You are assuming that this must, ultimately, be a distribution with a single mode. Given the survey, and the broad nature of the population answering the survey, there really isn't any reason to expect that to be the case.

I selected one of the low incident rate answers, but I am mostly a recreational cyclist and I commute only rarely. Many of your respondents appear to be dedicated commuters. As noted by one, nova I think, they will be riding during peak drive times, when drivers are rushed and stressed and probably at their worst.

I think that it is possible for a "calibrated" rider, who has a specified riding style and specified tolerance for near misses, to have the kind of bimodal distribution of responses given the distribution of environments.

Speedo


well i tend to ride at totaly diffrent times of the day. If you realy wanted to get accurate results in a study similar to the poll youd need a min of 4 sets of 3 cameras per mile of road way. Then see how each car passed other vehicals how often they breaked etc. You cant base the safty of a style of riding on any one person or area. I ride with a effective technique for me. It basically involves changing how i ride based on where i ride. Id not be all that nervous riding on a full blown no bikes high way with 70 to 100 mph trafic. corse highways tend to have huge emergency break down lanes. Would be like haveing a entire road to my self lol.

I did ride on a small section of 585 once (between portage/wooster) and the rest of wooster. This was years ago and it was itehr that or ride god knows how far to get home. Some one had a pretty nasty accident oat the bridge basicaly lost their entire full tank of gas on the road way. Cops closed the bridge off and detoured every one along that section of 585.