evidence for the proposed measure of community strength?
I have to say, I've never been so keen on 5 edits per month being the definition of active. That's incredibly few edits. I know it's always really hard to draw lines and any number one comes up with is going to be arbitrary. But five edits is just a little over one a week. And, really, how much can be achieved making one edit a week? Five edits would barely constitute one sweep of a day's changes on a watchlist of a thousand articles.
Obviously ONE edit could be the drafting of an entire new article or an entire draft policy that becomes a tremendous boon to the project, but I suspect people that do that kind of thing are also people who make a large number of edits.
All that said, I don't propose we change it now!
There is some evidence that using the 5 edits per month will show less attrition, based on other attrition statistics.
Incidentally, Erik Zachte added some history on the "active" stat on the active contributors page.