Originally Posted by
John Forester
Both types of facilities were installed on the roads already thought to be safer and/or, for bike lanes, already with the width to install bike lanes. So, if one were considering a route to take, one might do better by selecting a route so marked (though that's not at all certain), but that is no evidence that providing the marking actually reduced car-bike collisions from what they were before.
The roads of the survey are those frequently used by cyclists, so they would be thought to be safer regardless. And the greater widths of the lanes are part and parcel of bike lanes - you can't feasibly just eliminate the extra space from considerations and say this way, "but the bike lanes were only safer because there was already room for them."
While I agree with Forester that a study comparing only the same avenues with the same traffic before and after installing bike lanes would be preferable, his objections fail to reasonably dismiss these studies.