1. Those that aim their talk at professionals in their field.
These talks, although seemingly impressive, are difficult to follow and/or get anything out of unless the audience is in the same field as the presenter. These talks generally have little introductory material, lots of equations, and no personal touches. I have actually seen a couple of these done well (or, I guess, what I could assume as being done well).
Benefits: it makes the presenter look incredibly smart; there are typically few questions; the examiners will appreciate the level; tons of data can be included.
Drawbacks: most of the audience will be lost; there might not be any questions (never good in a talk).
2. Those that try to aim their talk at the general public (or at least general science population).
These are my favorite, because the speaker is obviously enthusiastic enough to make sure that most of the audience will understand a good portion of the talk. They typically spend a good chunk of time giving background info and motivation for the project, and only focus on a few key results. The slides are well thought out with few equations and text, but many interesting and useful images (that aren't directly out of a paper).
Benefits: most people in the audience will come out with new knowledge; there will be lots of good questions at a more basic level (helps with confidence!); it's way more fun because personality can be put into it.
Drawbacks: there might be general questions that were not thought of beforehand; it takes a long time to perfect the slides because only select data/results are used.
3. Those that throw their talk together the night before.
We've all unfortunately been privy to these talks: there is no apparent transition between ideas, the slides are sloppy with a lot of text, and there is no overview of the research area or motivation for the project. Some people can pull this off...but not as many as those that actually try.
Benefits: um..not as much work?
Drawbacks: pretty much everything else.
For my public lecture, I'm sure you can figure out I will go with #2. I'm of the feeling that it's called a "public" lecture for a reason! I love having people understand what I'm talking about/doing. Plus, the part of my research that I really get a kick out of is the history and background, so I'll get to spend a decent amount of time on that part.
Because of my outreach experience, I'm almost worried I'll pitch it at too low a level. Although, I think a talk can never be too simple - people like to be able to follow along, even if they already know the subject well. It makes them feel smart!
If you have done a PhD public lecture, what was your plan of attack? Who did you aim the talk too? Is there anything you would have done differently?