9
   

Affirmative Action in the US: is it still needed?

 
 
Thomas
 
  0  
Reply Tue 25 Sep, 2012 02:13 pm
@ehBeth,
ehBeth wrote:
google is your friend

No. Google is not my friend when I want to know what you think the relevant biases are. I am not interested in rebutting everything anyone says on the internet, but I am interested in considering what you say.
Thomas
 
  0  
Reply Tue 25 Sep, 2012 02:13 pm
@ehBeth,
ehBeth wrote:
at least pretend you're interested enough to look into it

I'm looking into it. This is the first question I would ask if I were joefromchicago's mayor.
0 Replies
 
ehBeth
 
  1  
Reply Tue 25 Sep, 2012 02:20 pm
@Thomas,
look at the links I posted toward the beginning of the thread - particularly the second one - there is a lot of interesting research to follow from the references there
0 Replies
 
cicerone imposter
 
  0  
Reply Tue 25 Sep, 2012 02:24 pm
@ehBeth,
You depend on "research outcomes" that are biased. I depend on personal observation, and my personal opinion about discrimination. I don't have any tolerance for discrimination.

I'm against all forms of discrimination - even reverse discrimination.
0 Replies
 
Thomas
 
  2  
Reply Tue 25 Sep, 2012 02:48 pm
@ehBeth,
ehBeth wrote:
well now we know you didn't follow any of the links I posted

No you don't. I have skimmed the first of your links. (I'm not going to read an entire 100-page academic study for an A2K thread, sorry.) I have also read all of your articles in this post. Only the last of your links concludes that the discrepancies among SAT scores by race proves bias. And it does that by assuming its conclusion.

Quote:
For a test to be regarded as "unbiased" or "fair", differential item functioning should not take place. Mean scores on the test's items should be alike for subgroups of interests who are alike with respect to overall ability on the construct of interest.

This is simply false. There is an alternative possibility, which you can't just dismiss without evidence or reasoning: There actually might be systematic differences between subgroups, which the SAT score measures correctly. If the average Asian applicant actually writes, reads, and calculates better than the average Caucasian, and the average Caucasian better than the average Black, that's what the SAT test should find. You can't just ignore this possibility by fiat.

As to the first three links in your post, none of them makes a reasoned case that the SAT score is biased. Numbers one and two merely look at the discrepancies in scores and report that some people say the discrepancies are evidence of test bias. But they offer no evidence by which the readers might decide whether those people are correct. So your claim is in there only as part of a he-said-she-said story. Link #3, finally, plots a variety of gaps by gender and ethnical group. But it, too, offers no reasoning or evidence that those gaps are caused by a bias in the test.

So yes, I have read the stories behind your links. And while they are interesting, they fall short of supporting your claim that SAT tests are biased.
0 Replies
 
joefromchicago
 
  2  
Reply Tue 25 Sep, 2012 03:42 pm
@Thomas,
Thomas wrote:
I don't know if subsequent Supreme Court decisions addressed the issue.

They have, although they've been all over the map. Gratz v. Bollinger said that a college can't automatically boost a minority applicant's admission score, but Gratter v. Bollinger, decided on the same day, said that a law school could take an applicant's minority status into account when making admission decisions.

Thomas wrote:
But they are both standards, however imperfect, that do not discriminate by race and gender. No standard is ever going to be perfect, and if universities find better standards than those, good for them! For purposes of this discussion, I'll settle for the best standard universities can find that doesn't discriminate by race gender, and so forth.

Well, if that's what you'll settle for, then a rock-paper-scissors contest would work just as well.

Thomas wrote:
It depends. Why did the minority applicants fail the exam?

Why does that matter? It was the same test for everyone, just like the SAT. Being a conscientious mayor, however, you can be confident that the test itself is not biased.
Thomas
 
  2  
Reply Tue 25 Sep, 2012 04:09 pm
@joefromchicago,
joefromchicago wrote:
Well, if that's what you'll settle for, then a rock-paper-scissors contest would work just as well.

It would, as far as the affirmative-action issue is concerned. I would have various problems with rock-paper-scissors, but discrimination against suspect classes would not be among them.

joefromchicago wrote:
Thomas wrote:
It depends. Why did the minority applicants fail the exam?

Why does that matter? It was the same test for everyone, just like the SAT. Being a conscientious mayor, however, you can be confident that the test itself is not biased.

If the test fails applicants who would make good police officers, it's a bullshit test, regardless of the applicant's race, just as rock-paper-scissors would be. But if it's a valid, no-bullshit test, the minority applicants' failure would indicate that they wouldn't make for good police officers. I would refrain from hiring bad police officers just because they belong to minorities. That's why it makes a difference, and that's why my first question would be: How did the police officers fail the test? What exactly did they fail at?
joefromchicago
 
  2  
Reply Tue 25 Sep, 2012 10:48 pm
@Thomas,
Thomas wrote:
If the test fails applicants who would make good police officers, it's a bullshit test, regardless of the applicant's race, just as rock-paper-scissors would be.

Why so? The police department doesn't have an unlimited number of jobs available. If there are 20 slots to fill, then wouldn't it be fair to take the 20 top-scoring applicants on the exam, even if the top 50 would have made good officers? After all, that's how you'd handle college admissions.

Thomas wrote:
But if it's a valid, no-bullshit test, the minority applicants' failure would indicate that they wouldn't make for good police officers. I would refrain from hiring bad police officers just because they belong to minorities.

I see. It's not that you support tests, you just support good tests. It's a fine distinction, to be sure. The trick, though, is to distinguish bad tests from good tests.

For instance, if candidate A fails the test to become a police officer, how would you determine if he would nevertheless have made a good police officer? If you need to see how that candidate would perform as a police officer, that would be problematic, given that you can't see him do that so long as he can't pass the test. And if there's a way to determine how well he would perform as an officer without having first passed the test, then isn't that the better way to determine if he would make a good officer rather than the test?

Thomas wrote:
That's why it makes a difference, and that's why my first question would be: How did the police officers fail the test? What exactly did they fail at?

I'm still not seeing why that would make a difference. You can assume that thy failed because they did not fulfill the requirements for passing.
cicerone imposter
 
  1  
Reply Tue 25 Sep, 2012 10:52 pm
@joefromchicago,
Good points; that applies to most professions. Testing can only reveal so much of a candidate no matter what profession they are testing for.
0 Replies
 
Thomas
 
  0  
Reply Wed 26 Sep, 2012 05:06 am
@joefromchicago,
joefromchicago wrote:
Why so? The police department doesn't have an unlimited number of jobs available. If there are 20 slots to fill, then wouldn't it be fair to take the 20 top-scoring applicants on the exam, even if the top 50 would have made good officers? After all, that's how you'd handle college admissions.

But that's not all you stipulated earlier. You stipulated that the majority of crime happens in minority neighborhoods, and that the people in those neighborhoods mistrust majority police officers. Consequently, the rejected minority applicants would have made superior officers if hired, meaning that the test may not be a valid predictor of good service. This apparent mismatch between what the city tests for and what the city needs is what gives me probable course to search for flaws in the test. Not unfairness to the applicants.

joefromchicago wrote:
I see. It's not that you support tests, you just support good tests.

Yes I do. Am I sensing a tone of sarcasm in your voice here, or is that just me?

joefromchicago wrote:
For instance, if candidate A fails the test to become a police officer, how would you determine if he would nevertheless have made a good police officer?

I don't have an algorithm for answering that question in the abstract and in all generality. If you have a specific case in mind, now would be a helpful time to talk about the specifics. What was the test, and what did the minority applicants fail at?

joefromchicago wrote:
I'm still not seeing why that would make a difference. You can assume that thy failed because they did not fulfill the requirements for passing.

Apparently I'm not getting across my distinction between "it makes no difference" and "it makes a difference for reasons other than racial fairness". In the scenario you stipulate, my problem with your test is that it seems to be dumb policy for the police department, not that it seems to be unfair to the applicants. I'm the mayor, so it's my job to enact policies that are not dumb. That's why it makes a difference.
Thomas
 
  2  
Reply Wed 26 Sep, 2012 05:40 am
@joefromchicago,
joefromchicago wrote:
For instance, if candidate A fails the test to become a police officer, how would you determine if he would nevertheless have made a good police officer?

I wouldn't. I would look around for better tests --- after all, I'm not the only mayor in America. If I found some other test, and if the cities who use it reduced their crime rates by switching to it from my city's current test, that would persuade me to ditch our test and adopt theirs.

And as a last resort, if all the tests I look at strike me as dumb in my professional opinion as a mayor, I might devise my own. As Justice Brandeis observed so astutely, "It is one of the happy incidents of the federal system that a single courageous state may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country." I'm confident the principle can work for cities, too.
0 Replies
 
joefromchicago
 
  2  
Reply Wed 26 Sep, 2012 08:20 am
@Thomas,
Thomas wrote:
This apparent mismatch between what the city tests for and what the city needs is what gives me probable course to search for flaws in the test. Not unfairness to the applicants.

Well, the city needs minority police officers. How do you test for that?

Thomas wrote:
Yes I do. Am I sensing a tone of sarcasm in your voice here, or is that just me?

No, that's sarcasm.

Thomas wrote:
joefromchicago wrote:
For instance, if candidate A fails the test to become a police officer, how would you determine if he would nevertheless have made a good police officer?

I don't have an algorithm for answering that question in the abstract and in all generality. If you have a specific case in mind, now would be a helpful time to talk about the specifics. What was the test, and what did the minority applicants fail at?

It's more of a logical conundrum. The test, presumably, is designed to weed out the bad applicants from the good applicants. According to you, a test is bad if it weeds out good applicants along with bad applicants. But how do you determine if the test is weeding out good applicants if the only means you have to measure a candidate's goodness is the test? You need something else -- a "meta-test" -- to test the test. And if the meta-test does a better job of determining who is good and who is bad, then shouldn't you be using that test instead?

Thomas wrote:
I wouldn't. I would look around for better tests

Let me get this straight: you would replace the current test with a test that allows more minority applicants to pass? How is that not affirmative action?
Thomas
 
  2  
Reply Wed 26 Sep, 2012 08:42 am
@joefromchicago,
joefromchicago wrote:
But how do you determine if the test is weeding out good applicants if the only means you have to measure a candidate's goodness is the test? You need something else -- a "meta-test" -- to test the test.

The meta test is to compare the crime rate in my city with the crime rates of cities that use different tests. If the statistics on those crime rates --- by race, gender, time series, and whatnot --- suggests that police forces with different tests get better results, that tells me I need to improve the test in my own city.

joefromchicago wrote:
Thomas wrote:
I wouldn't. I would look around for better tests

Let me get this straight: you would replace the current test with a test that allows more minority applicants to pass? How is that not affirmative action?

Because the policy objective is better law enforcement for the city's residents, not fairer job opportunities for the city's would-be police officers. If I could provide the citizens better law enforcement with an all-White police force, I would. But that's not the scenario you stipulated --- realistically so, I think --- and I'm running with your stipulated facts.

Just because a broken clock happens to be right twice a day, that doesn't make it an accurate clock. And just because good policy happens to meet the demands of affirmative-action advocates every now and then, that doesn't make it affirmative action. You know that. You've been giving Occom Bill the correlation-is-not-causation speech for years.
joefromchicago
 
  2  
Reply Wed 26 Sep, 2012 10:12 am
@Thomas,
Thomas wrote:
Because the policy objective is better law enforcement for the city's residents, not fairer job opportunities for the city's would-be police officers. If I could provide the citizens better law enforcement with an all-White police force, I would. But that's not the scenario you stipulated --- realistically so, I think --- and I'm running with your stipulated facts.

I'm astounded. After saying that you opposed affirmative action, not only have you taken a stand in favor of affirmative action, but now you're actually coming up with pretextual arguments to deny having supported affirmative action.

But from the point of view of a disappointed white applicant, who would have been hired under the old test, it makes very little difference whether your goal is to increase minority representation on the police force or to provide citizens with better law enforcement. To him, you just substituted a test that was fair with one that wasn't.

Thomas wrote:
Just because a broken clock happens to be right twice a day, that doesn't make it an accurate clock. And just because good policy happens to meet the demands of affirmative-action advocates every now and then, that doesn't make it affirmative action. You know that. You've been giving Occom Bill the correlation-is-not-causation speech for years.

This isn't an instance of correlation-versus-causation. When universities take student diversity into consideration when making admission decisions, you oppose that. But when a city takes crime-prevention into consideration -- and that indirectly involves considerations of race -- then you're in favor of that. But there's not a whole lot of difference between the two situations. If a city thinks that it would be a better place with more minority police officers, then it can take race into consideration when hiring officers. But, for some reason, if university thinks that it is a better place with more minority representation on campus, it can't take race into consideration when admitting students. I fail to see the distinction.

I'm not sure how you can oppose affirmative action while supporting programs that work exactly like affirmative action but just aren't called "affirmative action." Why does calling something "affirmative action" make a difference if, in the end, it achieves the same result?
Thomas
 
  1  
Reply Wed 26 Sep, 2012 11:50 am
@joefromchicago,
joefromchicago wrote:
I'm astounded.

I have my moments.

joefromchicago wrote:
But from the point of view of a disappointed white applicant, who would have been hired under the old test, it makes very little difference whether your goal is to increase minority representation on the police force or to provide citizens with better law enforcement. To him, you just substituted a test that was fair with one that wasn't.

It's a good thing, then, that I don't care about the White applicant's disappointment, just as I don't care about the Black applicant's disappointment with falling short on the SAT test. If that coincides with what an equal-opportunity disappointer would do, I don't care about that either.

joefromchicago wrote:
This isn't an instance of correlation-versus-causation. When universities take student diversity into consideration when making admission decisions, you oppose that. But when a city takes crime-prevention into consideration -- and that indirectly involves considerations of race -- then you're in favor of that. But there's not a whole lot of difference between the two situations.

The difference is that in the scenario you're describing, minority police officers are in a better position to win trust in the neighborhoods it's policing. That has direct consequences for how well they're doing their job. But in universities, in general, there is no such benefit to studying. Being white rather than Asian, or Black rather than White, will not make you a better mathematician, scientist, or engineer. There might be exceptions around the humanities, but as a general rule, I see no benefit comparable to the police-force scenario as you described it.

joefromchicago wrote:
Why does calling something "affirmative action" make a difference if, in the end, it achieves the same result?

That's not the distinction. I've explained my distinction several times; I won't explain it again.
joefromchicago
 
  2  
Reply Thu 27 Sep, 2012 08:48 am
@Thomas,
Thomas wrote:
It's a good thing, then, that I don't care about the White applicant's disappointment, just as I don't care about the Black applicant's disappointment with falling short on the SAT test. If that coincides with what an equal-opportunity disappointer would do, I don't care about that either.

Maybe you should care. After all, racial discrimination is illegal.

Thomas wrote:
The difference is that in the scenario you're describing, minority police officers are in a better position to win trust in the neighborhoods it's policing. That has direct consequences for how well they're doing their job. But in universities, in general, there is no such benefit to studying. Being white rather than Asian, or Black rather than White, will not make you a better mathematician, scientist, or engineer. There might be exceptions around the humanities, but as a general rule, I see no benefit comparable to the police-force scenario as you described it.

Sez you. Many universities would disagree. But then if your only objection to affirmative action is that it shouldn't be used where it does no good, then you're not really opposed to affirmative action per se, you're just opposed to it in certain situations.
Thomas
 
  1  
Reply Thu 27 Sep, 2012 09:53 am
@joefromchicago,
joefromchicago wrote:
Maybe you should care. After all, racial discrimination is illegal.

That's a matter of interpretation Affirmative action, in particular, can be viewed as racial discrimination against Whites and Asians. Yet it isn't just legal, it's the law.

joefromchicago wrote:
But then if your only objection to affirmative action is that it shouldn't be used where it does no good, then you're not really opposed to affirmative action per se, you're just opposed to it in certain situations.

I think that's an odd way of framing the issue. You might as well say: "But then if your only objection against broken clocks is that they shouldn't be used when they tell the incorrect time, then you're not really opposed to broken clocks per se. You're just opposed to them in certain situations." I'm not going to deny it's 11:53 am EDT right now just because my broken clock says it's 11:53. Likewise, I am not going to oppose a good policy just because affirmative-action principles approve of it, too. But that doesn't change my opinion that broken clocks are bad timekeepers --- and that affirmative action is bad social policy.

But hey, this is not the first time I think you frame political questions in odd ways, and I'm sure the feeling is mutual. I guess we're pretty much even in that department. Smile
0 Replies
 
ehBeth
 
  2  
Reply Thu 27 Sep, 2012 10:10 am
@Thomas,
Thomas wrote:
But in universities, in general, there is no such benefit to studying. Being white
is the best choice, if you've got a choice of what colour to be.

http://www.insidehighered.com/news/2010/06/21/sat

Quote:

But what Freedle found in 2003 has now been confirmed independently by the new study: that some kinds of verbal questions have a DIF for black and white students. On some of the easier verbal questions, the two studies found that a DIF favored white students. On some of the most difficult verbal questions, the DIF favored black students. Freedle's theory about why this would be the case was that easier questions are likely reflected in the cultural expressions that are used commonly in the dominant (white) society, so white students have an edge based not on education or study skills or aptitude, but because they are most likely growing up around white people. The more difficult words are more likely to be learned, not just absorbed.

While the studies found gains for both black and white students on parts of the SAT, the white advantage is larger such that the studies suggest scores for black students are being held down by the way the test is scored and that a shift to favor the more difficult questions would benefit black test-takers.


interesting results. Studying helps if you can't manage being white.


http://voices.washingtonpost.com/class-struggle/2010/06/new_evidence_that_sat_hurts_bl.html#more

Quote:
On average, he said, black students were performing only slightly above matched-ability whites on hard questions. But averages did not submit applications to colleges. Individual students did. Some of those individuals, he discovered, would have gotten a boost of a hundred points or more on the SAT if the score was weighted toward the hard items. He proposed that the College Board offer a supplement to SAT scores, called the Revised-SAT, or R-SAT, which would be calculated based only on the hard items. This, he said, would "greatly increase the number of high-scoring minority individuals."

In their paper, Santelices and Wilson rule out Freedle's suggestion that the bias he found in the test might affect all kinds of multiple-choice questions, or minorities other than blacks. But they did find it in sentence completion and reading comprehension sections of the SAT.

Saul Geiser was the director of research in Atkinson's office originally given the assignment to look into Freedle's theory. Eventually he arranged for Santelices, then a doctoral candidate at UC Berkeley, to do the research as her PhD thesis, working with Wilson, a UC Berkeley psychometrician who had also been asked to look at Freedle's work.

Geiser said he thinks the two researchers did a good job. He does not agree with Bunin's criticisms of their work. He said he, like Freedle, wants more more research on why blacks and whites answer these questions differently, so that any unfair disadvantages for blacks can be removed.

He said he thought the College Board, in particular, should "get over the denial" of any merit to what Freedle has discovered. That may take a while. The College Board, after all, may be right that the SAT is unbiased.

But the new paper means more researchers are likely to go more deeply into what Freedle has found, and eventually settle the question of what should be done about it.



cicerone imposter
 
  1  
Reply Thu 27 Sep, 2012 10:20 am
@ehBeth,
I'd like to see some samples of those SAT questions that favor white or black test takers.

0 Replies
 
Thomas
 
  1  
Reply Thu 27 Sep, 2012 10:49 am
@ehBeth,
Scott Jaschik, as quoted by ehBeth, wrote:
On some of the easier verbal questions, the two studies found that a DIF favored white students. On some of the most difficult verbal questions, the DIF favored black students.

But notice what they're not saying. They're not saying that, as currently scored, the Black advantage on hard questions and the White advantage on easy questions add up to a disadvantage for Black test takers overall --- which is what they would need to establish an anti-Black bias in the SAT test. And while the authors diplomatically say their results "don't invalidate" the findings of an old study, which claimed that its data did establish that, they did not say they confirmed it either. They would have said so if they had. Scientists aren't shy about these things in my experience.

Consequently, your article did not change my opinion, which coincides with what the article calls "the standard theory" behind SAT-score discrepancies: The SAT is basically fair. It's society that has been unfair to those students while they were still in the K-12 system. Hence, as I said in the beginning of this thread, I expect that most of the discrepancies in test scores will go away with policies that decrease after-tax income inequality among parents, and that increase the budgets of poor-neighborhood schools.

That said, if a net disadvantage for Black students remains after accounting for income inequality among parents and schools, it makes sense to correct them. And statistical instruments like your study's "Differential Item Functioning" are, in principle, the right way to spot them. But from the facts presented in your article, I do not get the impression that the authors of your study have established that the SAT has an overall bias against Blacks.
 

Related Topics

 
Copyright © 2019 MadLab, LLC :: Terms of Service :: Privacy Policy :: Page generated in 0.03 seconds on 10/15/2019 at 10:25:36