Lars Aronsson, 20/07/2013 02:19:
But that seems to be about as hard as getting sports
people to edit Wikipedia.
That's hard on sv.wiki? Oh, the envy! </irony>
I don't know if asking gender to people on a website is really more
accepted/acceptable than other demographics questions, but that would
seem wrong in itself to me. (Apart from the highly-connotated word
"race" which is basically unusable unless you're talking with someone
you know very well.) I don't see why gender would be more relevant than
the social context of origin; it's definitely the opposite IMHO, gender
is largely irrelevant in our case (unlike Facebook or a dating website).
We have a focus on gender just because that's a problem so easily seen,
not because it's actually more important. From a statistical point of
view, Ryan's interest is definitely legit.
We have some statistics from other resources like comScore or Google I
think, but their focus is different and reliability always mysterious;
Google is mostly interested in income and so on to tailor ads. It would
make sense to put some effort in getting useful data ourselves if there
are no other good sources.
If asking about ethnicity is impossible, we could invest more efforts
in measuring correlates, for instance:
1) find out what other metrics correlate best to ethnicity and focus on
those, infer possible underrepresentations (is there more than income
and education which we currently ask? do we have enough focus on those?
we also don't ask about education of the family of origin IIRC and we
were only able to conclude that wikimedia editors are more educated than
2) ask the question directly, "Why didn't you edit/register? ... x-1)
Thought it was a male-only club, x) Oh, isn't Wikipedia a nerd cabal?,
x+1) Because Wikipedia feels like a WASP-only thing" etc. This would
require a lot of effort to come up with a good phrasing to cover all
"discrimination" feelings and to avoid leading/loaded/biased questions
which would skew results, but doesn't sound impossible. (Profs in my
university regularly do such things for sexual harassment and other
discrimination surveys in order to assess the scale of the problem.)