4 ms·

> If, as the manifesto’s defenders claim, the population averages do not have anything to say about individual Googlers, who are all exceptional, then why is Go

> If, as the manifesto’s defenders claim, the population averages do not have anything to say about individual Googlers, who are all exceptional, then why is Google the subject of the manifesto’s arguments at all? What do averages have to do with hiring practices at a company that famously hires fewer than one percent of applicants? In the name of the rational empiricism and quantitative rigor that the manifesto holds so dear, shouldn’t we insist that it only cite studies that specifically speak to the tails of the distribution — to the actual pool of women Google draws from? This is a great point.

throwaway39559 9y ago
This is a strained, nitpicky point: if you want to get ultra-specific, and look only at the pool of Google applicants, you’re out of luck: only Google HR has tail accept/reject percentages, and they are not sharing. Further, the data is meaninglessly skewed, as Google tries much harder to find women applicants than men. Usually, the mean, when combined with standard deviation, is an excellent predictor of relative abundance at the tails. Google is like a co-ed NBA team who is struggling to find enough seven foot tall female players, and then firing the person who attempted to explain why.
unityByFreedom 9y ago
> shouldn’t we insist that it only cite studies that specifically speak to the tails of the distribution — to the actual pool of women Google draws from? Yes. If women fall on a bell curve, and Google pulls only from the far right (x-axis is skill, y-axis is number of women), then the research used to change Google's hiring practices should be based on that selected portion of the curve, not the whole population. If Damore was really just questioning Google's hiring practices, he should have sought out research that looks only at the pool of candidates from which Google draws. And, if he was questioning all affirmative action policies for women, then he was being disingenuous by targeting Google.

throwaway39559 9y ago

This is a strained, nitpicky point: if you want to get ultra-specific, and look only at the pool of Google applicants, you’re out of luck: only Google HR has tail accept/reject percentages, and they are not sharing. Further, the data is meaninglessly skewed, as Google tries much harder to find women applicants than men. Usually, the mean, when combined with standard deviation, is an excellent predictor of relative abundance at the tails. Google is like a co-ed NBA team who is struggling to find enough seven foot tall female players, and then firing the person who attempted to explain why.

unityByFreedom 9y ago

> shouldn’t we insist that it only cite studies that specifically speak to the tails of the distribution — to the actual pool of women Google draws from? Yes. If women fall on a bell curve, and Google pulls only from the far right (x-axis is skill, y-axis is number of women), then the research used to change Google's hiring practices should be based on that selected portion of the curve, not the whole population. If Damore was really just questioning Google's hiring practices, he should have sought out research that looks only at the pool of candidates from which Google draws. And, if he was questioning all affirmative action policies for women, then he was being disingenuous by targeting Google.