[yummyfajitas]

All this paper shows is that men compete just as hard even if you give preferences to women.

It does not address the question of whether lowering the bar for some people will lower the overall average or the average for that subgroup (hint: it will - this is almost a mathematical identity).

[roguecoder]

Who said anything about lowering the bar for anyone? There is no bar to attend most hackathons, no qualifying heats or even resume screening.

You also assume that lowering the bar inherently attracts less-qualified people than the marginal alternative participant. That is not true unless the outcome you care about is whatever bar you are using to measure and people accurately self-assess (or universally apply). For example, there is a 1992 study that found that SAT scores were equally good at predicting success of women and men, but only within those groups. Women performed as well during college courses as men with SAT scores 50 points higher (http://her.hepg.org/content/1p1555011301r133/). In such a case in order to maximize total academic performance, you would need to compensate for that systematic discrepancy and lower the SAT bar for women: what you are actually doing is normalizing the predicted-college-performance bar. That would not maximize total admitted SAT scores, but might maximize the outcome the college actually cares about.

[yummyfajitas]

Who said anything about lowering the bar for anyone?

Cletus, in the post you replied to.

As for using gender as a predictor in admissions, you'd also need to penalize high scoring women (and reward the low scoring ones). I have no particular objection to any of this.

http://www.overcomingbias.com/2008/08/variance-induce.html

[anthonyb]

Mean SATs, not their variance. I don't think that overcomingbias link is really adding anything to the discussion.

The paper that roguecoder referenced is just pointing out that SAT scores are not a perfect predictor, and adjusting the intake based on gender is probably a good idea if you want to maximise the real effect (academic performance), rather than the predictor (SAT scores).

[yummyfajitas]

The Robin Hanson blog post points out the exact same thing. Why is a positive correction for women useful, but a negative one "not adding anything to the discussion"?

If you want to use gender as a predictor, it could be positive or negative.

[anthonyb]

Counter hint: It doesn't. Look at figure 4 in the paper.

I suspect that the reason for this is that when you exclude women in general, you're also excluding women who are just as good as men are.

Which is the unspoken assumption in your post and Cletus' original post: if you include more women, those women are going to be dolts who couldn't make it in on their own steam. I don't see that that's the case (eg. what if they just don't want to participate in crap hackathons where they're going to be harassed?). Neither you or Cletus are providing anything other than the usual hearsay.

[yummyfajitas]

The paper did not lower the bar for entry, it lowered the bar for women winning. The correct test would be to compare the winners of the competition with the lower bar to the winners of the competition without the lower bar.

Which is the unspoken assumption in your post and Cletus' original post: if you include more women, those women are going to be dolts who couldn't make it in on their own steam.

The unspoken assumption is that lowering the bar has any effect at all. If the bar is normally at 10, but you lower it to 7 for women, then any extra women this policy lets in must be in the range [7,10). Since all numbers in the interval [7,10) are below the previous minimum, they therefore lower the average.