The Signpost

News and notes

Research project banner ads run afoul of community

Contribute  —  
Share this
By Resident Mario

CentralNotice research project taken down over community opposition

For several hours on Thursday of this week, 10,000 experienced editors and 30,000 new contributors were exposed to a CentralNotice message advertising a 25-minute survey of Wikimedia participants. Put together by the Berkman Center for Internet & Society at Harvard University and Sciences Po in Paris, the study, aimed to "better understand the dynamics of interactions and behavior in online social spaces", was sanctioned by the Wikimedia Foundation, and had been on the CentralNotice calendar for the past month.

Nonetheless, as Wikimedia Foundation Senior Research Analyst Dario Taraborelli explained on the foundation-l mailing list, the banner was poorly received by Wikimedians, who found it confusing and reminiscent of a commercial venture. It was taken down hours later, having incited much discussion at the Administrators' noticeboard, the village pump, foundation-l and internal-l mailing lists, and even an RfC.

The study's research team first approached the Wikimedia Foundation and Wikipedians in January 2010, proposing to post messages on user talk pages; editors suggested they scrap the method, and make a minimally invasive CentralNotice banner to registered editors instead. The proposal went through a review by the Wikimedia Research Committee in July of this year, which resulted in the project's current parameters, namely, only for logged in users, only on the English Wikipedia, and only for at most a week.

Soon after, the Wikimedia Foundation invested in added functionality for the CentralNotice extension; the new backend now extends to editor databases, allowing its users to target specific subsets of users. Using this functionality, the banner was created, put on the CentralNotice calendar a month ahead of launch, and finally launched on 8 December at 11:00 UTC, garnering 800 responses before being shut down.

While the notice was live, it garnered both negative and positive response from the Wikipedia community. According to ErrantX, "One of the criticisms was the lack of discussion/input the English Wikipedia community was granted – and the lack of notification prior to the launch. Feedback on the mailing lists seems to suggest that the Foundation and various committees are not aware of the communities preferences regarding Central Notices, partly because it has not been discussed before."

As explained by Taraborelli, "We realize that despite an extensive review, the launch of this project was not fully advertised on community forums. We plan to shortly resume the campaign ...after a full redesign of the recruitment protocol in order to address the concerns raised by many of you over the last 24 hours." The Foundation is now moving to provide better information on the project, by creating an FAQ and linking to the study directly from the banner, redesign the banner to be less "ad-like", and make privacy terms more transparent, as participants were not aware that they were sharing their usernames, edit counts, and user privileges with the study team.

Brief notes

Lori Phillips, the WMF's new United States Cultural Partnership Coordinator, pictured 2010

+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.
  • "as participants were not aware that they were sharing their usernames, edit counts, and user privileges with the study team." Wait...what? O_o I don't remember the survey saying that anywhere. Now I feel uncomfortable about having took it. I assumed it was going to be anonymous, as that would be the point of such a survey in the first place. That's also why I had the money be donated to the Red Cross, so that there wouldn't be any break in the anonymity with an actual Paypal account being required. SilverserenC 18:29, 13 December 2011 (UTC)[reply]
    • The information gathered needs to be thrown out and the people who wrote the survey interface need to have their asses dragged before the human testing board and have some basic concepts explained to them. I am fully aware that all of those things are publicly accessible, however if personal information is going to be connected with responses, it damn well better fucking say that. As for the WMF not insuring that something that they were endorsing on Wikipedia had adequate protection for Wikipedia users, well it's not like I had any confidence in them before this, but now I have one more egregious error to put on the list. Sven Manguard Wha? 19:52, 13 December 2011 (UTC)[reply]
      • Huh, I wouldn't go as far as to scorn the researchers, but I agree that not being explicit about sharing usernames AND the responses is, in my view, a privacy violation. I would not mind if the researchers were allowed to link each questionnaire with some basic statistical data (edit count, Wikipedia privileges, etc.), but linking it to username definitely goes too far. This makes this research non-anonymous to wide extent, basically (for more experienced editors). Pundit|utter 00:24, 16 December 2011 (UTC)[reply]
  • Is it possible to take this survey? It sounds interesting. BTW, ads like this on our wiki are outrageous Bulwersator (talk) 18:34, 13 December 2011 (UTC)[reply]
    • It's not altogether clear that calling this an "ad" is reasonable, although I do share the concerns voiced by others, I think it better to call this "a mistake" (or several mistakes) rather than "outrageous". Rich Farmbrough, 21:22, 13 December 2011 (UTC).[reply]
  • I'm an admin; but until I read this report, I was unaware that there was any such thing as "CentralNotice", much less a "CentralNotice/Calendar". I still am unclear as to what they are and what their purpose might be (though I will look into it). What deluded anybody into thinking that posting a demolition notice on Alpha Centauri would constitute sufficient warning to the residents of Earth that we were about to be demolished to build a galactic bypass? --Orange Mike | Talk 19:40, 13 December 2011 (UTC)[reply]
    • "Oh, for heaven’s sake, mankind, it’s only four light years away, you know. I’m sorry, but if you can’t be bothered to take an interest in local affairs, that’s your own lookout." Protonk (talk) 23:03, 13 December 2011 (UTC)[reply]
      • "Energize the demolition beam. I don't know, apathetic bloody planet, I've no sympathy at all." SD (talk contribs) 00:32, 14 December 2011 (UTC)[reply]
        • Dunno about those of you with the issues of sharing Usernames, edit counts and privileges details, but considering I was getting paid for it, I don't give a rats. As long as they didn't get my real name, address, phone number or my SSN/NiNo, I don't care :) I'm not turning my nose up at someone giving me free money for filling out some forms, blow that for a lark!  BarkingFish  01:51, 14 December 2011 (UTC)[reply]
          • Well, I donated all of the money, so... SilverserenC 01:59, 14 December 2011 (UTC)[reply]
          • I wouldn't have a problem with the information shared, as long as they were clear about that, up-front. It's a mistake, and it's been picked up, so all is good now... I hope. (Btw, love the Hitchhiker's segue above.) --Chriswaterguy talk 03:20, 14 December 2011 (UTC)[reply]
I like to think of it as the one time Wikipedia's physically payed us back for our work ;) ResMar 03:12, 14 December 2011 (UTC)[reply]
Can someone point where the survey (or the FAQ on meta) state the username, edit count, and user access level is shared with the researchers? Was it there but I missed it? OhanaUnitedTalk page 13:40, 14 December 2011 (UTC)[reply]


The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0