There's a moment most of us instinctively avoid: being asked a question we can't possibly answer. A teacher poses it before the lesson has started. A textbook opens a chapter with a quiz on material you haven't read. It feels pointless, even a little cruel. What good is a test you're guaranteed to fail?
Quite a lot, it turns out. One of the stranger findings in memory research is that attempting to answer a question before you've learned the material — and getting it wrong — makes you more likely to remember the correct answer later. Psychologists call this the pretesting effect, and it quietly rearranges some of our deepest assumptions about how learning is supposed to work.
What the pretesting effect actually is
The classic demonstrations come from work by researchers including Lindsey Richland, Nate Kornell, and Elizabeth Ligon Bjork in the late 2000s. In one line of experiments, participants studied a reading passage. Some of them first tried to answer questions about it — questions they had essentially no way of getting right, because they hadn't read anything yet. Others spent that same time simply studying. On the final test, the pretested group remembered the material better, despite having spent part of their time failing.
Kornell and his colleagues found the same pattern with weakly related word pairs: people who guessed at an answer before seeing it later remembered the correct pairing better than people who just studied the intact pair for longer. The failed attempt wasn't wasted time. It was preparation.
This is worth sitting with, because it contradicts the intuition that errors are contaminants — that a wrong answer leaves a smudge on memory that must be scrubbed away. For decades, education was shaped by exactly that fear; behaviorist-era teaching tried to prevent students from ever producing errors, lest the errors be 'stamped in.' The pretesting research suggests the opposite: an error made while genuinely searching for an answer, followed promptly by the correct one, leaves memory better prepared than passive study does.
Why wrong guesses help
Nobody fully agrees on a single mechanism, but several well-supported ideas converge here.
The first is search. When you attempt to answer a question, you can't help but activate the neighborhood of knowledge where the answer ought to live. Asked what causes color blindness before you've read the chapter on vision, you start pulling on threads — something about cones, maybe genetics, something about the retina. That activated network becomes scaffolding. When the real answer arrives moments later, it doesn't land on bare ground; it lands on a structure you just built, and it connects.
The second is error-driven learning. A great deal of learning, in brains and in machine-learning systems alike, runs on prediction error: commit to a guess, compare it against reality, and update in proportion to how wrong you were. A pretest forces a commitment. Without one, the correct answer is just information passing by. With one, it's a correction — and corrections are encoded with a kind of priority that mere exposure never gets.
The third is curiosity and attention. Trying and failing to answer a question opens a gap you can feel. When the answer appears, you're not neutrally reading it; you're resolving something. Studies of prequestions consistently find that people attend more closely to material that answers a question they've just wrestled with. The failed guess turns you from a passive reader into someone who wants to know.
This is not quite the generation effect
If you've read about memory before, this might sound like the generation effect — the finding that producing an answer strengthens memory more than reading one. The two are cousins, but the pretesting effect is the stranger cousin, and the distinction matters.
The generation effect concerns successful production: you retrieve or construct the right answer, and the act of producing it strengthens the trace. The pretesting effect concerns unsuccessful attempts — cases where you couldn't have known the answer, guessed, missed, and still benefited. It's also distinct from the hypercorrection effect, which describes what happens when a confident belief turns out to be wrong. Pretesting involves no confident belief at all. You know you're guessing. The benefit comes not from being corrected in your convictions but from the search itself.
That's what makes the finding so useful: it means the value of testing yourself doesn't begin once you know something. It begins before.
How to use pretesting when you study
The practical translation is simple, though it requires tolerating a moment of discomfort.
Guess before you look. Before reading a chapter, skim its headings and try to answer the questions they imply. Before a lecture, spend two minutes writing what you think the answers to the syllabus topics might be. You will be mostly wrong. That's fine — that's the point.
Use end-of-chapter questions at the beginning. Textbook authors put review questions at the end. Read them first, attempt them cold, then read the chapter. The questions now function as prequestions, and the chapter reads differently: passages that answer your failed guesses will snag your attention in a way they otherwise wouldn't.
Get feedback soon. This is the one hard constraint in the research. Pretesting works when the correct answer follows the attempt reasonably promptly, so the correction can attach to the still-warm search. A guess with no resolution is just a guess. Work by researchers such as Rosalind Potts and David Shanks found benefits even for entirely unfamiliar material — invented word meanings — as long as the right answer came soon after the attempt.
Don't grade yourself on the guesses. The failure is doing its job invisibly. If you treat every wrong pretest answer as evidence you're behind, you'll abandon the technique before it pays off. The whole trick is reframing: the wrong guess isn't a measurement of your knowledge. It's a treatment applied to it.
The caveat worth respecting
Pretesting is a supplement, not a replacement. It prepares memory for encoding; it doesn't replace the later retrieval practice that keeps knowledge alive over weeks and months. A guess before learning and a review after learning are two different medicines. The first helps the material get in. The second — spaced, effortful recall over expanding intervals — keeps it from leaking back out. Students who use one without the other are leaving half the benefit on the table.
There's also an emotional prerequisite: you have to be willing to be wrong on purpose, repeatedly, in private. For anyone whose schooling taught them that errors are shameful, this is genuinely hard. But the research is consistent — the learners who guess, miss, and check outperform the ones who wait until they're sure. Fluency is something you build by failing toward it, not something you're obligated to have before you start.
Where flashcards fit
A well-made flashcard is a pretest you can run on demand. The first time a new card appears, you haven't learned it yet — and the honest move is to attempt an answer anyway, feel the miss, then flip the card and let the correction land on the search you just performed. From that moment on, spaced repetition takes over the second half of the job: resurfacing the card at expanding intervals so the memory that pretesting helped you form doesn't quietly dissolve. Recall is built for exactly this rhythm — fast, beautiful flashcards scheduled by modern FSRS spaced repetition, so the algorithm decides when each memory needs you next. It works fully offline, and if your material already lives in Anki or Quizlet, you can import your decks and keep going without starting over.
If you'd like a study tool that honors both halves of the science — the productive failure at the start and the spaced retrieval that follows — you can try Recall at recall.lumenlabs.works.