I have nothing against anecdotes. The problem is data gathered in a
selective manner. Suppose one person in a hundred who is sexually abused
develops some disorder. That person goes to a therapist. From the
therapist's perspective abuse can lead to this disorder, she has direct
evidence. But how many people who have not been abused develop the disorder?
No way to know. Nor is there any way for the therapist to know what
percentage of the abused population winds up with the disorder. I know that
you know this, but it is the reason we have to be very careful about drawing
conclusions from non-random sample. By the way, I own a copy of Birkhoff and
MacLane that once belonged to you. So there... (I know, I know... You've
wondered for years where the book had gone, and now you want it back.)


The Insufferable One