Nope, it's just the impression that I got when I cared enough to look around the web for information about the relative risks of cycling. A lot of numbers posted on websites with varying degrees of coherency, so take from that what you will.

You're right, but I was referring to the preference of most people to minimize their commute; a three hour commute is hardly representative of what most people are willing to do. There's a point where most of us will switch to a different mode of transport if we feel that cycling takes too long, and this will limit risk exposure. For short trips where travel times are the same, risk is probably about the same, and for longer trips the risk isn't that high anyway.

My school commute takes 20 minutes by car (easily half of which is spent sitting at stop lights =/ ) and, if I take the road and not the MUP, about 35 by bike; so I guess that I'm almost doubling my exposure risk when I cycle on the road. But the risk that I'm doubling is very low, and by cycling I mitigate other risks, like those for heart disease and obesity.

I feel (no data or science or whatever here, just a hunch) that low speed roads offer less risk, regardless of mode of transport, and when I drive I eschew them for higher speed arterials because I just want to get back out of the car as soon as possible; so by driving I tend to take more dangerous roads, which might equalize the risk exposure with cycling where I avoid high speed and high traffic roads whenever possible.