They logged in to do something, and have presumably done so other times as well. I agree that the data is not that detailed, but that's what I had to work with. My point was that perlmonks appears to have a pretty good addiction rate, and I think that is a good thing.

Other statistics I thought of gathering, that perhaps someone else will take up:

• Number of monks who joined and shortly thereafter asked a question who stuck around and contributed more (to see if perlmonks offers more than just a one-time help desk to these folk). Obviously, not having AM's included in this is a flaw.
• Some kind of evidence of progression from asking questions to also answering them.