Karma YogiKarma Yogi
Start
The Journal
JEE Preparation

11 September 2026
11 min read

A
Anish Guruvelli
Karma Yogi

Worked Examples vs Solving Cold: What JEE Prep Gets Backwards

Cognitive load theory has a genuinely counterintuitive answer to "should I study the solved example or attempt the problem cold": it depends entirely on how new the topic is to you, and the right approach on day one is the wrong approach three weeks later. A 2019 review of 20 years of research explains why, and names the specific point where you should switch.

Whether you should study a worked solution before attempting a JEE problem or attempt it cold depends on one thing: how new that specific topic still is to you, and the answer flips as you gain mastery on it. Sweller, van Merriënboer and Paas (2019, Educational Psychology Review, 31, 261-292) reviewed twenty years of cognitive load theory research and describe two effects that together explain this directly: the worked example effect, first reported by Sweller and Cooper (1985) in algebra, shows that studying a fully worked solution builds more transferable understanding than solving the equivalent problem cold - for a genuine novice on that material. The expertise reversal effect (Kalyuga et al., 2003, 2012) shows that as expertise on that same material grows, the benefit "first decrease[s] in size, then disappear[s], and can eventually reverse" - at which point cold problem-solving becomes the more productive use of time. JEE's syllabus is dozens of topics, each starting from zero at a different point in your two-year prep. The right study method is not one answer; it is a moving target per topic.

Why worked examples help - at first

The mechanism is about working memory, not motivation. A conventional problem forces a novice to hold the current problem state, the goal state, the difference between them, and every candidate operation, all at once - the review calls this "means-ends analysis," and it is expensive precisely when you have no prior structure in long-term memory to lean on. A worked example removes that load: it "focus[es] the learners' attention on problem states and associated operators (i.e. solution steps), enabling them to induce generalised solutions." You are not spending your limited working-memory capacity searching for a path: you are watching one, which frees capacity to actually notice the pattern in it.

This is the worked example effect, and the review calls the empirical evidence for it "very strong." It has been replicated across product-oriented examples (just the final solution), process-oriented examples (the reasoning steps), and modelling examples (a person visibly generating the solution) - different formats, same underlying mechanism.

Why the same advice becomes wrong later

The review is explicit about the constraint most summaries of this effect leave out: "Worked examples are less effective for high-expertise learners." The theoretical account is the expertise reversal effect. As you build real expertise on a specific topic, the many separate pieces of a problem - a formula, a substitution step, a sign convention - stop being separate pieces you have to juggle in working memory. They compress into a single retrieved chunk. At that point, an instructional method built to manage a novice's working memory load is managing a load that no longer exists, and it starts costing you something instead: time you could have spent actually retrieving and applying the solution yourself.

The review states the practical implication as plainly as a cognitive science paper states anything: "For example, worked examples benefit novices. With increasing knowledge, practice at solving problems becomes increasingly important rather than having negative effects." This is not "worked examples are good" or "worked examples are bad" - it is a genuine reversal, tied to your own current level on that specific topic, not to the topic's difficulty in general or to your overall JEE preparedness.

The bridge: completion problems, and the guidance-fading effect

The review names a specific middle step for the transition, developed by van Merriënboer and Krammer (1987, 1990): the completion problem - a partial worked solution the learner must finish, rather than a full solution to study or a blank problem to solve alone. "Worked examples are completion problems with a complete solution and conventional problems are completion problems with a partial solution" - so a course, or a single topic's practice sequence, can start with near-complete solutions and gradually withdraw more of the working before you are solving fully cold.

This progression - full worked example, then completion problems with steadily less filled in, then a full cold attempt - is what the review calls the guidance-fading effect: "For novices, additional information or particular activities such as studying worked examples may be essential. With increases in expertise, these same activities may become redundant and impose an unnecessary cognitive load. Past a certain point, studying worked examples may be counterproductive and they should be faded out and replaced by problems." That is a direct, topic-by-topic prescription, not a one-time study-method decision you make for JEE prep as a whole.

Stage on a given topic What to use Why
Just introduced, no prior structureFull worked examplesWorking memory is fully occupied holding problem state + goal + operators; a worked example removes that load
Recognise the pattern, shaky on executionCompletion problems (partial solutions to finish)Forces engagement with the steps you have not yet chunked, without re-imposing full novice-level load
Reliably solve it correctly, just slower than idealFull cold problems, timedContinued worked-example study at this stage is the expertise reversal effect working against you

Applying this to a JEE topic list

JEE Physics, Chemistry and Maths are each dozens of distinct topics with almost no shared timeline - you might be a genuine novice at Rotational Mechanics in October while already fluent at Kinematics from June. Cross-linking this with JEE Maths shortcuts and formulas: a shortcut is only safely usable once you already understand the full derivation it is compressing, which is exactly the "cross the expertise line first" sequencing this research describes for worked examples generally - a shortcut studied before the underlying method is understood is a worked example you never earned the transition past.

The practical read for a JEE study session: treat "should I look at the solution first" as a per-topic, per-week question, not a global study philosophy. A reasonable, though self-constructed, heuristic - not something the paper itself states as a number - is to track your own cold-attempt accuracy on a topic's practice problems: once it is reliably high on problems you have not seen the solution to first, the worked examples for that specific topic have likely stopped helping and started costing you retrieval practice you should be doing instead.

Where this is weak, or does not directly apply

  • The foundational experiments are decades old and mostly narrow in domain. The worked example effect was first demonstrated in algebra (Sweller & Cooper, 1985); the split-attention and redundancy effects came from geometry and biology diagrams. The review is a synthesis of that body of work, not a JEE-specific study, and applying it to JEE's actual syllabus is this post's extension, not a tested claim.
  • There is no universal number for "how much expertise is enough" to switch. The review describes the direction and mechanism of the reversal clearly; it does not give a measurable threshold (like "60% accuracy") at which a learner should stop using worked examples on a given topic. The cold-accuracy heuristic above is a reasonable operationalisation, stated as such, not a finding.
  • Good worked examples are harder to build than they look. The review notes they must avoid two other documented pitfalls - the split-attention effect (forcing a learner to mentally combine a diagram and separate text) and the redundancy effect (restating the same information twice, which the review says can actively hurt learning even though it feels harmless). A poorly designed worked example does not get the benefit this post describes at all.
  • The review also reports a "reverse worked example effect" in one specific domain (learning mathematical definitions): students required to generate their own response learned more than students shown the correct one. The general pattern is not universal across every kind of content.

Frequently asked questions

End of essay

- Anish Guruvelli

Common questions

What is the worked example effect, in one sentence?
For a genuine novice on a topic, studying a fully worked solution builds more transferable understanding than attempting the equivalent problem cold, because it removes the working-memory cost of searching for a solution path and lets attention go to the pattern in the steps instead.
What is the expertise reversal effect?
As a learner gains real expertise on specific material, instructional support that helped as a novice (like worked examples) first stops helping, then can actively hurt - because the working-memory load it was designed to reduce no longer exists at that expertise level, and the support now costs time better spent on retrieval practice.
So should I look at the solution before attempting a JEE problem, or not?
It depends on your current expertise on that specific topic, not on JEE problems in general. New topic: study the worked solution first. Topic you already recognise but execute shakily: use completion problems (partial solutions to finish). Topic you reliably solve correctly: attempt it cold - continuing to lean on worked examples at that stage is the expertise reversal effect working against you.
What is a completion problem?
A partial worked solution that the learner must finish, rather than a full solution to study or a blank problem to solve alone. It is a deliberate bridge between a full worked example and a full cold problem, developed by van Merriënboer and Krammer (1987, 1990) specifically to force engagement without reimposing novice-level cognitive load.
How do I know when I have crossed from "novice" to "should solve cold" on a topic?
The underlying research does not give a specific threshold - this is a genuine gap, not something this post is hiding. A reasonable, self-constructed proxy is your own cold-attempt accuracy: once you reliably solve a topic's problems correctly without having seen the solution first, continued worked-example study on that same topic has likely stopped helping.
Does this apply differently to JEE Main versus JEE Advanced?
The underlying cognitive mechanism does not distinguish between the two exams - it is about your expertise on a specific topic, not which paper you are ultimately sitting. In practice, JEE Advanced's more novel problem framing may mean you stay in the "recognise but execute shakily" stage longer on some topics even after Main-level fluency, which argues for completion problems rather than assuming full cold-solving readiness just because Main-level questions on that topic already feel easy.
Is this the same as the testing effect or retrieval practice?
Related but distinct. Retrieval practice research (like Karpicke and Blunt's work) shows that self-testing beats re-reading for retention generally. This research is more specific: it is about the right sequence of study formats (worked example -> completion problem -> cold problem) as expertise on one topic grows, which is really a claim about when retrieval practice should start on that topic, not whether it works.
What is the split-attention effect, and why does it matter for worked examples?
It occurs when a worked example forces you to mentally combine two separate sources - like a diagram and separate solution text - that only make sense together. That mental integration itself costs working memory and can cancel out the worked example's benefit. A well-designed example integrates the diagram and steps physically so no mental combining is required.
Can overusing worked examples actually hurt my JEE prep?
Yes, per the expertise reversal effect specifically - once you have genuine expertise on a topic, continued worked-example study is described as potentially counterproductive, not merely neutral, because it takes time away from the cold problem-solving practice that would actually be building further skill at that stage.
Is there real evidence for this, or is it just theory?
The worked example effect itself has, in the review's own words, "very strong empirical evidence" behind it, built up since Sweller and Cooper's original 1985 study. The expertise reversal effect is credited to a specific body of work by Kalyuga and colleagues (2003, 2012). Both are summarised, not originated, in the 2019 review this post draws from.
Does this mean I should stop using solved-example books for JEE?
No - it means their usefulness is front-loaded to when you are new to a specific topic, and should taper as you gain fluency on that topic specifically. The same book of worked examples is well-used on a topic you just started and poorly used as a crutch on a topic you have already mastered.