When Congress reauthorized the law in 2006, one amendment proposed to use data from more recent elections. It failed by a wide margin.

While the places with low turnout now are very different from those with low turnout in the 1960s and 1970s, a trial court ruling in the case considered by the Supreme Court, Shelby County v. Holder, suggested the more recent data was irrelevant.

The years used as triggers, Judge John D. Bates ruled, “were never selected because of something special that occurred in those years; instead, they were chosen as mere proxies for identifying those jurisdictions with established histories of discriminating against racial and language minority voters.”

During oral arguments, Chief Justice John G. Roberts Jr. noted that Massachusetts, a state that is not covered, had the greatest disparity in registration between whites and blacks, referring to data from 2004.

While that data comes with a high margin of error, five states — none covered — had registration gaps of at least five percentage points in all of the last three elections.

Some members of the court dismiss these sorts of statistics. According to Justice Ruth Bader Ginsburg, “Congress said up front: We know that the registration is fine. That is no longer the problem. But the discrimination continues in other forms.”

Lawsuits brought under Section 2 of the Voting Rights Act, which applies to the entire nation, offer one way to compare discrimination across the country.

A study of cases that were published on two legal databases, Westlaw or Lexis, between 1982 and 2005 found that jurisdictions covered by Section 5 lost more such cases than the rest of the country, even though being covered might seem to deter discrimination.

Judge Stephen F. Williams, who dissented from the federal appeals court decision reviewed by the Supreme Court, considered this data in his opinion. Based on it, he found that “a more narrowly tailored coverage formula — capturing only Mississippi, Alabama, and Louisiana, and possibly the covered portions of South Dakota and North Carolina — might be defensible.” But, he continued, “beyond these, the covered jurisdictions appear indistinguishable from their uncovered peers.”

Some voting discrimination suits, including those with a court-approved settlement, are resolved without a published opinion.

Judge David S. Tatel, writing for the majority in the appeals court's decision, found that “the difference between covered and non-covered jurisdictions becomes even more pronounced,” when including cases settled in favor of minority voters.

Critics of this metric, as well as one that includes only published cases, argue that adjusting for population is unfair to small states and that patterns have changed since the data was assembled.

Law professors at the University of California, Davis, and the University of Connecticut have created an estimate of prejudice using survey data from 2008.

The National Annenberg Election Survey asked people to rank the intelligence, trustworthiness and work effort of different groups of people, on a scale from 0 to 100. Christopher S. Elmendorf and Douglas M. Spencer estimated prejudice based on how people rate their own ethnic group, compared with how they rate blacks.

“It may be argued that Virginia and perhaps South Carolina should not be covered, but the rest of the covered states in the Deep South top the list of the most prejudiced states by anti-black stereotyping,” the professors found.