you “knew” won’t come. What happened? You didn’t misjudge the subject—you misjudged yourself. The single most useful skill a learner can build isn’t a memory trick at all. It’s an accurate sense of what you actually know and what you only think you know. Psychologists call this metacognition: thinking about your own thinking. Get it right and you study the gaps instead of the parts you’ve already mastered. Get it wrong and you pour hours into a feeling.

The fluency trap: why rereading fools you#

Here’s the uncomfortable truth about the most popular study method on earth. When you reread a page for the third time, the words flow easily, nothing snags, and your brain reads that smoothness as a signal: I know this. But the smoothness—psychologists call it fluency—is tracking how easy the text is to read, not how well you’d do without it in front of you. You are confusing “this is familiar” with “I can produce this.”

This is the illusion of competence, and it has been demonstrated again and again. Koriat and Bjork showed that learners systematically overrate how much they’ve learned when study feels easy, precisely because the easy, fluent experience whispers reassurance it hasn’t earned (Koriat & Bjork, 2005). Rereading and highlighting are the worst offenders: they maximise that comfortable familiarity while doing very little for durable memory, which is exactly why they sit near the bottom of the rankings of techniques that actually work (Dunlosky et al., 2013). The cruel part is that the methods that feel best are often the ones building the least.

So the first rule of judging your own learning is to distrust the warm glow. Ease of reading is not evidence of memory. If a study session felt frictionless, that’s not a sign it worked—it’s a reason to check.

Calibration: are you confident in the right places?#

Knowing something is rarely the problem. The problem is being confident about the wrong things. The technical word for the match between your confidence and your actual performance is calibration. A well-calibrated learner is sure of what they truly know, unsure of what they don’t, and—crucially—can tell the two apart. A poorly calibrated learner feels roughly equally confident across the board, which means their confidence carries no information.

Two patterns make calibration hard. The first is plain overconfidence, and it isn’t harmless. Dunlosky and Rawson followed students across a course and found that those who overestimated their knowledge studied less and stopped too early—and as a direct result, underachieved. The miscalibration wasn’t just an inaccurate self-image; it caused the lower grades, because the students called it quits the moment they felt ready rather than the moment they were (Dunlosky & Rawson, 2012). The feeling of knowing pulled the trigger on stopping too soon.

The second pattern is sharper and more humbling: the less you know about a topic, the worse you are at judging how little you know. This is the popularised but much-debated idea from Dunning and Kruger—the suggestion that weak performers don’t just do poorly but grossly overestimate themselves, because the very knowledge needed to do the task well is the same knowledge needed to recognise you’re doing it badly (Kruger & Dunning, 1999). How much of the original effect is real versus a statistical artefact is still argued over, so don’t lean on it too hard. But the cautious version is enough: when you’re a beginner, your sense of “I understand this” is at its least trustworthy, not its most. Treat early confidence with suspicion.

How to actually check: predict, then test#

The good news is that you can replace the feeling with a measurement. The tool is simple and a little brutal: stop asking “does this feel familiar?” and start asking “can I produce it right now, with the book closed?” The most fluent rereaders and the most accurate self-assessors aren’t different in talent—they’re different in which question they ask themselves.

Here are the monitoring habits I lean on:

Predict, then test. Before you check an answer or flip a flashcard, say out loud what you think you’ll score, then find out. The gap between your prediction and the result is your calibration error made visible—and the act of noticing it trains your judgement to be sharper next time.
Use retrieval as your honesty check. The moment you try to recall something and the page isn’t there, the illusion shatters instantly. Trying to retrieve and failing is the single fastest way to discover a gap you didn’t know you had. (This is why testing yourself does double duty—see retrieval practice .) Kornell and Bjork found that even unsuccessful retrieval attempts—guessing before you’re shown the answer—improve later learning, so a wrong prediction is never wasted (Kornell & Bjork, 2008).
Delay your self-judgement. Don’t decide “I’ve got this” the instant you finish reading; your judgement then is contaminated by what’s still echoing in short-term memory. Wait a while—an hour, a day—and then test. A judgement made cold, after a delay, is far more accurate than one made warm, on the spot (Nelson & Dunlosky, 1991).
Hunt for the holes, not the comfort. When you reread, you naturally gravitate to the parts you already know because they feel good. Flip it: deliberately go looking for the bits you stumble on, and spend your time there. The discomfort is the map.

The takeaway#

Your sense of how well you know something is a guess, and when study feels easy that guess is usually too high. So stop trusting the warm glow of familiarity and start demanding proof: before you decide you’ve learned something, predict how you’d do, then close the book and find out. The gap between what you expected and what you managed is the most honest feedback you’ll ever get—and chasing it down, over and over, is how you run the review phase of the learning cycle on yourself and learn to know what you actually know.