The New York Times Freakonomics folks had an interesting run down of studies on cycling safety a couple of days ago, under the title Will Bicycling to Work Get You Killed?. Timely considering another recent fatality here on the West Coast.
The intuitive answer is that the more riders there are, the higher the chances of a fatality. The article sites a counter-intuitive result from the Safety in Numbers study (Full text): “An individual’s risk while [cycling] in a community with twice as much [cycling] will reduce to 66%”.
Weirdly, the Freakonomics article morphs this study (and a few other links) into a discussion of cyclists obeying traffic laws and helmets. The cited study actually draws a very different conclusion: more pedestrians or cyclists cause motorists to behave differently.