I used to think that a great talk started with a somewhat brief introduction that described only the context and main questions of the work. It wouldn't have any extraneous information before the data. That way, the data would seem puzzling to the audience until the speaker revealed the lynchpin concept in the discussion that would draw everything together. Aha! the audience would think, and they would perceive the speaker as a brilliant, creative mind
But, I've started to notice that a great talk makes the audience feel smart and important. It should be about the audience, not the speaker. Like entertainment. If they feel good at the end, they will think well of the speaker.
Hence, I have changed my mind about the structure of a great talk. I now think that the introduction should be a bigger part of the presentation. It should include enough background information that the audience can attempt to interpret the data as they see it. That way, they can make their own journey of discovery as the talk progresses. They can feel so clever if they figure out the interpretation before the speaker presents her own conclusion. Or, if the listeners had been forming an interpretation that is different from what the speaker suggests, they can think oh, now I get it or be ready to ask good questions.
I think this approach may help the audience stay engaged through the data, since they have the information they need to understand it. Without this, I think it's easy to lose people through data slides they think are boring and then they don't pay attention to the big reveal anyway.
What do you think? What kind of structure makes the best talk?
11 comments:
Yes, I think it's definitely a mistake to leave the data interpretation mysterious until the end. I like it when the speaker sets up a question and says there could be 2-3 possible answers to this; to figure out which one is right, we did this experiment which would give this result if (a) is correct and that result if (b) is correct. And boom, (a) is correct!
Great post. I think you're right that it's important to give enough of an introduction to allow the audience to follow your story. I also like the suggestion from Andrew to list a number of possible explanations in the beginning. I'll bookmark this post and hope I'll remember to look at it before I'm preparing my next talk. I usually structure talks the boring way, and think I should make an efort to do better.
I completely agree with you. When I present my research, I try to convey enough background information to show why I'm really excited about this line of research and what the unanswered questions are (that I am addressing.) I think when your audience sees your research as important, they pay more attention. I've been to super boring talks where I thought "yeah, so? Why are you doing this?" And completely tuned out.... (even though this was my former advisor, yikes!)
I completely agree with this approach. By giving enough background you answer the "So what?" question right away. Also, I tend to do what Andrew does with giving the audience choices as to what the results mean. It has the added benefit of teaching the audience how to interpret your data (I use some techniques that my department is largely unfamiliar with).
Andrew & Amanda, that's a great idea! I don't think I've seen that approach before, but I can imagine that it would be very effective.
It's much easier to see what's boring about someone else's talk than to weed the boring out of your own presentation.
This is a great post. As you have described it I was reminded of how a good manuscript should go. Talk and papers are very similar in context and organization, however talks are more condensed and you don't always give enough details in the methods section (way too boring for a talk).
I recently attended a talk where the speaker was trying to get the audience to guess the meaning of the results by asking pointed questions and calling on anyone that made eye contact. Perhaps if the talk provided more intro like you mentioned, that part would have gone smoother ;)
I agree that giving talks is so much about entertaining your audience. I feel this way about being in the classroom too. I often feel like I am half entertainer and half educator - and that's what I love about it. I like taking a group or audience from being a bunch of individuals to an intimate gathering of intellectual friends.
Excellent post!
I always like to say a talk should be given as if you are taking the audience's hand and leading them down the golden path. Which is to say that you should provide a good introduction so that they know why what you are doing matters, its relevance etc. Each slide is giving them more information and making them feel more connected to the data. Like a good novel, you want them to be thinking about what is happening, giving them hints (forshadowing results). Above all it should easy to understand / follow.
I saw a very interesting talk yesterday where the speaker hung her results from a framework of different kinds of model systems, and what each can tell us. So rather than "here's the background, here's what I did, here's what it means", she described why she chose each model system for each set of experiments, with a discussion of the advantages and restrictions of each type of model. Parts of it worked better than others, but I really appreciated seeing a fresh structural approach to the concept.
After today's talk, I'd just like to remind everyone to be consistent in their slide transitions (or not use any at all). It's very distracting when one new slide glides in from the left, the next one emerges from the centre of the previous one, the next one folds down from the top, and the next one appears from a mass of bubbles.
Seriously.
I like your revised version too. I often give talks to people outside my field so I do try to answer the "so what?" question right up front.
Post a Comment