Why Affirmative Action is Justified

[redemption]

I am going to try, perhaps very naively, to propose a sensible discussion about Affirmative Action.

I am going to post a series of facts  that I believe are: persuasive; that may be interesting for all of us to think about; and that may re-orient the way in which we think about affirmative action on this Board.

The order in which I will post will in the end constitute a narrative, and it is probably better that I outline that storyline here. It goes something like this:

---

1. There is a phenomenon known as ‘stereotype threat’. It exists, it is real, and its performance effects are -  for motivated, able and accomplished URMs - substantial .

2. The LSAT is a test in which ‘stereotype threat’ flourishes

3. Holding UGPA + UG institution +  academic department + Major + socioeconomic status all constant,  African Americans receive on average 6 scaled points lower on the LSAT  than white; Latino/as about 4 points lower, and Native Americans about 2 points lower.

4. For high-GPA URM applicants to the top schools, this gap in average LSAT scores may be largely explained by the presence of stereotype threat.

5a. Law Schools have known of this gap, and of its causes, for at least 25 years

5b. That they continue to rely on a test that penalizes groups taking it under a stereotype threat, and that they rely on it so heavily, despite its weakness in predicting law school performance (i.e. a test that discriminates between URMs and the majority, and that is of dubious utility) is .... puzzling

5c.  Using a test known to discriminate against URMs is not, in practice, all that different from an intent to dicriminate; and a system in which the biased results are forseeable is not all that different from a system in which the biased results are intended.

6.  Race-based Affirmative Action can and should -- in principle -- be fully justified on the basis of this present-day systematic bias against accomplished URMs applying to law schools.

7. When combined with the diversity rationale, and with the need to counteract the demonstrable evidence of implicit bias even in the case of two candidates with exactly the same credentials, the case for Affirmative Action becomes overwhelming.

8. Affirmative Action is not a remedy for the supposed under-qualified academic standing of URMs. It is a corrective, instead, for the strong implicit societal bias against well-qualified, well-motivated URMs.
------------

I shall now try to post something that demonstrates the first point in my narrative, and will, I hope, generate some discussion. Please contribute any thoughts that you may have.  Constructive, thoughtful and intellectually honest criticism is particularly welcome. 

Before moving on to point 2, I shall attempt to summarize the agreements and points of disagreement on the first point.

I should state explicitly that I myself am open to persuasion by stronger counterevidence or counternarratives.

This is an experiment -- if it fails, so be it.

[redemption]

Re: the First Point


1.1. There is a psychological phenomenon known as ‘Stereotype Threat’

Steele and Aronson (“Contending with Group Image: The Psychology of Stereotype and Social Identity Threat,” by Claude M. Steele, Steven J. Spencer, and Joshua Aronson, Advances in Experimental Social Psychology, 2002.) & very many later studies have found that stereotype threat affects performance.

In their initial study, after adjusting for initial differences in SAT scores, black students at Stanford University who took a challenging verbal test answered approximately 10 percent fewer questions correctly than whites did—but only if they believed that the test was a measure of their ability. If they were told that the test measured “psychological factors involved in solving verbal problems,” the black-white test score difference was eliminated.

These studies have been replicated many times, and are undiputed, both in terms of their results and in terms of their methodology.


1.2. Stereotype Threat is a Real Psychological State for which there is Direct Evidence


“what we needed next was direct evidence of thesubjective state we call stereotype threat. To seek this, we looked intowhether simply sitting down to take a difficult test of ability was enough to make black students mindful of their race and stereotypes about it. This may seem unlikely. White students I have taught over the years have sometimes said that they have hardly any sense of even having a race. But blacks have many experiences with the majority "other group" that make their race salient tothem.

We again brought black and white students in to take a difficult verbal test. But just before the test began, we gave them a long list of words, each of which had two letters missing. They were told to complete the words on this list as fast as they could. We knew from a preliminary survey that twelve of the eighty words we had selected could be completed in such a way as to relate to the stereotype about blacks' intellectual ability. The fragment "__ce," for example, could become "race." If simply taking a difficult test of ability was enough to make black students mindful of stereotypes about their race, these students should complete more fragments with stereotype-related words. That is just what happened. When black students were told that the test would measure ability, they completed the fragments with significantly more stereotype-related words than when they were told that it was not a measure of ability. Whites made few stereotype-related completions in either case.”


1.3. Stereotype Threat is not the Same Thing as Self-Doubt

(Aronson J, Lustina MJ, Good C, Keough K, Steele CM, Brown J. When white men can’t do math: Necessary and sufficient factors in stereotype threat. Journal of Experimental Social Psychology, 35, 29-46. 1999).

The researchers told white male students who were strong in math  (they all had nearly perfect scores on the SAT Math) that a difficult math test they were about to take was one on which Asians generally did better than whites.

White males should not have a sense of group inferiority about math, since no societal stereotype alleges such an inferiority. Yet this comment would put them under a form of stereotype threat: any faltering on the test could cause them to be seen negatively from the standpoint of the positive stereotype about Asians and math ability. If stereotype threat alone--in the absence of any internalized self-doubt--was capable of disrupting test performance, then white males taking the test after this comment should perform less well than white males taking the test without hearing the comment.

That is just what happened. Stereotype threat impaired intellectual functioning in a group unlikely to have any sense of group inferiority -- high-achieving white men who are very strong in math.

“In science, as in the rest of life, few things are definitive. But these results are pretty good evidence that stereotype threat's impairment of standardized-test performance does not depend on cueing a pre-existing anxiety.”

Alternatively

Studies (e.g.  Kray, Laura, Reb, Jochen M., Galinsky, Adam D. and Thompson, Leigh, "Gender Stereotype Activation and Power in Mixed-Gender Negotiations")  show that women do worse on challenging tests of mathematical and scientific material, both when they are primed to think that the test demonstrates gender differences in math ability and when they are not primed about the test’s content (and thus are reacting purely on their knowledge that society expects women to be bad at math). The male-female gap is eliminated only when women are led to believe that the test is gender- neutral.

Or

White athletes did worse than black athletes in a golf exercise when they thought their scores demonstrated “natural athletic ability” (a stereotypically black trait), whereas blacks did worse than whites when they thought it tested “sports strategic intelligence” (a stereotypically white trait).


1.4. Stereotype Threat Affects the Most Able, most Qualified, and Most Motivated Members of a Group


Claude Steele -- “Is everyone equally threatened and disrupted by a stereotype? One might expect, for example, that it would affect the weakest students most. But in all our research the most achievement-oriented students, who were also the most skilled, motivated, and confident, were the most impaired by stereotype threat. This fact had been under our noses all along--in our data and even in our theory. A person has to care about a domain in order to be disturbed by the prospect of being stereotyped in it. That is the whole idea of disidentification--protecting against stereotype threat by ceasing to care about the domain in which the stereotype applies. Our earlier experiments had selected black students who identified with verbal skills and women who identified with math. But when we tested participants who identified less with these domains, what had been under our noses hit us in the face. None of them showed any effect of stereotype threat whatsoever.”

Ironically, and poignantly, the data show that

“what exposes students to the pressure of stereotype threat is not weaker academic identity and skills but stronger academic identity and skills. They may have long seen themselves as good students--better than most. But led into the domain by their strengths, they pay an extra tax on their investment--vigilant worry that their future will be compromised by society's perception and treatment of their group.”

What exactly is happening under Stereotype Threat?

“In some of our experiments we administered the test of ability by computer, so that we could see how long participants spent looking at different parts of the test questions. Black students taking the test under stereotype threat seemed to be trying too hard rather than not hard enough. They reread the questions, reread the multiple choices, rechecked their answers, more than when they were not under stereotype threat. The threat made them inefficient on a test that, like most standardized tests, is set up so that thinking long often means thinking wrong, especially on difficult items like the ones we
used.”

and, it has physiological manifestations

A study (Blascovich, J., Spencer, S., Quinn, D., & Steele, C. (2001)). African-Americans and high blood pressure: The role of stereotype threat. Psychological Science, 12, 225-229) found that the blood pressure of black students performing a difficult cognitive task under stereotype threat was elevated compared with that of black students not under stereotype threat or white students in either situation.

[redemption]

1. Is the degree of difference in LSAT scores attributable to stereotype threat and other factors that affect minority performance equivalent to the 10+ points lower URMs seem to be able to score without experiencing a narrowing in the schools available to them?

2. Also, does stereotype threat affect GPA? I've seen the discussions on standardized testing and stereotype threat but nothing on its effect on GPA.

Looking at LSN, there seems to be a strong possibility that two students--one a 3.8/172 white applicant and one a 3.1/163 URM applicant, will end up at the same school. I understand if the LSATs are taken as equivalent, considering stereotype threat and less access to resources (such as test prep companies) that URMs face. However, the GPAs are strikingly different so why the added consideration there?

1. Nationwide estimates are that it the total gap in LSAT performance is 9 points on average, and the inference is that an average of six points or so are accounted for by the stereotype threat (and the remainder by SES differences, motivational differences, etc).

Since the sterotype threat disproportionally affects those students who are most able and most motivated, I think we can infer that the gap will be larger than 6 points among those applying to higher-ranked schools than among those applying to lower-ranked schools. The motivation gap will, on the other hand, likely be smaller.

Would it reach 10 points? I don't know, but one could imagine that it very easily could. At the very least, we wouldn't be talking about a 10-point gap between LSAT scores, but a 3-point gap: something that happens frequently enough that it would be unremarkable. (It may be useful to remember that at the 95% confidence interval, the score band for a particular individual's LSAT score is 11 scaled points wide).

2. Stereotype threat affects GPA, but much much less than it affects LSAT. This is so because of the pervasive view that the LSAT is a measure of intelligence.

There may be other variables that affect GPA -- major, school, SES (having to work), hostile environment, etc. What is known, however, is that the schools that African Americans are more likely than average to attend have tougher grading curves than the schools that white law school applicants are likely to attend. HBCUs, for example, have  much tougher grading curves than the Ivies.

3. As for your third question, it's hard for me to make an intelligent comment on individual cases without looking at their files. What would be interestiing, though, would be to see if URMs with a higher index than the 3.1/163 were rejected from that school. That may point to some strengths in that individual's application that are invisible to us.

[WarrenG]

I’ll play devil’s advocate here even though I’m a supporter of AA.  First the definition of stereotype threat:

Stereotype threat – the threat, perceived by persons who are the target of stereotypes, that they will be evaluated in terms of these stereotypes (Steele, 1997).

As you can see it has 3 basic components: 1) the stereotype itself, 2) the person’s perception of how that stereotype relates to them in a given situation and, 3) a persons fear of being judged by that stereotype and the negative effect that this has on performance.

One could say that AA strengthens a more enduring stereotype threat: the incompetent minority.  The minority who has completed law school now has to worry about the stereotype threat that they are incompetent for the rest of their working lives, which should, in turn, make them less competent lawyers than they would be if AA, and the stereotype threat that it introduces, did not exist.

If the goal is the help minorities then the detrimental effect of this stereotype threat needs to be measured, amongst other things, against the benefits that AA brings.

[Quintana]

  Great post, thanks.

 As far as the stereotype threat affecting GPA, it seems that the mechanism (over-thinking) would still be in place, but the negative results would be less pronounced than they would be in a standardized test designed for quick thinking.  

 

[redemption]

One could say that AA creates a new, more enduring stereotype threat: the incompetent minority.  The minority who has completed law school now has to worry about the stereotype threat that they are incompetent for the rest of their working lives, which should, in turn, make them less competent lawyers than they would be if AA, and the stereotype threat that it introduces, did not exist.

Yes, I think that's a fair argument to make.

I think that AA is a second-best solution in response to law schools' failure to abandon the LSAT. The best solution is to revamp the admissions process entirely. Given, however, that they haven't and that they presumably won't, it is perverse and unfair to place the burden and the cost of this system on URM applicants' admissions prospects. It is, in fact, the very definition of unfair.

As a factual matter, it has been shown, using longitudinal survey data, that applicants who had been admitted under Affirmative Action at Michigan Law went on to lead careers that were as productive, as accomplished and as well renumerated as those of white students. The only significant difference was that URMs made a significantly better and greater contribution to the communities in which they lived.

The effects of the sterotype effect, therefore, seem to b potentially deadly in the applications process, in law school exams, but not thereafter.

[redemption]

  Great post, thanks.

 As far as the stereotype threat affecting GPA, it seems that the mechanism (over-thinking) would still be in place, but the negative results would be less pronounced than they would be in a standardized test designed for quick thinking. 

Yes. I think that it is entirely possible to design a test that eliminates the systematic gap caused by the stereotype threat. It seems that the psychometricians at LSAC test individual test items in the context of the test as a whole -- a process which will, of course, retain the systematic bias in outcomes, rather than work to reduce it.

[Quintana]

One could say that AA creates a new, more enduring stereotype threat: the incompetent minority.  The minority who has completed law school now has to worry about the stereotype threat that they are incompetent for the rest of their working lives, which should, in turn, make them less competent lawyers than they would be if AA, and the stereotype threat that it introduces, did not exist.

Yes, I think that's a fair argument to make.

I think that AA is a second-best solution in response to law schools' failure to abandon the LSAT. The best solution is to revamp the admissions process entirely. Given, however, that they haven't and that they presumably won't, it is perverse and unfair to place the burden and the cost of this system on URM applicants' admissions prospects. It is, in fact, the very definition of unfair.

As a factual matter, it has been shown, using longitudinal survey data, that applicants who had been admitted under Affirmative Action at Michigan Law went on to lead careers that were as productive, as accomplished and as well renumerated as those of white students. The only significant difference was that URMs made a significantly better and greater contribution to the communities in which they lived.

The effects of the sterotype effect, therefore, seem to b potentially deadly in the applications process, in law school exams, but not thereafter.

  Do you know of data concerning the comparative career success of AA admits from other schools?  If the results of the Michigan survey hold across the board, and I have no reason to think they wouldn't, then this speaks volumes towards the motivation to provide AA.  This of course begs the question: Does the end justify the means?  

[John Galt]

Red:

I'm saving this spot for a response in the near future. This topic requires more thought and detail than normal. Thank you for this thread.

[Hank Rearden]

If the LSAT measures ability to succeed during the first year of law school, and if blacks are adversely affected by a "stereotype threat" when taking the LSAT, is there any reason to think that they would not suffer this same sort of mental lapse during their first-year exams?  When would the stereotype threat expire?  Would they perform worse at law firms or on the BAR as well?  

I don't deny that blacks (especially high-achieving ones) may suffer because of stereotypes regarding their racial group's intelligence.  However, I don't see how the solution to this problem could reasonably involve artificial inflation of black test scores.  Such a measure would just further this "stereotype threat."