Definitions

I think that three categories might be needed, "New", "Occasional", and "Active Editor". Or two dimensions: "activity" and "wiki knowledge/expertise". A regular editor of some Wikipedia who only makes 1 edit a month in English Wikipedia is very different from a new or occasional editor across Wikimedia projects.

The trouble is finding the proxies to measure "wiki knowledge/expertise". Expertise comes from experience in other wikis (especially MediaWiki instances for formattig, or other Wikimedia wikis for rough process/policy, though that does vary). Also from *reading* and contributing to Talk pages, policy, Wikipedia namespace discussions, XfD, Village Pump,... Or from off-wiki discussions on mailing lists or in person. Some of this comes from Readers are not free-riders: Reading as a form of participation on Wikipedia, which I highly recommend.

As for activity--I think it also matters whether the person has made a series of related edits (e.g. I often correct my edits and it could take three 'tries' to make one intended change) or a variety of different edits. Finding social roles in Wikipedia (PDF) may give some ideas for thinking about HOW activity varies, besides volume.

Details on the papers suggested above. Antin, J., & Cheshire, C. (2010). Readers are not free-riders: Reading as a form of participation on Wikipedia. In Proceedings of the 2010 ACM conference on Computer Supported Cooperative Work (pp. 127-130). Presented at the CSCW 2010, Savannah, Georgia, USA: ACM. doi:10.1145/1718918.1718942

Welser, H. T., Cosley, D., Kossinets, G., Lin, A., Dokshin, F., Gay, G., & Smith, M. (2008). Finding social roles in Wikipedia. In Proceedings of the American Sociological Association 2008. Presented at the American Sociological Association, Boston, MA, USA.

Dear Jodi,

1) I agree with you that a more granular roles differentiation is possible, but the question we are asking is which type of editors are leaving: the new ones or the more senior ones. Thus, we want first to shed light on who is leaving. A subsequent step could be a further categorization as you suggest to paint a more detailed picture.

2) Thanks for all the references, I will download those papers and have a look.