There is another potential solution, not discussed here and maybe not generally applicable, but nonetheless coherent with these very interesting findings: make as many domains as possible *less* competitive, and/or provide multiple avenues for success.
Simply put, if the avenue to success and wealth are few and far between, with insufficient capacity for the population, then minimal levels of bias will have an outsize influence. This would call for example for having many more colleges and universities with far less endowment disparity, and therefore effect of the college brand on the professional success. It really should not be the case that to become a US supreme justice you almost need to study at Harvard, or to become a UK MP it really helps to have gone to Eton and then Oxford/Cambridge.
Italy (which however has a lot of different issues with its academic sector) solved this problem with a law which forces the public sector to *not* consider the school awarding the degree for public positions. The older I get, the more I think it is in general a good idea.
IMO, your notion of all "acts" (versus experiences) as the denominators is not workable. It makes the apparent frequency of discrimination depend on the baseline frequencies of positive and negative decisions. Academic hiring, for example, will never be worse than 1 act of discrimination out of, say, 200 applications (ignoring interviews).
I have different views about the ultimatum game vs employment audit studies vs the UW Madison experiments. It's about what's at stake in the decision and how. In the ultimatum game, the player is trying to predict how the opponent will react. It is not clear that, for example, a racist will be stingy toward a black. Maybe he'll expect the black to be irrationally angry and self-destructive.
In the audit studies, the employer is deciding whether to invest 20 minutes in a preliminary interview. Since the CV is limited information, his prior, racial expectations will play a role in estimating the benefits of the interview.
In the Madison studies, the subject is just trying to treat people appropriately (according to whatever standards he has). One may be a total racist, but think that even an inferior person should have a door held open or receive directions when needed.
Baseline discrimination *is* the baseline frequencies of positive and negative behaviors. Discrimination is something done by potential perpetrators. To know how often they are actual perpetrators, you need to take their acts of discrim and divide them by their acts of nondiscrimination. NOT hiring either a Black or White applicant, when both are similarly qualified, is an act of nondiscrimination.
Your other points are mere explanations for why discrim did or did not occur. But the explanation is irrelevant to the assessment of whether discrim did or did not occur in those studies.
This contrast of act and experience is enlightening. It could be expressed more starkly as act and outcome – both being objective, reserving "experience" for subjectivity. British adjudication of discrimination cases frequently states "a detriment plus a protected characteristic does not imply discrimination – causation needs to be established." The prevalence of overriding beliefs in isms, which bridge the gap of causation, implies that the experience of isms is far higher than the fact.
The word "discrimination" means the ability or power to see or make fine distinctions (American Heritage Dictionary) or the act of doing so. To misuse it to mean "prejudice" or "bias" inverts the real meaning, leaving it meaningless.
Now dear fellow, you see: calling language standards to the muster in support of denying even the meaning of something that offends you is a common behaviour of people. Nevertheless it is almost always a failing strategy and pretty silly all in all.
Languages evolve and words take on further meaning and lose some (otherwise, we would still be using "weird" to mean "something touched by Fate"). This happens whether we like it or not, by the force of usage over time. To be blind to it is pretty vacuous, and unhelpful.
To go down to the specifics -- so sorry, but I am one of those the new generations call "grammar nazis" -- you are incorrect even with respect to your quoted dictionary.
This is the entire entry in the American Heritage Dictionary about the word "discrimination":
dis·crim·i·na·tion (dĭ-skrĭm′ə-nāshən)
n.
1. The act of discriminating.
2. The ability or power to see or make fine distinctions; discernment.
3. Treatment or consideration based on class or category, such as race or gender, rather than individual merit; partiality or prejudice.
You seem to have (deliberately) missed definition 3.
My own OED (towards which I am biased, being British and inclined to look down my nose at the language developed by the Rebels in the former Colonies) not only gives me all of those three definitions, but also the time in which they first appeared in printed writing.
Discrimination in sense 3. is originally a US usage (and stands today as a testament to the massive influence of US culture in the English-speaking world and beyond -- a state of things that I cannot change much though it irks me mightily).
Further. It appeared the first time in the field of economics, in 1789, in a paper of your Congress. And it appeared, in the social/sociological sense used also today, in 1819 in the Analectic Magazine of Philadelphia.
Now, it is pretty meaningless to accuse a usage that has been common course for over 200 years of betraying the original meaning (which also includes separating things into categories, which is what drove its adoption to mean "prejudiced attitude or behaviour).
I'd suggest finding better rhetorical weapons to express your distaste of the too broad and frequent use of a concept.
I made no other comment. You ascribed the motive. I neither agreed nor diagreed with the author in any other particular to a substantive degree. Although I might quibble whether taking the difference in percentages as a percentage is ever justifiable.
There is another potential solution, not discussed here and maybe not generally applicable, but nonetheless coherent with these very interesting findings: make as many domains as possible *less* competitive, and/or provide multiple avenues for success.
Simply put, if the avenue to success and wealth are few and far between, with insufficient capacity for the population, then minimal levels of bias will have an outsize influence. This would call for example for having many more colleges and universities with far less endowment disparity, and therefore effect of the college brand on the professional success. It really should not be the case that to become a US supreme justice you almost need to study at Harvard, or to become a UK MP it really helps to have gone to Eton and then Oxford/Cambridge.
Italy (which however has a lot of different issues with its academic sector) solved this problem with a law which forces the public sector to *not* consider the school awarding the degree for public positions. The older I get, the more I think it is in general a good idea.
IMO, your notion of all "acts" (versus experiences) as the denominators is not workable. It makes the apparent frequency of discrimination depend on the baseline frequencies of positive and negative decisions. Academic hiring, for example, will never be worse than 1 act of discrimination out of, say, 200 applications (ignoring interviews).
I have different views about the ultimatum game vs employment audit studies vs the UW Madison experiments. It's about what's at stake in the decision and how. In the ultimatum game, the player is trying to predict how the opponent will react. It is not clear that, for example, a racist will be stingy toward a black. Maybe he'll expect the black to be irrationally angry and self-destructive.
In the audit studies, the employer is deciding whether to invest 20 minutes in a preliminary interview. Since the CV is limited information, his prior, racial expectations will play a role in estimating the benefits of the interview.
In the Madison studies, the subject is just trying to treat people appropriately (according to whatever standards he has). One may be a total racist, but think that even an inferior person should have a door held open or receive directions when needed.
Baseline discrimination *is* the baseline frequencies of positive and negative behaviors. Discrimination is something done by potential perpetrators. To know how often they are actual perpetrators, you need to take their acts of discrim and divide them by their acts of nondiscrimination. NOT hiring either a Black or White applicant, when both are similarly qualified, is an act of nondiscrimination.
Your other points are mere explanations for why discrim did or did not occur. But the explanation is irrelevant to the assessment of whether discrim did or did not occur in those studies.
This seems similar, although from a different perspective, to this thought experiment about hate crimes: https://medium.com/@inconvenientnumbers/inconvenient-models-of-bias-70e7562ca965
This contrast of act and experience is enlightening. It could be expressed more starkly as act and outcome – both being objective, reserving "experience" for subjectivity. British adjudication of discrimination cases frequently states "a detriment plus a protected characteristic does not imply discrimination – causation needs to be established." The prevalence of overriding beliefs in isms, which bridge the gap of causation, implies that the experience of isms is far higher than the fact.
Such a gift, Jussim. This is a problem that has plagued me for a long time. Thank you for so clearly unravelling it.
The word "discrimination" means the ability or power to see or make fine distinctions (American Heritage Dictionary) or the act of doing so. To misuse it to mean "prejudice" or "bias" inverts the real meaning, leaving it meaningless.
Now dear fellow, you see: calling language standards to the muster in support of denying even the meaning of something that offends you is a common behaviour of people. Nevertheless it is almost always a failing strategy and pretty silly all in all.
Languages evolve and words take on further meaning and lose some (otherwise, we would still be using "weird" to mean "something touched by Fate"). This happens whether we like it or not, by the force of usage over time. To be blind to it is pretty vacuous, and unhelpful.
To go down to the specifics -- so sorry, but I am one of those the new generations call "grammar nazis" -- you are incorrect even with respect to your quoted dictionary.
This is the entire entry in the American Heritage Dictionary about the word "discrimination":
dis·crim·i·na·tion (dĭ-skrĭm′ə-nāshən)
n.
1. The act of discriminating.
2. The ability or power to see or make fine distinctions; discernment.
3. Treatment or consideration based on class or category, such as race or gender, rather than individual merit; partiality or prejudice.
You seem to have (deliberately) missed definition 3.
My own OED (towards which I am biased, being British and inclined to look down my nose at the language developed by the Rebels in the former Colonies) not only gives me all of those three definitions, but also the time in which they first appeared in printed writing.
Discrimination in sense 3. is originally a US usage (and stands today as a testament to the massive influence of US culture in the English-speaking world and beyond -- a state of things that I cannot change much though it irks me mightily).
Further. It appeared the first time in the field of economics, in 1789, in a paper of your Congress. And it appeared, in the social/sociological sense used also today, in 1819 in the Analectic Magazine of Philadelphia.
Now, it is pretty meaningless to accuse a usage that has been common course for over 200 years of betraying the original meaning (which also includes separating things into categories, which is what drove its adoption to mean "prejudiced attitude or behaviour).
I'd suggest finding better rhetorical weapons to express your distaste of the too broad and frequent use of a concept.
As Carlin might say, its third for a reason.
I made no other comment. You ascribed the motive. I neither agreed nor diagreed with the author in any other particular to a substantive degree. Although I might quibble whether taking the difference in percentages as a percentage is ever justifiable.