What was the question?

Joseph Fishkin

Cary Franklin

Numbers. Boy, do journalists in the post-538 era love numbers. They are objective! They can generate interactive graphics! Best of all they give us answers—the true objective fact-based answers that mere narratives (and, you know, actual reporting) can so often obscure.

On the top of the NYTimes.com front page this morning, we see a beautifully rendered, interactive red and blue time-slice chart describing “The Roberts Court’s Surprising Move Leftward.”  The numbers are unimpeachable—who can argue with numbers? Over 50% of this term’s decisions, as we head into the final days, have been “liberal.” The authors of the post, who include the Times’ Supreme Court correspondent Adam Liptak (although he’s not the first author listed, that’s Alicia Parlapiano) add breathlessly: “If that trend holds, the final percentage could rival the highest since the era of the notably liberal court of the 1950s and 1960s led by Chief Justice Earl Warren.”

Wow! The most liberal term since the Warren Court! Now that is news! Especially after Justice Alito replaced Justice O’Connor, causing a deep and fundamental shift to the right on a broad range of issues, from race to abortion, how is it even possible to have such a liberal Supreme Court?

It is not.

This is the point in this blog post where one of us in particular should admit up front a love for numbers and graphics: I read Nate Silver and I read The Upshot, the Times’ post-Silver answer to Silver; I harbor an unabashedly nerdy love for the work of Edward Tufte; indeed I am the type of reader who often clicks on the graphs and infographics before reading the text of the article, a tendency the other of us often makes fun of, and not undeservedly.  But all this is just to say, I am (and indeed both of us are) hardly a hostile audience for this kind of stuff.  We're a very friendly audience.  But come on, people. This is the kind of “story” that gives quantitative reporting a bad name.

Imagine a tale of two Supreme Court terms. In the first, at Time One, the Court decides (a) yes, affirmative action is constitutionally required, (b) yes, the death penalty is unconstitutional, and (c) no, the government is not obligated to pay compensation to the descendants of former slaves. Two liberal decisions, one conservative decision, for a liberal rate of 66.66667%. In a term some decades later, at Time Two, the Court (having by the way overturned the first two decisions from Time One long ago) decides the following: (a) no, affirmative action is not even constitutionally permitted, but is prohibited by the Fourteenth Amendment (b) yes, the death penalty by the method of burning at the stake raises some Eighth Amendment problems, and (c) yes, despite the rulings of some lower courts, currently enslaved people can indeed bring claims under the Thirteenth Amendment in narrow circumstances. Between Time One and Time Two, the Court moved far to the right. Yet that liberal percentage remains precisely 66.66667%. How odd and how newsworthy!

The difference is not the answers. The difference is the questions.


The Court generally takes cases that raise hard questions. Cases with easy questions, as a general rule, were decided correctly below; taking them often seems pointless. Hard cases generate circuit splits and deciding them will give useful guidance to lower courts. But—what’s an easy case versus what’s a hard case depends on where the Court is. As the Court moves right, and lays down new precedents that move right, the boundaries of what any responsible circuit judge will hold also move right. Eventually, cases that would have seemed cuckoo-land crazy bits of right-wing fantasy in the Warren Court era not only must be taken seriously, but will likely wind up divided 5-4, whichever way they come out, and even if the end split looks about 50-50 “liberal”-“conservative.”

That is why Justice Stevens, who was a Republican appointee in the middle of the Court when appointed, could say truthfully in the Parents Involved case that “no Member of the Court that I joined in 1975 would have agreed with today’s decision,” by a five-justice majority, to strike down voluntary school integration.  He was right.

Or take the Inclusive Communities Project case, about disparate impact claims under the Fair Housing Act, to be decided probably later this week. To the Ed Meese Justice Department, the idea of getting rid of disparate impact claims under the FHA was a kind of pie-in-the-sky aspiration. Now that’s either about to be the law, or will miss it very narrowly. Either way, the fact that we are asking that question at all is how you can tell how enormously far to the right the Court as a whole has moved.  (Not on every single issue, of course.  On gay rights, the Court has followed the country to the left.)

To say that the Court is roughly 50-50 in “liberal” vs “conservative” outcomes says almost nothing about where the Court is, overall, on the spectrum of high politics. It is like trying to determine where you are on a map by asking what directions your most recent ten steps pointed.  It's not quite a non sequitur, but it's close.

The raw percentage might say something more interesting, which the authors of this Upshot post do not explore at all, about the relationship between the Supreme Court and the lower courts.  If the Supreme Court is coming out “liberal” or "conservative" a high percentage of the time, as it did in the Warren Court era on the "liberal" side, this might mean that some of the Circuit Courts are generating a lot of decisions the other way that the Supreme Court Justices feel the need to take and reverse. In other words, it suggests a period in which the Supreme Court and the Circuit Courts might be a bit out of sync.

But ordinarily, since presidents nominate the judges to all the federal courts, one would not expect the “liberal” versus “conservative” raw outcome numbers to move much, whether the Court turns into a court with a median justice of John Roberts or one with a median justice of Elena Kagan. As long as lower court nominations and Supreme Court nominations are very roughly in sync, the raw outcomes will probably hover around 50-50-ish. To tell how liberal or conservative a court’s decisions actually are, one needs to do some actual reporting, and produce a narrative.

The Times is generally pretty good at that.  Time to get to work.
Posted 5:22 PM by Joseph Fishkin [link]