VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

How Redundancy Works in Learning | When More Information Creates More Load

Three students study around open books and worksheets in a shared classroom.

eduKateSG Learning Node Series · 0011

More explanation feels safer. More labels feel clearer. More examples feel more helpful. More copies of the same message feel as though they should make learning easier.

Sometimes they do.

Sometimes they create a second job: the learner must now coordinate information that adds no new meaning.

A diagram appears on screen. A narrator explains it. The exact same paragraph appears beside it. The learner reads, hears and looks simultaneously. Nothing is missing. Yet attention is divided among three representations whose verbal content partly duplicates itself.

The redundancy principle asks a counterintuitive question: when does additional information stop being support and start becoming processing overhead?

Quick Read: Redundancy Is Not Repetition

The 2021 Cambridge Handbook of Multimedia Learning describes redundancy broadly as unnecessary information, including information repeated in multiple forms or elaborated beyond what the task requires. Cognitive load theory predicts that coordinating redundant material with essential material consumes working-memory resources that could otherwise support learning.

A classic multimedia example compares graphics plus spoken narration with graphics plus narration plus identical on-screen text. In many studied conditions, the extra printed copy can reduce learning because the eyes must process both the graphic and the written verbal stream while the same words are also being heard.

Redundancy is not “seeing something twice.” It is processing something again when the second representation contributes too little new value for its cognitive cost.

The Photocopier Problem

Imagine a student who does not understand a difficult explanation. The natural response is to give the student the same explanation again in more places.

The teacher says it. The slide contains it. The worksheet repeats it. The revision guide repeats it again.

If the problem is that the explanation is structurally weak, duplication does not repair the structure. It multiplies the same weakness.

The student may need a different representation, a worked example, a diagram, a simpler prerequisite, a contrast case or a retrieval prompt—not another copy of the same paragraph.

Why Redundancy Can Feel Helpful

Repeated information creates familiarity. Familiarity feels fluent. Fluency feels like understanding.

This is one reason redundancy can survive unnoticed. A learner sees the same wording in the notes and hears the teacher say it. The sentence becomes easy to recognise, and the learner interprets that ease as mastery.

But if the learner cannot explain the relationship without the supplied wording, the duplicated information may have created recognition rather than independent knowledge.

Useful Repetition Has a Different Job

Education needs repetition. Retrieval practice, successive relearning and spaced return all depend on repeated encounters.

So why is repetition sometimes good while redundancy is sometimes harmful?

The answer is timing and cognitive job.

Spaced retrieval repeats the target across time while requiring the learner to reconstruct it. Redundant presentation often duplicates the same information concurrently, asking the learner to process multiple versions at once.

One creates a new learning event. The other may create a coordination cost.

See How Successive Relearning Works for the distinction between repeated exposure and repeated successful reconstruction.

The Narrated Slide Trap

One of the clearest redundancy failures appears in presentations.

A slide contains a diagram and a paragraph. The presenter reads the paragraph aloud word for word.

The learner can listen, read or inspect the graphic, but trying to do all three simultaneously can create unnecessary competition. The eyes move from sentence to diagram and back again while the spoken stream continues whether the learner is ready or not.

A cleaner design may use the visual representation plus concise labels while the presenter speaks the explanatory relationships. Detailed text can remain available in a handout or transcript for later inspection.

Redundancy and Modality

Redundancy is tightly connected to modality.

If spoken explanation allows the learner’s eyes to stay on a meaningful visual, modality can reduce visual competition. Add an identical paragraph on screen and the learner may again have to divide visual attention.

But this does not mean on-screen text is always wrong. The correct design depends on whether the text is necessary, whether the learner can process the narration reliably, and whether accessibility requires a persistent verbal representation.

See How Modality Works in Learning.

Accessibility Is Not Redundant

A design principle should never be used to remove access.

Captions may duplicate spoken words, but for a deaf or hard-of-hearing learner they are not redundant. A transcript may repeat narration, but it may be essential for a learner processing speech less reliably or studying in a noisy environment.

The definition of redundancy depends on the learner’s route to the content. Information is redundant only if it is unnecessary for that learner in that task.

This is why accessibility must be treated as part of instructional architecture rather than an optional layer added after cognitive design.

The Expertise Reversal Problem

Information that helps a novice can become redundant for an expert.

A beginner solving equations may need every transformation explained. A more advanced learner may find those explanations intrusive because the underlying moves are already automated. The extra guidance now occupies attention without adding meaningful structure.

This is one expression of the expertise reversal effect: instructional support should change as learner knowledge changes.

The lesson is not “less help is better.” It is “help has to retire when its job is finished.”

Redundancy in Mathematics

Mathematics can become overloaded by explanations that repeat what the notation already makes obvious to the target learner.

For a novice, the line “subtract 3 from both sides to preserve equality” may be essential. For an expert, narrating every arithmetic micro-step can conceal the strategic decision that actually matters.

Good mathematics teaching therefore changes explanatory resolution. Early instruction can expose the mechanism. Later instruction compresses routine steps and restores attention to problem structure, method selection and verification.

Redundancy in Science

Science diagrams often suffer from text duplication.

A labelled diagram may already show that blood moves from the right ventricle to the pulmonary artery. A nearby paragraph that restates only that visible fact adds little. A better explanation might instead answer the causal question the diagram cannot show: why pressure changes, how valves constrain direction, or what oxygenation state changes next.

The verbal layer should add interpretation, causality or evidence—not simply narrate what the eye can already read directly.

Redundancy in English

English teaching can become redundant when students are given a model answer, a teacher paraphrase of the same answer, a summary of the paraphrase and a checklist that repeats the same points.

The learner may need variety instead: one model, one explanation of why it works, one contrasting weak version and one independent attempt.

Each representation has a distinct job. Duplication becomes transformation.

Redundancy in Vocabulary Learning

A vocabulary card can contain the word, dictionary definition, simplified definition, translation, synonym, image, mnemonic, etymology, pronunciation guide and three example sentences.

That richness may help one learner and overwhelm another.

The question is which representations add distinct information. A pronunciation recording adds sound. A good example adds usage. A contrast word sharpens the semantic boundary. Three definitions that say almost the same thing may not add enough value to justify their processing cost.

Redundancy in Note-Taking

Students often create multiple notebooks containing the same content because rewriting feels productive.

Copying can support attention, but repeated transcription without transformation may become redundant work. The student owns three beautiful versions of the same notes but has not retrieved, reorganised or applied the knowledge.

A better second pass changes the cognitive job: turn notes into questions, compress them into a map, reconstruct them from memory or apply them to a new problem.

Redundancy in Homework

Twenty nearly identical questions can be useful when fluency is the bottleneck. The same twenty questions can become redundant when the learner already performs the operation accurately and the real weakness is method selection or transfer.

Practice volume should follow the learning state.

If the learner needs automaticity, repetition may be the job. If the learner needs discrimination, vary the problem. If the learner needs transfer, change the surface. If the learner needs diagnosis, choose questions that separate competing hypotheses about the weakness.

Redundancy and Feedback

Feedback can also become redundant.

If a student already knows the error and knows how to repair it, repeating the same explanation after every attempt may add little. If the learner continues making the error for different reasons, the same feedback may be too generic.

Good feedback changes with evidence. Early feedback may explain. Later feedback may only cue. Eventually, the learner should self-detect and self-correct.

The Difference Between Reinforcement and Duplication

Reinforcement strengthens a route. Duplication merely adds another copy.

A second example can reinforce a principle if it changes the surface and preserves the structure. A second retrieval can reinforce access if it happens after spacing. A second explanation can reinforce understanding if it uses a genuinely different representation.

The key question is: what new operation does the second encounter require?

Redundancy and Coherence

Redundancy is one route into incoherence.

A page can be cluttered not because it contains unrelated information but because it contains the same information too many times. The learner must keep checking whether each version is identical, complementary or contradictory.

Removing redundant material can therefore sharpen the central route without reducing essential content.

See How Coherence Works in Learning.

Redundancy and Spatial Contiguity

Sometimes information looks redundant because the learner cannot tell which explanation belongs to which part of a visual.

A diagram may have a legend far away, forcing repeated search. Designers then add extra labels, arrows and explanatory text to compensate. The underlying problem is not missing information; it is poor placement.

Series 0012 examines this in How Spatial Contiguity Works.

The AI Redundancy Problem

AI can produce unlimited explanation at negligible marginal cost.

That changes the bottleneck. The learner no longer struggles to obtain another explanation. The learner struggles to decide whether another explanation is needed.

A useful AI tutor should therefore not respond to every uncertainty with more prose. It can ask for the learner’s current model, identify the missing link, give the smallest sufficient repair and then return the problem to the learner.

More output is not automatically more teaching.

The Parent Redundancy Problem

Parents can accidentally add a second teaching system on top of the first.

The school has one method. The tuition centre has another. The parent introduces a third shortcut from a video. Each may work, but the child must now coordinate multiple representations before the underlying concept is stable.

Sometimes the better move is not another method. It is to understand one method deeply enough to use it reliably, then compare alternatives when the learner has a stable base.

The Tutor Redundancy Problem

An experienced tutor can see five ways to explain the same idea.

That expertise is valuable, but giving all five explanations at once can overwhelm the learner. Use one representation. Test whether it repairs the weak link. Add another only if the first leaves a specific gap.

Explanations should be diagnostic interventions, not a display of explanatory range.

A Teacher Redundancy Audit

  • Does the on-screen text duplicate the narration word for word?
  • Does the diagram explain something the paragraph already states clearly—or vice versa?
  • Does each example expose a new case or merely repeat the same surface?
  • Are explanations still present after the learner no longer needs them?
  • Could one representation be moved to a later reference resource?
  • Are accessibility supports being misclassified as redundant?
  • Would a different representation add more value than another copy?

A Student Redundancy Audit

  • How many versions of the same notes have I created?
  • Am I rereading because I need information or because recognition feels comfortable?
  • Does this second resource explain the idea differently?
  • Can I replace another pass of reading with retrieval?
  • Can I replace another identical question with a varied one?
  • Can I compress three similar explanations into one model?

The Compression Test

Try removing one representation.

What learning value disappears?

If the learner loses a necessary label, example, access route or causal explanation, the material was not redundant. Put it back.

If nothing important disappears and the remaining material becomes easier to process, the removed element was probably costing more than it contributed.

This is a practical definition of redundancy: removable information whose absence improves the learning route without removing essential meaning.

What the Research Does and Does Not Say

The redundancy effect has a substantial research history, but it is not a universal command to remove duplicated information. The 2021 Cambridge Handbook chapter by Kalyuga and Sweller explicitly discusses conditions of applicability. Learner expertise, task complexity, presentation format and necessity of the duplicated source all matter.

Modern multimedia research also operates in a world of captions, replay controls, mobile screens, second-language learning and assistive technologies. These conditions can change whether information is genuinely redundant.

The right use of research is not “never repeat.” It is “make every representation justify its cognitive cost.”

The Deeper Principle: Addition Should Create New Information or New Action

Whenever you add something to a lesson, ask what becomes possible that was not possible before.

Does the new diagram reveal spatial structure? Does the second example expose a boundary? Does the label reduce search? Does the narration free the eyes? Does the transcript provide access? Does the retrieval question force production?

If the addition changes nothing except the amount of material, it may be redundant.

Education does not improve by adding indefinitely. It improves when each addition has a job.

Use This Tomorrow

Choose one slide, note page or revision resource. Identify every repeated idea. For each repetition, ask whether it adds a new representation, new access route, new example, new decision or new retrieval event. If it does not, remove or relocate it and see whether the central model becomes easier to follow.

Research and Further Reading


eduKateSG Learning Node Series · 0011 of the continuing series. Previous: 0010 — How Coherence Works in Learning.