Originally Posted by I-Like-To-Bike
You can start with his comparing the alleged safety record of 8 year olds with the alleged safety record of middle aged experienced cyclists to draw conclusions about the effectiveness of his training program. Or try his cock-eyed risk analysis of cycling "crashes" and "falls" and "accidents" as identical issues with no consideration of exposure, probability or severity consequences; another Forester 10 bagger of Sophistry.
Well, from what I can tell, he repeats the analysis that others present as evidence. Or sometimes uses the same techniques that others use as evidence. Or simply works with the limitations of the data.
I recall reading the example you reference (vaguely) ... this is on his website, right? But I don't recall him over-stating the evidence nor obfuscating the problems with the data. Could you explain your point in more detail?
More generally, I think that there are important differences when somebody writes "this proves X" versus "this is evidence for X" versus "this is suggestive of X" versus "this does not conflict with X" and so on.
In the past, I wrote that I found it hard to support many strong statements about cycling safety because the data is generally bad. But it also means that if you reverse the null hypothesis, it will be difficult to disprove many statements. Like other fields, my anecdotal experience is that this leads to theory being the basis of many actions without enough consideration to the opposing view.
Hmmmm, I think that we are getting a bit far from Diane's original post. So I will stop here until I read her response.