How to give a talk

It was a great idea to link to tips on academic talks, especially with so much criticism being heaped on scientists' communication styles. I hope all the N/M criticism doesn't discourage scientists from reaching out to the public where we need them.

Some good ideas.

Tufte's Visual Explanations has a great section on the same topic on pages 68 to 70. Here is one piece of advice I take to heart:

'Give the talk and finish early: "People will be pleased with a nice short speech. I believe that Paul Halmos, a very great lecturer, noted that in a lifetime of giving and attending mathematics lecures he had never heard complaints about a seminar ending early." [Mosteller, "Classroom and Platform Performance]'

Not so much a comment on the topic, but a fine academic performance occurred on Art Bell's show last night. Janna Levin put the smackdown on his typical listener. She spoke succinctly, eloquently, and professionally.

http://www.coasttocoastam.com/shows/2007/04/21.html#recap

She held up well on Colbert:
http://www.comedycentral.com/search/search_index.jhtml?searchtype=all&p…

The trick is *practising* these skills. After all, we all know the simple rules of chess, but it takes many years to become even remotely good.

The University of Cambridge had some absolutely stellar professors, but many of them were hopeless at lecturing, which was about the biggest frustration I came across there (alright, 2nd biggest) - so much knowledge to impart, so many obstacles in the way...

I have quite a collection of links to advice on giving talks, and will add this one to them.

Simon Peyton Jones, a CS prof in the UK, has a video on giving a good talk that's well worth a look.

Also, although it's focused more on posters (and poster talks) than on talks in general, Swarthmore's biology department (home of the Darwin has a Posse meme) has a good page.

In a previous thread I mentioned that most scientists lack empathy with their audience. The interesting thing about these guidelines - which are excellent, by the way - is that they are so obvious. Yet the same scientists who have to sit through other people's crappy talks (presumably without really thinking about why they are so bad) get up and give equally crappy talks themselves. It's quite sad.

When I go to any academic conference, I am always disappointed by the poor quality of so many of the presentations. I wonder if the presenters are as bad in the classroom; for their students' sakes, I hope not. I hope they are just nervous in front of an unfamiliar audience of peers, but have doubts about that hope.

Andy - What makes you think that most scientists lack empathy with their audience? I haven't found that to be true when I attend events for the lay audience (e.g., Café Scientifique). The speakers I've seen have been excited to share their passion for science with the lay audience and are very patient answering basic questions.

Paguroidea - my comment about scientists lacking empathy comes from personal experience. I've been in science for getting on 22 years now (if you count my undergraduate years), and have had plenty of opportunity to see scientists in action in front of their peers, their students or the general public. I also do a fair amount of peer review of grants and journal articles.

I will grant you that some scientists are very gifted at communicating with lay people, and I would guess that at least some of the scientists who are motivated enough to address Café Scientifique meetings are going to be better at communicating than most of their peers. But I feel strongly that they are in the minority.

PZ has suggested that faculty who spend more time teaching than doing actual science-in-the-lab/field are better communicators than full-time scientists. He may have a point, although again, my own experience would indicate otherwise. I'm sure practice helps, and he is correct in saying that teachers have to give far more presentations than researchers.

One tip he missed and a problem I have seen everyone from undergrads to PHD's make is don't read the slide. They can read it for themselves, that is why your wrote it. Add to the slide.

Good in some respects, but one key point is dead wrong, in my opinion ...

Instead, try to express all your ideas in illustrations. Resort to text where figures fail you.

My experience over a good number of years and presentations is the exact opposite. Always draft your point in words first. At the very least, that will help crystallize what you're trying to say more than futzing around with graphics. I've seen far too many elegant graphics which require far too much spoken explanation of just what point is being made.

If the point can then be better and more powerfully delivered by supplementing or replacing the draft text, then go for it.

This approach saves time overall, and certainly improves clarity and precision.

One mistake that research scientists frequently make is to take figures from their recently published papers and insert them into their Powerpoint presentations. This fails to take into account that paper figures tend to have much more information in them than the average audience member can take when the figure is flashed up during a seminar.

Speakers also have a habit of showing a complicated figure from a paper during a talk and then only talking about part of it, with the result that the rest of the figure is a huge redundant distraction.

This behavior is borne both out of laziness but also out of the lack of empathy I mentioned above.

I teach "Scientific Presentation Skills" to graduate students at POSTECH in Korea. I truly recommend the book "How to give a Scientific Presentation" by Michael Alley. It contains numerous examples (good and bad) of slides and organization strategies.

The tips in the link given here are all good, but one more that Alley gives, and that I have bought 100%, is the idea that slides should have Sentence Headings (of no more than 2 lines). For Example, instead of "Results", the title could be "Genetic mapping reveals homologies at BCD and WXY". This emphasizes your point and gives the listener a reference point while you elaborate.

Got the name of Alley's book wrong! it's "The Craft of Scientific Presentations".

sorry (blush)

These are great tips to begin with. What most of us need, though, is to not relax and think were so good just because we've given lots of presentations before. Every now and then, I will ask people I know in the audience for feedback on the delivery of the presentation. Using a simple scorecard, I usually don't have to ask more than 3-5 people before I see a pattern on what needs to be improved. It's so easy to slip into bad habits...

A very bad thing you can do is to give your talk while sitting behind a desk. Audiences are much more receptive if you stand up in front of them. And I really mean "in front of them", not "hiding behind a lectern". I'm still amazed at the number of people who don't catch this.

John Roth has a great page up describing tips for giving science talks. In classic Roth style he starts with this great caveat;
"Public speaking may share some features with making love. There's no absolute standard of how to do it well. There are probably some universally accepted ways to do it badly. Different people have distinct opinions on how it should be done. What passes as excellence for one situation may be inappropriate in another context. Some people who do it well won't talk about it and some people who talk about it all the time are failures in practice. Thus -- it's risky to presume to tell someone else how to do it. "

Sorry, that link should be:
http://rothlab.ucdavis.edu/protocols/GivingATalk.html

Thanks for that Roth link.

I find it interesting, he spends a good amount of space discussing the WHY of giving good talks (enlightened self-interest, essentially), which leads quite naturally into the HOW.

Thanks for the link, PZ... I was just about to sit down to prepare my next job talk.

Well, actually, I was already sitting down and was well into my 'procrastinate for 4-5 hours before writing' routine, an important part of a healthy breakfast. Or something.

(OK, blogs checked. Hmm.. quick game of nethack before I begin writing? Why not?)

While not specifically about scientific talks, a great talk on giving a talk is Conference Presentation Judo. It's for computer programmers, but I think most of the info applies to just about everything. You might just not get his jokes if you aren't ingrained into the Perl Programming culture.

Most of these suggestions are common sense, and it's a good source. Thanks for posting that.

I'll provide my counter-intuitive presentation software (some of you know this as PowerPoint though that is NOT the only alternative!) suggestion:

For the "look and feel" of your presentation: Use a calming blue background with white lettering. Have a logo representing your topic on exactly the same spot on every slide. Have something like a horizontal line underneath the heading, and make sure there is a heading in the same typeface and font for every slide. Any other embellishments such as a slice border, etc. should be on every slide and always the same. When working on your presentation, spend a LOT of time on this sort of style, rather than on content. Never change type face or font. Make sure that any pictures you show do not take up the whole screen, but rather, are displayed in a box that is smaller than the screen, so that the audience can see your common look and feel (the heading, horizontal lines, etc. mentioned above) on every single slide. The rule of thumb you should use is to make sure each slide looks exactly the same to the extent that if you were to look at them from a certain distance you would hardly be able to tell that they are different. And when choosing the style, make sure you ......

........... zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz.

Uh-oh. I'm putting together a new talk for a meeting, and I'm going to be doing it a bit differently. The slides have no words, except for a few with some data tables. It's an hour talk, and I'm aiming for about 15 slides. I'm going to put up great big pictures and then talk about them for four or five minutes each. I'm consciously making an anti-powerpoint powerpoint presentation.

Will there be riots in the streets?

Anyone else remember the days before powerpoint, when slides had to be made photographically, and they cost money and time to put together?

"Data tables"? Don't you mean data dumps?!

Zounds, PZ!

(-;

I remember those days well, and if you wanted to put up a table or graph you had to somehow photograph it (that is if you don't want to use an overhead).

Some of us had knowledge of how to photograph a table/graph and make it look good on a slide, and we became unwilling gurus. (Remember the self-developing Polaroid slide system?)

Remember 35mm black-and-white reversal films? I'd go in the darkroom with a camera stand and shoot my graphs directly from paper! And develop it! With chemicals! At the crack of dawn! And then I'd trudge back 5 miles through hip deep snow to the lab! And I liked it!

We used to carve our negatives out of stone. And the data tables had to be in Roman numerals. And all the U's had to be V's.