Dual Coding

Dual coding is the learning-science principle that information encoded through two channels simultaneously — verbal and visual — produces stronger and more durable memory than information encoded through either channel alone. First formalized by Allan Paivio in 1971 and extended through decades of research into multimedia learning, dual coding is the reason a well-designed diagram paired with spoken explanation outperforms either a pure lecture or a slide full of text read aloud. For workshop creators, it is the structural principle that turns slides and visuals from decoration into a second retention engine running alongside the presenter's voice.

Why It Matters for Workshops

Most workshops run on a single channel of information. A presenter talks; the slides behind them contain bullet points that more or less summarize what the presenter just said. Some sessions are all verbal and no visual — just a voice in a video call. Others are all visual with minimal verbal — slides overflowing with content the presenter paraphrases. Both ends of this spectrum are single-channel delivery, and both leak retention for the same reason: one of the two memory systems that durable learning depends on never got activated.

The brain is built with two parallel systems for encoding new information. One processes language — the words you hear and read. The other processes imagery — the pictures, diagrams, shapes, and spatial relationships you see. These systems run concurrently and store memories in partially independent ways. When a concept is encoded through both systems, the brain builds two memory traces instead of one, and the two traces cross-reference each other during retrieval. This is what makes a participant able to recall the specific diagram they saw when they're asked about a concept they last heard about three weeks ago — the visual trace acts as an anchor for the verbal one.

Dual coding reframes what slides, visuals, and handouts are for. They are not decorations, and they are not handouts to refer back to later. They are a second encoding pass on the same material, happening in real time, alongside the presenter's voice. A workshop that ignores the second channel is a workshop that handed the brain one hook when it could have handed it two. The content doesn't have to change. The channel coverage does.

Composing Across Two Channels

The workshop creator most people recognize is the presenter — the person at the front of the room, talking. What gets overlooked is that every session is also a composition. At any given moment, participants are receiving information through two separate and simultaneously active channels in their brains: the verbal channel (what they are hearing) and the visual channel (what they are seeing). A session that only attends to one of those channels is doing half the job, whether the presenter knows it or not. The move this article points at is deliberately designing both.

Most workshop creators are fluent in the verbal channel. They can explain the material, field questions, tell stories, and hold a room on voice alone. The visual channel is typically where the apprenticeship lives. Slides become afterthoughts, diagrams get added as decoration, and the second channel plays a backup role at best. This is not a flaw of the presenter. It is the predictable result of a workshop-design culture that treats the visual as an accessory to the verbal rather than as its equal. Dual coding makes the second channel mandatory, not optional.

The discipline is to decide what each channel carries — separately. In any given moment of the session, what should the verbal channel be delivering, and what should the visual channel be delivering at the same time? Sometimes the two reinforce each other: the verbal names the concept; the visual diagrams its structure. Sometimes they divide labor: the verbal provides narration; the visual provides data the narration can't carry in real time. What they never do, in a well-composed session, is duplicate. Verbal-visual duplication — the presenter reading the slide text out loud — collapses dual coding back into single-channel delivery and wastes the second channel entirely.

You are a master of your content. The second craft — composing the two channels so they complement rather than compete — is what this article is about, and it multiplies the retention of the same material without changing the content at all.

The Learning Science

Allan Paivio, a Canadian psychologist at the University of Western Ontario, introduced dual coding theory in his 1971 book Imagery and Verbal Processes and refined it across three decades of follow-up research, culminating in his 1991 paper Dual Coding Theory: Retrospect and Current Status. Paivio proposed — and then demonstrated empirically — that the brain has two separate, functionally independent systems for encoding new information. The verbal system processes language: words read, words heard, linguistic meaning. The visual system processes concrete imagery: pictures, diagrams, scenes, spatial relationships. Both systems can store memories independently, and both can be activated during retrieval, which means a memory encoded through both channels has two pathways back to the surface instead of one.

The retention consequence of dual coding is not additive. It is multiplicative. A memory with two independent traces is more resistant to decay than a memory with one, because even if one channel's trace weakens, the other can still retrieve the memory and reinforce the missing side. This is why studies comparing verbal-only instruction, visual-only instruction, and dual-channel instruction consistently find the same pattern: the dual-channel condition outperforms either single-channel condition, and the effect is largest on delayed tests — exactly when retention actually matters for behavior change.

Richard Mayer built on Paivio's work to develop the cognitive theory of multimedia learning, summarized in his Multimedia Learning (2nd ed., 2009). Mayer's research identified specific design principles that amplify dual-coding gains. Visuals and narration should present complementary information, not the same information in two forms (the multimedia principle). They should appear at the same time (the temporal contiguity principle). And on-screen text should not duplicate what the presenter is saying (the redundancy principle). The last one matters most for workshop design. Slides packed with bullet points that the presenter reads aloud produce worse retention than either the slide alone or the narration alone — because the verbal channel is running over itself, processing the same content twice in two slightly misaligned versions, instead of being paired with a genuinely complementary visual.

What Paivio and Mayer documented is something good multimodal composition has always operated on intuitively: two channels running in parallel can do more than two channels running the same content. The research gives the principle; the practical move it suggests is a better starting question. "What should the second channel carry that my voice can't?" outperforms "how do I make better slides?" every time.

How to Recognize the Problem

Observable signs that the session is composed for one channel when the brain was ready to receive on two:

• Your slides contain bullet-point lists of the same content you're saying aloud
• Participants describe the session as "a lot of slides" or "a lot of PowerPoint" rather than "a lot of content" — they noticed the visuals but the visuals didn't carry their weight
• When a participant tries to explain the concept weeks later, they can recall words but not any mental image that anchors those words
• Sessions that go remote (camera-only or audio-only) produce noticeably worse retention than the same material delivered in person — a clue that the visual channel was doing something the pure-audio version can't replicate
• Your slide deck reads as a standalone document, complete without you; if the deck and the narration can each stand alone, they are duplicating, not composing
• You spend hours building elaborate slides and then rarely point at them during the session — they function as visual wallpaper rather than active channel
• Recap decks and reference PDFs get sent after the session, but participants rarely open them, because the material inside is the same as the talk in document form

These aren't signs of bad design taste. They are signs of a session composed for one channel when the brain was ready to receive on two.

Common Mistakes

Reading the slide aloud. This is the most common composition error, and it is the audio-visual equivalent of putting both instruments on the same note. The presenter's voice says the exact words the slide displays, which is not dual coding — it is the verbal channel delivered twice in slightly misaligned timing. The visual channel gets spent on text, which is processed by the verbal system anyway, so the brain ends up with two slightly different versions of the same verbal content and no visual trace at all. Mayer calls this the redundancy effect. The fix is simple: if it is text, don't read it. If you are going to say it, the slide should show something else.

Treating visuals as decoration. Stock photos, logos, and icons used to "make the slides look better" without carrying information the verbal channel isn't also carrying leave the second channel effectively silent. Pretty visuals that don't teach are, from a dual-coding standpoint, indistinguishable from blank slides — the visual channel was available and nothing was sent down it. The discipline is that every visual should carry something: a diagram, a comparison, a spatial relationship, a quantity, a shape. Aesthetic alone is not enough.

Running both channels with identical content. Some presenters intuit that two channels is better than one and respond by showing visually what they're saying verbally — the full transcript of the talk appearing on screen, often highlighted as each sentence is spoken. This is an extreme version of reading the slide aloud, with the same core problem: the visual channel duplicates rather than complements. The brain does not encode in two channels when both channels are running the same word-stream.

Using only one channel throughout. On the opposite end, some sessions default to all-verbal (no slides, pure talk) or all-visual (slides with no narration). Story-heavy workshops often skew verbal-only because the presenter feels slides "break the flow." Reference-heavy workshops often skew visual-only because the presenter wants to "let the material speak for itself." Either default leaves the other channel silent for long stretches, halving the encoding capacity of the session during those stretches.

Composing only one side of the composition. Many presenters write their talk first, then build slides to accompany what they already wrote. Others design slides first and improvise the narration. Either sequence treats one channel as primary and the other as derivative. The move that produces real dual coding is designing both channels together, asking at each moment what the visual carries and what the verbal carries, so the two can be complementary rather than sequential.

Removing the second channel when the session goes remote. Many workshops that are dual-channel in person collapse to single-channel on a video call. The presenter's face replaces the slide, and the visual channel disappears. Remote workshops especially benefit from dual coding precisely because they lose the ambient visual information a physical room provides. A well-composed remote session uses the screen-share channel with complementary visuals as a continuous second track, not as occasional reference.

How to Design for Dual Coding

1. Decide what the visual channel carries before you write the talk. Start by asking, for each section of the session, what participants should be looking at while you are talking. The answer should not be "bullet points of what I'm about to say." It should be the thing the visual channel can do better than the verbal — a diagram that shows structure, a graph that shows quantity, a comparison that shows difference, a before-and-after that shows transformation. When the visual answer is "nothing specific," that is a signal the composition hasn't been designed yet, not that the section doesn't need a visual.
2. Keep the two channels complementary, not duplicate. The rule of thumb: if the slide could be read as a standalone document and still convey everything, the slide and the talk are duplicating. The slide should be something that needs the narration to fully make sense, and the narration should be something that needs the slide to fully land. Complementary composition means neither channel is complete without the other.
3. Match the timing deliberately. Visuals should be present when the corresponding verbal narration is happening — not thirty seconds earlier, not thirty seconds later. This sounds obvious and is routinely violated. Slides advance ahead of the narration or lag behind it, which causes the brain to encode verbal and visual content that don't actually correspond. Tighten the timing or slow the advance so each moment of audio lands against the visual it was composed for.
4. Design for retention, not for the deck. A slide deck is not the goal. Retention is. Slides that look elegant as a standalone document but fail to pair with the narration are decoration. Slides that look sparse out of context but carry exactly the complementary load the narration needs are doing their job. The question to ask at every moment is "does this work as a composition?" not "does this deck look good?"

Dual coding doesn't require more slides, better design taste, or new tools. It requires a compositional stance — deciding, deliberately, what each channel carries in every moment of the session, and making sure the two are working together instead of on top of each other.

Seeing and Saying: The Participant-Side Flip

Everything to this point has described dual coding from the presenter's side — what you build, what the screen shows, how the two channels harmonize when information is delivered. There is a deeper tier available, and it flips the picture. The most powerful use of dual coding in a workshop is not what the presenter does with their two channels. It is what the participant does with theirs.

Participants have the same two channels the presenter does. Neuroscientist Alan Baddeley, working with Graham Hitch, named the working-memory subsystems that make this possible: the phonological loop holds and manipulates verbal and auditory information, and the visuospatial sketchpad holds and manipulates imagery and spatial relationships (Baddeley & Hitch, 1974). Both subsystems can be actively engaged or left idle depending on what the session asks the participant to do. When a participant is writing in their own words, diagramming a process, sketching a plan, or laying something out on a page, their visuospatial sketchpad is active — they are seeing their own thinking take shape in front of them. When they are talking something through aloud, teaching it back to a partner, reasoning out a decision in words, or reading their own draft aloud, their phonological loop is active — they are saying their own thinking into the room. Encoding through both channels at the same time, on the participant's own side, is dual coding from the inside.

This is where a common misconception needs addressing. The "learning styles" idea — that some people are visual learners, others are auditory learners, and instruction should be matched to each — has been substantially debunked. Pashler, McDaniel, Rohrer, and Bjork's (2008) review found no reliable evidence that matching instruction to a preferred style improves outcomes. What is well-supported is the underlying architecture: every brain has both a phonological loop and a visuospatial sketchpad, and durable learning benefits from engaging both. The implication is not that you teach differently to different participants. The implication is that you design activities that engage both systems in every participant.

Practically: at every major encoding moment in the session, the participant should be asked to do both. Write it and talk it out. Diagram it and walk a partner through the diagram. Draft it and read the draft aloud. List it and teach it back. The pairings are flexible; the discipline is that both channels, on the participant's side, are active during the moments that matter. This is what separates a session where dual coding was used to present information from a session where dual coding was built into the work the participant actually does — and it is where dual coding compounds with retrieval practice, the 80/20 engagement ratio, and the producible-outcome discipline that runs through the rest of this series.

Where this goes next: Dual coding closes the Stage-2 encoding toolkit. Value opens the attention gate; cognitive load keeps the session within the capacity it opened; warm, active, scaffolded, dual-channel composition gets the material encoded into long-term memory. All of that is upstream of a problem every workshop faces the moment the session ends — decay. Even material that was perfectly encoded starts disappearing within hours. Stage 3 of the memory journey begins with naming that opponent. See The Forgetting Curve in Workshops.

Related

Read this next:

• The Forgetting Curve in Workshops — Stage 3 of the memory journey; the decay problem every workshop faces the moment the session ends, and the opponent the next two articles are about countering

Foundations of this concept:

• How People Actually Learn — the full memory pipeline dual coding runs inside
• Cognitive Load — the constraint that determines how much each channel can carry before the system overflows

Where this leads:

• Scaffolding and the Zone of Proximal Development — the sequencing architecture dual-channel composition sits inside
• The 80/20 Workshop Engagement Ratio — the in-session structure where composed dual-channel moments actually land
• Warm vs. Cold Information — visuals are often the fastest way to warm cold content before narration arrives

References

Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. 8, pp. 47–89). Academic Press.

Mayer, R. E. (2009). Multimedia learning (2nd ed.). Cambridge University Press.

Paivio, A. (1971). Imagery and verbal processes. Holt, Rinehart and Winston.

Paivio, A. (1991). Dual coding theory: Retrospect and current status. Canadian Journal of Psychology, 45(3), 255–287.

Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning styles: Concepts and evidence. Psychological Science in the Public Interest, 9(3), 105–119.