QUOTE(Newsfeed @ Thu 11th August 2011, 12:56pm)
Researchers reveal Wikipedia gender biasesPhysOrg.comComputer science researchers in the University of Minnesota's College of Science and Engineering are leading a team that has confirmed a substantial gender gap among editors of Wikipedia and a corresponding gender-oriented disparity in the content.
...View the article This study, for which you can find full .pdf file
here merits a look, as it's actually new data, and
NOT the 2009 user survey from which Jimbo based his Wikimania talk conclusions. Much data from THIS (different) study used male/female userbox self-reporting of gender, then looked at actual editing patterns.
Some of it is interesting. Userbox women were 16% of edits, very similar to the survey of 2009. Women are twice as likely (relative to men) to edit in people and arts articles, than history and science articles. 10% vs. 5%. Women aggregate to the social interaction areas (mainspace TALK and WT), more or less validating my comparison of WP to Facebook and its > 50% female/male ratio. There's a fun Kaplan-Meyer "surivival curve" for males and female editor, from first edit to last. It's about 20% for women at 100 days and 30% for men.
Odd fact: women on WP are TWICE as likely to edit controversial articles (as defined by whether the articles were protected)! Female editors are drawn to conflicted articles (no way to tell cause and effect-- perhaps the reason the articles are conflicted is the same reason women are attracted to them, so there's not direct connection).
At the beginning of editing, women are far MORE likely to leave after contributing something as NEWBs, and been reverted. I guess that last one doesn't surprise. However, once they attained experience, women were MORE likely then men to become administrators.
Here's also an oddity:
QUOTE
We found that 4.39% of female users (673) and 4.52% of male users (4,449) have been blocked at some point in their Wikipedia tenures, which is not a significant difference, c 2(1,N = 113848) = 0.545, p = 0.460. Looking only at users who were subject to an indefinite-length block, we found a 3.85% rate for females (592), and 3.32%formales (3,274), c 2(1,N =113848)=11.2, p<0.001. So, while males and females appear to be blocked at similar rates, females are significantly more likely to be blocked indefinitely.
It turns out female-users (as reported by userbox) are more likely to vandalize than males. The rate of vandals who managed to figure out how to put UP a userbox was (understandably) low, but once they'd done this, females were more likely to be blocked for vandalism than males. Which made the writers of the paper (as well as myself) question the validity of these things, at least as concerns vandalism-blockees.