{"course":{"slug":"ai-for-students","title":"AI for Students","url":"https://addaly.com/learn/ai-for-students","level":"beginner","track":"work"},"lesson":{"slug":"worked-examples-and-fading","title":"Worked examples, and when to stop using them","url":"https://addaly.com/learn/ai-for-students/worked-examples-and-fading","position":25,"of":73,"minutes":8,"answer":"Worked examples help beginners and hinder people who already have the method, so fade from full example to blank problem by removing steps from the end.","keyIdea":"Worked examples help beginners and hinder people who already have the method, so fade from full example to blank problem by removing steps from the end, and cap how many examples you read before moving on."},"body":{"format":"markdown","text":"## Beginners and experts need opposite things\n\nWhen you are new to a type of problem, studying a fully worked example teaches you more per minute than attempting the problem yourself. This is well established, and it is the one place where being shown the answer is the right move.\n\nThe reason is capacity. Working memory holds only a few items at once. A beginner attempting an unfamiliar problem spends all of it on searching (what do I do first, is this the right formula, what was that rule) and has nothing left for noticing the structure. A worked example removes the search, so attention goes to the pattern.\n\nBut this reverses. The same worked examples that help a beginner actively slow down someone who already knows the method: they now have to process an explanation of something they can already do, which costs capacity and gives nothing. This reversal has a name, the **expertise reversal effect**, and it is why study advice that ignores your current level is unreliable.\n\n## Fading, which is the actual technique\n\nThe right sequence is not *examples* or *problems*. It is a fade from one to the other:\n\n1. **Full worked example.** Every step shown, with the reason for each.\n2. **Completion problem, last step removed.** You do the final step.\n3. **Last two steps removed.**\n4. **Only the first step given.**\n5. **Nothing given.**\n\nEach stage is a small generation attempt at the point you are most ready for it. This structure, *backward fading*, because the removal starts from the end, outperforms both pure examples and pure problem solving in controlled comparisons.\n\nModels are unusually good at generating this, and almost nobody asks them to:\n\n> Give me a worked example of this type of problem with the reasoning at every step. Then give me the same type of problem with the last step missing for me to complete. Then one with the last two missing. Then one with only the first step given. Then one with nothing. Wait for my answer at each stage before continuing.\n\nThat is a complete, well-evidenced practice sequence, generated in one message.\n\n## Study the example properly, which most people do not\n\nReading a worked example passively gets you very little. The version that works:\n\n- After each line, cover the rest and predict the next step before revealing it.\n- Write, beside each step, *why* that step is legal. Not what it does: why it is allowed.\n- At the end, close the example and write the sequence of moves from memory, in words, without numbers. This is the pattern you are actually trying to acquire.\n\nThat last one is the difference between learning to do *this* problem and learning to recognise this *kind* of problem.\n\n## Knowing when to stop\n\nThe signal that you have crossed out of the worked-example stage is that you can predict the next step correctly before revealing it, twice in a row. From that point, examples cost you time. Move to completion problems immediately, and to blank problems soon after.\n\nStudents overstay in this stage badly, because reading examples is comfortable and pleasant and feels like work. The number of worked solutions somebody has read is close to uncorrelated with what they can do; a solution read is a solution you can follow.\n\n## The trap specific to AI\n\nA model will produce an unlimited supply of worked examples on demand. That makes it easy to spend an entire evening in stage one, which feels productive and is barely study at all. The fix is procedural: **decide in advance how many worked examples you will read before moving to completion problems.** Two is usually right; three if the material is new. Then move, whether or not you feel ready, because feeling ready is not a reliable signal here.\n\n## Checking that the example is right\n\nOne caution the classic research did not have to deal with. Worked examples in a textbook were checked. A generated one was not. In any subject where the steps matter, verify the example against a source before you learn from it: a textbook problem, a past paper with a mark scheme, a symbolic tool like SymPy or Maxima for the algebra. Learning a wrong worked example is worse than learning nothing, because you will apply it confidently.\n\nThe safest arrangement: take the worked example from your course materials, and use the model to generate the *fading sequence* of practice problems around it."},"check":{"question":"A student who can already solve a class of problem is given detailed worked examples of it. Why does this measurably slow them down?","options":[{"text":"Because it reduces motivation, and material that feels too easy lowers the effort they put into the session.","correct":false,"tempting":"Substitutes a motivational story for a capacity one; the effect appears even when students are fully engaged and trying."},{"text":"Because worked examples always show one particular method, which interferes with the method the student had already learned.","correct":false,"tempting":"Interference is real when methods genuinely differ, but the reversal effect appears even when the example uses exactly the student's own method."},{"text":"Because reading a solution suppresses the retrieval attempt, and retrieval is the only mechanism by which practice strengthens memory.","correct":false,"tempting":"Correctly names retrieval but overgeneralises it into the sole mechanism, which would wrongly predict that worked examples never help beginners either."},{"text":"Because processing an explanation of something already known consumes working memory and returns nothing.","correct":true}]},"prev":{"slug":"explaining-it-back","title":"Explaining it back","url":"https://addaly.com/learn/ai-for-students/explaining-it-back"},"next":{"slug":"a-session-that-works","title":"A study session that works","url":"https://addaly.com/learn/ai-for-students/a-session-that-works"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Worked examples, and when to stop using them', in AI for Students. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-for-students/worked-examples-and-fading (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Worked examples, and when to stop using them","courseTitle":"AI for Students","module":{"index":3,"slug":"study-that-raises-difficulty","title":"Study that raises the difficulty"},"answer":null,"keyIdea":"Worked examples help beginners and hinder people who already have the method, so fade from full example to blank problem by removing steps from the end, and cap how many examples you read before moving on.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"beginners-and-experts-need-opposite-things","runs":[{"text":"Beginners and experts need opposite things"}]},{"t":"p","runs":[{"text":"When you are new to a type of problem, studying a fully worked example teaches you more per minute than attempting the problem yourself. This is well established, and it is the one place where being shown the answer is the right move."}]},{"t":"p","runs":[{"text":"The reason is capacity. Working memory holds only a few items at once. A beginner attempting an unfamiliar problem spends all of it on searching (what do I do first, is this the right formula, what was that rule) and has nothing left for noticing the structure. A worked example removes the search, so attention goes to the pattern."}]},{"t":"p","runs":[{"text":"But this reverses. The same worked examples that help a beginner actively slow down someone who already knows the method: they now have to process an explanation of something they can already do, which costs capacity and gives nothing. This reversal has a name, the "},{"text":"expertise reversal effect","b":true},{"text":", and it is why study advice that ignores your current level is unreliable."}]},{"t":"h2","id":"fading-which-is-the-actual-technique","runs":[{"text":"Fading, which is the actual technique"}]},{"t":"p","runs":[{"text":"The right sequence is not "},{"text":"examples","i":true},{"text":" or "},{"text":"problems","i":true},{"text":". It is a fade from one to the other:"}]},{"t":"ol","start":1,"items":[[{"text":"Full worked example.","b":true},{"text":" Every step shown, with the reason for each."}],[{"text":"Completion problem, last step removed.","b":true},{"text":" You do the final step."}],[{"text":"Last two steps removed.","b":true}],[{"text":"Only the first step given.","b":true}],[{"text":"Nothing given.","b":true}]]},{"t":"p","runs":[{"text":"Each stage is a small generation attempt at the point you are most ready for it. This structure, "},{"text":"backward fading","i":true},{"text":", because the removal starts from the end, outperforms both pure examples and pure problem solving in controlled comparisons."}]},{"t":"p","runs":[{"text":"Models are unusually good at generating this, and almost nobody asks them to:"}]},{"t":"quote","runs":[{"text":"Give me a worked example of this type of problem with the reasoning at every step. Then give me the same type of problem with the last step missing for me to complete. Then one with the last two missing. Then one with only the first step given. Then one with nothing. Wait for my answer at each stage before continuing."}]},{"t":"p","runs":[{"text":"That is a complete, well-evidenced practice sequence, generated in one message."}]},{"t":"h2","id":"study-the-example-properly-which-most-people-do-not","runs":[{"text":"Study the example properly, which most people do not"}]},{"t":"p","runs":[{"text":"Reading a worked example passively gets you very little. The version that works:"}]},{"t":"ul","items":[[{"text":"After each line, cover the rest and predict the next step before revealing it."}],[{"text":"Write, beside each step, "},{"text":"why","i":true},{"text":" that step is legal. Not what it does: why it is allowed."}],[{"text":"At the end, close the example and write the sequence of moves from memory, in words, without numbers. This is the pattern you are actually trying to acquire."}]]},{"t":"p","runs":[{"text":"That last one is the difference between learning to do "},{"text":"this","i":true},{"text":" problem and learning to recognise this "},{"text":"kind","i":true},{"text":" of problem."}]},{"t":"h2","id":"knowing-when-to-stop","runs":[{"text":"Knowing when to stop"}]},{"t":"p","runs":[{"text":"The signal that you have crossed out of the worked-example stage is that you can predict the next step correctly before revealing it, twice in a row. From that point, examples cost you time. Move to completion problems immediately, and to blank problems soon after."}]},{"t":"p","runs":[{"text":"Students overstay in this stage badly, because reading examples is comfortable and pleasant and feels like work. The number of worked solutions somebody has read is close to uncorrelated with what they can do; a solution read is a solution you can follow."}]},{"t":"h2","id":"the-trap-specific-to-ai","runs":[{"text":"The trap specific to AI"}]},{"t":"p","runs":[{"text":"A model will produce an unlimited supply of worked examples on demand. That makes it easy to spend an entire evening in stage one, which feels productive and is barely study at all. The fix is procedural: "},{"text":"decide in advance how many worked examples you will read before moving to completion problems.","b":true},{"text":" Two is usually right; three if the material is new. Then move, whether or not you feel ready, because feeling ready is not a reliable signal here."}]},{"t":"h2","id":"checking-that-the-example-is-right","runs":[{"text":"Checking that the example is right"}]},{"t":"p","runs":[{"text":"One caution the classic research did not have to deal with. Worked examples in a textbook were checked. A generated one was not. In any subject where the steps matter, verify the example against a source before you learn from it: a textbook problem, a past paper with a mark scheme, a symbolic tool like SymPy or Maxima for the algebra. Learning a wrong worked example is worse than learning nothing, because you will apply it confidently."}]},{"t":"p","runs":[{"text":"The safest arrangement: take the worked example from your course materials, and use the model to generate the "},{"text":"fading sequence","i":true},{"text":" of practice problems around it."}]}],"check":{"question":[{"text":"A student who can already solve a class of problem is given detailed worked examples of it. Why does this measurably slow them down?"}],"options":[{"text":[{"text":"Because it reduces motivation, and material that feels too easy lowers the effort they put into the session."}],"correct":false,"tempting":[{"text":"Substitutes a motivational story for a capacity one; the effect appears even when students are fully engaged and trying."}]},{"text":[{"text":"Because worked examples always show one particular method, which interferes with the method the student had already learned."}],"correct":false,"tempting":[{"text":"Interference is real when methods genuinely differ, but the reversal effect appears even when the example uses exactly the student’s own method."}]},{"text":[{"text":"Because reading a solution suppresses the retrieval attempt, and retrieval is the only mechanism by which practice strengthens memory."}],"correct":false,"tempting":[{"text":"Correctly names retrieval but overgeneralises it into the sole mechanism, which would wrongly predict that worked examples never help beginners either."}]},{"text":[{"text":"Because processing an explanation of something already known consumes working memory and returns nothing."}],"correct":true,"tempting":null}],"order":[0,3,2,1],"why":null},"prev":{"slug":"explaining-it-back","title":"Explaining it back","minutes":8},"next":{"slug":"a-session-that-works","title":"A study session that works","minutes":9}}}