(Not a lawyer) I would have thought Google would be able to argue other causes.
Damore said Google policies "lower[ed] the bar for “diversity” candidates by decreasing the false negative rate"
That statement is (a) factually inaccurate, you can't lower the bar by descreasing the false negative rate, and (b) slanders his colleagues hired through that policy by saying the bar had been lowered to allow them in.
You could argue that the word "effectively" means he actually meant something slightly different in point (a), but he still made the claim in point(b).
In the original memo, that statement also linked to an internal Google discussion to back up his claim - which presumably detailed what he meant. That's context that we are all missing. It would be interesting to see what that discussion was about.
> (b) slanders his colleagues hired through that policy by saying the bar had been lowered to allow them in.
No he didn't. He talked about reducing false negatives, not increasing false positives.
The "bar" is an approximated objective assessment of ability, not an assessment of someone's ability percentile... as demonstrated by the fact that Google policy hires 'people that pass' rather than 'top x% of candidates' or 'top y% of the population according to some measure'
To reframe the "bar" as being a percentile is inaccurate and excuses Damore's misleading and inflammatory terminology.
Honestly not sure what could be on the other end of that hyperlink that would make it suddenly okay to suggest that a portion of your colleagues aren't worthy enough to be there.
I appreciate your comment. Damore's statement made little sense to me and I chalked it up to another example of imprecise wording. Your explanation was thorough and clear. I'm going to reply to that comment here if you don't mind.
You state:
> Because of that and because of Google's standards, let's say that only 25% of all candidates have the skills and qualifications necessary to work at Google, which breaks down to 5 women and 20 men.
> Once again assuming a normal distribution with neither men nor women over or underrepresented in skill level, then without any sort of correcting for diversity, candidates would need to be in the top 10% of all candidates to receive a job offer (10 / 100). , this is 2 women and 8 men, leaving us with a false negative rate (i.e. people skilled enough to work at Google but who get declined) of 3 women and 12 men.
So if there 25 candidates who technically meet the requirements. But Google has a sorting process that is able to pick the best 10 of those 25. Why are those 15 considered a "false negative"? I would've thought a false negative would be when Google's evaluation puts someone who should be in the top 10 somewhere in the lower 25. Maybe it's semantics (and the generally complex problem of defining job qualifications and evaluating candidates). But I also wonder if such ambiguous/unoptimal hiring results are the fault of the job requirements. That is, could these requirements be tightened so that there's just 20, not 25 possible candidates to filter from.
(None of the above questions likely has anything to do with Damore's memo).
But back to Damore's memo. You think he isn't slandering colleagues because he's just stating how the mathematical distribution works out. Whether it involves decreasing false negatives or increasing false positives, isn't the real-life result the same? Given female Google employee hired in the same cohort, the likelihood that the female was let in because of lower standards (i.e. the top 3 of the bottom 15) are substantially higher than that she was hired for being among the 2 women of the top 10.
In a standard cohort, there are 3 women who have been hired who wouldn't been hired if Google hadn't stuck to non-diversity-prioritized system. That doesn't imply that at least a few women lucked out for being women, and not because of merit?
My understanding is that if you take positive as being offered a position and negative as being denied a position, and true or false as whether that was a correct decision then you have:
True Positive - Someone who was offered a position (positive) and who was suited for the position (the positive assessment turned out true).
True Negative - Someone who was not offered a position (negative) because they were not suited to for it (the negative assessment turned out true).
False Positive - Someone who was offered a position (positive), but who was not suited for the position (the positive assessment turned out false).
False Negative - Someone who was not offered a position (negative), but who would have been suited to the position (the negative assessment was false).
So those 15 are considered false negatives because they were not offered a position despite being qualified for it - e.g. perhaps there were only a limited number of positions, and the cut had to be made somewhere.
> isn't the real-life result the same?
Not quite. Increasing false positives is saying those people were given offers but they were not qualified for the job. Decreasing false negatives is saying they were qualified for the job and were picked ahead of other also qualified candidates.
> That doesn't imply that at least a few women lucked out for being women, and not because of merit?
When put like that, then relative to other applicants in the cohort yes, but not because they lack the merit to work at Google - and the latter is what many people were claiming he had stated, and what I was trying to refute.
I can see how the former could also upset people, but if the claims he makes about priority queues for diversity hires, and reconsidering sets of candidates if they are not diverse enough (but not the other way around) are true, then it's not so much that he's claiming people are unqualified, rather he's pointing out a mathematical consequence of those hiring decisions. Insensitive? Perhaps, but to the level worthy of all the internet drama it has caused and to the level of being fired in such a public fashion? I'm not so sure, especially considering many people seem to have misinterpreted what he was saying - both in this and his other statements.
You're leaving off half the sentence. I'm interested to hear your interpretation of how 'decreasing the false negative rate' plays in to this.
If he had said 'hiring practices which lower the bar for “diversity” candidates', then I would agree with your interpretation.
But he didn't say that, he said 'hiring practices which can effectively lower the bar for “diversity” candidates by decreasing the false negative rate'.
It's an inaccurate interpretation of his comments to leave off the last half of that sentence, and I'm not sure what other way you can interpret 'by decreasing the false negative rate'.
Damore said Google policies "lower[ed] the bar for “diversity” candidates by decreasing the false negative rate"
That statement is (a) factually inaccurate, you can't lower the bar by descreasing the false negative rate, and (b) slanders his colleagues hired through that policy by saying the bar had been lowered to allow them in.
You could argue that the word "effectively" means he actually meant something slightly different in point (a), but he still made the claim in point(b).