Hacker News Reader

Comment by cuteboy19 | Hacker News Reader

>some white people who were affected will believe it was because of their race

But is it actually true? Did they fire white people at a higher rate?

They apparently laid off white people at a higher rate than they would have. Otherwise the statement about using an anti-racist lens would be incredibly misleading.

"Than they would have" if what though? They just allowed subconscious racist biases to affect their decision making? It's not like we lack evidence that this does in fact happen.

If an "anti-racist lens" means anything other than "processes to help avoid the effect of subconscious racial bias" I'd agree it's a problem.

The opposite of anti-racist isn't racist.

You are assuming the decision makers have subconscious racist biases. And you are assuming those biases favor white people, men, etc.

In the world I live in there is a good chance that each of these is false. The fact that they made this statement indicates as much.

There are actual studies showing 95-99% of us have subconscious racist biases. And so what which races they favour? I'd call it a massive fail for any "anti-racist lens" if, say, only whites were fired. (You also mention "men" - subconscious sexist biases are just as universal. For highly skilled roles unfortunately they very much typically favour men.)

The field of subconscious biases suffers from several fatal defects. The scores people get on subconscious bias tests vary wildly over time, even short periods of time. This is a fundamental flaw and should be the end of the inquiry. If we can't measure the thing, we can't draw any conclusions about it.

Also, there has never been a linkage drawn between scores on subconscious bias tests and real world actions. Are racists probably "subconsciously biased"? Sure. But the reverse has never been shown.

Actually we can and do measure it and such linkages have been demonstrated. There are probably more succinct references, but here was one of the first I found:

http://kirwaninstitute.osu.edu/wp-content/uploads/2014/03/20...

Can you share a section reference? Looks like it's 90 pages long.

Chapter 4 (pp. 44-) is a good place to start, particularly "UNCOVERING BIAS BY USING FICTITIOUS RESUMES" (p. 47).

Another study found that anonymizing resumes actually hurt minority candidates. [1] This is because the companies were trying hard to help people with minority sounding names and no longer could.

Anonymizing seems like an appropriate way to handle layoffs as well. It may be hard to do because some managers know the people involved. But they could force all decisions to be justified and externally reviewed by others who don't know the demographics.

No one would have had a problem if they described a rigorous process of anonymizing the layoff decisions. It swings the pendulum too far to go full on anti-racist. Others point out it's also illegal.

1: https://www.povertyactionlab.org/case-study/unintended-effec...

"Anti-racist" to me simply means "against racism; actively ensuring racism doesn't occur". Maybe it means something different in the US. The idea of a company trying hard to "help people with minority sounding names" sounds pretty absurd to me.

true or not, those statements would create that perception.

Unless they lied in their statement, yes.

Reply on news.ycombinator.com