"logged-in only" editing

"logged-in only" editing

Yes, anonymous editing again. (I've cross posted this a couple of places.) I've been really skeptical of this, because I believe in openness, almost as much as I believe in privacy. But I found some new evidence that is worth weighing:

I did a lot of positive work as an anon/IP. But I was never treated that well. Definitely never became integrated into the community. Once I slipped into my username, I understood why. An IP feels like a ghost -- a username gives you the faintest outline of a person. And people were much more welcoming and supportive. I became part of a dialog and contributed more. And enjoyed contributing more. ... but I've never wanted to force anyone to sign up a user name if they wanted to edit... until now?

Honestly, if we want people to have a better first experience with Wikipedia, and users have better experiences than IPs... this starts to make more and more sense. The expert in the interview also seems to think it leads to less frustration for experienced users, and better content overall.

Randomran01:14, 23 January 2010

Right now on ru.wp we are running a vote on whether anonymous article creating (not editing!!!) should be discontinued, and it looks like the suggestion is going to be voted down despite the factual evidence presented that most of the articles created by ips is the speedy deletion material. They have of course all kinds of arguments like "creating inequality", "undermining the basic principles" and "nobody invited you to be an admin, please do not complain or quit" which I am sure must be familiar in English Wikipedia as well. Discontinuing anonymous editing at this stage is a no-go, I believe.

Yaroslav Blanter16:46, 23 January 2010

I am strongly in favor of anonymous and pseudonymous editing, but I think allowing unregistered users to create articles is a drain on administrator time and should be avoided.

There are disadvantages to editing as an IP, but there can also be advantages. Most of them are of the "no pain, no gain" variety and have to do with the ability to determine truth without knowing authorship. Also, privacy is important for people who have been threatened., 24 January 2010

Truthfully, the "privacy" advantage of an IP is overrated. Do you really want people to be able to see that you're operating using AT&T/BellSouth, from Richardson, TX?

But I agree with the "no pain" assessment. If you've always been an IP and never been a user, it's hard to get sucked into the drama.

Randomran17:40, 24 January 2010

Unless they think it will improve the community more in the long run, compared to some of the editors who would be demoralized (and even leave) right now.

... so yeah, you're probably right.

Randomran20:51, 23 January 2010

I too support is the "logged in only editing". But if anonymous users starts making accounts and do bad edits then ? Also to note that nice users will also suffer.

Extra99911:37, 31 January 2010

It would be wonderful if someone were willing to do an experiment on disallowing some forms of anonymous editing. I'm glad to see that Russian Wikipedia is at least putting it to vote. (Have others considered that as well?)

Strategy wiki is, in its own right, a project, and we could certainly consider running an experiment like this. If we were to do this, what metrics should we track? And what metrics could we point to now that would argue for disallowing certain forms of anonymous editing?

Eekim18:33, 2 February 2010

The conclusion to the discussion in Russian Wikipedia was drawn yesterday - to keep the option of anonymous article creation.

Here, on one hand, I do not see any point in allowing anonymous editing, on the other hand, a number of anonymous comments have been added to the treads, which means some users have chosen not to register for whatever reason.

Yaroslav Blanter22:28, 2 February 2010

We'd want to compare IPs to new users, and look for differences on a few metrics:

  • Reversions (e.g.: are non-anonymous users less likely to be reverted?)
  • Number of talk page edits, or time of first talk page edit (E.g.: are non-anonymous users more likely to be drawn into community-building activity?)
  • RVVs -- Reversions due to vandalism (e.g.: are non-anonymous users less likely to engage in vandalism?)
  • Blocks / bans / incidents (e.g.: are non-anonymous users less likely to engage in disruptive activity?)

The jackpot metric would be whether new users are more likely to continue editing than anonymous users. But because anonymous users have no consistent IP or username, it would be impossible to measure if they are editing for a long period of time. I wish we could figure out a way to get at this, though, in some roundabout way.

Randomran19:21, 2 February 2010
Bodnotbod00:34, 3 February 2010

I'd reject it too, off hand. But that's without any factual basis. I'm very interested to see what metrics follow from logged-in only editing.

Randomran04:53, 3 February 2010

For anonymous creation, we collected some statistics. Apparently, 55 to 75% of articles newly created by anonymous users are speedy deletion material, but the percentage is the same for the users who just registered. In other words, the usefulness depends only on the experience of the user, not on whether he/she is logged in. There are of course other issues like for instance impossibility to communicate with a dyunamical ip and others.

Yaroslav Blanter08:11, 3 February 2010

That is what I want to tell?

Extra99906:32, 7 February 2010

Do you have a link to that data? I'd like to have it bookmarked for further reference.

That's one nail in the coffin... suggests to me that trying logged-in only editing isn't worth the trouble.

Randomran13:50, 3 February 2010

Would Russian be ok? Otherwise, I can translate it, but the translation will take some time.

Yaroslav Blanter17:58, 3 February 2010

Now also read Eric's post below. I will try to translate quickly.

Yaroslav Blanter18:00, 3 February 2010

I'd really love to read that as well... and since the amount of Russian I know will fit in a thimble (mostly names of food), a quickie translation would be great... even if you just ran it through google translate for a first pass...?

~Philippe (WMF)18:41, 3 February 2010

A translation would help. But even just the raw numbers should be understandable, no? Either way, we'd appreciate any information you can give us.

Randomran05:52, 4 February 2010

OK, I post it here, please feel free to move to an appropriate place.

For the record, the whole discussion (in Russian): ru:Википедия:Обсуждение правил/Запрет на создание новых статей анонимными участниками. The statistics are summarized in the preamble.

Statistics 1: ru:Участник:Skydrinker/Анонимы и новые статьи. Summary: User Skydrinker monitored over the weekend 12/13.01.10 newly created articles (six times, different times of the day). In total, he looked at 300 articles; of these 300, 72 articles were created by unregistered users (24%). Out of these 72, 61 (84.7%) were speedy deletion material. To be precise, 8 - nonsense, 6 - empty, 19 - copyvio, 7 - obvious vandalism, 12 - not notable, 7 - spam or ads, 2 - not in Russian. Eventually, 6 out of these 61 have been improved by experienced users; 3 speedy deletion reviews were contested and sent to usual deletion review. Another 11 (=72 - 61) articles did not qualify for speedy deletion and eventually have been improved as a part of regular flagged revision procedure.

(to be continued)

Yaroslav Blanter22:02, 4 February 2010

Statistics 2: ru:Участник:Шуфель/Статистика (новых статей). 460 articles created in the period 22.12.09-31.12.09 have been monitored. Eventually, 122 articles have been deleted (8 be deletion review, others by speedy deletion process). Here is the summary:

user class articles created % of total deleted % of created flagged %(flagged/all)
ip 127 28% 80 63% 29 62%
not-autoconf 42 9% 29 69% 8 62%
user 106 23% 12 11% 48 51%
autoeditor 83 18% 1 1% 81 99%
editor 95 21% 0 0% 95 100%
sysop 7 2% 0 0% 7 100%
Total 460 100 % 122 27 % 268 79 %

The conclusion is that the quality of articles created by unregistered users is virtually the same as the quality of articles created by non-autoconfirmed users.

Yaroslav Blanter22:10, 4 February 2010

This is great stuff. As Bodnotbod pointed out, it's a perennial proposal, and it's also been proposed here twice:

Would people here like to take a pass at:

  • Merging the above two proposals
  • Adding case study information (for example, Yaroslav, it would be great if you would copy and paste your posts about Russian Wikipedia's statistics and votes on the issue)
  • Start brainstorming the metrics we should use to evaluate the possibility. Randomran, your list is a great start, and it would be great to copy it into a merged proposal and build on it.

If we can get a strong proposal written up, we can (and should) definitely consider implementing this proposal on strategy wiki.

Eekim08:29, 3 February 2010

Actually, that is completely wrong. These two proposals aim at something different completely, namely a move towards using real names (and with real personal data). Thus a move towards converting Wikipedia into Citizendium.

There was a proposal towards logged-in only editing: Proposal:Create a reputation system and require logging-in to edit.

But it is not wise to mix up these two (quite separate) issues. - Brya 17:56, 3 February 2010 (UTC)

Brya17:56, 3 February 2010

Thanks for the clarification, Brya. In general, I'd rather work off of an existing proposal than create a new one, but if the modifications are going to dramatically change the tenor of the original proposal, then we would need to engage with the author on the Talk page first.

If we need to, we should create a separate proposal.

Eekim00:05, 4 February 2010

PS. Wikipedia could be clearer in the information supplied about IP-numbers, and the consequences of showing these.

Also, it is not a given that the present policy will be allowable in future. Under Dutch law there are severe restrictions to publishing personal data, and this includes IP-numbers. It is a fairly new law and there is as yet no legal track record; so far everybody is ignoring it, but if somebody took Wikipedia to court to force a stop to showing IP-numbers they are very likely to win. - Brya 18:12, 3 February 2010 (UTC)

Brya18:12, 3 February 2010

I'm not certain about that, Brya - my understanding is that the WMF operates under US law and that we don't fall under the jurisdiction of the Dutch courts.

~Philippe (WMF)00:55, 4 February 2010

Of course US laws apply, but Dutch laws applies to Dutch citizens working on the Dutch Wikipedia and certainly also to the Dutch Chapter and any servers on Dutch soil. Other countries may have similar laws, so it is likely that from a global perspective the legal picture may be quite complicated.

Forbidding to publish IP-numbers does not have all that many consequences, as each IP-number can be assigned a random number, which can be published as long as the WMF keeps it a close secret which random number belongs with what IP-number.

Actually this may have more consequences for logged-in users. Although they are promised that a username guarantees a degree of protection, their IP-numbers are traceable: if IP-numbers were kept secret this would make a substantial difference. - Brya 05:24, 4 February 2010 (UTC)

Brya05:24, 4 February 2010

I want to tell something:

1. If we will block anonymous users they will create new account and start editing.

2. So for that we will only allow auto-confirmed users which means no more new users will come. Gradually, constructive editing will decline which is root of our success.

3. There are many good anonymous users and they will be also blocked including the good faith ones also. Most viewers are anonymous, and gradually, we will also lose popularity.

4. If you want to keep good anonymous users, it will create a havoc among the admins for blocking only those anonymous which do bad edits. If we will block good anonymous also then there will no problem, as we will remove the system of anonymous users directly, but good faith and constructive anonymous editors will be gone.

5. If you will allow only half articles or some to be edited by new users or anonymous, it will definitely lead to edit wars in these articles.

--[[User:Extra999]] 04:22, 19 March 2010

So blocking anonymous users have serious disadvantages

--[[User:Extra999]] 04:29, 19 March 2010