My (so far untested, as far as I know) advice is to start off with a talk designed to give people the sense more is possible, i.e. that this stuff is deeply interesting and relevant and worth paying attention to, rather than trying to start teaching them something. Suppose you manage to teach them about a single bias; maybe base rate neglect, or availability bias, or so on. I'm having trouble thinking of a single bias that's exciting enough for people to see benefits from knowing about it, but I know there are several that can be detrimental.
If the audience includes kids, hindsight bias is probably the easiest and most dramatic place to start - you can even do a demonstration with them as the subjects.
At my local Harry Potter fanclub (Bogotá, Colombia) some members teach "classes" on subjects they're passionate about. We've had informal courses on history, creative writing, English, etc. But recently some other classes have appeared that have made me worry seriously: astrology, divination, ancient runes, and all other sorts of nonsense. They're not taught as folklorical pieces of the past, but as serious practices that are supposed to actually work. I think this is particularly dangerous for the small kids that comprise the majority of the fanclub and still need help learning that magic doesn't exist.
So I proposed the fanclub chief that I could teach a Muggle Studies class: logic, critical thinking, philosophy of science, etc. In two weeks we'll have our first class, and I intend to begin talking about the most common biases. I already downloaded this website's PDF guide to holding a Less Wrong meeting. Aside from that, what can you suggest for a successful Muggles Studies course?