Remove information that can be inferred once the concept is known

Once a learner can process information independently, adding redundant support actually increases load rather than reduces it.

Why it works

The redundancy effect is a counterintuitive CLT finding: for learners who already have relevant schemas, presenting the same information in two formats (text + narration of the same words) imposes more load than presenting it once, because the learner must allocate capacity to reconciling or suppressing the redundant source. Scaffolding that helps novices becomes noise for more advanced learners, and removing it frees capacity for deeper processing.

How to do it

1. Identify any supporting material (labels, annotations, captions, step-by-step prompts) that was added for a beginner state.
2. Test whether you can now process the material without that support.
3. Remove the support and check whether comprehension holds; if it does, the support was redundant.
4. Progressively strip scaffolding as expertise grows, rather than keeping all aids indefinitely.

Evidence

The redundancy effect has been directly tested in multiple CLT experiments: learners with some prior knowledge performed worse — not better — when presented with integrated text+narration versus text alone, because the narration added no information and consumed processing resources. (rct)

The redundancy effect depends strongly on learner expertise; for novices, the same material may not be redundant. Misapplying this principle by stripping scaffolding too early impedes learning.

Sources

• Sweller, Ayres & Kalyuga (2011), Cognitive Load Theory, Springer
• Kalyuga et al. (1998), Levels of expertise and instructional design, Human Factors

Common mistake

Keeping full scaffolding indefinitely after the learner no longer needs it, creating the illusion of thorough instruction while actually increasing cognitive overhead.