←back to thread

219 points thisisit | 3 comments | | HN request time: 0.001s | source
Show context
lostmsu ◴[] No.16126641[source]
There's one important datapoint in this article: "The Bamboo Ceiling".

When the whole fuzz about gender discrimination started, Microsoft and Google published numbers, claiming women got the same pay at the same positions as men. Knowing there's discrimination from personal experience/feeling, I theorized, that women are discriminated in a different way: they don't receive promotions.

Under otherwise similar circumstances having children does not feel to be enough to explain why of 100 women hired in tech on professional roles less are promoted to higher positions, than of 100 men. That trend is (at least anecdotally for me) observable even before people become parents.

This "Bamboo Ceiling" shows the same effect for another potentially discriminated group of people.

replies(5): >>16126781 #>>16126794 #>>16127061 #>>16127979 #>>16138507 #
geofft ◴[] No.16126794[source]
This is the allegation of the Ellis, Pease, and Wisuri lawsuit against Google - that Google does okay at hiring women, but slots them into lower positions and gives them fewer promotions than men. https://www.nytimes.com/2017/09/14/technology/google-gender-... The NYT's report on the leaked #talkpay spreadsheet seems to show that pattern: https://www.nytimes.com/2017/09/08/technology/google-salarie...

The neat thing about this form of discrimination is that you can claim to be fixing "the pipeline" all you want and you can still maintain the discrimination, because the leak is after the pipeline. The dominant group isn't threatened by competition if they fund efforts to increase the number of underrepresented groups in grade school or college STEM education, as long as those college graduates aren't later competing for senior jobs on a level playing field.

replies(3): >>16126943 #>>16127042 #>>16127247 #
ThrustVectoring ◴[] No.16127042[source]
Your post kind of assumes that a fair process would promote women at Google at the same rate as men. If reality is sexist, are we obligated to discriminate against men to fix it?
replies(4): >>16127199 #>>16127225 #>>16127243 #>>16127479 #
krastanov ◴[] No.16127243[source]
First of all, discriminating against the currently dominant demographic is indeed a stupid way to "fix" anything.

But "reality" does not seem to be sexist, rather our biases and tribalism is sexist (and racist and plenty of other -ist). When high-quality reproducible research has observed phenomena like "stereotype threat" and "implicit bias" it is worthwhile to spend some of our idle time on thinking how to address this unfairness. Even if we do it simply so that we have a wider applicant pool from which to pick high-quality employees.

replies(1): >>16127620 #
tomp ◴[] No.16127620[source]
Links to high-quality reproducible research on "implicit bias"? Last I heard, it was more-or-less debunked.
replies(1): >>16127822 #
1. krastanov ◴[] No.16127822[source]
If you just want a big list: https://www.projectimplicit.net/papers.html

For more digestible information look around their website.

I guess it is my turn: What debunking are you talking about besides the more extremist men's rights advocates (which are different from the moderates that have very valid concerns)? "Implicit bias" is indeed only the start of a discussion, as one needs to consider its predictive value in non-test conditions, but if you are sincerely interested in pursuing this conversation, the website above is a good starting point.

replies(1): >>16127894 #
2. tomp ◴[] No.16127894[source]
https://news.ycombinator.com/item?id=15843167

This was a discussion a month ago. I'm mostly basing my opinion on it.

replies(1): >>16128390 #
3. krastanov ◴[] No.16128390[source]
That is a good point, but as I mentioned, this is only the start of the conversation. Implicit bias is an easy thing to measure, but it has predictive issues in non-test situations. However, as it is always in science, it is the flawed experiment that led to better experiments (I am picking up the "popsci" stars, for this conversation to be rigorous we have to include actual meta studies, but you will have to contact the professionals for that):

- the John/Jennifer study http://www.yalescientific.org/2013/02/john-vs-jennifer-a-bat...

- the chairs study http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.465...

- the police shooting armed people in VR study http://www.washington.edu/news/2003/07/08/blacks-more-likely...

- the general idea of "stereotype threat" (which becomes unrelated, not as much of an offshoot)

I am not expecting you to spend the time to vet every single of those links (and admittedly I used google, so some of the links might be overly editorialized), however I do believe these are good resources to consider for inclusion in your intellectual toolkit when you have the time.