What Happens When the Supreme Court Gets the Facts Wrong?

For every ruling made by the U.S. Supreme Court, some will think the justices made the wrong decision. But a recent report shows that the court also gets the facts wrong in an alarming number of cases.

ProPublica, a nonprofit organization that produces investigative journalism, examined a sample of Supreme Court cases from 2011 to 2015.

In the 24 cases that relied on real-world facts, the group found seven errors. Some mistakes were arguably minor, but others concerned a central aspect of the case.

In the 2013 case of Shelby County v. Holder, the Supreme Court ruled that Section 4(b) of the Voting Rights Act of 1965 was unconstitutional. Section 4(b) contains the formula that determined which states needed preclearance before changing their voting laws or practices, and the Supreme Court, in a 5-4 decision, determined that section was unconstitutional because the formula was based on decades-old data.

Chief Justice John Roberts, writing for the majority, published a chart in the body of the opinion showing the gap between black and white voter registration levels in the six states covered by the Voting Rights Act. The chart was meant to illustrate the undeniable fact that black voter registration surged between 1965 and 2004. But some of the numbers were wrong.

ProPublica discovered that Roberts relied on a 2006 report generated by staff members of the Senate Judiciary Committee. Those staffers used a reliable source, the U.S. Census Bureau’s 2004 post-election survey, but they used the wrong numbers, perhaps due to unfamiliarity with how the census bureau handles race and ethnicity.

The staffers mistakenly pulled numbers from the “white alone” category, which includes Hispanics, rather than “white non-Hispanic alone,” which would accurately show the registration levels for white voters.

As a result of the error, the chief justice’s chart showed that black voter registration in Georgia in 2006, which had reached 64.2 percent, exceeded the white voter registration level of 63.5 percent. This was incorrect, as the actual white voter registration level was 68 percent. Roberts’ chart also showed that white voter registration in Virginia in 2004 exceeded black voter registration by only 10 percent, when the actual gap was 14.2 percent.

In the 2011 case NASA v. Nelson, the Supreme Court unanimously held that independent contractors working for the federal government could be asked about their private lives in background checks, including questions about drug use. Justice Alito, writing for the unified court, wrote that such questions are “part of a standard employment background check of the sort used by millions of private employers.”

Roberts cited an amicus brief submitted by background check industry groups that claimed “more than 88% of U.S. companies … perform background checks on their employees.”

When ProPublica questioned the National Association of Professional Background Screeners about the source of its “more than 88%” figure, the association was not able to provide one, and ProPublica found that the association featured the 88 percent figure in lobbying materials months before the survey that supposedly provided the data was even conducted. ProPublica warns that its sample was too small to draw any conclusions about the rate of factual errors in Supreme Court decisions, but it is troubling that significant errors were uncovered even in such a small sample.

The errors are not really surprising, as neither the justices nor their clerks usually have any background in statistics, and there is not an established research service or group of fact checkers to assist them. The court often relies on amicus briefs that cite factual assertions, some of which are unsubstantiated. The justices sometimes acknowledge their own lack of expertise. During oral arguments in a recent gerrymandering case, Chief Justice Roberts referred to some of the evidence being presented as “sociological gobbledygook.”

Legal expert Kenneth Culp Davis proposed a solution to the problem in the 1980s. Davis said that the Supreme Court should have its own research organization similar to the Congressional Research Service, to perform needed research and analysis. The idea did not take hold at the time.

Of course, the justices could choose to employ clerks with training in statistics instead of just law degrees, or hire experts with such training. But some solution must be implemented. The justices of the nation’s highest court, deciding cases of the utmost importance, need to take seriously the need for accurate empirical facts in their decisions.