Tips for Speakers

You definitely need some text on a slide for people who walk in late and especially for non-English speakers. And although I don't do it, I like seeing conclusion slides with full sentences and all text. Since it's the last slide of the presentation, it's helpful for that slide to summarize as much as possible. I think of it as opposite of everything else. During the talk, your slide should briefly cover what you're saying in more detail. For the conclusion, you should briefly cover what's covered on the slide in more detail. It's also nice to have something more substantial to look at during the question and answer portion.

As for figures, in my engineering communications class, during our presentations, we were required to introduce all figures and plots with "What you're looking at is..." It was a very helpful practice technique for making you always adequately explain a figure.

And although I don't do it, I like seeing conclusion slides with full sentences and all text. Since it's the last slide of the presentation, it's helpful for that slide to summarize as much as possible. I think of it as opposite of everything else. During the talk, your slide should briefly cover what you're saying in more detail. For the conclusion, you should briefly cover what's covered on the slide in more detail. It's also nice to have something more substantial to look at during the question and answer portion.

Oh, yeah. I forgot that one.
One thing that students here tend to do, that drives me absolutely nuts, is to end with a slide of references, just a big long list of bibliographic citations. Or, worse yet, click through that, to a big slide that just says "Questions?"

End on a summary slide, always. That way, people leave the talk thinking about what you talked about, not "Boy, he sure did cite a lot of papers in tiny little type..."

These days, I tell my students to plan their talk as if they're telling a story: each slide should, in essence, raise a question that the next slide answers or addresses, and so forth. It also helps to plan a few 'connecting words' between slides: I find it quite jarring in a talk when a speaker finishes their explanation of one slide and leaves a pregnant pause before discussing the next one.

The one specific tip I would add to their list is this: When you put up a graph, you should clearly identify what is being plotted on what axes. The first thing you say should be "Here we have YYYY on the vertical axis versus XXXX on the horizontal axis," so that everybody knows what you're showing, without having to squint to read the labels.

One of my biggest peeves here is people giving presentations where the graphs aren't even labelled, so there's nothing to squint at. Christ, this is something we ostensibly learned as undergraduates-- no graphs without labelled axes.

I have gotten to the point where I will stop a presentation every single time I see that, and ask for clarification. Apparently, being a big jerk is the only way to get that point across over here. If we could get to that point, then I'd worry about the proper verbal technique to introduce those slides.

(For myself, I've found that if I don't have at least one sentence I want to say out loud about the graph, it gets left out, or put in the supplemental slides in case someone asks about it. I see far too many design reviews here where graphs are thrown in because they're expected and the engineer tries to let them speak for themselves. Um, no.)

http://www.scs.uiuc.edu/suslick/seminaronseminars.html

These are all very good comments. It amazes how poorly many job candidates for faculty positions doe when it comes to the scientific job talk. I only have a few comments about my personal preferences.

1. Do not use a distracting background. I personally prefer a solid, dark color background for dark rooms and a solid, light background for light rooms.

2. Complete sentences should be used sparingly and for effect only. I use complete sentences on the slide only for stating the hypothesis that drives a series of experiments and for the conclusions.

3. If you create the presentation on a mac, please check it on a pc prior to giving the presentation. It does not take long and will prevent you from looking extremely unprepared if the presentation does not render correctly.

Echo gg. The way I put it is the talk is primary, the slides secondary. Create the talk, then build slides to illustrate or emphasise or augment the talk. If you do it the other way round -- first build the slides then come up with things to say about the slides -- you end up with the pregnant pauses between them (and an incoherent presentation).

The problem is people giving talks try to give the impression that they're doing it extempore, just riffing off the slides, which leads the young and impressionable to try to emulate them.

The first thing you say should be "Here we have YYYY on the vertical axis versus XXXX on the horizontal axis," so that everybody knows what you're showing, without having to squint to read the labels.

Useful in many situations, but I'd like to comment on "without having to squint to read the labels." If you've done a proper job of preparing your graphics, your labels are big enough that looking at them is not too much of a strain. I always try to do that in my presentations, and when a co-author has given me a figure that resists that treatment I'll add text boxes as appropriate to annotate. (This goes for poster presentations as well as talks.)

Of course, I'm constantly amazed at conferences how many of my colleagues are clueless from a graphic arts standpoint. There is a reason why the PowerPoint default font sizes are as big as they are.

And speaking of PowerPoint: It's one of the most evil tools ever foisted on our profession. One of the reasons is that it makes it too easy to make slides (and too many of them) in search of a talk, as Jim points out. When I was using viewfoils, a scarce resource, I would always plan the talk to use exactly a certain number of slides (usually one per minute in a shorter talk), and I would know what each slide was going to be before making any of them. (Similar to Jim's approach of creating the talk first.) However, this takes more effort than the other approach. Since there is no longer a disincentive for winging it, winging it becomes very tempting.

Thanks Chad for highlighting our video. From the comments it appears that many scientists/engineers feel the same way as we do about preparation, not reading to your audience, labeling your figures well, etc.

Jim (comment #7) makes a great point about students emulating their professors' presentation style. I know professors are often time-crunched and throw together presentations at the last minute and as such end up making all these errors Greg and I highlight. I really don't know how to solve the problem other than to try to make the best presentations I can to hopefully influence my audience to spend more time on their presentations as well.

As a scientist-in-training in my late 20's, I have always lived in a powerpoint world. I wonder if some of your readers, who may have been around before powerpoint took hold, could comment on the quality of scientific talk prior to presentation software (in the days of slide projectors and overheads). Were talks generally better then? Or are the issues we highlight simply basic public speaking issues that have been around forever?

As the old fogey here, I'll mention that Physics Today has run articles periodically on this subject. One of the best ones was a reprint of an article from the 50s or 60s when transparencies first appeared on the scene (replacing actual slides or actual chalk). Every sin and suggestion mentioned here was addressed in that first article! Powerpoint did not create the problem, it only made it worse. (The case of the speaker reading from ppt notes on the laptop was preceded by notes written on the cardboard frame of a transparency.)

Maybe one of your readers can track it down. It is a must read for any professor (some of the worst offenders) or student.

By the way, know what you are talking about is high on that list, followed closely by that old standby, practice. I saw a speaker recently who was able to continue her talk, almost without pause, when the computer she was using crashed with a BSOD. She not only knew what was on the next slide, she could describe it perfectly (including numbers) without looking at any notes. She is the CEO of a major company.

CCPhysicist, I found a mention of such an article (but unfortunately, no link) on Peter Norvig's site, which hosts the infamous Gettysburg PowerPoint presentation (http://norvig.com/Gettysburg/making.html):

Note: a reader pointed out that John S. Rigden had an article in the March 1990 issue of Physics Today entitled "The Lost Art of Oratory: Damn the Overhead Projector" that also used the Gettysburg Address concept.

In response to Tim: CCPhysicist correctly notes that most of the problems with PowerPoint existed in the viewfoil era. With a quick Google search you can find the classic "Guidelines for Giving a Truly Terrible Talk", which dates from that era; in my grad student days the American Geophysical Union would include a copy with the notification that you were assigned an oral presentation at a meeting. What has suffered is the quality of discussion, particularly at workshops. The problem is that PowerPoint is designed to foreclose rather than stimulate discussion (another reason why I consider it one of the most evil tools foisted on our profession). There are times in the corporate world where that may be desirable, but for scientists it is definitely not wanted.

I found a copy of "Damn the overhead Projector" that CCPhysicist mentioned. Its on a course website here:

http://www.columbia.edu/itc/seas/freeman/e3011-01x/client_edit/readings…

It seems Peter Norvig's famous Gettysburg speech powerpoint was inspired by the opening of this article.