STEM-ing the Tide of Scientific and Mathematical Illiteracy in the Law

Attorneys need to understand how numbers work. It’s time we teach them.

[Editor’s note: In recent weeks, the Martin Center has published articles on the intersection of science, technology, and higher-ed reform. Please click here to read David Randall on the importance of digital archiving and here for J. Scott Turner on federal research oversight.]

In 1897, before becoming a justice of the Supreme Court, Oliver Wendell Holmes wrote, “For the rational study of the law … the man of the future is the man of statistics and the master of economics” (“The Path of the Law,” Harvard Law Review). Although Holmes’s prediction was blind to the entry of women into the practice of law, he accurately foresaw the intrusion of technical, scientific, and statistical issues into the legal realm, whether concerning legislation, regulation, or litigation.

In 1959, C.P. Snow, in his book The Two Cultures, described universities as polarized and isolated camps of the humanities and the sciences. Traditionally, law had been a field for students of literary abilities who received their undergraduate training in the humanities, often in history or English literature. To this very day, admission to law school in the United States is largely based upon grade point average and performance on a standardized exam, the Law School Admission Test, which assesses applicants’ ability for critical reasoning, reading comprehension, and persuasive writing. Judges can still be heard to joke in unbuttoned moments that, had they been better at science, they would have gone to medical school. Snow’s The Two Cultures still lingers over the legal profession, and Holmes’s prediction has not been fully realized.

Disincentives to enroll in STEM courses create a serious problem for law-school education.To be sure, for lawyers involved in patent law, membership in the patent bar requires an undergraduate degree in science or engineering. Otherwise, applicants to law school need not take any courses in science, mathematics, or statistics. The rigor of such courses, and perhaps steeper grading curves, are substantial disincentives for undergraduates bound for law school to take them.

These disincentives to enroll in science, technology, engineering, and math (STEM) courses create a serious problem for law-school education. The basic law degree, the Juris Doctorate, follows upon a three-year academic program. Law schools face significant pressure to include clinical courses to prepare students for trial and appellate work, as well as for counseling clients. There is a core curriculum, which will be tested on the bar examinations and which must be taught. Law students have little time and opportunity to extend their basic education in the STEM disciplines while pursuing the J.D.

For almost 10 years, I taught a course in statistics and probability in the law at the Columbia Law School. The course is an outlier; only a handful of law schools offer such a course, and none requires an undergraduate course as a prerequisite. And yet, statistical issues, as Holmes predicted, have become omnipresent in the law. Civil cases involving claims of discrimination, election irregularities, antitrust, securities fraud, product disparagement, trademarks, class actions, deceptive advertising, environmental damage, health effects, and many other areas regularly involve statistical evidence. On the criminal side, virtually all forensic analyses, including DNA testing, require application of scientific and statistical principles.

To be sure, lawyers will have some access to expert witnesses, either to guide them or to testify for them, but lawyers must have sufficient scientific and statistical acumen to make careful choices in selecting their witnesses. Perhaps more important, lawyers must be able to take those technical concepts and calculations and present them effectively to lay judges and jurors in advocating for their clients. Correlative to their duty to advocate zealously, lawyers also have a duty of good faith and candor toward the tribunal. Lawyers without STEM training are at serious risk of failing their clients, their obligations to the court, and their duty to the public.

There seems to be no lack of examples of scientific and mathematical illiteracy among lawyers, but two high-profile instances may be helpful to illustrate the seriousness of the problem. Harvard law professor Alan Dershowitz was a member of O.J. Simpson’s defense team in what was perhaps the highest-profile murder trial of the 20th century. Simpson faced several damning pieces of evidence in his trial for the murder of his former wife, Nicole Simpson, including a well-documented history of having physically abused her. Dershowitz notoriously argued that, since fewer than one in 1,000 abused married women are subsequently murdered by their abusive spouse, Simpson’s history of abuse was legally irrelevant. Dershowitz’s stunning error was to ignore the obvious; Nicole had been murdered. Elementary probability theory emphasizes the conditional nature of many probabilities. Given that the victim had been physically abused by a spouse and had been murdered, the probability that Nicole Simpson had been murdered by the abusive spouse was actually greater than 80 percent. The principle of charity requires us to conclude that Alan Dershowitz advanced his argument out of ignorance and not out of a malevolent intent to deceive.

Effective and ethical advocacy of scientific and statistical issues is now an indispensable desideratum in the American legal system.A similarly fallacious distortion of probabilities occurred on an even larger scale in 2020, when the attorney general of Texas, Ken Paxton, sued the states of Pennsylvania, Georgia, Wisconsin, and Michigan to overturn their election results. With the support of 17 other state attorneys general and 126 members of Congress, Paxton sought to file a complaint directly in the Supreme Court. Candidate Donald Trump, represented by John Eastman, sought to intervene on the side of Texas.

The Texas lawsuit turned in large part on a statistical analysis conducted by an econometrician, Charles Cicchetti. That analysis assumed that voters had the same party preferences in both 2016 (Clinton vs. Trump) and 2020 (Biden vs. Trump) and that the distribution of votes counted before 3 a.m. on November 4, 2020, was the same as those counted after that time. Given those assumptions, Cicchetti calculated that the probability that Biden had defeated Trump in the four contested states was less than one in a quadrillion. In advancing these calculations, Cicchetti showed two utterly irrelevant (and uninteresting) conditional probabilities. The assumption that party preferences and votes are immutable is, of course, stunningly stupid. If it were true, the Federalists would still control the White House. We have elections precisely so that citizens can update and change their preferences. When a null hypothesis is meaningless and absurd, we should not be surprised that statistical tests reject it with a high level of statistical significance. Furthermore, the electorate is not static. People die, and young people come of voting age. Aliens become citizens and are thus eligible to vote. Turnout varies from election to election. In 2016, roughly 60 percent of eligible voters participated, whereas, in 2020, the turnout increased by 10 percent, with 66 percent of eligible voters participating. Similarly, there is no basis to assume that late-counted votes have the same partisan distribution as early-counted votes.

Although Cicchetti’s analysis was both silly and irrelevant, Paxton went further by misrepresenting Cicchetti’s calculations, which were conditional probabilities. In the Texas brief, Paxton dropped the antecedent conditions and expressed the probabilities as absolute: “The probability of former Vice President Biden winning the popular vote … is less than one in a quadrillion.” Although lawyers have a duty to advance their clients’ causes zealously, they also have an ethical responsibility to act in good faith and to eschew invalid data and arguments, even when those arguments involve scientific and statistical evidence. The California State Bar brought disciplinary charges against Mr. Trump’s lawyer, John Eastman, for his role in developing Trump’s strategy to retain office, including misrepresentations of Cicchetti’s statistical analyses. Earlier this year, a judge ordered that Eastman be disbarred, a judgment that is now on appeal to the California Supreme Court.

Effective and ethical advocacy of scientific and statistical issues is now an indispensable desideratum in the legal system. Law schools cannot fully address the need. Undergraduate students with degrees or even minor concentrations in the STEM disciplines have an important opportunity to step into the breach—to fill a social and political need while enjoying a fair prospect for material and spiritual success.

Nathan Schachtman is a practicing lawyer who has defended against claims of health effects in products-liability litigation for over 40 years.