Hints for giving good talks (by Scott Keogh, ANU):
Do keep the message of your talk very simple, have only a few main points.
Do think and plan carefully about the structure of your talk - make sure it follows a logical progression.
Do be very prepared for your talk, practice, practice, practice.
Do plan on about four minutes each for intro, materials and methods, and results/discussion.
Do plan on about one slide per minute.
Do speak slowly and clearly.
Do have a joke or two ready if you are feeling confident, people like to laugh but be prepared to go on if they don't.
Do have very good text slides, keep them simple.
Do have good study species slides, at least 20% of total slide number.
Do have conclusions slides.
Do make sure that the people in the back of the room can see what's on your slides/overheads
Do state your aims clearly and explain WHY you have done the research.
Do try to reach as wide an audience as possible and reflect this attitude in the way you present statistics and complicated results.
Do make sure that every slide/overhead means something.
Don't read your talk if you can help it. If you are well prepared and have practiced you won't have to.
Don't use a laser pointer until you have practiced using one, be slow and deliberate with your movements or it annoys everybody.
Don't apologise for mistakes in your slides, just explain.
Don't show complicated tables or DNA sequences - tables are death. If you have to show tables, keep them very simple and only show the data that you are actually going to talk about.
Don't ever start explaining a slide by saying "I know you can't read this but....".
Don't go overtime, it's disrespectful to everyone and unprofessional.
Don't ever say "Now this is really interesting.....", just make it self-evident.
Don't mix slides and overheads to much - try and have one or the other, preferably slides.
Don't use more than three colours, and make sure they are highly contrasting.
Don't say "that's it" at the end, have an ending prepared.
Don't say "more work is required" because it always is.
Don't show raw data.
Don't go too much into well known methods.
Don't use the shadow of the pointer unless you say that's what you're doing.
Don't end with a slide of a sunset because that always annoys me!
