giving-talks.md

Giving and attending talks

Talks are important: They show your command of your research and your ability to stitch together a story worth listening to. Most professors will agree that the new faculty hired into the school are judged primarily on their job talk.

What are you trying to achieve?

Giving a talk is presumably associated with some goals, otherwise, why give it? You should organize your talk to achieve these goals. Here are some examples of things you might want people to think after your talk:

• I saw a nice talk about X from this person Y at Georgia Tech. We should look more into what they’re up to (or talk to them after the talk)
• This talk was about X and the main results were Y. (You will find many talks do not meet this seemingly low bar)
• This work was about X and I’m also working in X! They’re doing Y and we’re doing Z. Maybe we can collaborate and improve X together.
• This talk was great — who is this person? (then they remember your name and will use it at a later date, perhaps for hiring purposes)

All of the above are very real things, and I’ve thought about each of them several times before.

How to

The guidance I usually give around talks is somewhat traditional: plan, make, and practice.

1. Plan: I write down what I want the audience to learn from the talk. Then, I look at the time I have available to speak and subtract 5 minutes. From that number, I break down how much time I want to spend on the different parts of the talk.

2. Build: After the above, I start building slides. Use the group template, which is available in this repository. Acknowledge the sponsor of your research on the last slide or two, depending on how the slides are arranged.

3. Practice: After you have your slides, practice the talk by yourself: Pull up your slides and start going through the talk aloud on your laptop with no one else around. Fortunately, the talks you're likely to give early in your academic career are also the shortest. This means you can practice them, in their entirety, a dozen times within an hour or two. I guarantee that a talk practiced front-to-back 12 times will be delivered more cleanly than one never practiced, and everyone in the room can tell the difference. If I'm nervous about a talk or am giving one on a new topic, I practice even more than this. Sometimes I spend hours practicing over a few days.
   Especially early in your Ph.D., you should then plan on giving at least two practice talks to an audience (one of them with me). Later, you won't need as much practice for a conference talk, let's say; but a job talk may require even more practice and iteration. After you synthesize and iterate on all the feedback, you're ready to go. This process often takes many days, even for a short talk. This means that you should be ready to start practicing at least one week before you give your talk.

Feedback

The most common feedback I give for talks is:

• No "table of contents" for short talks (15 minutes or less)

  • This just wastes otherwise valuable time

• The talk was not structured for the audience.

  • Who are you talking to? It matters! In fact, it might be the thing that matters most.

• What was your message?

  • If the message is unclear to me, it will definitely be unclear to others. If your talk doesn't have a message, consider why, come up with an appropriate one, and restructure your slides.

• You did not practice enough.

  • You know your talk is well practiced when you can deliver it without looking at the slides/projector most of the time. Why? Because you already know what's on them, and what's coming next! This allows you to engage with the audience.
  • You know your talk is not well-practiced when you are reading the slides: I, too, can read your slides. If that's what you were going to do, you could have just sent a PDF of the slides to the audience to peruse at their own leisure.

• Stop using superlatives.

  • Words like very can pollute an otherwise excellent talk. "The relationship is very complicated." "This equation is very nonlinear." What does one intend with such words?

• You use many filler words.

  • Avoid using "umm" or "uhh". You will likely be guilty of this to some degree until the day you graduate, but you will be better than when you started working on research. I am guilty myself, that's for sure. I think the best remedies are
    1. Slowing Down. If you have to speed through your talk, it's far too long anyway.
    2. Practicing
    3. Having someone tell you when you do this.

• Don't let the first words out of your mouth be your talk title, verbatim

  • You will almost always have someone introduce you and your talk. They will almost surely recite your talk title during this. Don't start your talk by repeating what the introducer/session chair just said, it suggests you are reciting your talk from some kind of script.
  • An above example of what not to do looks something like:
    • Introducer: "We are pleased to have Mr/s. X giving a talk! They will be talking about "your talk title here".
    • You: "Thanks for having me! Today I am going to be talking about "your talk title here".

• There is too much text on your slides.

  • Figure out what your message is and then remove text and speech that does not help deliver this message. Here is a simple example.
    • Example: You list, to five decimal places, the coefficients of the stiffened gas equation of state you used to model the properties of water.
    • Problem: No one cares and they assume you are sufficiently competent to use reasonable ones. If your results are sensitive to this information then you would most certainly be giving a very different talk indeed.
  • The point: This is not a journal article, and not everything needs to be explicitly stated or referenced. Only the things that contribute to your story.

• I can't hear you.

  • This isn't usually technically true, my hearing is pretty good, but I recommend you project your voice into the room and speak clearly. Not doing this is one sure way to lose the attention of the audience.
  • Example: Many times we acknowledge that we could barely understand what a speaker was saying. Rarely do we say "man, they were way too loud, that was obnoxious."
  • It can sometimes be hard to judge if you are speaking too quietly. One way to know for sure is having someone tell you, which I will do. Another reason students speak quietly or are simply unclear is because they are nervous, bringing me to...
  • Nervous about public speaking? You are in good company, 77% of the US population feels the same. This isn't to undermine the issue, but support it. The best therapy is likely exposure therapy. This is, in part, why you will stand up to give your group updates each week, give practice talks, and give low-stakes talks at various departmental/student seminars. If you struggle with anxiety, consider seeking a mental health specialist.

Q&A

After a talk, there will usually be a Q&A period. I like to think of it as discussing with a colleague, just in front of a large audience. That’s the tone I’m aiming for in my questions and answers.

If you are answering the questions (you're the speaker)

1. Remember that you almost certainly know more about the details of your work, especially the very specific work you conducted, than anyone else in the room.

2. If I'm unsure what the questioner is asking, I make sure I understand it before I start answering. Typically, I will restate the question in my own words and ask if that’s what was meant. This allows them to correct and reword. Note that many questions are misunderstood on first listen, even if both the questioner and answerer are well-versed in the topic. An interesting side effect is that it also gives me a few more seconds to think about a possible answer. This is an even better trick if English is not your first language.

3. It is okay to take a second or two to think before you start talking. After all, the asker has been sitting there for 15 minutes considering the question; you only just heard it! And after all that, it's entirely okay to say you're not sure, and that you need to think about it harder and get back to them later. I have done this many times during Q&A, and there is nothing embarrassing about it.

4. If someone asks what about X?, and you didn't perform a study involving X and aren't entirely sure what would happen if you did, then you should:

   1. Say "we did not study X directly" (you want it to be crystal clear what you did and did not do), and then
   2. Follow up with a hedged hypothesis, if one comes to mind. For example, "I would expect that if we looked into X, we would see an outcome something like (outcome here), because (reason here), but I cannot say for sure." If you don't have a plausible hypothesis, say so, as in 3. above.

5. Good etiquette: It's generally good etiquette to tack on to the beginning of your response, "Good question!". Some broad guidance: Do not do this for every question, and, if appropriate, consider using a variation of this remark; for example, after doing some thinking while answering, one could say "...yeah, that's quite interesting, I hadn't considered that before. Thanks."

6. Wait until the questioner is done asking the question before you start answering.

If you are asking the questions (you're in the audience)

1. Do not ask "questions" that do not end in question marks. Use Q&A to learn or clarify something. If you have a comment, bring it up with the speaker one-on-one. If you listen carefully, you'll notice that some people ask "questions" that are not questions. This is generally frowned upon.

2. Do not ask two questions at once. This means two things:

   1. Do not ask two questions without letting someone else interject their question after your first one. It is impolite to take up all of the Q&A time.
   2. Do not ask two questions in a row (without letting the speaker answer the first one).