The reasons for this inverse correlation are many, according to Peck, and include the likelihood of better cycling infrastructure in areas where more people ride, the fact that if car drivers also occasionally cycle they are likely to be more careful with bikes, and the statistical quirk that a higher proportion of riders in low-cycling areas tend to be young men with a higher than average threshold for risk.
I love the “statistical quirk”. It’s good to put a scientific survey behind something that seems intuitively obvious. I would have also thought that drivers who are more used to cyclists on the road would drive more carefully, but that didn’t seem to make the cut. Boston/Cambridge is just going to keep getting safer and safer the more people ride.