Research talk:How do editors work anonymously?

Terminology

It appears that "anonymous editor" and "IP editor" are the same thing; if so, that should be explicitly stated, and one term or the other used consistently.

My sense is that "anonymous editor" is the preferred term at Wikipedia, rather than "IP editor", but it does have a certain irony, since registered editors are generally still anonymous; most have not identified themselves in any way (as to who they are in the "real world").

Relatedly, I can't make any sense of this question: "How have registration patterns of IP editors changed?" That doesn't make sense to me because IP editors are not registered editors; the two are in fact exact opposites. (If the question has to do with people who started out editing as IP editors, then registered, then there are real privacy issues involved in generating and handling such data, since any publication of such detail would be a massive violation of privacy. Wikipedia has an extensive system of controls to allow a very select, trustworthy few people to look at the IP addresses of registered editors, to address the issue of prohibited multiple registered editor accounts.) John Broughton 14:46, 7 June 2011 (UTC)

In my perspective, I see non-registered editors referred to as both anonymous and IP editors. Example use of IP editor[1]. Example use of anonymous[2]. If you'd like them to be consistent, please feel free to make the change. I'm not sure who decides what a "preferred term" is, so I suggest you change it to whichever you like.

As far as registration patterns, you're right and the question doesn't describe what a registration pattern might be. Again, we are still filling in details and expanding questions. Thanks for pointing out the confusion. I'll add an explanation. --EpochFail 16:59, 7 June 2011 (UTC)

John, I just to answer the last part of your question: our standard practice in all the research is to publish anonymised data only. The Contribution Taxonomy Project may eventually publish lists of users who participate heavily in certain public activities, but I can firmly say that any non-public data the Wikimedia Foundation has access to will remain non-public regardless of our research. Steven Walling at work 20:40, 7 June 2011 (UTC)

Anonymous editor may well be the more common term, but those of us who think it misleading and incorrect will tend to use IP editor instead. As for "the registration patterns of IP editors" I took that to mean "to what extent is IP editing a precursor to registering an account" and what proportion of IP editors are:

1. logged out registered editors
2. casual editors who might go on to create accounts
3. active longterm IP editors with no intention of creating accounts
4. complete one offs who've fixed one error they came across but may never do so again and wouldn't have bothered if they'd had to create an account.

I'm not sure the best ways to collect all of that, but the first and second could be answered by surveying a sample of registered wikipedians and asking them if they edit logged out and how many edits if any they did before creating an account. I suppose all those questions could be answered by surveying a sample of IP editors. Some active longterm IP editors could be identified by searching for IPs with a large number of edits and no blocks, but that won't find them all as not everyone has a stable IP, and you can't extrapolate because I've come across at least one IP who has said he doesn't need an account because he has a stable IP address.

As for statistical research using checkuser level of access, which I think is what Steven Walling might be alluding to, then I would like to suggest the following would be contentious or otherwise:

1. Statistics on new registered editors, showing how many were created by IP addresses which had just edited. Possibly with stats by amount of editing. (uncontentious)
2. Statistics on new user names created which don't then edit, showing how many were created by IP addresses which had just edited. Possibly with stats by amount of editing. (uncontentious)
3. statistics on prolific IP editors showing the proportion of edits reverted and ratio of blocks to unreverted edits. (uncontentious, and wouldn't actually need checkuser data)
4. Statistics on the amount of IP editing done by editors after they've just logged out. (interesting but probably far too contentious to do) WereSpielChequers 14:15, 8 June 2011 (UTC)

If "Develop model for predicting sameness of an IP editor and a newly registered user" meant working out the probable IP address of newly registered users based on related IP edits then that would be a clear breach of our privacy policies. I suggest rephrasing to something clearly statistical such as "develop model to predict the proportion of newly registered editors who were formerly IP editors and chart the results. Apocryphally many registered editors start as IP editors, it would be good to know whether that route to becoming an active editor has or has not been affected by whatever process has reduced the flow of new editors. WereSpielChequers

IPs who register

We have an overlap here with Research talk:Alternative lifecycles of new users#Policy callers, with citing of policy posited as one possible sign of a "newbie" who has been IP editing for a while. WereSpielChequers 09:21, 9 June 2011 (UTC)

RQ5 from Wikimedia Summer of Research 2011/Questions

The RQ5 summary at Research:Wikimedia Summer of Research 2011/Questions says the following:

A substantial proportion of the edits to Wikipedia are from editors who choose not to create an account and instead have their edits logged to their current IP address. What do we know about this part of the community, are they changing? Are they experiencing a decline that is similar to the tail off of new editors with registered accounts?

Since it is explicitly in the charter, shouldn't the last sentence about the tail-off be added as a 4th question to the Questions section? 72.244.200.204 20:27, 9 June 2011 (UTC) (contributions)