OMNILIBRIUM
  Rational Discussion of Controversial Topics


GO TO THE MAIN THREAD Sort By:


melian 3 September 2015 12:43 PM
70%

do you mean that you're an X, and when you take the test as an X it says you're biased against Ys, but when you check the Y box the results say you're biased against Xs?

No, they had separate tests for different ethnicities. Before starting a test for each ethnicity you were supposed to check the box on whether you belong to that ethnicity.
I would find it very surprising if greater unconscious racism/sexism/etc. did not on average reduce the effectiveness of collaboration with members of the biased-against groups

If unconscious biases were so important, I would expect companies in countries where such biases are generally much stronger (e.g., Singapore) not being able to compete with their western competitors.
it may seem scary to judge individuals for something in their heads they can't control

What made “1984” so scary to me when I first read it is precisely the opposite - what people think in their heads can be controlled. Obviously, physical torture is more effective, but the threat to employment prospects is also a pretty strong motivator.


stars0
Reply


Silent Cal 4 September 2015 02:03 PM
67%

I'm still not understanding what happened on the IATs you took. Are you saying that the box you checked beforehand influenced the results? Or just that it's implausible that it found you were biased against your own ethnicity?

If unconscious biases were so important, I would expect companies in countries where such biases are generally much stronger (e.g., Singapore) not being able to compete with their western competitors.

'So important' doesn't have to be that important--even modest gains could justify such testing. Singapore has a lot of positives in its favor that could be compensating for its bias issues.

Maybe we can agree that studies are called for? Get a volunteer pool, measure IQ or a more task-specific ability measure, also measure bias, then form ability-controlled groups of varying bias levels. See if more biased groups do worse or not.

On the 1984 side, let's think about other mental things companies could test for. Trustingness of others' intentions? Trust of others' abilities? Or trust of authority? I have to admit that last one creeps the hell out of me, no matter how justified it may turn out to be from a performance-maximizing standpoint.

stars0
Reply


melian 4 September 2015 03:42 PM
70%

Are you saying that the box you checked beforehand influenced the results?

I strongly suspect so.
Or just that it's implausible that it found you were biased against your own ethnicity?

That too.
Maybe we can agree that studies are called for?

This is an interesting question to research, but I don’t see how it can be done in practice. First, if you use volunteers for the experiment you are already skewing the results due to self-selection effect (volunteer “bigots” and “non-bigots” are likely to be different from “bigots” and “non-bigots” in the general population). More importantly, due to the highly controversial nature of the question I would not trust people in charge of the experiment to conduct it in a truly scientific way. As with most sociological experiments, the published results are likely to reflect the biases of the researchers rather than the truth.

A somewhat related question – what do you think of allowing the use of polygraphs?


stars0
Reply


Silent Cal 7 September 2015 01:21 PM
69%

You're probably right that, if unconscious bias doesn't actually affect team performance, that result would have trouble making it to publication. Maybe the team would need an unmeasured scapegoat explanation handy so that if their results are negative, they can say 'Unconscious bias isn't the problem, so it must be X' and save some face.

I have no objection to the use of polygraph as an instrument. I would place some restrictions on what questions can be asked, though. My first impulse is that any question you aren't allowed to ask on polygraph, you shouldn't be able to ask at all.



stars0
Reply


melian 14 September 2015 08:44 AM
70%

Based on my experience with sociological studies, I’m afraid something like the following scenario is more likely. A bunch of college students are mixed with high school dropouts. After being tested for racial biases, they are split in pairs and given some g-loaded tasks. The results show that pairs of people without strong biases have much better performance. Experiment organizers publish a paper -“Racial biases hurt productivity”.

stars0
Reply