Some considerations (1)

Some things that might be considered in doing the survey:

• There are two types of editors: registered and anonymous. Edits by the latter are recorded as IP addresses. Some IP addresses belong to multiple people (for example, a couple of years ago, a single IP that represented an entire country - because that's the way the ISP had set it up - was temporarily blocked by a Wikipedia administrator); some are dynamic. Any survey ought to be distinguish between the two types of editors. (And yes, a few long-time editors do use only an IP address, though this is quite rare.)

• While the Wikipedia database contains all edits (with the possible exception of oversighted edits), when an article is deleted, that edit no longer is visible via the Special:Contributions page. It's not clear that this makes that much difference, population-wise (when sampling), but since more than a thousand articles are deleted every day, and since the majority of these (I'm fairly sure) were created by new editors, it's possible that there are a lot o cases where what looks like (say) an editor's second edit is really his/her twelfth, if (for example) he/she did ten edits in one or two new articles that was/were speedily deleted. John Broughton 00:59, 16 February 2009 (UTC)

• You definitely need to make sure that bots aren't in the sample; they often have very high edit counts, but since they aren't actually human beings, they are irrelevant to the quality of the user interface. More problematical is editors who use specialized tools, typically for vandal fighting or doing mass cleanup of typographical or other common errors. These tools (Twinkle, Huggle, WikiCleaner, etc.) have their own user interface. (Such edits usually do contain a link to the tool that was used.) John Broughton 00:59, 16 February 2009 (UTC)