zlacker

Sorry, all the statements in the study would need to be qualified like that (and most of them much more), and virtually none of them are.

I am assuming the study has taken some care to be representative. Otherwise, there's no need to talk about the findings at all.

>>mpweih+(OP)
It's not a matter of representativity. It's a matter of measuring a different thing.

The article measures shares of reasons between the people that left, but those are not correct by the share of people that left on each subgroup.

>>mpweih+(OP)
(See my edit if you haven't - I agree that the blog post is trying to claim something that the Kapor study is not saying, or at least is phrased poorly.)

>>geofft+H8
I see the same sort of thing in the "study", all over the place.