This page presents the design principles and results of a survey (now closed and being analyzed) to understand expert motivations for or against contributing to Wikipedia and other collaborative projects. This survey is an initiative of the Wikimedia Research Committee and is run by Dario Taraborelli, Daniel Mietchen and Panagiota Alevizou.
The first results from the survey were presented at Wikimania 2011 (Haifa, 4-6 August, 2011).
Expert participation: slides from the talk at Wikimania 2011
Main outcomes:
A majority of respondents considers Wikipedia as useful for educational purposes (sl. 14)
About 75% of respondents would consider rating or reviewing Wikipedia articles (sl. 16)
Wikipedia contributors are more likely than non-contributors to have most or all of their published research articles available by way of Open Access (sl. 17)
The target group of this survey is experts, broadly defined (e.g. scientists, academics and scholars as well as other expert professionals outside the academia) and on the barriers that keep them from contributing, either as editors or as reviewers.
The survey consists of: (1) a short introductory section to collect basic demographic information of participants (age, gender, professional status, field of expertise, geographic location, languages spoken and prior participation to Wikipedia); (2) two main sections follow, focusing on the general perception of Wikipedia participation by experts and on individual motivations respectively; (3) a final short section is devoted to collecting further information about participants, including their ICT literacy and their participation to other collaborative projects. We explicitly designed the survey so as not to force participants to consider themselves as either contributors or non-contributors: we want to capture reasons that keep existing contributors from contributing more or reasons why previous (expert) contributors decide to leave or decrease their involvement.
The survey is open, meaning that we cannot control upfront for potential biases in the sample and responses will need to be debiased a posteriori.
