That's actually a good point.
How can we assess if editors are leaving because they reached the
limit of *their
*knowledge, or for other reasons?
Facebook, for example, has a rather annoying feature that if you try to shut
down your account, it asks you why (and tries to guilt you into staying).
Wikipedia does not allow you to shut down your account, but a similar (not
so annoying, hopefully) feature could provide some more tangible data.
The absence of this data should not however ilude us to the fact that the
editing interface is one of the major hurdles stopping people from
contributing, followed closely by a lack of or aggressive feedback.
But, we are approaching some limits.
Let me explain it through one simple example.
Let's say that there is Wikipedia in X language with just one editor.
That editor is expert in, let's say, medieval history and has passion
toward chess. That person would spend years in: (1) writing basic
articles -- although he is not astronomer, he knows that it is
important to have articles like "Sun", "Earth", "Jupiter"
writing articles in medieval history; (3) writing articles about
chess; (4) and, finally, writing articles about surrounding areas of
medieval history and chess (let's say, ancient history and go).
If that person didn't stop because of lack of time or lack of
satisfaction, it is reasonably to expect that he will at some point
come to the situation where all articles are written according to his
level of knowledge. (That's the ideal situation, but it also assumes
the ideal systematization of the work on articles, which is not