First-year calculus students tend to fall into two camps. One says proudly, “I never look at the answers; if I’m stuck, I keep thinking.” The other opens the solution the moment a new problem type appears. On unfamiliar question types, the second group often does better early on; later, once the types are familiar, the first group pulls ahead. Both are half right.
Research on problem-solving shows that the value of a worked example depends on where the learner is. Very useful at the start, a burden later. This article helps you tell which stage you are at and how to use solutions at each one.
Why beginners learn more from worked examples
Faced with a completely new problem type, beginners usually resort to trial and error: try this formula, no, try that one. All their attention goes into searching for a route, leaving almost none for noticing the structure of the problem: which clue in the question points to the method, which step is the crucial one.
A good worked example removes the search. The learner can spend their effort on understanding why each step was chosen. This is known as the worked example effect, and it has been observed in mathematics, physics, programming and accounting.
And why it eventually backfires
Once you know a problem type, reading full solutions becomes redundant: you process steps you already know, while what you actually need is practice in producing the method yourself. The finding that worked examples lose their advantage, and can even do worse than independent solving, for more experienced learners is called the expertise reversal effect.
So the right question is not “should I look at the solution?” but “for this problem type, am I a beginner or already fluent?”.
Four ways to use solutions, by stage
| Stage | Signs | How to use examples |
|---|---|---|
| New problem type | No idea where to start | Study two worked examples, self-explaining each step |
| Starting to recognise it | You follow solutions but get stuck alone | Example–problem pairs: one example, then one similar problem immediately |
| Can solve with hints | Stuck on one or two familiar steps | Fading: solutions missing the last step, then more |
| Fluent | You predict nearly every step | Solve independently, check answers only; move to mixed practice |
Studying a solution actively: three non-negotiables
- Predict the next step. Cover the rest with a sheet of paper, guess the next move, then reveal. Wrong guesses are fine; not guessing is the problem.
- Write the reason in the margin. One short line per step: “substitution because of the square root”, “debit because the asset increased”, “dictionary because we look up by key”.
- Find the trigger in the question. Underline the words in the problem statement that made the solver choose that method. This is what lets you recognise the type when it is phrased differently.
Passive reading, nodding from top to bottom, is close to useless. It creates a powerful “this is easy” feeling that evaporates the moment you face a blank page.
Compare two solutions to see the structure
An underused but powerful move: put the solutions to two problems that look different but share a structure side by side and ask where they are the same. Or do the reverse: take two problems that look nearly identical but need different methods, and ask which detail makes the difference.
In probability, for instance, drawing with and without replacement look almost the same on the page and are calculated quite differently. Ten minutes comparing the two solutions often teaches more than ten more problems of one kind.
Common mistakes when learning with answers
- Opening the answer after 30 seconds of being stuck. Once a type is familiar, set a threshold of a few minutes of genuine effort, then reveal only the step you are stuck on.
- Checking only the final answer. Matching the last number does not tell you where you went wrong or whether you were right by luck.
- Watching video walkthroughs straight through at speed. The instructor solves smoothly and the viewer assumes they could too.
- Never returning to problems you looked up. A problem you saw the answer to only counts as learned once you can redo it alone a few days later.
In practice: a 60-minute session on a new chapter
- 15 minutes: study two worked examples of one problem type. Predict steps, write reasons in the margin.
- 10 minutes: after each example, solve a similar problem straight away without looking back unless you must.
- 15 minutes: two problems with the last step or last two steps hidden. Complete the hidden part yourself.
- 15 minutes: two problems entirely on your own. Open the answer only when finished or truly stuck.
- 5 minutes: log in your error notebook which step stopped you and which clue in the question you missed.
Two days later, redo one of the problems from step one without opening anything. If you can, you have moved from “I understand when shown” to “I can do it”.
For tutors and teaching assistants
If you tutor or run problem classes, avoid solving in one smooth run and asking “any questions?”. Work one problem fully, stop halfway through the next and let students finish, give only the first move on the third. Withdraw support at the pace students start predicting steps correctly, not at the pace of the syllabus.
Something to do now: open the problem sheet for a course you are taking and label each problem type with one of the four stages above. Where you are still at stage one, use worked examples tonight without guilt; where you are at stage four, put the answers away.
Câu hỏi thường gặp
Is looking at the solution before trying a problem cheating yourself?
Not if you are new to that problem type and you study the solution actively. For beginners, struggling alone with an unfamiliar problem often costs a lot of effort and teaches less than analysing a good solution.
When should I stop relying on worked examples?
When you can correctly predict most of the next steps before looking. At that point the full solution is redundant, and you should switch to solving on your own and use answers only to check.
What does fading a worked example mean?
Start with a complete solution; on the next problem hide the last step and do it yourself; then hide the last two, until you are solving from scratch. Each hidden step becomes a retrieval attempt.
Do video walkthroughs work like written worked examples?
They can, provided you pause before each step and predict it. Watching straight through produces a strong feeling of understanding that often does not translate into solving a similar problem alone.
Should I copy worked solutions into my notes?
Verbatim copying does little. It is better to rewrite the solution your own way after closing the source, adding a short note on why each step was chosen.
