In my experience, the majority of conference paper presenters fall in one of two categories: (1) Ramblers: these people attempt to hit the highlights of their work off the top of their heads and end up rambling on in no particular direction for the majority of their time. (2) Readers: in a sense, these people are the opposite of ramblers, so petrified to stray from their prepared ideas-or too lazy to prepare a presentation-that they end up just reading their papers (often in monotone, without looking up from the lectern).

Given that the bar is set so low in terms of presentation quality at most conferences, I think someone who puts in even a modest amount of serious preparation can come away looking very good by comparison. Here are my main tips on how to do this:

(1) Present for no more than 15 minutes: Even if you are given more than 15 minutes in your panel, 15 minutes is plenty of time to cover the main point of your work-and the audience, which often has sat through paper presentation after paper presentation before yours, will thank you for your brevity.

(2) Summarize the main point of your work: A conference presentation isn't the proper forum to deal with all of the nuances of your argument and subtlety of your thought. Hit the main point, hit it hard, and hit it in a compelling way. For a 15 minute talk, I would do the following:

(A) Hook (3-5 minutes): At the outset, you have to capture the audience's attention. Although they are at your session, you still need to convince them that what you have to say is important. You need a hook: "This work is important because..." The hook is usually broader than the study's topic, and sets the stage for the study. Roll immediately from the hook to the particular research question or problem you're working on. Show the audience how your work helps to inform the issue of broader significance which was your starting point.

(B) Line (5-7 minutes): From the research question/problem move immediately into your data and methods, which you should cover only enough so that the audience understands where your findings come from. Then spend the bulk of this time summarizing your key findings. Don't get lost in the details; again, hit the main point and hit it hard.

(C) Sinker (5 minutes): Draw out the conclusions/implications of your work. Suggest further avenues for research. Theoretical implications. Policy implications. Future directions for society. Whatever you can say to take the audience beyond the narrow confines of your particular study to some broader realm of significance.

(4) Bring copies of your paper: In addition to the handout, you might also bring 5-10 copies of your paper to distribute at the conference. Sometimes, after your panel is over, people who are interested will ask for copies of your paper. It's nice if you can hand one over rather than promising to send one when you get home. Some people are now posting their papers on the web and giving out their URL instead of handing out papers. If you do this, you can note on your handout that "the complete paper is available at http://...etc."

(5) Don't read your presentation: I remember as an undergraduate at UC-Berkeley going to hear the eminent anthropologist Clifford Geertz give a series of lectures. I was so excited going in, but couldn't have been more disappointed going out. He read verbatim the prepared text of his lectures, which were written to be read by his audience not to be heard. That is, the sentences and paragraphs were so complexly constructed that it was impossible to follow the logic of his talk, even if you could stand the constant barrage of reading. As an undergraduate who had been subjected to endless hours of droning lectures, I was secretly pleased to see a number of eminent Berkeley professors sleeping through Geertz's lectures, as I had wanted to do so often during theirs!

At the other extreme is the British social theorist Anthony Giddens. As a graduate student at Wisconsin, I saw Giddens deliver an hour long lecture that was totally coherent, intelligent, engaging, and witty--and delivered without any text at all, not even written notes!

My approach to public presentations falls somewhere in the middle. For fear of completely blanking out before my audience, I always write out completely the text of the presentation I am going to deliver, like Geertz. But, more like Giddens, I write my text with the understanding that it is going to be heard not read, and I practice my presentation enough that I can deliver it not read it. I may read parts of the text--e.g., specific quotes from authors or interviewees, precise wording of survey questions or statistics--but never the entire presentation.

Although there is no scientific basis for my advocacy of these practices, they have worked for me in the past. And, after all, "data" is just the plural of anecdote, right?

I welcome your feedback on these comments.
e-mail David Yamane