{"id":6725,"date":"2026-07-16T12:15:03","date_gmt":"2026-07-16T12:15:03","guid":{"rendered":"https:\/\/edu-link.de\/?p=6725"},"modified":"2026-06-29T19:02:21","modified_gmt":"2026-06-29T19:02:21","slug":"active-recall-vs-rereading-memory","status":"publish","type":"post","link":"https:\/\/edu-link.de\/de\/active-recall-vs-rereading-memory\/","title":{"rendered":"Active recall vs re-reading: How to memorise so it stays"},"content":{"rendered":"\n

You read the chapter. You highlighted it in three colours. You read it again. The day before the exam you read it once more and feel ready. Then in the exam you stare at a question on exactly that material and nothing comes. This isn’t a memory failure. It’s a misallocation of effort. Re-reading builds familiarity (the comfortable sensation of „I recognise this“) but barely touches recollection (being able to produce the answer under demand). This post explains what to do instead, with a concrete TestAS-style example, and why this is the single most evidence-backed technique in the entire study-research literature.
<\/p>\n\n\n\n

\n\n\n\n

Why re-reading feels like learning but isn’t<\/h1>\n\n\n\n

Robert Bjork at UCLA gave this phenomenon a name in the 1990s: the illusion of competence. When you re-read familiar material, your fluency goes up \u2014 the words feel easier, the sentences flow, you have the warm sense that the material is „in there“. That sensation is a measurement of familiarity, not of memory.<\/p>\n\n\n\n

The classic demonstration comes from Karpicke and Blunt’s 2011 paper in Science<\/em>. They compared elaborative concept-mapping (the kind of „study harder“ approach teachers often recommend) against retrieval practice (self-testing). In immediate post-tests, the concept-mappers felt more confident in their learning. On the delayed test a week later, the retrieval-practice group remembered significantly more. The students consistently believed re-reading and mapping were working better. They were measurably wrong.<\/p>\n\n\n\n

Henry Roediger, who has spent decades studying this at Washington University, calls it the testing effect: the act of retrieval \u2014 not the act of input \u2014 is what builds durable memory. The mechanism is straightforward. Each time you successfully pull a piece of information out of your head, you strengthen the pathway from cue to answer. Re-reading skips this entirely; you’re just re-exposing yourself to the cue with the answer already attached.<\/p>\n\n\n\n

The cost of getting this wrong is large. Most students study the way they were taught: read, highlight, re-read. They feel productive. The material slides out of their head over the next week. They blame their memory.<\/p>\n\n\n\n

\n\n\n\n

The Karpicke & Roediger finding \u2014 50% less forgetting<\/h1>\n\n\n\n

The headline number from this literature: one self-test reduces forgetting by roughly half on a delayed exam compared to re-reading the same material the same number of times. This is the strongest single finding in the entire learning-science archive. Roediger and Karpicke’s 2006 paper in Psychological Science<\/em>, „The Power of Testing Memory,“ is the modern landmark; Gates 1917 was the first demonstration; the Cepeda 2006 meta-analysis covered hundreds of replications.<\/p>\n\n\n\n

For a TestAS or Leuphana candidate, the implication is direct: every chapter you study must end with closed-book retrieval. Every past-paper question is worth more than re-reading the same explanation. If you have two hours and a chapter, spend 45 minutes reading it once carefully and 75 minutes testing yourself on it. Not the other way round.<\/p>\n\n\n\n

The reason this isn’t more widely known is that students are bad at predicting which study method will work. In Karpicke’s experiments, even after being shown that retrieval beat re-reading on their own delayed test scores, many students continued to prefer re-reading because it felt productive. We’re calling this out so you don’t fall for it.<\/p>\n\n\n\n

\n\n\n\n

What active recall actually looks like<\/h1>\n\n\n\n

Active recall is closed-book retrieval. Anything that forces you to produce information from memory rather than recognise it on a page counts. The simplest version:<\/p>\n\n\n\n

    \n
  1. Read a chapter or watch a video.<\/li>\n\n\n\n
  2. Close the book or the tab.<\/li>\n\n\n\n
  3. Write down \u2014 by hand \u2014 what you remember. Three to five bullets minimum.<\/li>\n\n\n\n
  4. Open the book and check. The gap between what you wrote and what’s there is your next study target.<\/li>\n<\/ol>\n\n\n\n

    That’s it. There’s no app needed for the basic version, no special technique, no learning curve. The only barrier is that it feels harder than re-reading. That difficulty is the point.<\/p>\n\n\n\n

    Mueller and Oppenheimer’s 2014 paper in Psychological Science<\/em>, „The Pen Is Mightier Than the Keyboard,“ showed that hand-writing notes outperforms typed verbatim notes for conceptual retention. The mechanism: typing tempts you to transcribe; hand-writing forces selection and paraphrase, which itself is a form of retrieval. Use a pen.<\/p>\n\n\n\n

    One underrated point: even guessing BEFORE you read works. Charan Ranganath at UC Davis has shown this pre-testing effect \u2014 read the chapter heading, generate a guess about what’s inside, then read. Memory for the actual content is significantly higher than passive reading. The wrong guesses create the prediction-error signal that flags the correct answer for encoding.<\/em><\/p>\n\n\n\n

    \n\n\n\n

    Five formats: which to use when<\/h1>\n\n\n\n

    Different formats of active recall suit different material. Here’s how we choose:<\/p>\n\n\n\n

    1. The blank-page test.<\/strong> Best for: dense theory chapters, conceptual material, anything you’d be asked to explain. After a 60-90 minute block, take a blank sheet and write down everything you remember from the chapter \u2014 structure, key claims, examples, formulas. No headings, just write. Then check.<\/p>\n\n\n\n

    2. Flashcards \/ Anki.<\/strong> Best for: vocabulary, definitions, formulas, country-pair facts, dates. Anything where the cue and the answer are short. Don’t use Anki for material that needs an essay-length explanation; the format mismatches the content.<\/p>\n\n\n\n

    3. Teaching out loud.<\/strong> Best for: material you think you understand but haven’t tested. Explain the concept to an imaginary 14-year-old (or a real one) without looking at your notes. The gaps will surface within 30 seconds. This is sometimes called the Feynman technique; the research base is thinner than for blank-page testing but the logic is identical \u2014 production reveals where memory fails.<\/p>\n\n\n\n

    4. Past papers (timed).<\/strong> Best for: TestAS, Leuphana SFT, Deltapr\u00fcfung, KL Krems \u2014 any exam where the question format itself is the test. Real past papers are higher-yield than re-reading the prep book. Use papers from the last 3-5 years as your primary study artefact, not your secondary one.<\/p>\n\n\n\n

    5. Closed-book short-answer questions.<\/strong> Best for: building recollection rather than recognition. Multiple choice tests recognition; short answer tests recollection. The exam will demand recollection, so we train on the harder format. Practising harder than the real test is the edulink house rule for a reason.<\/p>\n\n\n\n

    A useful rule of thumb: if a study session ends with you re-reading, it ends weak. End every session with one of these five.<\/p>\n\n\n\n

    \n\n\n\n

    A worked example with a TestAS-style passage<\/h1>\n\n\n\n

    Let’s make this concrete. Suppose you’re working through a TestAS Quantitative Problems chapter on percentage change. You read the explanation, you do five practice problems, and you get four right.<\/p>\n\n\n\n

    The conventional move: read the explanation again, redo the problems you missed.<\/p>\n\n\n\n

    The retrieval-practice move:<\/p>\n\n\n\n

      \n
    1. Close the book.<\/li>\n\n\n\n
    2. On a blank page, write the rule for percentage change in your own words.<\/li>\n\n\n\n
    3. Write the formula from memory.<\/li>\n\n\n\n
    4. Generate one example with realistic numbers (e.g., a Bundesliga club’s revenue going from \u20ac280m to \u20ac315m).<\/li>\n\n\n\n
    5. Solve your own example.<\/li>\n\n\n\n
    6. Now open the book. Did your rule match? Did your formula match? Did your worked example use the rule correctly?<\/li>\n<\/ol>\n\n\n\n

      This takes maybe 10 minutes. It produces a stronger memory than 30 minutes of re-reading because three retrieval events happened \u2014 the rule, the formula, the worked example. Each retrieval was effortful. Each one strengthened the pathway.<\/p>\n\n\n\n

      For a TestAS Verbal Analogies chapter, the move is different but the principle holds: after reading the strategy, do a stack of 10 questions cold (without checking the strategy mid-question), then review the wrong ones, then close the book and write the strategy from memory before bed.<\/p>\n\n\n\n

      For the Leuphana SFT Tabellen-und-Diagramme section, the move is to take a real past-paper graph, close every explanation, and answer the questions under time pressure first \u2014 then read the model solution to fill in what you missed.<\/p>\n\n\n\n

      \n\n\n\n

      Free tools we recommend<\/h1>\n\n\n\n