An instructional video adds on-screen text captions that exactly match the narrator's spoken words, an addition made with the reasonable-seeming intention of reinforcing the material through two channels simultaneously — and controlled studies comparing this version against an otherwise identical video using narration alone often find the narration-only version produces better learning outcomes, a counterintuitive pattern called the redundancy effect.
Why this runs counter to a very natural, intuitive assumption
The intuitive expectation that presenting the same information through two simultaneous channels should reinforce and strengthen learning, rather than undermine it, is reasonable on its face — more exposure to the same content, delivered redundantly, seems like it should help, not hurt. The redundancy effect research specifically challenges this intuition for the particular case of simultaneous, fully matching text and narration, finding that this specific combination can measurably reduce learning relative to narration alone.
What mechanism is thought to actually produce this counterintuitive result
Cognitive load theory's explanation centers on working memory's limited capacity — when a learner receives identical information simultaneously through both spoken narration and matching on-screen text, working memory has to process and attempt to reconcile these two redundant streams, consuming cognitive capacity on this reconciliation process rather than directing that same limited capacity toward actually comprehending and integrating the underlying content, which is precisely the germane cognitive load that produces genuine learning.
Why this specifically applies to identical, redundant information, not complementary information
The redundancy effect research specifically concerns fully duplicative information presented through both channels simultaneously — a distinct and different situation from complementary multimedia, where narration and visuals present different, mutually reinforcing information (narration explaining a process while a diagram shows the process's structure), which doesn't create the same redundant-processing burden and is generally associated with improved, not reduced, learning in the broader multimedia learning research literature.
Why this specifically matters for how instructional videos and e-learning materials get designed
A common, well-intentioned instinct in instructional design — adding captions or on-screen text that exactly mirrors spoken narration, intending to accommodate different learning preferences or simply reinforce the material — can inadvertently work against the goal it's meant to serve, specifically when the text and narration are fully redundant rather than complementary, a distinction that's easy to overlook when captions are added as a seemingly straightforward accessibility or reinforcement feature without considering this specific research finding.
What this means for designing multimedia instructional content
- Avoid presenting fully identical information simultaneously through both narration and on-screen text in instructional video design
- Use visuals and narration to present complementary, mutually reinforcing information rather than duplicative content covering the identical material
- Recognize that accessibility captions serve a genuinely different, important purpose from redundant on-screen reinforcement text, and shouldn't be removed for accessibility reasons based on this research
- Test instructional video formats directly where feasible, since the redundancy effect's size can vary depending on the specific content and audience involved
The redundancy effect is one of the more genuinely counterintuitive findings in multimedia learning research — a well-intentioned attempt to reinforce learning through repetition across two channels can instead overload the exact cognitive capacity that genuine comprehension actually depends on, making narration alone, in this specific case, the more effective design choice.