Guidelines for Better Slides
First off, thanks to Jane Kokernak for directing me to these resources. Most of the content below is taken from these sources.
- Michael Alley, The Craft of Scientific Presentations (2nd ed, Springer). This inexpensive paperback is a good manual, covering the entire process from organization, preparation, slides, and delivery.
- https://www.assertion-evidence.com/. The website for the
book. Lots of good material here, especially if you want to avoid
crawling through all the prose in the book. In particular, the
following links are good starting points:
- https://www.assertion-evidence.com/tutorial.html. A high-level overview.
- https://www.assertion-evidence.com/teaching.html. Several sets of slide decks for teaching the Assertion-Evidence method. Check out the Peer Slides and the Instructor Slides.
- https://www.assertion-evidence.com/templates.html Lots of sample slides. I wish he had supplied better Powerpoint Master Slide templates. (Anybody want to contribute such a set to the College?)
- Jonathan Schwabish, Better Presentations: A Guide for Scholars, Researchers, and Wonks. More emphasis on good use of color, graphics, and visualizations.
- https://policyviz.com/. The website for Schwabish's books. Lots of useful links, though many are affiliate links with paywalls. However, there is a lot of useful free stuff at .
(adapted from Schwabish, p. 15). See also
- What kind of presentation are you giving?
- Classroom lecture
- Research group presentation
- Departental colloquium
- Conference presentation
- Hiring talk
- Keynote address
- Slides for students to read offline
- How long is the talk supposed to be?
- Who is your audience?
- Departmental Audience
- Scientific/Technical Professionals
- Students (grads? undergrads?)
- Thesis or hiring committee
- What is the goal of your presentation? What is the message you
want your audience to come away with?
- They should get an in-depth understanding of my results (not likely!!)
- They should get a general understanding of my work.
- They should be motivated to read my paper
- They should pass my proposal, thesis, etc.
- They should hire me.
Each slide should have a role
We've talked about the role of each paragraph in a paper. Most of the time, we've been able to explain the point of each paragraph in a simple declarative sentence that summarizes the emotional impact that the reader should get from the paragraph, such as "X is bad" or "Y is not enough to solve problem P" or "Z solves problem P". I think of these as slogans: if you had to carry a protest sign to communicate the message of this paragraph, what would you write on it?
Each slide should have a message
Alley presents what he calls the "Assertion-Evidence" organization. In this organization, a slide should be organized like a paragraph in a text: it should have a topic sentence or thesis, which is the message of the slide. This message should communicate the slogan or role of the slide, but in polite and specific language.
The rest of the slide should then present evidence that supports the thesis. The evidence should be visual whenever possible.
Some slides, like those that present definitions, have a role that doesn't fit neatly into this format. Even these can sometimes be usefully adapted into this format. Ask the question: what words would I use to explain this concept, before getting into the details. Examples (sorry, Peizun):
- A pushdown system is a finite automaton that manipulates a stack
- At each step, a pushdown system reads the top of the stack to decide on its next action.
- A Concurrent Pushdown system is a set of pushdown systems that share a finite global state.
What about code? Probably the right answer is not to display code unless it's vital for your argument. (If you think it's vital to your argument, lie down until the urge passes.) I haven't looked at guidelines for displaying code if you absolutely must. If anyone has found such guidance, I'd appreciate it if you shared it.
The Assertion-Evidence Slide Structure(From Alley 2013)
- Begin each body slide with a sentence-assertion headline that is left justified and no more than two lines. Capitalize the headline as if it were a sentence, but do not end it with a period. If the headline goes more than one line, break the first line at the end of a natural phrase.
- Support the assertion headline with visual evidence (photographs, drawings, graphs, films, or words and equations arranged visually). Avoid lists unless the connection between list elements is obvious.
- In the body of the slide, use words only when necessary—design your slides so that the audience reads no more than 20 words per minute.
- Use a bold sans serif typeface such as Calibri. An alternative
is to use a serifed font for titles and sans-serif for the body.
NU's new branding uses Sentinel (alas, not free) and Lato in this way.
I think Alley's example slides use too much bold type. I'd lean towards using more non-bold type.
- Use 28 point type for the headline, 18–24 point type for the body text, and 12-14 point type (not bold) for reference listings.
- Avoid setting text in all capital letters, in italics, or with underline. Use color when emphasis is desired.
- In graphs and charts, make sure the important thing, the thing you want your audience to look at, is highlighted in color, line style, etc.
- Use flowcharts or similar diagrams to visualize multi-step processes.
- Make sure that everything on the slide is readable from the back of the room, under the worst anticipated lighting conditions. Try it out on a real projector in a real room. Don't assume that because you can read it on the screen, your audience can read it when it is projected.
- Keep blocks of text, especially the headlines, to no more than two lines.
- Keep lists to two, three, or four items. Avoid bullets, which only add clutter. Use extra space between list items instead.
- Use small margins on the sides, so that you can insert sufficient white space between element. For instance, leave at least a half-inch of white space below the headline.
- Are structure slides recognizable as such? Do they differ enough, at a glance, from content slides (eg different layout, color, graphical elements?)
- Can the structure of your talk be seen at a glance? Limit the preview to one level, with short labels.
- Avoid titles that add little, like Preview or Outline. Use the title of the talk instead when appropriate. (If the title is too long, make up a short title instead)
- Your audience will see your last slide longer than any other. Make sure it carries the message you want to linger in your viewers' minds. A conclusions slide will often work well for this purpose.
Last modified: Wed Nov 7 13:17:37 Eastern Standard Time 2018