Common Myths about Cognitive Load Theory

The cognitive load theory (CLT) is not a plea for effortless lessons. It is a design lens for deciding where effort should go.

Photo by Bret Kavanaugh on Unsplash

Cognitive Load Theory (CLT) is often invoked in teaching, sometimes in ways that drift from what learning sciences actually claims. CLT is a theory of how limited working memory interacts with long term memory during learning, so the practical goal is not “low load” but “the right load in the right place” for a given learner and task (Sweller, 1988; Sweller et al., 2011).

Educators hear “cognitive load” and often get a simple takeaway: “make things easier”. The trouble is that “easier” can mean two very different things. You can remove pointless friction so students can think about the right thing or you can remove the thinking itself and end up with smooth lessons that do not stick. In practice, CLT is most helpful when it helps you spot avoidable burdens, then decide what challenge you actually want students to face.

Myth 1: “High cognitive load is bad”

This is the most common misunderstanding. CLT does not say mental effort is harmful. It says some mental effort is wasted. Sometimes “higher” load is unavoidable and often necessary because the material itself is complex.

A worksheet that forces students to flip pages, decode messy instructions or hunt for the key information creates the wrong kind of load. A challenging question that makes them compare, justify and connect ideas creates the kind of load that can produce learning.

In digital and online learning, this distinction becomes sharper. Some features can raise students’ reported load yet still support learning, depending on whether they help alignment to the learning goal or distract from it (Skulmowski & Xu, 2022). The point is not “keep load low”. It is “keep the load relevant”.

If students are so-called tired, the fix is not always to simplify the content. Sometimes the fix is to tidy the delivery so the difficulty sits in the concept, not the formatting. Even a so-called “crowded” infographic could work well if your instructions are clear for students to know what to do with it.

Myth 2: “The solution is always to simplify”

Teachers are told to chunk, scaffold and “reduce intrinsic load”. Sound advice, until it becomes a habit of flattening a subject. Some ideas are hard because many elements interact. Strip away the interaction and you can end up teaching a version of the topic that does not behave like the real one.

Basically, not everything has to be “explain-to-a-five-year-old” content. Even Richard Feynman does not superficially make his explanation plain. He spent time thinking about the best ways to deliver the concepts. A better framing is to keep the intellectual work, remove the accidental work. You can teach a complex concept but you design so the complexity lives in the concept, not in split attention resources or unclear steps (Sweller et al., 2011).

Myth 3: “Scaffolding is always good, so keep it”

Scaffolding is meant to be temporary, yet classroom scaffolds often become permanent furniture. CLT predicts an uncomfortable pattern, in which guidance that helps novices can hinder learners who already have some knowledge. At that point, the scaffold turns into redundancy, slows decision-making and reduces the need for retrieval and self-checking.

Recent synthesis work on expert scaffolding in visual problem-solving highlights scaffolding as regulation of information flow in working memory, which matters but also signals a design responsibility. Scaffolding needs timing, not just presence (van Nooijen et al., 2024).

In other words, scaffolds should expire. If they never fade, students never have to organise knowledge for themselves.

Myth 4: “AI reduces cognitive load, so students will be less critical”

This claim treats cognitive load as one thing and assumes that any reduction must weaken thinking. If AI removes extraneous friction (clarifying a brief, translating jargon, organising steps) it can free capacity for evaluation and explanation rather than replacing them. Experimental evidence (Deng et al., 2025) synthesised across studies suggests that ChatGPT-supported interventions often reduce reported mental effort while also improving learning outcomes, including higher-order thinking propensities.

The real risk is not “AI lowers load” but “AI becomes the thinker”. When students use AI mainly to obtain finished answers, they shift core generative work out of working memory and reduce opportunities for retrieval, self-explanation and error checking. That is a dependence story, not a “AI inevitably kills thinking” story.

A practical CLT-aligned stance is to use AI to reduce extraneous load, then deliberately reinvest the freed capacity into the desired thinking. One clean design move is to require an initial student attempt, then use AI for critique or counterargument, so the student still has to justify, revise and reconstruct reasoning. When AI is configured to ask questions rather than supply answers, it can support this role.

Myth 5: “You can measure cognitive load with one quick rating”

Single-item mental effort scales are popular and sometimes informative, but they are not diagnostic. A learner’s “this felt hard” can reflect intrinsic complexity, extraneous design problems, low prior knowledge, low confidence or even productive struggle. The classic one-item approach (mental effort ratings derived from Paas’s work) is widely used, yet even careful measurement papers emphasise its limits as a standalone indicator of *which load is present and why *(Klepsch et al., 2017).

This is why later instruments explicitly separate types of load (intrinsic, extraneous and germane) rather than treating load as a single quantity to minimise. More broadly, research such as Makransky et al. (2019) comparing subjective ratings with objective measures (for example eye tracking and EEG) shows that different measures can diverge, so triangulation matters if you are making design decisions.

A practical classroom implication is that “high reported load” should trigger a diagnosis, not an automatic simplification. Look for where attention is being spent (instructions, navigation, decoding representations, uncertainty about steps). Then check learning indicators that reflect schema building such as error types, transfer, delayed quiz performance or students’ ability to explain their choices. If those indicators are improving, the load may be doing useful work even if it feels effortful.

Final Thoughts

Taken together, these myths point to a more disciplined use of CLT: diagnose first, then design. It’s not enough by saying “Oh that looks crowded or dense, therefore cognitive load is high”. Start by checking where working memory is being spent. If attention is leaking into navigation, unclear instructions or decoding messy representations, fix the delivery. Then decide what cognitive work you want students to do and design so that effort lands there, even if it feels demanding in the moment.

Let scaffolds fade as knowledge grows and treat AI as a tool for removing friction, not as a substitute for generative work. Finally, treat “high load” reports as a signal to investigate, alongside learning evidence such as error patterns, transfer and delayed performance. The goal is not low load. It is purposeful load that moves learners from fragile performance to durable understanding.

Further Readings

Deng, R., Jiang, M., Yu, X., Lu, Y., & Liu, S. (2024). Does ChatGPT enhance student learning? A systematic review and meta-analysis of experimental studies. Computers & Education, 105224. https://doi.org/10.1016/j.compedu.2024.105224

Klepsch, M., Schmitz, F., & Seufert, T. (2017). Development and validation of two instruments measuring intrinsic, extraneous, and germane cognitive load. Frontiers in Psychology, 8. https://doi.org/10.3389/fpsyg.2017.01997

Makransky, G., Terkildsen, T. S., & Mayer, R. E. (2019). Role of subjective and objective measures of cognitive processing during learning in explaining the spatial contiguity effect. Learning and Instruction, 61, 23–34. https://doi.org/10.1016/j.learninstruc.2018.12.001

Skulmowski, A. & Xu, K. M. (2022). Understanding cognitive load in digital and online learning: A new perspective on extraneous cognitive load. Educational Psychology Review, 34(1), 171–196. https://doi.org/10.1007/s10648–021–09624–7

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.

Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. Springer.

Tetzlaff, L., Simonsmeier, B., Peters, T. & Brod, G. (2025). A cornerstone of adaptivity: A meta-analysis of the expertise reversal effect. Learning and Instruction, 98. https://doi.org/10.1016/j.learninstruc.2025.102142

van Nooijen, C. C. A., de Koning, B. B., Bramer, W. M., Isahakyan, A., Asoodar, M., van Merriënboer, J. J. G., Kok, E. M. & Paas, F. (2024). A cognitive load theory approach to understanding expert scaffolding of visual problem-solving tasks: A scoping review. Educational Psychology Review, 36, Article 12. https://doi.org/10.1007/s10648-024-09848-3