The Testing Effect: Why Quizzing Yourself Beats Rereading (Even Though It Feels Worse)
Published: September 12, 2026 | By: Roman Vasylyshyn, founder of Lambdio | Reading time: 8 min
Imagine two students preparing for the same exam. Student A reads the chapter, then rereads it two more times before the test. Student B reads the chapter once, then closes the book and tries to write down everything they remember, checking their answers afterward. Ask either student, right after studying, how well they know the material — Student A will almost always feel more confident.
Student B will almost always score higher on the actual exam.
This gap between how confident studying makes you feel and how much it actually helps you remember is one of the most consistently replicated findings in cognitive psychology. It's called the testing effect, or retrieval practice, and it's the specific mechanism behind why quizzing yourself works so much better than rereading — not a study tip, but a measurable property of how memory is formed.
Table of Contents
- The Experiment That Named the Effect
- Why Retrieval Changes Memory, and Rereading Doesn't
- The Confidence Trap
- How Big Is the Effect, Actually?
- It Works Even When You Get the Answer Wrong
- Why This Isn't Just "Studying Harder"
- How Lambdio Applies This
The Experiment That Named the Effect
The clearest demonstration of the testing effect comes from a 2006 study by Henry Roediger and Jeffrey Karpicke at Washington University in St. Louis. Students studied short passages of text, then were split into groups: some restudied the passage multiple times, others were tested on it (recalling what they could, without feedback) instead of restudying.
The twist was in the timing of the final test. When students were tested five minutes after studying, the repeated-restudy group did better — rereading gives you a short-term boost, which is exactly why it feels productive. But when the same students were tested one week later, the result flipped entirely: the group that had practiced retrieval outperformed the rereading group by a wide margin, despite spending the same amount of study time.
The students in the study were also asked to predict how well they'd do on the delayed test. The rereading group predicted they'd do better than the testing group did — the opposite of what actually happened. That mismatch between predicted and actual performance is a core part of why the testing effect is so easy to miss in everyday studying: the technique that works better is the one that feels like it's working worse.
Why Retrieval Changes Memory, and Rereading Doesn't
Rereading is passive. The information is right there on the page; recognizing it as familiar requires very little cognitive effort, and that low effort is exactly why it produces a weak, short-lived memory trace.
Retrieval practice is different because it forces your brain to reconstruct the information from scratch, using whatever cues and associations it has stored. That reconstruction process is effortful, and — this is the key mechanistic point — the act of retrieving a memory appears to strengthen it more than the act of re-encoding it does. Each successful retrieval seems to reorganize and reinforce the underlying memory trace in a way that passive re-exposure simply doesn't trigger.
This connects directly to the forgetting curve: a memory that's been actively retrieved decays more slowly afterward than one that's only been reread, because retrieval — not exposure — is what the brain treats as a signal that this information matters and should be preserved.
The Confidence Trap
The gap between how a study method feels and how well it works isn't a minor footnote — it's arguably the main reason retrieval practice remains underused despite decades of supporting research. Rereading produces a fluent, familiar feeling ("I recognize all of this") that gets mistaken for actual mastery. Testing produces a effortful, sometimes frustrating feeling ("I can't remember this") that gets mistaken for the technique not working.
Here's roughly what that divergence looks like over time, based on the pattern found across retrieval-practice studies:
Immediately after studying, rereading looks like the winner. A week or a month out — which is the timescale that actually matters for an exam, a certification, or a skill you need to keep — retrieval practice has pulled well ahead. Anyone judging their study method by how it feels in the moment is, in effect, reading the wrong end of this chart.
How Big Is the Effect, Actually?
This isn't a fragile, one-study result. A widely cited meta-analysis by Roediger and Karpicke, along with follow-up work by Andrew Butler and others, has replicated the testing effect across age groups (children through older adults), material types (word lists, prose passages, foreign language vocabulary, science concepts), and retention intervals ranging from minutes to months.
Effect sizes in this literature typically fall in the moderate-to-large range (Cohen's d commonly between 0.4 and 0.9 depending on the study design), which puts retrieval practice among the more reliably effective learning interventions that cognitive psychology has identified — well above many more popular study techniques like highlighting or summarizing, which meta-analyses have generally found to be weak.
It Works Even When You Get the Answer Wrong
One of the more counterintuitive findings in this area: attempting to retrieve an answer and getting it wrong, as long as you see the correct answer afterward, still produces a stronger memory than simply rereading the correct answer would have. The act of searching for the answer — even unsuccessfully — appears to prime the memory system to encode the correction more deeply once it arrives.
This has a direct practical implication: a quiz you sometimes fail is doing more for your long-term retention than a quiz you always pass. If every self-test feels easy, you're probably testing material you've already secured, not material on the edge of being forgotten — which is exactly the material that benefits most from a review in the first place.
Why This Isn't Just "Studying Harder"
It's worth being precise about what the testing effect is and isn't claiming, because it's sometimes flattened into generic "effort is good" advice. The effect is specific to retrieval — actively producing an answer from memory — not to effort in general. Rereading a passage slowly and carefully takes real effort too, but it doesn't involve retrieval, and the research shows it doesn't produce the same benefit. Underlining, highlighting, and note-taking-while-reading have all been studied directly and consistently underperform retrieval practice, even when they take just as long.
The mechanism is retrieval specifically, which is why the format of a self-test matters less than whether it requires you to actually produce or select an answer from memory rather than recognize it on a page you're currently looking at.
How Lambdio Applies This
This is the research basis for Quiz mode in Lambdio. Quiz mode isn't available immediately after you learn a chapter — it unlocks once the spaced repetition engine estimates you're approaching the point where you'd otherwise start forgetting the material, based on that chapter's individual difficulty and stability scores. That timing isn't incidental: retrieval practice is most effective precisely when recall requires real effort, not when the answer is still fresh enough to feel obvious.
When you do take the quiz, your score feeds back into the same FSRS engine that scheduled the quiz in the first place — a strong result (85%+) counts as an "Easy" rating and pushes the next review further out; a weak result (below 50%) counts as "Again" and brings it back sooner. So the quiz is doing double duty: it's the retrieval event that strengthens the memory, and it's also the data point that recalibrates when you'll need to revisit that chapter next.
Quiz mode sits alongside Standard, Socratic, Chat Review, and Feynman modes as one of several ways Lambdio can turn a chapter into a retrieval event rather than a rereading session — we'll cover the research behind the conversational modes in an upcoming post.
If you want to feel the difference between rereading and retrieval for yourself, courses are free to start at lambdio.com.
Sources: Roediger & Karpicke (2006), "Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention," Psychological Science; Karpicke & Roediger (2008), "The Critical Importance of Retrieval for Learning," Science; Butler (2010), meta-analysis on retrieval practice and transfer.