Thread by @AnechoicMedia_

District Court of Massachusetts in SFFA v. Harvard:

~"diversity is axiomatically good and we don't need to prove this"~

All the usual vague appeals to the goodness of the diverse human soul or whatever lie within. These are either unproven, spiritual assertions, or misinterpretations of actual social science.

Fact check: diversity is consistently correlated with less cohesion, not more.

Even research which claims this is not so, does so in the context of group members already being matched on other merits -- so if you lower standards to achieve diversity, you're just worse, period.

"[Plaintiffs] do not contest the importance of diversity"

This is why the right lost this fight before it began.

They're not even attempting to counter the foundational, "axiomatic" assumption of the virtues of racial discrimination in American society.

~"Harvard investigated itself and concluded that racial diversity is great"~

Don't worry, Rakesh Khurana is on the case -- you can count on him to be unbiased!

Harvard: "every applicant evaluated as a unique individual" with "personality, curiosity, character ..."

Reality: The content of their character takes a backseat to skin color. The estimated boost to merely being black is larger than being on the Dean's List.

~"we don't have any racial animus against you, we just want fewer people who look like you here, and penalize you for it"~

This is an official document issued by an American court.

Harvard has an official procedure in which race is used to "lop off" otherwise admitted, fully qualified white/asian candidates to reach the desired class size.

Typical American doublespeak:

~"we don't have any race quotas, but our racial proportions are the same every year, and we bias admittance to keep it that way"

Having actual, publicly-stated race quotas would be preferable to our bureaucratic dystopia of unspoken discrimination.

This is an outrageous lie:

"race is only intentionally used as a positive attribute"

It's a literal zero-sum game - awarding bonus points to favored groups is assigning a penalty to the non-favored groups (whites, asians).

Here is the same doublespeak again: ~"we don't discriminate"~ even though all the math says they do.

Liberals say this because they assume that "discrimination" is something mean people do, and since we're goodthinking liberals achieving diversity, we can't be discriminating!

More doublespeak. These are sentences that are literally *impossible by construction*.

Can you imagine any white person getting away with employment discrimination, by saying "I'm not penalizing blacks; I'm rewarding whites! Totally different!"

The court handwaves away regression models showing persistent racial disparities on the merits by vague reference to SES/family confounds.

Dean Khurana is "puzzled" by the suggestion that Harvard admissions should even be subject to regression models in the first place.

Here, the background of massive anti-white discrimination clouds the waters, in a way that clearly confuses the judge.

The asian coefficient is "relatively modest" because it's dummy-coded *relative to whites*, who are already at a massive disadvantage relative to favored races.

A glaring problem with this suit is its framing of the discrimination as being against asians, relative to whites, when the largest divide is really "whites + asians" vs everyone else.

By ignoring the largest racial gap, the smaller extra penalty for Asians can be quibbled over.

The decision contains absurd characterizations of the data, like implying a "modest" regression coefficient on asians does not suggest discrimination.

By my calculations, had Asians been admitted at the rate of white comparables, their numbers would increase 57%! "Modest?!"

This is weasel language. The "asian coefficient" is "modest" only in comparison to the far larger privileges awarded to other applicants relative to the white baseline.

If you want to say the asian penalty is "modest" in comparison to the black or legacy bonuses, uh, sure!

The opinion's description of the Arcidiacono evidence doesn't align with his recently-published paper.

Including ALDCs, interaction terms, and personal ratings fails to eliminate the asian bias. This was all in table B.7.2R of trial document 415-9!

This case may seriously have been lost due to the court's strained/errant interpretation of regression models.

Card probably massaged the defense's asian coefficient into non-significance, in part by limiting statistical power by analyzing each admissions year separately.

The Court finds it "persuasive" evidence that Harvard swears it sees no bias in the assignment of personal ratings ... because these officers are "considerate, diligent, intelligent" people who would surely notice their own bias and report it if present!

Are you kidding me?!

~"I mean they all say they're not racially biased, and if there's one thing we can count on people with racial bias to do, it's reliably confess this bias before a court of law, while testifying for the defense!"~

Well, I'm certainly convinced!

~"decent regression model, but does it control for the romantic appeal of recounting triumph over adversity? Didn't think so; can't say this evidence is persuasive."~

This is definitely what science looks like; Justice certainly being done here, yep.

~"It's not discrimination if you outsource racial bias to third-party reviewers!"~

Disclaimer: do not attempt this if you are not Harvard

Contrast the moral separation displayed above with the dominant disparate impact standard applied when judging the use of allegedly biased credentials in employment.

In any other context, if you give weight to a biased credential, the locus of discrimination is still on *you*!

~"if you control for the appliants' parents running dry cleaners and noodle shops, bias against asians goes away"~

Why is this relevant? This is just laundering the bias being alleged!

If Harvard gives extra weight to non-asian parental occupations, that's still discrimination!

Another absurd statement - The court says without qualification that Harvard makes no attempt to achieve any particular racial composition.

This is laughably false and contradicted by previous descriptions that racial targets are vigilantly policed at every step in the process.

Harvard admits to racially selective "lopping off" to meet targets, to disparate racial standards in seeking applicants, and to guiding applicant ratings in the context of class racial composition.

What more does there need to be before we can call it a quota?

Apparently, all you need to do to elude getting caught for "impermissible racial balancing" (aka quotas) is add a bit of year-to-year jitter in your otherwise constant racial proportions.

This is an absurd argument; This pattern is completely consistent with discrimination.

Finally, a statement I wholly agree with:

There is no way for Harvard to maintain its current racial proportions without engaging in racial discrimination. No race-neutral policy can justify the status quo.

Access to elite institutions in the current year: Gotta have a "compelling narrative" that you can tie to your identity or you're just not worth it.

It's all just storytelling. Failing that, be a donor.

The court concedes eliminating racial preferences would cut black/hispanic admittance in half. By my estimates, it's probably even more than that.

Here, the Court/Harvard can define the bar for who is "qualified" as low as is needed for sufficient minority representation, then claim that any race preferences applied within this space is merely deciding between facially similarly "qualified" individuals.

Isn't that devious?

All of this wordplay and backhanded tricks to avoid the Bakke prohibition on having categorically reserved slots by race.

What is the moral distinction, when the end result is an implicit quota of slots that whites/asians will never have access to under any "diversity" goal?

Is there any scenario under which Harvard would tolerate 100% of its admits being white and asian?

No? Then you have a quota. It might not be one you wrote down anywhere, but like any biased process, everyone involved is tuned in to what the correct outcome is to work toward.

America's attempts to tap-dance around explicit racial preferences are repugnant and obtuse.

There is no moral difference between awarding direct extra points to favored groups, vs applying that same average bonus through reviewer ratings that explicitly incorporate race.

Racial discrimination is constitutionally foul-smelling, so SCOTUS ostensibly subjects it to "strict scrutiny" and limits it to the "compelling state interest" of securing the benefits of a diverse student body ...

... but what are those benefits, and how are they measured? This is defined nowhere!

In Fisher II, SCOTUS leaves the definition of the benefits of diversity ... up to the universities! Because they're the experts, and the court assumes their "good faith".

So, to sum up:

- The university is not given deference in deciding whether their use of race was necessary to achieve their diversity goals

BUT

- They *are* given deference in deciding for themselves what the diversity goals are, and what their benefits are.

Neat, huh?

For an idea of how bonkers this is:

Imagine if the government strictly monitored whether your factory had appropriate emissions scrubbers to achieve your emissions limits ... but they left it up to you to decide what those limits were!

Does that sound like oversight to you?

So when a university says they need diversity to "cultivate global leaders" and "destroy racial stereotypes" for their educational mission, that sticky ball of vague pablum never gets subjected to any legal tests, or quantified, or peer-reviewed, or approved by any regulators.

"Broadening perspectives" and "immersion in a diverse community" are exemplars of an "elusory or amorphous" policy goal.

If the court thinks such mission-statement tripe is sufficiently "concrete and precise" a target within strict scrutiny, then what the hell isn't?

Notice that, in both the Harvard and UT cases, the college was permitted to invoke without question the consequences of racial discrimination - "increasingly diverse society", "cultivating legitimate leaders" - as justification for continuing state-backed diversity enforcement!

So:
- elite institutions push for a diverse society
- institutions then point to our more diverse society as something they must train and propagandize people for
- the state permits institutions to keep up discrimination in furtherance of this goal!

What a mockery of justice!

The Court permits universities to define for themselves what they think "legitimate leaders" for our country look like, call that their "educational mission", then blesses their use of racial discrimination to achieve this mission!

A coup by any name!

The court makes clear it's basically impossible to trigger the Bakke quota prohibition.

Harvard has racial "dashboard" metrics that are closely watched throughout the admission process, but as long as they never get caught denying any one person exclusively on race, they pass.

Bakke's prohibition on "mechanical" bias is similarly evaded, by applying bias imprecisely and at different stages.

The racial boosts are as large as bonus given to a heavy donor or legacy student. If such a bonus isn't a "defining feature" for an application, what ever can be?

More obfuscatory language about the "modest" preference, and how every group of admits is of similar potential.

The difference in potential at +2σ, vs +2.5 to +3σ, is quite large, and is the kind of bar-lowering we're talking about.

Almost all blacks admitted to Harvard are substantially above-average, and would succeed at most American universities.

But they're not geniuses, and they're not on par with their non-black classmates, on a scale that matters at the ultra-elite colleges.

In the end, what else to expect? Plaintiffs never contested diversity goals, only the implementation, which we have seen are impossible to nail any school down on.

By splitting students into many fuzzy metrics, none can ever discretely claim to have been denied their just place.

By electing to use asians as the vehicle for their suit, rather than the whites they were aiming to help implicitly, plaintiffs ensured this specific defeat.

Refusing to defend white Americans, right-wing donors exhausted two years and millions of dollars on a futile gesture.

Victory on the merits was unlikely; challenging the principle more so.

But by ignoring the overwhelming racial disparity faced by whites, to instead focus on the comparatively small asian gap, the plaintiffs enabled the defense to play games to deny any such gap existed at all.

"Show me a single person who was wrongly denied entry" is a tricky, multivariate line-drawing exercise to begin with -- but it's a challenge that the defense could plausibly sustain in front of this judge for asians with respect to whites.

For whites? Not a chance.