Open main menu

Wikimania05/Paper-JT1

This page is part of the Proceedings of Wikimania 2005, Frankfurt, Germany.


The Motivation of Wikipedia ContributorsEdit

  • Author(s): Jeremy Tobacman
  • License: GFDL and cc-by-sa
  • Slides: {{{slides}}}
  • Video: {{{video}}}
  • Note: Presentation, 15 minutes

About the author: Jeremy Tobacman is a graduate student in behavioral economics at Harvard University. More information: http://kuznets.fas.harvard.edu/~jtobacma/

Abstract

{{{abstract}}}


AbstractEdit

Economists typically emphasize how incentives motivate self-interested individuals to work at productive tasks. In the context of Wikipedia, no contributors receive any pecuniary compensation. This paper analyzes why people nevertheless contribute substantially to Wikipedia, specifically by testing social psychological theories of intrinsic motivation. The paper also develops software tools for measuring the quantity and quality of individuals' contributions and establishes procedures for running "randomized evaluations" of changes in Wikipedia policy or page structures.

In more detail, this paper evaluates and compares several theories for why people contribute to Wikipedia by making subtle random variations to a small number of pages, all of which are intended to increase subsequent contributions. These interventions (or "treatments") affect (i) the amount of publicity for contributors' contributions, (ii) the amount of factual and positive feedback contributors receive, and (iii) whether or not identification with the Wikipedia community is primed in the feedback. Experimental psychologists have found that well-calibrated praise can increase motivation to supply effort. This is the case if future praise might be possible, contingent on high effort, in which case praise serves as a nonpecuniary incentive; and it is also the case if the possibility of future praise is not present.

The experiments' individual-level outcome variables are the quantity and quality of subsequent edits to the site by treated and untreated contributors. Quality will be measured by the amount of text contributed per edit, the amount of time that text persists, and the number of other pages that link to the edited page. All of these outcome measures are publicly available in the revision histories (page-by-page data on readership would be an additional useful outcome variable, since readership plausibly reflects the quality of a page). Ongoing efforts to survey Wikipedia contributors will complement this research.

Statistical analysis will be performed on a new machine which the National Bureau of Economic Research has agreed to host for this purpose. This paper will compare the importance of emphasizing publicity, praise, and social identity for increasing motivation among volunteers. It provides insight into the strengths of the Wikipedia community and suggests how to nurture the involvement of volunteer contributors.

PaperEdit

TBD