How is administrator attrition being addressed?

I agree with the general tenor of this conversation. From the strategic priorities perspective, the key is to acknowledge the importance of multiple types of contributors. Time invested is a valuable way of contributing, but it's not the only way. There may be opportunities to contribute in valuable ways that do not require such significant investments in time.

For those who do have that time, there should be mechanisms to encourage the growth from occasional to active contributor.

I also agree that policy creep is a huge problem, although this hasn't been explicitly stated in the movement priorities, and I'm not sure that's where it would belong.

At least on the English Wikipedia, policy creep seems to have more or less stopped dead in its tracks. There are far fewer policies being created now. That said, there are definitely opportunities to simplify redundancies. But there's a lot of resistance. For example, there were a bunch of people who merged "verifiability" with "no original research" into a simple and clear message: "attribution". But it was overturned by Jimbo. I think that kind of killed any impetus for simplification.

I think a quick survey of formerly active English Wikipedia administrators, to find out why they left, is the first step to determining what is going wrong and how it can be addresssed.

I think this is an excellent idea.

Hey Foundation! Can you all make something like that happen?

I agree!

I think the survey need to make the distinction between active as an admin and active as an editor.

Behind this there is the question of whatever adminship is full time job or an additional set of tools.

I think this would be good as well. We got some of this data on the Former Contributors Survey Results, but this was not focused entirely on admins.

I'd encourage people who are interested to organize such an effort, perhaps basing it on the other survey. It doesn't have to come from the Foundation.

We can definitely draft up a survey, and use the old one as a template so that it's easy to compare numbers. But actually contacting the users and compiling the data, we'd need a little bit of help.

I'm sure we can arrange that.

A draft (in-development) is up at Task force/Community Health/Former administrators survey and a discussion is started on its talk page as well.