I've seen reference to that study (the one that shows motorists driving closer to a cyclist wearing a helmet) on here for almost two years now. I even read the in depth study itself, though I can not find a link to it any more. But the study is somewhat one sided and really has no control to it.
Take a cyclist, slap a helmet on him, and stick him on a suburban arterial thruway in the middle of rush hour... yeah, people are going to pass closer just because of their frustration.
Now take the same cyclist, leave him helmetless, and place him on a residential street in an area why cyclists are plenty at thee o'clock in the afternoon... people are used to cyclists and give him a wide birth.
see the problem with this study?
They should have performed it at the same place and time using the exact same drivers under the exact same circumstances, though this is somewhat impossible.
My point being, this study, and all like it, have really never had any consistancy to them.