Race and gender can be (generally are) used by potential attendees to determine what comfort level they'd have at an event like that. The more people you can get to your event, generally the better the outcome will be - increased connections, networking and exchange of ideas from a broader range of people.
I made an earlier comment about my recent experiences organizing a conference, and linked to my blog which touched on this very reason.
Thanks, I hadn't considered that, and it's a good point.
Although perhaps I'm an oddity: I rarely see myself as among the speakers so I don't look at them as a guide. Also, I've been to "gender-equal" conferences where almost the only women there were speakers. So speakers != audience
Speakers != audience all the time - the audience won't be an exact match for the speaker demographics. But having done this for a bit, I've had enough feedback that indicates that having a more balanced/diverse set of speakers is a signal to a wider audience that there will more likely be a wider range of attendees to socialize and network with. Again, yes, not 100% all the time.
FWIW, at indieconf we had about 30% of the speakers were female. About half of the attendees were female as well. Interestingly 30% answered the 'gender' question as female - 30% answered 'male', 30% left blank - I'm assuming based on names given about half of the remaning were also female - that's where I'm getting the ~50% were female.
I made an earlier comment about my recent experiences organizing a conference, and linked to my blog which touched on this very reason.