Advice on giving a research talk
Compiled by Clay Scott. Created 8/22/07. Last modified 9/5/07.
- Research talks should contain certain key elements,
possibly but not necessarily in this order. Some of these
elements will be more developed than others depending on
the maturity and nature of your project.
statement, with motivation
related work, where your work fits in
your specific innovations and achievements
faced and overcome
experimental findings, simulations, theory, etc.
of your work (does it help solve other problems
besides the one that motivated your work?)
outline slides in short talks. They tend to waste time
and do not aid the audiences understanding. In
longer talks, outline slides might be justifiable, but
only after you have defined and motivated the problem. In
most cases your talk should be one seamless narrative
where each point flows naturally into the next, and an
outline should not be necessary.
- Your talk
is an advertisement for your research. It does not need
to convey every last detail. It should inspire the
audience members to read your paper or follow up with you
later in person. Give key ideas, motivation, intuition,
major findings, impact, but resist the temptation to
overload with information.
- A rate of
3 slides per 5 minutes is very reasonable. Any more than
that and you're likely to go over your time limit.
you are very experienced, its a good idea to
memorize what you plan to say on the first and last
slides. Your opening sentences should quickly turn to a
problem statement or motivating example. The last slide
should recap your contribution, and should ideally
present a memorable image or phrase or result that will
stick with audience. Do not end by asking if there are any
questions, because then the audience may not applaud, which
can be awkward.
graphics/images/figures in your slides frequently. Even
if a slide contains equations or a lot of text, its
a good idea have at least one figure. It gives the
audience something to stare at. Many people tune out when
faced with a slide full of text and equations. A figure
can be as simple as a cartoon made of simple
powerpoint graphics objects (dots, lines, etc.). Google
Images is also a very good source of images.
- Don't go
overboard with the animations. They can be distracting.
- Regarding pointers
and laser pointers: As a rule I try not to use them. If you are nervous,
the pointer will shake, which is distracting. You may unconsciously shoot
a laser pointer into the audience members' eyes, which is very unsettling.
If you are close enough to the screen that you can walk to it and point,
the best option. It also helps you be more active. If you do have to use a
laser pointer, use it sparingly. It is for emphasis, not to underline
every point on your slide as you talk about it.
- About 5%
of males are color-blind, with red-green colorblindness
being most common, so avoid graphs where differences
between red and green (or other subtly different colors)
are critical. Effective use of color will enhance a talk,
but use other methods of emphasis as well. Also keep in
mind that the colors on your computer screen and the
colors on a projector may be significantly different.
- It is
better to answer a question with I dont know
than to convey this information by talking about
something you dont understand. Pause no more than
5-10 seconds when thinking of how to answer a question.
- Give one
or two practice talks in front of a small audience (5-10
people). When you speak out loud, your ideas are likely
to come out in a different or less clear way than you
imagined when preparing the slides. Invite people who are
unfamiliar with your work, as they tend to give the best
- In a
practice talk, distribute hardcopies to the audience
- Arrive at
your talk early to make sure everything is set up
- Something to keep in mind as your presentation skills
develop: Your slides are not the focal point of your
talk. Your ideas are the focal point, and the
slides are only there to assist in conveying those ideas.
Therefore, let the flow of your talk be determined by
what best advances your argument, not by what is next
sequentially on your slide. In other words, if you find
that you want to jump forward a slide, even though you
have not addressed all the material on a current slide,
it is best to move forward. Then adjust your slides
before you next give the talk.
researchers and students by Michael Ernst
Ian Parberry's speaker's
to give a good research talk: Simon L Peyton Jones, John
Launchbury, John Hughes
tips posted by Craig Kaplan
writing and presentations page by Armando Fox
presentation advice by Mark Hill
good research talk by Nicholas Nethercote