Interesting Links for 01-05-2014

[andrewducker]

• Scotland's secret tunnel under the Forth, 50 years old and forgotten
  (tags: scotland tunnels Edinburgh mining )
• Dragon Age Goes Gender-Neutral!
  (tags: games gender fantasy )
• Want a young child to 'help' or 'be a helper'? Choice of words matters
  (tags: children language psychology )
• Why UKIP matters in the Scottish independence referendum
  (tags:

Comments

[alitheapipkin]

Depressingly enough that study about replies from professors doesn't surprise me in the least. I was only commenting to a colleague this morning that senior male academics frequently think they can be rude to everyone after we got a very rude email from one yesterday evening bitching about us using the departmental mailing list.

I'd be interested to see data on whether male and female professors differed in their response rate.

[andrewducker]

FTA:
"There's absolutely no benefit seen when women reach out to female faculty, nor do we see benefits from black students reaching out to black faculty or Hispanic students reaching out to Hispanic faculty,"

[alitheapipkin]

Yeah, I read it, I just don't understand quite what they mean by that. Probably it makes sense in terms of the whole article, but as a snippet, it isn't clear what 'no benefit' means.

[andrewducker]

Aaah. I can't think of any explanation other than "No improvement in the response rate for those people."

[andrewducker]

Oh, and study available here:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2063742

Let me know what you find!

[alitheapipkin]

The sentence directly following the quoted one in the actual article reads "Thus far,
we find essentially no evidence that discrimination against women and minorities is lower in disciplines with higher female and minority representation."

They found that private universities and higher salaries = more discrimination.

Interesting stuff. Probably says a lot about how the women and minorities who succeed in academia do so by fitting into the existing culture rather than changing it. The only thing I didn't notice was whether they had looked at the rank of the professors split by gender and racial lines and whether that had any influence, given professor just means you are tenured staff in the USA.

[steer]

Interesting. There's a lot hidden in the details here. So a lot of the results are within error bounds. That is, in layman's terms, they don't have statistically significant results for the individual groups -- they can't say for sure "black males" are discriminated against with respect to "white males" only that non "white male" is discriminated versus "white male". In fact many of the results were showing positive discrimination but at a non significant level. Which should not be taken to mean I don't think that there is discrimination -- there for sure is -- but this experiment doesn't seem to be able to pick it apart finely enough. (Although it does highlight that Chinese professors are more likely to see Chinese students --- in fact that's the strongest correlation they find -- for obvious reasons they don't harp on about this. In fact the strongest two correlations by race are that chinese profs talk to chinese students and (less strongly) Indian profs talk to Indian students (but the latter are of marginal stat sig

[erindubitably]

This article seems a little clearer to me, and states:

The study found no relationship between representation of any group among faculty in a given discipline and the degree of bias that students faced when trying to interact with them. This means the findings cannot be attributed to the largely white, male academy preferring to associate with others like them, says Milkman. “One of our hypotheses was that more diverse departments would be less biased and we just don’t see it,” she adds. The only exception was among Chinese faculty, who were less likely than other faculty to discriminate against Chinese students.

[alitheapipkin]

Ah, that is better, and it has a clearer summary figure than the ones in the study!

[steer]

That's rather a misstatement. In actuality what they found was this: they didn't really find enough evidence of any discrimination when you broke down so far as to look at "does race/gender X favour or not favour race/gender Y" -- either because that doesn't happen or because (more likely) they don't have enough results