Thursday, November 1, 2018

Harvard Admissions - Part 11

Harvard’s Star Witness Testified All Day. Here Are 4 Moments That Mattered.

By Eric Hoover Oct. 31, 2018, Chronicle of Higher Education

A key witness for Harvard University spent five hours testifying here in federal court on Wednesday. For most of that time, David Card appeared calm and composed while discussing numerous data points and explaining various charts. But by mid-afternoon, he seemed uncomfortable, shifting around in his chair. His seat, after all, was getting hotter.

In his second day of testimony, Card, an economist at the University of California at Berkeley, described many findings from his statistical analyses of Harvard’s admissions process. Students for Fair Admissions, an organization that opposes affirmative action, has alleged that the university discriminates against Asian-American applicants.

Yet Card found no such thing. “The evidence,” he said, “does not support that claim.”

Seth P. Waxman, a lawyer representing Harvard, asked Card hundreds of questions. In answer after answer, the researcher rejected the findings of another economist whose conclusions underpin the plaintiff’s lawsuit against the university.

Card’s elaborate testimony, accompanied by colorful bar graphs and pie charts, just might represent Harvard’s best chance of winning this high-profile case. That’s why the soft-spoken economist has spent so much time on the witness stand — and why a lawyer for the plaintiff seemed intent on raining doubt upon him.

Here are four key moments from Wednesday’s testimony:

Card said his model better captures the reality of Harvard’s admissions process.

Elaborating on testimony he gave on Tuesday, Card challenged the findings of Peter S. Arcidiacono, an economist at Duke University. Arcidiacono, a key witness for the plaintiff, found evidence of bias against Asian-American applicants in Harvard’s admissions process.

Card’s main objection to Arcidiacono’s model is that it omits recruited athletes, the children of alumni, the children of Harvard faculty and staff members, and students on a special list that includes children of donors. Excluding all those applicants, who are accepted at a relatively high rate, Card suggested, had skewed his counterpart’s results.

Moreover, Card said, Arcidiacono’s analysis fails to account for several contextual variables that admissions officers consider when evaluating applicants. Those include their parents’ occupations and intended majors.

“Harvard is thinking about trying to get a set of students who will have lots of diversity,” he said. “Having a large fraction of students who all, for example, intended to pursue a career in medicine would not accomplish that goal.”

Then there was the personal rating, one of the four ratings that admissions officers use to assess applicants. That rating, Card explained, is meant to capture evidence of integrity and leadership skills, among other factors. Such factors, he said, can go a long way toward explaining why some qualified applicants get in and others don’t.

Personal ratings, Card said, aren’t “a mechanism” of discrimination.

Card rejected what is perhaps Arcidiacono’s most significant claim: that Harvard’s admissions officers, by assigning relatively low personal ratings to Asian-American students, are expressing a bias against them.

Though Asian-American applicants have stronger academic and extracurricular ratings than white students do, Card found that white applicants have stronger ratings in the other two nonacademic measures (personal and athletic qualities) and are more likely to be “multidimensional,” with top ratings in three of the four categories.*

Why, Card asked, would admissions officers assign high ratings to Asian-American students in some categories but not others? “I find it extremely hard or impossible to reconcile [his] claim that the personal rating is the mechanism by which discrimination against Asian-Americans is operating,” he said. “Like there’s some kind of schizophrenia going on here.”

Race matters, Card said, but in a more nuanced way than the plaintiff has alleged.

Yes, an applicant’s race matters in Harvard’s admissions process, Card said. But race alone won’t punch his or her ticket to Cambridge, Mass. For that variable to come into play, he explained, an applicant must be highly competitive to begin with.

Three-quarters of Harvard’s applicants are, as Card put it, “out of the money,” with no chance of getting in. Those in the top percentiles, who are “on the bubble,” he explained, excel in multiple dimensions of merit that the university considers.

Add race and — boom — the odds will increase for some applicants. But only if they are highly qualified in various ways.

For the most-competitive black students in the top two deciles, Card said, race can increase their chances of acceptance by 50 percentage points. For Hispanic applicants, there’s a similar but smaller effect.

That testimony echoed what admissions officials often say: A given attribute — such as race or geographic location — doesn’t really come into play unless a student already possesses many other desirable qualities.

“Race, in isolation, has almost no contribution,” Card said, to an applicant’s chances of admission.

One notable aside: Card said that there is no “tip,” or advantage, for Asian-American applicants in his model.

Later, Card rejected Students for Fair Admissions’ claim that Harvard engages in racial balancing. He discussed one chart showing how the racial breakdown of admitted applicants at Harvard varied from year to year. One year, for instance, the number of Asian-Americans dropped 11 percent. A year later, they rose by 17 percent.

Another chart showed a similar pattern among students who matriculated.

If Harvard officials were trying to balance classes by race, Card said, “they’re not doing a very good job.”

Card got an earful of tough questions.

Adam K. Mortara, a lawyer for Students for Fair Admissions, wasted no time in challenging Card’s findings. As soon as the cross-examination began, the mood in the courtroom changed.

Mortara was clipped and assertive. He challenged the economist’s finding that race wasn’t a major factor in admissions decisions. He said that there were “inaccuracies” in his slides. He contended that Card had mislabeled charts.

A trial over the race-conscious admissions policy of Harvard University could have lasting implications for selective colleges. Here is detailed background on the case and coverage of the trial as it unfolded, in a federal court in Boston.

Some of Mortara’s questions seemed to fluster Card, who paused several times before answering his questions. Once, after the economist gave a somewhat long-winded answer, Mortara asked, “Are you done?”

That prompted Waxman, one of the Harvard lawyers, to stand up and object. “It’s one thing for counsel to be asking questions,” he said. “Snide remarks are another.”

Judge Allison D. Burroughs of the Federal District Court told Mortara that Arcidiacono, the plaintiff’s witness, had also given some longer answers to questions.

The tense moment served as a reminder: The stakes in this case are high, with both sides girding for a long legal battle that could end up at the U.S. Supreme Court.

The plaintiff’s lawyers will continue to question Card on Thursday. And if the conclusion to Wednesday’s cross-examination is any indication, the gloves are most certainly off.

*Note: The more you "accept" Harvard's procedures as valid, the less likely it is that you will find statistical symptoms of discrimination. Harvard's procedures ultimately lead to an overall rating which determines whether you get it. If you start from the rating, i.e., if you accept everything that went into that rating as unbiased, then there will be no discrimination. The more you depart from that assumption, the more likely it is that you will find some discrimination. In the Harvard case, the personality ratings have been an issue. If you accept the personality ratings as unbiased (and then it just so happens that Asians get lower personality ratings for some unbiased reasons), then you won't find discrimination.

No comments: