Slides are a 'visual aid,' not a script: the glance medium + the redundancy principle
First, correct the most fundamental role confusion: slides exist to aid you (the speaker), not to move everything you want to say onto the screen.
Garr Reynolds (Presentation Zen) and Nancy Duarte (slide:ology) both stress that a slide is a 'glance medium' — a slide should be graspable in about 3 seconds, after which the audience hands their attention back to you, the person.
So the principle is: **put what should be seen on screen (an image, key data, a single claim), and say what should be heard (the explanation, the story, the connections).** Piling whole speaker notes or big blocks of text onto the screen lets the slide steal your role; anything that needs detailed text should be handed out as a post-talk handout, not stuffed into the live slides.
The redundancy principle (Mayer 2009): don't turn on-screen text into the words you're already saying.
Cognitive psychologist Richard Mayer found that people learn better from 'graphics + narration' than from 'graphics + narration + on-screen identical text.'
Why: when the on-screen text nearly repeats what you're saying, the audience's visual and auditory channels are occupied twice by the same information, and they end up reading the screen instead of listening to you — which actually splits attention and raises cognitive load.
What to do: keep only keywords / images on screen, not full sentences; deliver the explanation with your voice, and don't force the audience to choose between reading and listening.
Coherence + signaling: cut the irrelevant decoration, then guide where to look
The coherence principle (Mayer 2009): cut every irrelevant 'decoration.'
Mayer found that people learn more deeply when irrelevant material is excluded rather than added.
Applied to slides — every element unrelated to the core message is competing for cognitive bandwidth with what actually matters and adding 'extraneous load': flashy templates, 3-D shadows, meaningless clip art, showy animations, optional text… all of it is noise.
Principle: subtract for the sake of the content; whitespace beats filling every inch. For each element you add, ask: did it help the audience understand the core message? If not, delete it.
The signaling principle (Mayer 2009): use cues to guide 'where to look.'
Subtracting alone isn't enough — what's left also needs a hierarchy. Mayer found that people learn better when cues are added that highlight the material's organizational structure.
On a slide, this means actively guiding the eye:
- use contrasting color / bolding / circling / arrows to highlight the part to look at right now;
- use progressive reveal (build / animation, used sparingly) to expose one point at a time;
- use a consistent layout so the audience knows the hierarchy.
A slide should make it clear what to look at first, and what next. Don't lay everything out at equal weight filling the whole slide — that's the same as having no focus, and the audience won't know what you're pointing at.
Assertion-evidence + one-idea-per-slide + data-ink: don't die by PowerPoint
The assertion-evidence structure (Michael Alley / Penn State): a single-claim headline + one piece of visual evidence.
It replaces the old 'topic word + a string of bullets (bullet list)' routine:
1. Write the headline as a full claim (assertion) — e.g., 'The new process cut the defect rate in half,' not a noun topic (like 'Results');
2. Support that claim with visual evidence (a figure / photo / chart / diagram), not text bullets;
3. On the spot, compose your sentences live to explain that visual with your voice.
The evidence: a study of 110 engineering students showed the assertion-evidence group had better comprehension, fewer misconceptions, lower perceived cognitive load, and stronger recall on a delayed post-test. In one line: each slide states one claim, then proves it with one image.
One idea per slide + Tufte's data-ink.
One idea per slide: rather than cram five points onto one densely packed slide, split them into five slides, one point each, each with a large visual. Why — it fits the glance medium (a single point is graspable in 3 seconds), it controls cognitive load, and it gives you a handle to control pace (whatever you flip to is where the audience's attention goes). Better to flip through a few clean slides than to use one crowded one.
When showing data, use Tufte's two principles: (1) maximize the data-ink ratio — every drop of ink in the chart should present real data; (2) kill the chartjunk — extra gridlines, 3-D, flashy gradients, and meaningless decoration only interfere with reading the numbers. In practice: use the simplest chart possible (often one line / a few bars is enough), label directly next to the data (don't force the audience to cross-reference a legend), and let one chart make only one point and put that point in the title (echoing assertion-evidence).
自测 · 学完检查一下
想真正动手做题、记进度、攒连胜？到互动课里练。
Reynolds (*Presentation Zen*) and Duarte (*slide:ology*) say a slide is a 'glance medium.' What does that most precisely mean?
答案：A slide should be graspable in about 3 seconds, after which the audience hands their attention back to you, the speaker
A 'glance medium' means: a slide should be graspable in about 3 seconds, after which the audience hands their attention back to you, the person. A slide's role is to aid the speaker — put what should be *seen* on screen (an image, key data, a single claim), and *say* what should be heard (explanation, story, connections). It is neither about being as flashy as possible, nor a teleprompter to read from, nor a place to pile all detailed text — text that needs detail should be handed out as a post-talk handout. (Sources: Reynolds, *Presentation Zen* 2008; Duarte, *slide:ology* 2008)
True or False: According to Mayer's redundancy principle, putting the exact words you're saying as text on the slide lets the audience 'both listen and read' and learn better.
答案：False
False. Mayer's redundancy principle says the opposite: people learn better from 'graphics + narration' than from 'graphics + narration + on-screen identical text.' When the on-screen text nearly repeats what you're saying, the visual and auditory channels are occupied twice by the same information, and the audience ends up reading the screen instead of listening — which splits attention and raises cognitive load. The fix is to keep only keywords / images on screen and deliver the explanation with your voice, not forcing the audience to choose between reading and listening. (Source: Mayer, *Multimedia Learning* 2009, redundancy principle)
By Mayer's coherence principle, which practice is correct?
答案：Delete flashy templates, 3-D shadows, meaningless clip art, showy animations, and every other element unrelated to the core message — subtract for the content
Coherence principle (Mayer 2009): people learn more deeply when irrelevant material is excluded rather than added. Every element unrelated to the core message — flashy templates, 3-D shadows, meaningless clip art, showy animations, optional text — competes for cognitive bandwidth with what matters and adds 'extraneous load.' The principle is to subtract for the content and prefer whitespace over filling every inch; for each element you add, ask whether it helped the audience understand the core message — if not, delete it. So piling on decoration or filling up whitespace is backwards. (Source: Mayer, *Multimedia Learning* 2009, coherence principle)
What is the correct on-slide application of Mayer's signaling principle?
答案：Use contrasting color / bolding / circling / arrows to highlight the part to look at now, so the audience knows at a glance 'what to look at first, and what next'
Signaling principle (Mayer 2009): people learn better when cues are added that highlight the material's organizational structure. On a slide that means actively guiding the eye — use contrasting color / bolding / circling / arrows to highlight the part to look at now, use progressive reveal (build / animation, used sparingly) to expose one point at a time, and use a consistent layout so the audience knows the hierarchy. A slide should make clear what to look at first and next; laying everything out at equal weight equals having no focus. Note that signaling is not an excuse to pile on animation — animation should be used sparingly. (Source: Mayer, *Multimedia Learning* 2009, signaling principle)
By Michael Alley's 'assertion-evidence' approach, which slide's headline is done right?
答案：Written as a full claim, e.g., 'The new process cut the defect rate in half,' with a piece of visual evidence below it
The assertion-evidence approach (Alley, Penn State) replaces the old 'topic word + a string of bullets' routine: (1) write the headline as a full claim (assertion), e.g., 'The new process cut the defect rate in half,' not a noun topic (like 'Results'); (2) support that claim with visual evidence (a figure / photo / chart / diagram), not text bullets; (3) compose your sentences live to explain that visual with your voice. In one line: each slide states one claim, then proves it with one image. (Source: Alley — The Craft of Scientific Presentations / assertion-evidence.com)
True or False: A controlled study of 110 engineering students showed that assertion-evidence slides led to better comprehension, fewer misconceptions, lower perceived cognitive load, and stronger recall on a delayed post-test than the traditional 'topic word + a string of bullets' slides.
答案：True
True. This is the key evidence behind the assertion-evidence approach: a study of 110 engineering students (Garner & Alley) showed the assertion-evidence group had better comprehension, fewer misconceptions, lower perceived cognitive load, and stronger recall on a delayed post-test than the traditional 'topic word + bullet list' group. That's why the advice is to state one claim in the headline of each slide and support it with one piece of visual evidence. (Source: Garner & Alley — the assertion-evidence controlled study, Penn State / assertion-evidence.com)
What does 'one idea per slide' tell you to do when you have five points?
答案：Split them into five slides, one point each, each with a large visual — better to flip through a few clean slides
One idea per slide: rather than cram five points onto one densely packed slide, split them into five, one point each, each with a large visual. Why — it fits the glance medium (a single point is graspable in 3 seconds), it controls cognitive load (don't make the audience face a wall of information at once), and it gives you a handle to control pace (whatever you flip to is where attention goes). Use more images, charts, and keywords, fewer full sentences; big enough font, strong enough contrast. Better to flip through a few clean slides than to use one crowded one. (Sources: Duarte, *slide:ology* 2008; Reynolds, *Presentation Zen* 2008)
Following Edward Tufte's two principles for showing data, which practice is right?
答案：Maximize the data-ink ratio and kill the chartjunk: use a simple chart, label directly beside the data, and let one chart make only one point
Tufte's two principles: (1) maximize the data-ink ratio — every drop of ink in the chart should present real data, removing whatever conveys no information; (2) kill the chartjunk — extra gridlines, 3-D, flashy gradients, and meaningless decoration only interfere with reading the numbers. In practice: use the simplest chart possible (often one line / a few bars is enough), label directly beside the data (don't force cross-referencing a legend), and let one chart make only one point and put that in the title. A simple chart is read faster and trusted more than a flashy one. (Source: Tufte, *The Visual Display of Quantitative Information* 1983)
True or False: A slide packed with bullet points looks 'substantial and professional'; if there's too much content to fit, the right move is to shrink the font and cram a few more bullets in.
答案：False
False. Treating a wall of bullets as 'substantial / professional' is precisely the root of 'death by PowerPoint.' Two rebuttals: (1) Mayer's redundancy principle — on-screen text that echoes your narration splits attention and raises load, so people learn worse; (2) Tufte criticizes PowerPoint's nested bullets for fragmenting reasoning and hiding logical relationships. The fix is not 'shrink the font and cram in more,' but cut text, add images, one idea per slide, and leave the details to your voice or a handout. (Sources: Tufte, *The Cognitive Style of PowerPoint* 2003; Mayer, *Multimedia Learning* 2009)
You have a complex argument to include in a talk and need to give the audience detailed text to keep. Combining this lesson's principles, the best handling is?
答案：Turn the detailed text into a post-talk handout; keep only an image, key data, and a single-claim headline on the live slides
Combining this lesson's principles: a slide is a glance medium (put what should be seen on screen, say what should be heard); the redundancy principle says don't turn your spoken script into on-screen identical text; the coherence principle says delete irrelevant text; and Tufte criticizes nested bullets for fragmenting reasoning. So when you need detailed text, the right move is to make it a post-talk handout and keep only an image, key data, and a single-claim headline (assertion-evidence) on the live slides, delivering the details with your voice. Copying it all onto the slide, dense bullets, or shrinking the font to fill the screen are all death-by-PowerPoint moves. (Sources: Reynolds, *Presentation Zen* 2008; Mayer 2009; Tufte 2003)