I admire many things about Judge Sonia Sotomayor, especially her deep compassion for underprivileged people. I may well support her confirmation to the Supreme Court if her testimony next month dispels my concern that her decisions may be biased by the grievance-focused mind-set and the "wise Latina woman" superiority complex displayed in some of her speeches.

But close study of her most famous case only enhances my concern. That’s the 2008 decision in which a panel composed of Sotomayor and two Appeals Court colleagues upheld New Haven’s race-based denial of promotions to white (and two Hispanic) fire-fighters because too few African-Americans had done well on the qualifying exams.

The panel’s decision to adopt as its own U.S. District Judge Janet Arterton’s opinion in the case looks much less defensible up close than it does in most media accounts. One reason is that the detailed factual record strongly suggests that — contrary to Sotomayor’s position — the Connecticut city’s decision to kill the promotions was driven less by its purported legal concerns than by raw racial politics.

Whichever way the Supreme Court rules in the case later this month, I will be surprised if a single justice explicitly approves the specific, quota-friendly logic of the Sotomayor-endorsed Arterton opinion.

Judge Jose Cabranes, Sotomayor’s onetime mentor, accurately described the implication of this logic in his dissent from a 7-6 vote in which the full U.S. Court of Appeals for the 2nd Circuit refused to reconsider the panel’s ruling.

"Municipal employers could reject the results of an employment examination whenever those results failed to yield a desired racial outcome — i.e., failed to satisfy a racial quota," Cabranes wrote.

The Sotomayor-endorsed position allowed such a "race-based employment decision," Cabranes added, even though the New Haven exams were "carefully constructed to ensure race-neutrality" and even though the city had neither found nor tried to find a more job-related test.

The Cabranes dissent and the voluminous factual record that was before the Sotomayor panel flatly contradict the widely stated view that her position was justified by evidence that the exams were not job-related and that they discriminated against blacks in violation of the "disparate-impact" provisions of federal civil-rights law.

In fact, neither Sotomayor nor any other judge has ever found that the exams — one for would-be fire lieutenants, one for would-be captains — were invalid or unfair. Nor has any judge found that allowing the promotions would have violated disparate-impact law.

Rather, the Sotomayor-endorsed position was that under 2nd Circuit precedents, New Haven’s discrimination against high-scoring whites must be upheld based solely on the fact that disproportionate numbers of blacks had failed to qualify for promotion and might file a disparate-impact lawsuit — regardless of whether they could win it.

The decision to kill the promotions was driven less by purported legal concerns than by raw racial politics.

Indeed, the evidence shows that the promotions would have been lawful and that any disparate-impact suit by blacks was doomed to fail.

Disparate-impact law — as codified by Congress in 1991 — specifies that an employer whose qualifying exam or other selection criterion produces a racially disparate impact can be held liable for unintentional discrimination only if (1) the test is not "job-related … and consistent with business necessity," or (2) the employer is presented with and refuses to adopt another, similarly job-related test with less disparate impact.

But the unmistakable logic of Sotomayor’s position would encourage employers to discriminate against high-scoring groups based on race — no matter how valid and lawful the qualifying test — in any case in which disproportionate numbers of protected minorities have low scores, as is the norm.

Such logic would convert disparate-impact law into an engine of overt discrimination against high-scoring groups across the country and allow racial politics and racial quotas to masquerade as voluntary compliance with the law.

Speaking of racial politics, even the Sotomayor-endorsed Arterton opinion found that there was evidence from which a jury could infer that, as the plaintiffs contended, "city officials worked behind the scenes to sabotage the promotional examinations because they knew that, were the exams certified, the mayor would incur the wrath of … influential leaders of New Haven’s African-American community."

The victims of the city’s discrimination included lead plaintiff Frank Ricci. Like other plaintiffs, he studied for months, for as many as 13 hours a day, in 2003 to prepare for the combined written and oral exam that he hoped would win him a promotion. He also spent more than $1,000 buying the books that the city had suggested as homework and paying to have them read onto audiotapes. (Ricci is dyslexic.) And he got one of the highest scores.

I sketch below some of the evidence belying the Sotomayor panel’s assertion in its own strangely sketchy opinion that the city "was simply trying to fulfill its obligations" under disparate-impact law when it blocked the promotions. This and other evidence show that the exams were fair — although not perfect (no exam ever was or ever will be) — and that the city’s decision was driven by racial politics, not by any desire to comply with the law.

• Connecticut and New Haven law require objective, job-related tests as a basis for promotions to counteract the race-tinged abuse of more-subjective promotion criteria in which — state courts have repeatedly found — the city has often engaged in order to facilitate political patronage, cronyism, and nepotism.

• Industrial/Organizational Solutions (IOS), the professional testing firm from which New Haven bought the two tests (which were 60 percent written and 40 percent oral) to measure the knowledge, skills, and abilities of would-be fire lieutenants and captains, based the exams on copious research to ensure that the questions were job-related and administered them in a scrupulously race-neutral way. The nine interviewing teams that actively administered the oral portion of the tests each had one white, one black, and one Hispanic member.

• Contrary to reports that no blacks passed, nine did. Their pass rate was about half that of the white test-takers, a common phenomenon in New Haven and elsewhere. Although no black candidate scored high enough to qualify for immediate promotion, three would have been eligible when, it turned out, new vacancies opened up over the next two or three years.

The mayor said he would block promotions even if the city’s civil service board approved the tests as valid and fair.

• Politically powerful African-Americans led by the Rev. Boise Kimber demanded loudly that no high-scoring whites or Hispanics be promoted. He has long been a key vote-getter for the city’s (white) mayor, John DeStefano.

• The mayor told aides that he would block promotions of the high scorers even if the city’s civil service board approved the tests as valid and fair.

• Kimber warned civil service board members of a "political ramification" if they approved the test results.

• The city presented three expert witnesses at the hearings. The only one who had read the exams closely, a firefighting and homeland-security expert named Vincent Lewis, who is African-American, found them to be valid and fair. The two others — Christopher Hornick, an IOS competitor, and Janet Helms, a professor with no public safety expertise — admitted they had not studied the tests, while vaguely suggesting that blacks might do better under some other selection process that wasn’t clearly described. But even Hornick recommended that the city proceed with promotions of the high scorers.

• The New Haven fire chief and the assistant chief, who is African-American, were involved in developing the tests and (the evidence suggests) believed them to be valid. But the city declined to call them as witnesses.

• IOS strenuously defended its tests and was ready to provide more-detailed written evidence of their validity. But the city — determined to block the promotions — refused to seek such information.

• Amid a political circus, the civil service board deadlocked 2-2, with two members expressing doubts about the exams. This killed the promotions.

• Neither the city nor anyone else identified any similarly job-related alternative test on which black firefighters might do better.

• When the high scorers sued, the city made no serious effort in the District or Appeals court to impugn the validity or fairness of the tests.

• Before the Supreme Court, the city’s new attorneys switched strategies and strained to make a silk purse out of their sow’s-ear case by coming up with a barrage of criticisms of the exams. But they shrivel on inspection. For example, the claim in the city’s brief that the tests were flawed by "irrelevant or contradictory" questions neglected to mention that black firefighters had faulted only two of the 200 written questions (100 on the lieutenant’s test, 100 on the captain’s test) — both of which IOS and the high scorers vigorously defended. The city’s attorneys also criticized at length aspects of the exams that clearly had no effect on which firefighters qualified for promotion.

If such belated, weak, and speculative criticisms — obviously tailored to impugn the outcome of the tests — are sufficient to disprove an exam’s validity or fairness, no test will ever withstand a disparate-impact lawsuit. That may or may not be Sotomayor’s objective. But it cannot be the law.

This article appeared in the Saturday, June 13, 2009 edition of National Journal.