Sociological Gobbledygook or Scientific Standard? Why Judging Gerrymandering is Hard

, Kendall Science Fellow | October 4, 2017, 3:55 pm EST
Bookmark and Share

In Tuesday’s historic Supreme Court case, the question asked was how to identify and remedy unconstitutional partisan gerrymandering, where electoral district boundaries are drawn so as to benefit one political party’s voters over others.  The phrase uttered during oral argument that is getting the most attention is Chief Justice Roberts’ assessment of the various techniques that have been proposed to measure it: “sociological gobbledygook.”  It’s a funny way to describe Roberts’ apparent distaste for mathematical, as opposed to legal, explanations, but it also reveals a serious problem for the use of scientific evidence in the court.

Let’s look at the evidence.

One of the core issues in these cases, as I’ve previously discussed, involves the discovery of “workable standards.”  To be workable, a standard must identify a constitutional (fundamental) harm, as opposed to a de minimus (minor) harm, so as not to inundate the court with cases.  Further, the standard must be capable of being practically applied by justices who are not themselves scientists.

Whether or not tests for the standard of partisan symmetry, the equal treatment of voters regardless of which party they support, are workable, was the primary point of contention when Justice Roberts made his remark.

In describing his concern about judicial overreach into the political process, Roberts proclaimed that “you’re taking these issues away from democracy and you’re throwing them into the courts pursuant to, and it may be simply my educational background, but I can only describe as sociological gobbledygook.”

On the one hand, Roberts is identifying a serious problem that needs to be addressed by scientists in the courtroom.  Statistics can be manipulated and are open to interpretation in ways that other forms of legal evidence are often not.

In many cases, both parties trot out potentially motivated “experts” to exchange criticisms in specialized language, leaving judges to make decisions based on evidence that their educational background does not train them for.  Consider two examples taken directly from yesterday’s argument.

The term “false positives” was used by the defense (the state of Wisconsin) to refer to the inaccuracy of one way to measure symmetry, the efficiency gap.  “False positive” refers to a Type I error, when the test for something (like pregnancy, using a urine test that measures levels of the hormone chorionic gonadotropin) turns up positive, but has not actually occurred (no fertilized egg embedded in the uterus, which produces the hormone).  Pregnancy tests have about a 3% false positive rate.  But back to gerrymandering.

In this case, the claim of “false positive” was misapplied, and expanded to describe any state with a significant efficiency gap, where the plan was not drawn by the state legislature.  That is, the defense implied that districting plans not drawn by parties (those drawn by courts through litigation or by commissions, etc.) could not be biased.  But the efficiency gap is not a test of who draws a districting map, it is a measure of bias.

Even randomly drawn maps using computer simulations can result in quite biased plans, depending on the underlying geographic distribution of voters.  None of the justices seemed to pick this up.  Justice Alito, responding to such claims, expressed grave concern about “the dozens of uncertainties about this whole process.”

Worse still was Chief Justice Roberts’ mistaking of symmetry for “proportional representation, which has never been accepted as a political principle in the history of this country.”

Partisan symmetry is explicitly not a test of proportionality in election results (where a party receives the same percentage of seats as its percentage of votes).  In fact, symmetry was intentionally designed as an alternative standard of testing the principle of political equality in U.S. elections, because proportionality is a higher standard than what the Constitution demands.

These mistakes might have been avoided through a more thorough reading of the many scientific briefs offered to the court for review (or the video above).  Nevertheless, the burden is on scientists to communicate our work clearly and concisely to non-experts, otherwise this problem will only persist.

On the other hand, several of the Justices had a strong grasp of how scientific standards operate within the voting rights framework.  Justice Kagan, for example, correctly noted that both partisan symmetry and the one-person, one-vote standard (prohibiting unequally populated districts) address the dilution of voting strength for individual voters as a function of statewide plans, not single electoral districts.

Moreover, Justice Sotomayor, responding to the defense’s claims about inaccuracies in estimating the impact of Wisconsin’s plan, pointed out that “every single social science metric points in the same direction.”  That is the sort of understanding about probabilistic estimates that scientists need to convey to judicial authorities.  It is how scientists forecast everything from economic growth to health epidemics and weather patterns.  The Justice continued, noting that the same types of statistical estimates were used to create Wisconsin’s maps in the first place, and that “it worked.  It worked better than they even expected, so the estimate wasn’t wrong.  It was pretty right.”

Judges have their work cut out for them if the Supreme Court finally provides a means by which political parties can be restrained from advancing their partisan interests at the expense of voters’ fundamental right to an equally weighted vote.  But it is up to the scientific community to work with the judiciary in the appropriate application of statistical evidence.  The consequences, which feedback through the entire policy making process, make it well worth the effort.

Posted in: Science and Democracy

Support from UCS members make work like this possible. Will you join us? Help UCS advance independent science for a healthy environment and a safer world.

Show Comments


Comment Policy

UCS welcomes comments that foster civil conversation and debate. To help maintain a healthy, respectful discussion, please focus comments on the issues, topics, and facts at hand, and refrain from personal attacks. Posts that are commercial, self-promotional, obscene, rude, or disruptive will be removed.

Please note that comments are open for two weeks following each blog post. UCS respects your privacy and will not display, lend, or sell your email address for any reason.

  • Brett_Bellmore

    From what I can see, the efficiency gap actually IS a test of proportional representation. It’s just done in a way that mathematically obscures what’s going on.

    You can prove me wrong, of course, by constructing a situation where the efficiency gap is zero and you don’t have proportional representation…Or where it’s high and you do have proportional representation.

    What it isn’t, is a test of gerrymandering. Because gerrymandering isn’t defined as “unfair” districts, it’s defined as districts (Maybe shaped like a salamander?) designed to produce a particular political outcome. Like, for instance, a low efficiency gap.

    The opposite of gerrymandering isn’t “fair” districts, it’s districts drawn without regard to political effect.

    • Michael Latner

      Hi Brett,
      Thanks for your comment. Two points, first the definition of partisan (as opposed to racial, or other forms) gerrymandering is designing districts that advantage one party over another, a very specific political outcome (as every particular outcome is political). Drawing districts “without regard” to political effect can produce very biased plans, typically because if compactness is used as a criteria they pack Democrats into dense urban districts.
      Second, the efficiency standard achieves maximum efficiency if a 1% increase in vote share produces a 2% increase in seat share above 50% of the vote, whereas proportionality is achieved when vote share equals seat share. An example of a fairly proportional assembly with a high efficiency gap is the Kentucky Senate. See the figure in this FairVote post, a pro PR group: http://www.fairvote.org/evaluating_the_efficiency_gap_as_a_measure_of_proportionality_and_competition