I was watching a documentary about Stephen Hawking the other week. Among the many interesting things I learnt was the fact that he doesn't use handouts at his talks. Or at least, he didn't at the talk that was shown in the film. I'm assuming that to be representative of his general practice.
For those not accustomed to the wonder that is the conference handout, an example is here. Its purpose varies. For me, it's a prop for your audience, a summary of what you're saying so that they can follow along while you're saying it and have a record afterwards. I think it should be clear enough that someone who wasn't there can read the handout and understand the basic idea of the talk. It might, if you use Powerpoint, be a printout of the slides (often this is what you will get).
For other people, it's simply a place to put the data. None of what you say is on there, it's just a list of examples, tables and whatnot so that you can say 'Now look at (3)'. This, to me, is a waste of time. Make a presentation of the data, or make your handout more informative. A pure data handout is not going to mean anything to an audience member a year or even a month later.
I think in other disciplines it's almost an accessory, like an added extra - I went to a literature talk given by a friend and she had a one-page handout that just contained two (long) quotes relevant to her subject.
And in other disciplines, a handout is never used at all. Generally, a presentation is given (with Powerpoint or similar), or even just a talk, with no visual material.
Within linguistics, whether or not handouts are used seems to depend very much on which branch of linguistics it is. I was once at a general postgraduate conference, and in my session, we had handouts, and in the parallel acquisition session, no one did. At a syntax conference, almost everyone has handouts. It's simply the done thing in syntax, to give a good, comprehensive record of your talk to the audience (even if, as I mentioned before, this is just the printout of the slides). I think this is because we often have examples from other languages, which take time to digest, and diagrams that one might need to refer back to. In reality I think it's just custom/tradition/what everyone else does. But either way, it's really good to be able to refer back to difficult examples or trees, or take a bit longer to study them than other people.
I would have thought that physics, with its equations and so on, would be similar to syntax in its attitude to handouts. I often think of theoretical physics as sort of equivalent to theoretical linguistics (though I know little about it, it seems logically similar). And I'd expect a physics talk to be reasonably difficult to follow without some kind of visual aid, so I was quite disappointed not to see Hawking's assistant dishing out nice thick wodges of paper. But we should be saving paper, so perhaps this is the right attitude to take.