She began with a couple of figures taken from someone else’s paper. They showed exactly what that other study measured. Sara Mitri told us what motivated that study. She then went on to clearly explain how her study would be based on the earlier work, yet extend it. Even though we are short-term visitors to Oxford, not in this lab group, Sara was clear.
One of the best things about giving a lab meeting talk is that it forces you to step back from pouring gels, watching videotapes, pipetting, plating, trapping, tagging, integrating, or whatever it is your empirical or theoretical work requires on a daily basis. You step back to think about what is really important, what you have discovered, and what challenges you face. You also think about how to present the material in a clear way, beginning with motivators and ending with unsolved problems.
You could just about use Sara’s talk from today in Kevin Foster‘s lab group as a formula for a great lab meeting talk, though other models also work. Here you go:
1. Show what went before, citing other people’s work or ideas, with a few figure slides.
2. Point out a natural extension or branch off the earlier work that you will take.
3. Discuss the plan of your work, but don’t make it too much like a boring outline.
4. In the heart of the talk make it clear what you did and how you did it, reminding us of the tie to theory.
5. Wherever possible precede showing of the data with showing what you expected the data to look like.
6. This is lab meeting, not a finished piece of work, so tell us about your problems. Show us your enigmas.
7. Keep to time, usually under an hour, and allow for questions, discussion from whether it was a good plan to start to what changes might be made, from minor to major. Wonder about fundamentally different interpretations of your data.
8. Remember NOT to use a laser pointer. They are so distracting.
It was really fun to hear her talk in Kevin’s group, and to see his own group around him, a former Rice University Huxley Fellow shining in his own career.