Wikipedia:WikiProject Wikidemia

This project, Wikidemia, provides a space for articles related to academic research about Wikipedia. Related pages include the Statistics Department, m:Research, and m:Statistics. This page and project are still very preliminary and will benefit from your contributions and insight. If you would like to help, please sign the Participants list below and introduce yourself on the talk page. The to-do list here is just a start...

Title

edit

WikiProject on Wikidemia

Scope

edit

This WikiProject aims primarily to design, implement, and discuss academic research about Wikipedia. We seek to better understand what promotes or circumscribes Wikipedia's success and why. We also seek to explore and rigorously evaluate new strategies for improving Wikipedia, and to archive research attempted by Wikipedians into related topics.

Descendant WikiProjects

edit

Similar WikiProjects

edit

Similar WikiProjects are:

Members

edit

Please add your name here!

Wikipedia Studies

edit

Completed Studies

edit

Data

edit

Ongoing Studies

edit

for Proposed Studies, see the Research Questions lists below; all content at /Studies has been incorporated into specific study pages.

Issues : How to identify classes of cases/user; what variations to impose (and how to implement them); then how to choose randomly among them to implement variation[s]; finally what data to collect [both primary and secondary metrics].

Research Questions

edit

Promoting Contribution

edit
  • Who contributes to Wikipedia? (demographics/ education / other volunteerism and community involvement)
  • How does feedback to contributors affect subsequent propensities to contribute?
  • What interventions can increase the quantity and quality of Wikipedia articles?
  • What makes contributors mad? stressed?
  • What makes contributors happy?
  • What is the effect of contributors' emotional reactions on their contributions?
  • What role do watchlists play in encouraging contributions and edit wars?
  • What is the effect of placing an article on the Main Page as a featured article on readership and contributions?
  • What steps are needed to secure accreditation on articles? Should there be any?

Promoting Readership/Consumption

edit

Multilingual Wikipedia

edit
  • What Wikipedia language editions exist, and why?
  • What factors make a language edition grow?
  • How are lacks of language planning (e.g. a lack of standardization) dealt with?
  • What kind of inter-Wikipedia collaboration exists?
  • How to compare language editions to each other?

Curtailing mischief

edit
  • How can disputes (e.g., edit wars) be resolved more efficiently?
  • How can vandalism be decreased or fixed more quickly?
  • What categories does vandalism fall into, and how much of total vandalism does each category (e.g. advertising) represent?
  • How long does vandalism typically remain visible before it is removed - statistical analysis needed. See for example Wikipedia talk:Don't protect Main Page featured articles/December Main Page FA analysis
  • How effective are bots in helping deal with mischief? What strategies can we use to further their effectiveness?
  • Who typically reverts vandalism? (figures for admins, regular editors, IP editors, bots)
  • What effects does semi-protection have on levels of contribution and vandalism? Several articles should be studied before protection, during and after.
    • What level of vandalism is acceptable; at what point is protection warranted?

Article quality

edit
  • How can our best articles be kept in pristine condition?
    • What policies and initiatives can we enact to prevent article deterioration
    • A case study of 'edit creep' is needed
  • What is the average quality of our articles?
  • Is the average quality improving? Does a typical article improve over its lifespan? How quickly? What trends do we see?
  • How can our article assessment system be improved?
  • What percentage of articles cite no references at all?

Networking

edit
  • How (much) are the pages linked together? (Paths, Meshing)
  • Which pages are visited together? How close are they in matter of content?
  • How important is #wikipedia to the administration of Wikipedia?
  • ...

Coherence and consistency

edit

Collaboration

edit

Methodology

edit
  • Many different methodologies would be possible and useful.
  • Some questions can be examined by direct analysis of existing field data.
  • Running randomized evaluations will facilitate drawing causal inferences about results. A standard way to pre-test possible large-scale innovations in a neutral way is to identify a class of visitors, editors, or pages; select a randomized subset of that class; and introduce a variation to the randomized subset. Then metrics can be evaluated for both the subset and the entire class, and inferences drawn about what effects the variation had. Stratification can increase the statistical power of the evaluation.
  • A user survey to which one could add important questions, would help inform background assumptions. Users who do not choose to be wholly anonymous in responding to such a survey could even partake in specialized control groups for some studies.
  • Pilot studies - running small, short initial studies to provide an example of how to run and evaluate a study; and to iron out implementation details specific to Wikipedia and its community.

Caveats?

edit
  • Privacy
  • Consent to participate
  • Interventions may have unpredictable results
  • Randomization may be difficult to sustain

General strategy and context

edit

Research on Wikipedia links naturally to topics currently under study in economics, psychology, and sociology. Specifically, focus is needed on:

Economists study markets in ideas; volunteerism; bargaining; and information. Psychologists study motivation; conflict resolution;.... Sociologists study networks of ideas and people; the culture of organizations; norms of behavior;....