Guidelines & Common Mistakes in Presenting an Academic Talk

 

Guidelines

For the Cognitive Science Department at RPI here are the Guidelines for a typical talk whether presented as a job talk or at our weekly Issues in Cognitive Science series.

We reserve a 90-min slot for the talk. This does NOT mean that the speaker has a 90-min presentation. Rather the time is ideally divided as follows:

10-min administrative stuff at the beginning -- this will include introducing the speaker and it also includes time for the audience to get to their seats and settle down. If yours is a lunchtime talk, then this time may include time for people to grab a sandwich or a slice of pizza and then sit down.

50-min prepared presentation

30-min question and answer period.


Caveats

We are a friendly crowd and treat our guests politely. However, we do like to ask questions and we will not sit silently through most talks. This means that although you should prepare for a 50-min presentation, the actual presentation period may be extended over a longer period than 50-min. The moderator (your host or the person who introduces you) is responsible for keeping control of the crowd and ensuring that in the 80-min actually allocated to your talk (50-min presentation + 30-min q&a) that you get in your full 50-min presentation period. But, this means that you should be prepared to handle questions during your talk. You also may need a certain amount of skill in turning off discussion on a side issue or in pointing out that you will get to that issue later in your talk.


Common Mistakes

We really believe that the Cognitive Science Department at RPI has the best and brightest faculty, graduate students, and undergraduates of any group of Cognitive Scientists, any place in the world. However, despite our brillance our cognition is bounded. We suffer the limits of bounded rationality as much as any other group does. These inherent limits to the human cognitive condition have certain implications for how you prepare your talk.

Avoid Lab Talk

Lab talk refers to the shorthand ways of talking that a small group of researchers who work with each other on a near daily basis have for talking about the basic assumptions that they all share. This may take the form of acronyms, special ways of using common words, or complex assumptions that are simply too long to bear repeating several times in a single laboratory meeting.

Some of our past speakers have given us the impression that the morning before they arrived, they had a lab meeting with their grad students that they had to interrupt to get on the plane. Unfortunately, for us, when they arrived at RPI they simply resumed the interrupted lab meeting, sometimes by completing the thought or statement that had been interrupted hours before.

Too Much Detail

You are a clever person and you have done many clever things. Indeed your book, dissertation, journal article that is in-press, in preparation, just published works thru the clever details of your proof, derivation, model at an exquisite level of detail.

That is nice.

BUT, do NOT try to walk us thru such detail in your talk. For the 90-min slot we will take everything you say at its face value. Although as individuals we like such details, as a group our backgrounds are very diverse. Some of us can follow a logic proof in real-time, some do Bayesian, some Markov, some swim in connectionist details, some love those ACT-R production rules, etc.

Tell us what you did at the "functional" level, not the "mechanism" level. Tell us why it is important. Tell us a bit about how clever what you did was as before you did it everyone else did things in a much more indirect, slower, stupid, whatever, way. Above all, tell us how what you did relates to the overall topic of your talk and to human cognition. (Duh? Yes -- people occasionally forget to do this.)

Too Much Text

Please do NOT transcribe your dissertation, journal article, book into powerpoint. There are no page limits in powerpoint and blank space does not consume any hard drive room. This means that you should have fewer words per page, not more words. If all of the words are really, really important then put them on 2, 3, 4, 5, whatever slides. Use large font so I can read it in the back of the room without my glasses. If, after you do this, you have 100 slides, you probably have prepared too much material for us.

Number of Slides

This is an individual difference variable with a wide range of parameter settings between individuals. Some people spend equal time on each slide regardless of how much or how little is on it. Some people take about 2-min per slide. Some people take about 1-min. Some take less, some take more. Some slides are walkthrus of experimental procedures and take about 5-s each.

Know what your setting on this parameter is and adhere to it. Do not prepare a 60-slide presentation for a 50-min talk if your parameter is 2-min per slide.

The Lamest Phrase Ever Uttered by Any Speaker: "Sorry that you can't read this"

Okay. It is lame and we all have said it. Why can't we read it? Very basic, visual acuity type issues and or a technology-based gotcha.

A. Font too small, too many words on slide, etc

B. The great colors that looked so good on your high-resolution computer monitor do not work so well on the computer projector (aka "beamer"). Reds tend to vanish, blues disappear, those great looking color line graphs all seem blank now. Hmmm.


The difference between what you see on your screen and what can be projected by a computer projector will get greater in coming years. Most universities, businesses, hotels that have projectors have a lot of money invested in old stuff. In our lab our projector is 4 years old but the computers we take on our talks are < 1 year old. Many of those large projectors installed in electronic classrooms are older than that. The best way to avoid the problem is to learn what looks good in low-res VGA and make your slides conform to that standard. The second best way is to test your slides out on an old projector before you leave home. The very worst way is to test your slides out on our projector during your talk.


Prepared by Wayne D. Gray

Last changed: 2006-01-13