Monday, July 22, 2013

Finding your friends on Wikipedia

A segment of a social network
A segment of a social network (Photo credit: Wikipedia)
In my opinion, the most notable missing features of Wikipedia is the ability to network socially . Lack of social features conspire to minimise participants' social capital (the expected collective or economic benefits derived from the preferential treatment and cooperation between individuals and groups).

Lack of networking capabilities is in line with a pervasive attitude that ownership of contributions should be suppressed, external authority ignored and that knowledge matters in the project.

Accordingly social costs of participation are growing. Participants in the project have great difficulty in acquiring social capital. The stakes of establishing world opinion tend to outweigh social stakes. Accordingly there is a growing gender gap, under-representation of cultural minorities and disrespect of real world authority.

Currently users will try to establish social capital by listing on their user page their many contributions;  merit badges called barn stars and other merit awards called wiki-love; listing participation within broad groups using "userboxes" and in narrow interest groups called projects many of which are not content related but have a more social character.

Perhaps a better remedy would be to
  1. allow participants to acquire social capital via the MediaWiki software.
  2. provide some real proof of expertise in selected areas.
My 5 cents is creating a friend recommendation system. This is a small offshoot from my edumetrics project I've developing a small tool to allow you to find users whose edits include a significant component which is similar to your more significant edits.

While not a full featured recommendation system it does have some merit already
Future directions would be to:
  • create and present a table of friends. 
  • define a friendship context (an explanation why someone is recommended similar to Amazon's recommendation system). 
  • differentiate friends from foe (currently extended conflict or even stalking can masquerade as friendship).
  • get a notification about other users' activity (particularly where they match your friendship context)
Enhanced by Zemanta

4 comments:

  1. My involvement in these questions dates back to the beginning of Wikipedia, including dialogue with Larry Sanger. Far from disrespecting external authority, Wikipedia editors rely on it; regularly citing scholarly research by recognized authorities. What we don't do is knuckle under to individuals who attempt to assert authority based on a doctorate or a position as a tenured professor. Often such assertions are not accompanied by citations of scholarly authority; Larry Sanger even posted the chatty lectures he was using in sophomore philosophy cases as articles.

    That said, way back, a decision was made to limit social media aspects of Wikipedia to those who actively edit; thus a user page cannot function as a webpage or Facebook page does. The userbox controversy was a watershed in that respect. I think we went too far.

    One problem is that many editors don't focus on particular subject areas as interesting research occurs over the entire field of human knowledge. Thus an editor following current developments may edit on DDT one day, and social mobility the next.

    ReplyDelete
  2. Well said, Fred. I agree with the general notion that it should be easier to engage socially, but such changes should be incremental and above all, informed by careful and broadly peer reviewed research. (Unfortunately, we don't seem to have much of a paradigm of this kind of feature development on Wikipedia…which I hope is a temporary condition!)

    One thing that we know, is that the only time in the history of humankind that an encyclopedia of this scale has been constructed, it has occurred on a platform in which "friendship" is deemphasized. We don't know a lot about the relevant cause and effect relationships, and so I believe caution is advisable.

    ReplyDelete
  3. Hi fred.

    I really enjoy hearing about these great debates from Wikipedia's history. I'm working on a visual history of the project for my educational program. I knew about the user box controversy but I thought the attitude on social media was just a Wikipedia meme.

    I agree about your suggested remedy, namely that to increase the longevity of an editor's edits - it pays to put in citations. At the other extreme I've seen what I consider COI editors ironclad their POV against outside interference using numerous citations. Yet most facts in Wikipedia require citations only challenged controversial ones. So long as experts complain that they cannot correct common knowledge mistakes in their areas of expertise, because their authority cannot be established the issue remains a significant point of failure. I'd love to cook up a tool that can tell the difference between real citations and COI citations but this is a task humans do effortlessly.

    About social features, they are not a panacea but some social features tend to keep users more honest than in dating sites or even in anonymous settings like wikipedia. Think of it as more eyeballs on the user keep him honest just like more eyeballs on an articles keep it accurate.

    You might recall an incidents with Ryan Jordan aka User:Essay who created a fake identity with fake PHDs to gain authority in fields he was unqualified highlight some of the complexity of the issue. I don't see this happening in a place like facebook. What is often forgotten is that Ryan Jordan
    not only got a job at wikia, he was also a steward and the director of project Esperansa - one of the 3 powerhouses in Wikipedia at the time. Ryan Jordan was able to gain massive social capital through talent and deception primarily because his actions did not get increased scrutiny that social media introduces.

    At one time I'd have told you to stick to working on your area of expertise (DDT?). But I now realise this is not needed unless the articles are approaching good/featured level.

    I wrote the script to for an educational tool and while I consider it primitive I can do some things fairly quickly.

    Your comment on editing patterns is quite correct, but the script can find find users within different social roles. Here is a non-technical explanation of how they are considered:

    My script only considers articles which have gotten significant editing. For your case It would be calibrated to only consider your top articles (which you have edited/commented on at least 20 times).

    To find you users in a mentor role it would then look for the top 30 editors and rank them according to who has worked on the most of those top articles again with a threshold of 60 edits on at least three of your top articles. (You could ask them to help as a 3rd opinion wikipedians, solicit for sources or request a FA peer review)

    To find you users in a socialiser role it would look for the top 30 editors and rank them according to who have worked on the most of those top talk pages with a threshold of 3 edits on at least six of your top article's talk page. (You could collaborate on more articles, help out in dispute, have a fun discussion do version reviews etc...)

    To find sympathetic wiki gnomes it would look at editors with a minimum of just 3 edits but on 8 of your top editors within the last 3 months. (You could give them some wiki love, return the favour or ask for help)

    I hope this also explains how social media can be adapted to Wikipedia settings by reinforcing its values and engaging editors on social capital and content related participation.

    ReplyDelete
  4. Hi Pete

    There is already a growing body of academic research showing that many if not most edits are now being performed by impersonal robots. As these robots have gained ascendency the community has become more balkanised and calcified against change. Strategically speaking the community is in decline both in editors and administrators. On the other hand the daily labour has gone up, not down and for all the talk no progress has been made on reducing the gender gap. I mention the gender gap because when it comes to social capital and social media features, it is where women have the advantages.

    I agree that new features should be handled diplomatically and even scientifically. I also think that this has repeatedly led a very vocal minority (of one) to successfully challenge most of the new features created by WMF developers such as Liquid threads, Moodbar, Article Feedback, Malayalam Fonts ... VE!?)

    So as I see it - it is becoming increasingly urgent to start implementing social features and suitable user metrics to measure their effects during limiter roll out.

    ReplyDelete