Skip to content

Quick tips on giving research presentations

Hi, I’m writing this so I can refer to it when covering “giving a presentation” in my statistical communication class. The general idea is for me to spend less time in class talking and more time helping out students with their ideas. So, if I have any general advice on presentations, let me give it here.

1. As always, think ahead of time about your goals and your audience.

2. When you start your talk, make your goals clear.

3. If you are given X minutes to talk, you may well have more than X minutes worth of stuff to say. That’s fine. Spend .8X minutes on your material and the final .2X minutes explaining how this fits in with the other stuff you don’t have time to say.

4. If you’re talking about your research, make your contributions clear. It’s not about bragging rights, it’s about making it clear to the audience why they should be listening to you on this topic.

5. Fractality: Every piece of your talk should contain all of it. OK, not really, but the key point is that most people won’t be paying attention all the time. So if someone tunes in at some random point in the talk, they should be able to follow along from that point. People always give the advice to start broad and then drill deep—and I agree, that’s good advice—but when you’re drilling deep, keep sending soundings out to the surface to remind people why you’re doing all this.

There’s lots of other good advice:
– Don’t put so many words on your slides; instead, write down some notes and say what you want to say.
– Move around, don’t stand still.
– If you have slides and they are displayed at a reachable height, stand in front of them and point at them with your hands. If instead you stand somewhere else (for example, wherever your computer happens to be), the trouble is that people in the audience won’t know whether to look at you (where the sound is coming from) or at the slides.
– If you know anyone—anyone—in the audience, plant a question or two. This isn’t “cheating,” it’s just a way to get the post-talk question period to go more smoothly.
– And here’s what Hal Stern told me before I presented my first professional talk: Don’t try to blow them away, it’s enough to just present what you did. If it’s good stuff, it can stand up unadorned and a discerning audience will realize its importance. And if it’s not good stuff, you shouldn’t be wasting people’s time on it anyway.

I could keep going but I think points 1 through 5 above are the most important.

But maybe I’m missing something big? Feel free to add your own suggestions in the comments.

P.S. And here’s a wiki on the subject from Jeff Leek.

P.P.S. And now I don’t use slides at all!


  1. Jeff L. says:

    Hey Andrew,

    This is great. Mind if I incorporate a few of these things into my talk guide?


  2. Todd Trautman says:


    Perhaps this applies more in the corporate environment where I work, but make sure the slides that you present are a visual aid, not simply writing out the actual presentation. Then if needed, include a detailed appendix or notes for anyone that wasn’t able to attend, but receives the slides.

    Too many people include every word they want to say in 8 pt. font that is projected at the front of the room.

    • Rahul says:

      I think that applies very well to Academic Talks too. Putting every word on the slides is a recipe for a horribly boring talk.

    • Peter Dorman says:

      Much agreed. I’m with Tufte on the notion that, if you have complex stuff that you want people to see visually, give them a handout. This goes especially for tables. My approach to slides is that their main use is giving the audience a visual metaphor for the structure of your talk. This is quite explicit with Prezi, although I find that my structures don’t match theirs in most cases. But regard each slide as a two dimensional space, and use this space to display the logical relationships (lists, causal processes, contrasts, etc.) you will be expressing verbally. The more economically you can do this, the better — so the audience will be listening to you rather than reading. Also, the slides as a whole can express logical hierarchy, which is difficult to keep track of when you’re listening.

      I typically advise *not* to turn to look at the slide, but familiarize yourself with them in advance so you can keep looking at the audience while you discuss what they’re seeing behind you.

      If it’s at all close, assume you will not get through all your material, and plan ahead for how you’ll handle it.

      I really like your point, Andrew, that the audience will zone in and out, so you have to touch base with re-entrants on a regular basis.

      An interesting question for instructors is whether to really teach this stuff — for instance, to require students to submit a draft of their presentation (slides/outline/handout) in advance so you can give feedback.

  3. Jordan Erickson says:

    Garr Reynolds (of has good tips of preparing, designing, and delivering presentations:

    They seem to apply to research presentations as well.

  4. Robin Morris says:

    For less experienced people, you missed the most important point:

    1. Practice. Out loud. It doesn’t matter if there is or isn’t an audience in the room when you practice, but find a room with a projector, stand up, and give your talk out loud at least once before the actual presentation.

    The best published advice that I’ve ever read (and now give to all my interns in the week before they give their final presentations) is “Who is listening? What do they hear?” which was published in Physics Today a while back,

    There’s three important things about a presentation. The audience, the audience, and the audience.

  5. Adam says:

    This is a joke, but also fairly accurately sums up what people expect you to present. You have to be a pretty good presenter to deviate successfully. I assume.

  6. Rahul says:

    Can I offer an example? There’s this Google-talk about “X-Box Security Mistakes” by Michael Steil that I thought was one of the best examples of technical presentations I’ve come across.

    It’s 60 minutes with rarely a slow point, high on information content, nicely illustrated, well explained, just the right bit of humor & coolness. Just the perfect technical presentation in my book.

  7. Great list! This is perhaps a subset of (1), but I think it’s very important to convey near the start of a talk why the audience should care about your topic. I’m always amazed at how many talks make no attempt to do this, carrying the implicit assumption that everyone in the audience walked in already fascinated by the speaker’s particular sub-sub-field or project. Of course there are very focused venues or conference sessions where this assumption may be warranted, but most of the time it’s not.

  8. Christian Hennig says:

    Some additional advice by me:

    Don’t show more on the slides than you can explain. Normally people need all the time to follow what you’re saying, so if they try to figure out what else is on the slides, they’ll get confused.

    Practice your presentation and know how long it is so you don’t have to panic and randomly leave out material when time is tight. (All too often speakers then say that they’re going to leave out material only to explain then why this can be left out and what is was in the same amount of time.)

    Ask yourself what’s boring and superfluous about other people’s presentations (including those rules for presentations that get boring because everybody follows them senselessly – the cartoon on introduction and conclusion linked above hit the nail on its head) and either don’t do it or do it in a funny way. (I loved the guy who in a four-minute presentation gave a two-minute overview of what would be in his presentation and then used the remaining two minutes to go through all items. That was funny and worked. Otherwise overviews seem pretty pointless to me in presentations of <=20 minutes and most don't contain anything useful at all.)

    You can ask the audience stuff, not only they can ask you.

  9. WB says:

    Point 6. Don’t simply paste tables of regression results from your paper into PowerPoint slides. Most of the numbers will be too small to read.

    You need to present findings in a way that an audience can grasp fairly quickly.

  10. MikeM says:

    Avoid like the plague mind-numbing tables!

  11. Rahul says:

    I don’t like the “plant a question or two” suggestion. Sounds too “fake” for my taste.

    If you get no questions, so be it. Maybe you just end earlier or chit chat with your audience to break the ice or something. Most places your host or someone senior will step in themselves anyways.

    • Dan Wright says:

      I have one talk where I write my “planted question” on a paper airplane and throw it into the audience. The person who grabs the airplane sees it tells them to stand up, raise their hand, and after being called upon (which I do!) to ask the question. That way, I am not hiding it being planted, but still get it asked.

  12. Alex Luedtke says:

    I enjoyed chapter 52 of

    Read it when you posted it awhile back.

    • Jordan Erickson says:

      My favorite one is the last one:
      13. Oh my, you are running out of time. Don’t skip anything, show every slide
      even if it’s just for a millisecond. Saying “This is really interesting stuff,
      I wish I had time for it” will make people grateful for getting “Chebychev”

  13. Bill Harris says:

    Mosteller wrote about the PGP technique, which sometimes seems like a good way to communicate in a business talk.

    Someone once told me to spend 10-20 seconds or so talking to and looking at a specific person in the room, and then do it again with another person. When I’ve done it, it seems to have had a favorable effect: people feel more included and engaged, and I lose the temptation to stare into nothingness.

    Practice, practice, practice. I once coached presenters on presentations, and, as far as I could tell, the better presentations were from those who had practiced more. One trick: don’t stop when you’ve done it well once. That may mean you’ve done a poor job 5 times and a good job once, but you’d probably do better if you’ve done it well more than you’ve done it poorly. If you can get feedback from a real person at least some of the time, that’s great. If you can video your presentation and then review the video, that’s good, too.

    To Peter: instead of having students submit presentations for review, what about having them work in pairs (“peer instruction) to deliver their presentations to each other and then give each other feedback? That way, it’s the full presentation of the material, not simply the slides, that get help.

    • Dario Boriani says:

      I agree. Making and holding eye contact with various people in turn as one moves around counteracts the tendency to develop the thousand-yard stare and can make it feel more intimate and less intimidating to the presenter, especially if the audience member smiles or is receptive.

      I would add that relatively few — and not overly cluttered — supporting slides are better than too many. This gives the audience time to digest the content and not feel overwhelmed or behind as the new slide is shown, which leads to losing focus. It also allows the speaker to pace him/herself better and perhaps point to some nugget they might otherwise have glossed over.

      Attention-grabbing slide headers also help — ‘In 2017, Switzerland produced more watches than all other European countries together” has more impact than “European production of watches for 2017 (by country)” on top of a column/bar chart.

      Absolutely, rehearse. And throw in anecdotes, if applicable.

      Do not forget to smile.

      If at all possible, ask for the audience to mute — or, better, put away — their cellphones.

      My $.02.

  14. Dan Wright says:

    Sometimes it is easier to tell people how to do a bad talk.

    On bad plots in a talk or paper, Howard Wainer’s How to Display Data badly ( is great. Here (“How to produce a bad results section”, Psychologist, 2003, on Sian Williams and I followed his lead in a brief article for a psychology magazine. Also see Efron’s chapter in He gives 13 ways to give a bad talk.

  15. Kellie O. says:

    Great tips, but I’m more excited about the fact that you’re designing a statistical communication course! That sounds like something every stat student (myself included) should take before getting their degree. Fantastic idea.

  16. Jeff says:

    Before designing the presentation (which is typically code for powerpoint), create a script of what you think the header on the slides will be, and how much specific time, to the minute you are devoting to it. You ideally know if the format is more talking head in which case if you plan 30 minutes you can deliver 30 minutes or if there is more back and forth in which you really can only talk 20 minutes.

    Practice. I’ve rarely seen an academic actually get to the end of their presentation in the time they expected. This is poor planning. Its where you realize the planned 6 mins on lit review is really 8.5, which means you just lost 2.5 mins on the ‘so-what’ and/or answering questions. The latter two are almost always the most interesting part of the presentation.

    Per Bill’s idea above I use pairs where they share/deliver the presentation/assignment to a partner, then receive feedback. Students then submit the first version and the second, with a 1-pager discussing which feedback they took action on and what they didnt and why. Not unlike the journal submission process.

Leave a Reply to Rahul