Duke University's Matthew Hirschey says that even if the department an interviewee is visiting says it's OK to use Powerpoint to give a chalk talk, it's not. "Don't," Hirchey writes at his lab's blog. "Powerpoint is a crutch." Because chalk talks are not meant to go as planned, it's best to "embrace the flexibility and use the board," Hirschey adds.
As for the talk itself, Hirschey says its best to "put your work into context." Start with a summary, he says, and then move on to outline future aims. All the while, it's important to keep the big picture in mind. "Experiments often fail. But the big picture will remain," Hirschey says.
Flexibility is also key. The chalk talk is "an important time for you to see how the department responds to your ideas," Hirschey says. "Clearly you can address a problem in several different ways, and the department will have their own way of thinking about your problem. This is a good thing: you can often get better ideas, avoid pitfalls, and come out the other side with a more clear idea of your project," he says, adding that "this will help you better realize how your [potential] new colleagues think, how they approach problems, and how useful they'll be to your science."
Finally, Hirshcey suggests highlighting the research of other faculty members in the department during the chalk talk. "A candidate should sell his/her fit in the department, but go one extra step and think hard about how to do this," Hirschey says. "The Department will likely be invited four [to] 12 people to interview and give a chalk talk. ... They will remember that great candidate who came in, has an interesting research theme ... wants to collaborate; and nailed the chalk talk."