Motion adds, never replaces

Claim: animation, video and live demos are added to an explanation, never in place of it.

Grade: supported with caveat.
Evidence: review, strand 7; position, Known and Not known.
Check: deck-judgment (judgment): for each clip, animation or demo, does an explanation come with it?
Reader styles: casual audience keeps it and adds a recorded fallback for a live demo.

What it asks

Say what the viewer should see before the clip, the animation or the demo starts, and what it showed after.
Name the learning goal of each visual. A visual without a clear goal can hinder (Bos 2025, abstract).
Keep embedded clips short, and place an educational clip in the middle of the talk (Ljubojevic 2014, abstract).
For a live demo, show only what has moving parts, rehearse it, and keep a recording ready (Nederbragt 2016 and LornaJane 2017, conventions).
Say aloud what you type (Nederbragt 2016, convention).

Why

In 105 randomised trials (7,776 students), video that replaced other teaching gave g = 0.28 and video added to it gave g = 0.80 (Noetel 2021, abstract).

Animation against static pictures, self-paced, gave d = 0.37 and g = 0.23. The gain was larger when the change is the thing to learn. That result is held in the learning science review, linked from the review above.

Limits

In one controlled study, live coding did not teach more than static code (Shah 2023). Active live coding did not beat traditional live coding, though engagement was higher (Shah 2025, abstracts).
No study of a live demo inside a conference talk was found, and none compares a live demo with a recorded one.
An animated explainer and a good slide video mostly differ in engagement, not comprehension (Pichler 2023; Li 2023, secondary). Higher engagement went with slightly lower comprehension in Li 2023.
Students learned from explainer videos and overestimated how much (Bos 2025, about 20 students, no control group).
Watching time is not learning. The 6-minute rule comes from engagement logs in four edX courses (Guo 2014).