On Thursday 30 November 2006 09:09, Andrea Forte wrote:
understand for people from outside the "camp." Clearly, Joseph and I are both strongly influenced by anthropological approaches; however, I also think that pluralism can be a strength within a research community, so I would not encourage those who enjoy survey research to abandon it.
Just to be clear, neither would I! I have participated in survey research in the past, aware of its limitations [1], and would like to see a good job done up in the Wikipedia case. However, it is a very hard problem and (fortunately?) not central to my present project. But the question remains, how do we move beyond the many small self-selecting surveys bouncing around towards something more robust? In my e-mail, I was pointing out that perhaps getting a representative sample of all of Wikipedia is an impossible goal, and perhaps we can focus on specific "local" communities. But maybe not! It'd be great to have something on the scale of Ghosh et al. (2002) on FLOSS which had 2784 (self-selected) respondents, a portion of which (487) validated/confirmed [2].
[1] http://reagle.org/joseph/2005/06/search.cgi?query=Cranor%2C_Reagle_and_Acker... [2] http://www.infonomics.nl/FLOSS/report/FLOSS-Final4a.htm