5 ms·

He specifically mentions in an interview that minority interviewees get assigned to a second interviewer if one interviewer doesn't like them in the first round

He specifically mentions in an interview that minority interviewees get assigned to a second interviewer if one interviewer doesn't like them in the first round. He saw this as a 'second chance' when the committee might just be controlling for interviewer biases. Though, the fact that he jumped to this 'lowering the bar' line of thinking shows to me that he was fishing for a conclusion.

waqf 9y ago
It's trading off the false positive vs. false negative rate in the hiring signal. The question is what your assumptions are about the sources of error in that signal. If you assume that the variance consists exclusively of false negatives due to discrimination, then the extra interviewers will work as Google claims and no bars will be lowered. If you assume that the variance consists of randomly distributed error that's the same for all interviewees, then the second interviewer is just some statistical sleight-of-hand that's, on average, equivalent to lowering the bar. Of course it's possible that the truth is in between.
- intopieces 9y ago
  >Of course it's possible that the truth is in between. In this case, isn't the finding the truth pretty straightforward? Pull the performance reviews of the second chance hires. Check to see if there's a clear correlation in review and find a pattern. Are we assuming that Google, which specializes in using data, does not bother to mine its own data with regards to its hiring practices?
  - micahbright 9y ago
    Furthermore, simply look for interviewers who tend to reject women and minorities more often. Remove them from the interviewer pool.
ErikVandeWater 9y ago
That definitely sounds like something that would lower the bar. If I was playing H.O.R.S.E. with an NBA player, if I got two chances to hit a shot for every one chance that an NBA player got, I might look competitive, even though I'm not very good at basketball.
jdoliner 9y ago
> Though, the fact that he jumped to this 'lowering the bar' line of thinking shows to me that he was fishing for a conclusion. Let's focus on things we can actually know rather than speculating about Damore's state of mind. I'd ask a few questions about this interview practice though. 1) Is this the entirety of Google's diversity practices in hiring? I'd be surprised if it is. So even if this isn't lowering the bar it still doesn't prove conclusively that's not what they're doing. Again I'd like to see a more complete accounting of what exactly it is they do. However, I'm certainly not saying that you need to provide this in order to have a legitimate argument, you don't have access to this information any more than I do. 2) This practice seems to have a somewhat narrow view of what a interviewer bias looks like. In particular it only tries to eliminate bias in the case of a minority being rejected. What would happen if we were to instead attempt to detect interviewers who were prone to bias by randomly giving rejected candidates second interviews and seeing which interviewers wound up frequently disagreeing with their peers? If the assumption that bias only effects minority candidates is true this would have much the same effect. 3) What if it wasn't a second chance but 100 chances? I.e. if you're a minority you get to interview for Google 100 times and if any of those say yes your in. White people only get 1 shot. Unless you think Google's false positive rate is 0, this would have to lower the bar wouldn't it?
- jshevek 9y ago
  Thank you for this. This is one of the most important comments on this page: " Let's focus on things we can actually know rather than speculating about Damore's state of mind." Or, at the very least, present speculation as speculation. It does "show" anything definitive.
- pj_mukh 9y ago
  "Let's focus on things we can actually know rather than speculating about Damore's state of mind." So let us extend him the benefit of the doubt that he didn't extend the hiring committee? How about we extend the benefit of the doubt to everybody involved, which would result in him never writing this memo and second guessing the hiring process. The rest of your points seem to be just a whole lot of speculation, which you just told me not to do. The article shouldn't have been written without clarification on these points and conclusive evidence. Maybe go talk to the hiring committee about their motives/state of mind first Mr. Damore?
moduspol 9y ago
The bar for all candidates at a non-Google company: Pass your first interview. The bar for white male candidates at Google: Pass your first interview. The bar for minority candidates at Google: Pass your first or second interview. Of course it lowers the bar for hiring. Maybe it's justified, but it's certainly not raising it. It's not keeping it the same. Maybe it's still tougher to get hired as a minority candidate than as a white one, but that doesn't change what the policy does.
- cbanek 9y ago
  Google seems to track which candidates, even if you're white and male, have done well enough in the interview process where they think it's a false negative, and give you another interview after 12 months. It's not just minority hires that interview more than once at Google, so there's no bar lowering. There's just trying again, and everyone can try again. So really it's what you'd expect: The bar for all candidates at Google: Pass your interview.
  - moduspol 9y ago
    Just some candidates get more attempts at that bar than others, systematically, as a result of explicit policy ensuring that’s the case. You can call it justified, leveling the playing field, or whatever else, but it’s still lowering the bar to getting hired.
joveian 9y ago
It is an attempt to get past the type of issues that Dan Luu wrote about almost three years ago[0], where he referred someone to Google as one of the most impressive engineers he ever met in any field and that person didn't get past the phone screening as "not technical enough". [0] https://danluu.com/tech-discrimination/ https://danluu.com/tech-discrimination/

waqf 9y ago

It's trading off the false positive vs. false negative rate in the hiring signal. The question is what your assumptions are about the sources of error in that signal. If you assume that the variance consists exclusively of false negatives due to discrimination, then the extra interviewers will work as Google claims and no bars will be lowered. If you assume that the variance consists of randomly distributed error that's the same for all interviewees, then the second interviewer is just some statistical sleight-of-hand that's, on average, equivalent to lowering the bar. Of course it's possible that the truth is in between.

intopieces 9y ago
>Of course it's possible that the truth is in between. In this case, isn't the finding the truth pretty straightforward? Pull the performance reviews of the second chance hires. Check to see if there's a clear correlation in review and find a pattern. Are we assuming that Google, which specializes in using data, does not bother to mine its own data with regards to its hiring practices?
- micahbright 9y ago
  Furthermore, simply look for interviewers who tend to reject women and minorities more often. Remove them from the interviewer pool.

ErikVandeWater 9y ago

That definitely sounds like something that would lower the bar. If I was playing H.O.R.S.E. with an NBA player, if I got two chances to hit a shot for every one chance that an NBA player got, I might look competitive, even though I'm not very good at basketball.

jdoliner 9y ago

> Though, the fact that he jumped to this 'lowering the bar' line of thinking shows to me that he was fishing for a conclusion. Let's focus on things we can actually know rather than speculating about Damore's state of mind. I'd ask a few questions about this interview practice though. 1) Is this the entirety of Google's diversity practices in hiring? I'd be surprised if it is. So even if this isn't lowering the bar it still doesn't prove conclusively that's not what they're doing. Again I'd like to see a more complete accounting of what exactly it is they do. However, I'm certainly not saying that you need to provide this in order to have a legitimate argument, you don't have access to this information any more than I do. 2) This practice seems to have a somewhat narrow view of what a interviewer bias looks like. In particular it only tries to eliminate bias in the case of a minority being rejected. What would happen if we were to instead attempt to detect interviewers who were prone to bias by randomly giving rejected candidates second interviews and seeing which interviewers wound up frequently disagreeing with their peers? If the assumption that bias only effects minority candidates is true this would have much the same effect. 3) What if it wasn't a second chance but 100 chances? I.e. if you're a minority you get to interview for Google 100 times and if any of those say yes your in. White people only get 1 shot. Unless you think Google's false positive rate is 0, this would have to lower the bar wouldn't it?

jshevek 9y ago
Thank you for this. This is one of the most important comments on this page: " Let's focus on things we can actually know rather than speculating about Damore's state of mind." Or, at the very least, present speculation as speculation. It does "show" anything definitive.
pj_mukh 9y ago
"Let's focus on things we can actually know rather than speculating about Damore's state of mind." So let us extend him the benefit of the doubt that he didn't extend the hiring committee? How about we extend the benefit of the doubt to everybody involved, which would result in him never writing this memo and second guessing the hiring process. The rest of your points seem to be just a whole lot of speculation, which you just told me not to do. The article shouldn't have been written without clarification on these points and conclusive evidence. Maybe go talk to the hiring committee about their motives/state of mind first Mr. Damore?

moduspol 9y ago

The bar for all candidates at a non-Google company: Pass your first interview. The bar for white male candidates at Google: Pass your first interview. The bar for minority candidates at Google: Pass your first or second interview. Of course it lowers the bar for hiring. Maybe it's justified, but it's certainly not raising it. It's not keeping it the same. Maybe it's still tougher to get hired as a minority candidate than as a white one, but that doesn't change what the policy does.

cbanek 9y ago
Google seems to track which candidates, even if you're white and male, have done well enough in the interview process where they think it's a false negative, and give you another interview after 12 months. It's not just minority hires that interview more than once at Google, so there's no bar lowering. There's just trying again, and everyone can try again. So really it's what you'd expect: The bar for all candidates at Google: Pass your interview.
- moduspol 9y ago
  Just some candidates get more attempts at that bar than others, systematically, as a result of explicit policy ensuring that’s the case. You can call it justified, leveling the playing field, or whatever else, but it’s still lowering the bar to getting hired.

joveian 9y ago

It is an attempt to get past the type of issues that Dan Luu wrote about almost three years ago[0], where he referred someone to Google as one of the most impressive engineers he ever met in any field and that person didn't get past the phone screening as "not technical enough". [0] https://danluu.com/tech-discrimination/ https://danluu.com/tech-discrimination/