Writing in 1901, Learned Hand asked this question: how can a jury decide between the conflicting opinions of two experts "founded upon an experience foreign to their own." 18 Harv. L. Rev. 405. As he noted, "it is just because they are incompetent for such a task that an expert is needed at all." Hand's question comes to mind when one reads the recent decision of the New York Court of Appeals in People v. Williams, 2020 N.Y. Slip. Op. 02123, a case involving a subject that Hand could not have imagined: the admissibility of expert testimony about an unconventional DNA test in a criminal case.

Background

Cadman Williams was convicted of the May 25, 2008 murder of Kenneth Sackey in the Bronx. The murder arose out of a dispute among teenagers, which Williams ended by firing four shots at Sackey, two of which proved fatal. Williams hid the gun in the wall of his girlfriend's apartment, where it was later recovered, telling her that he had "just shot someone." The girlfriend (by then a former girlfriend) testified for the prosecution, as did several eyewitnesses, who identified Williams as the shooter. The prosecution also offered DNA evidence linking Williams to the gun. Standard DNA testing could not find a link, but, using more sensitive low copy number (LCN) testing, scientists at the New York City Office of the Chief Medical Examiner (OCME) concluded that the DNA found on the gun almost certainly came from Williams.

The issue that reached the Court of Appeals in Williams was the admissibility of that expert testimony. More specifically, the question was this: Did the trial court abuse its discretion in denying Williams a Frye hearing to require the prosecution to show that LCN testing was generally accepted in the scientific community? (Williams also addresses the OCME's calculation methodology, but I focus here on LCN testing for ease of explication.)

The general principles for admitting novel scientific testimony under the Frye standard—the standard that applies in New York—are familiar to most lawyers. Frye v. United States, 293 F. 1013 (D.C. Cir. 1923). First, under Frye, such testimony is admissible only after the scientific principle or procedure on which it is based has "gained general acceptance" among scientists in its specified field. Although unanimity among scientists is not required, there must be a consensus in the scientific community that the methodology is reliable. Second, the burden of proof is on the proponent of the evidence to show general acceptance. And third, once a scientific procedure has been found to have received general acceptance (e.g., radar), a Frye inquiry is not required; courts can take judicial notice of its acceptance.

The court in Williams denied the defense's request for a Frye hearing on LCN testing. In doing so, it relied principally on a prior decision in Queens County, People v. Megnath, 27 Misc.2d 405 (2010), in which the court, after a Frye hearing, concluded that LCN testing was generally accepted as reliable. Notably, at the time, LCN evidence had been admitted 140 times in New York state, but Megnath was the only case in which a Frye hearing was held.

In Megnath, the court found that LCN testing was "simply a more sensitive form of DNA testing," which increases the number of times the DNA is amplified (from 28 to 31) and thereby allows for testing of smaller samples. The equipment used is the same as in traditional DNA testing, as is the basic process other than the increased amplification. Moreover, OCME had conducted extensive validation of LCN testing; the DNA Subcommittee of the New York State Commission on Forensic Science had approved LCN testing for forensic casework; and it was used in several countries abroad. The concerns that the defense raised, the Megnath court reasoned, went to the weight of the evidence, not its admissibility. Almost as an afterthought, the court added that LCN testing was not a novel scientific procedure; it was basically a refinement of what had been admitted routinely for years.

Before the trial began in Williams, a second court, this one in Brooklyn, conducted a second Frye hearing and reached the opposite result from Megnath. The opinion in People v. Collins, 49 Misc.3d 595, begins with the finding that LCN testing is, in fact, novel—that it is "crossing a border to a different world." Once one crosses that border, several facts stand out. The OCME is the only laboratory in America that produces LCN results for use in criminal cases. Other labs employ the procedure to develop investigative leads and identify persons after disasters like airplane crashes, but not to present results in court. (Among the labs declining to employ LCN testing is the FBI lab.) Moreover, Dr. Bruce Budowle, one of the nation's leading DNA experts (the Collins court called him the American father of DNA analysis) testified that LCN testing was not reliable enough for courtroom use. LCN enhancements, Dr. Budowle testified, increase the sensitivity of detection of genetic markers in the tested material and therefore may result in inaccurate genetic profiles and invalid results. Ten scholarly articles were also introduced that questioned the validity of LCN testing. Finally, the Collins court observed that, although the DNA Subcommittee of New York State Forensic Commission was comprised of "world class scientists," its views could not be equated with the general membership of the scientific community. Significantly, one member of the Subcommittee had since "defected" and testified for the defense at the Collins hearing.

The Collins court put its conclusion this way: "if the experts in the DNA field cannot agree on the weight to be given to evidence produced by [LCN] analysis, it would make no sense to throw such evidence before a lay jury and ask the jurors to give the evidence appropriate weight." Learned Hand would have appreciated that conclusion.

Thereafter, defense counsel presented a transcript of the Collins court's oral opinion (the written opinion followed later) to the Williams court and asked it to reconsider its decision that a Frye hearing was unwarranted. The court, however, was unpersuaded. The LCN evidence was presented to the jury; Williams was convicted; and the First Department affirmed the conviction.

Court of Appeals Decision

Which brings us to Judge Fahey's opinion for the Court of Appeals in Williams. Chiding the Williams trial judge for following "a single, questionable judicial determination [Megnath]," the court found that there was indeed "marked conflict with respect to the reliability of LCN DNA within the relevant scientific community at the time the LCN issue was litigated in this case." The fact that LCN evidence is used in foreign countries was deemed of "no consequence inasmuch as there [is] no indication that the threshold for admissibility in those bodies contains the same exacting standards to be applied in courts of this state." The conflict among the experts was simply too sharp to ignore. Thus, it was "an abuse of discretion as a matter of law" for the trial court not to have held a hearing.

That did not end matters. Having reached the conclusion that a Frye hearing was necessary, the court found that the error was harmless. The proof of Williams' guilt was overwhelming, and there was no significant possibility that he would have been acquitted if the LCN evidence had been excluded.

Chief Judge DiFiore, joined by Judges Garcia and Feinman, concurred in a separate opinion. The Chief Judge took the majority to task for its "pejorative view of the [OCME] LCN DNA typing technique" and for referring to the opinion in Megnath as "questionable." And she emphasized "the paramount importance of the … highly credentialed DNA Subcommittee and its approval of the OCME's validation techniques." In short, the Chief Judge agreed that a Frye hearing was required but believed that the majority's opinion implied wrongly that its result was preordained.

Analysis

Some observations about the court's decision:

• Chief Judge DiFiore is correct that the majority opinion suggests that the result of any Frye hearing would be an exclusionary ruling—that the scientific community is so divided on the reliability of LCN testing that the evidence should not be presented to a jury. One comes away from the majority opinion believing that the Megnath court was wrong and the Collins court right. Some of the Chief Judge's comments, however, seem questionable. Why are the views of the state's DNA Subcommittee of "paramount importance"? The Subcommittee's opinion should matter, but paramount means more important than anything else and that seems a stretch. Also problematic is a footnote in which the Chief Judge indicates that the Frye inquiry should focus on "whether the dissenting voices are from scientists who have empirical proof to refute the validated empirical evidence of the OCME." Surely, the voices of distinguished experts who challenge the principles and methods underlying a forensic test should count in the Frye calculus, even if they do not bring their own empirical evidence to court.
• A final portion of Judge Fahey's opinion on the Frye issue (Part E) deserves special note. In it, the court backtracks from its teaching that Frye hearings are "generally unwarranted absent a novel or experimental scientific theory." People v. Brooks, 31 N.Y.3d 939, 941 (2018). "Familiarity," Judge Fahey writes, "does not always breed accuracy," and a Frye hearing may be called for, even with respect to forensic tests that may have become standard courtroom fare. In that regard, Judge Fahey identified hair comparisons, fire origin analysis, comparative bullet analysis, bite mark matching and blood-pattern analysis as techniques that "have long been accepted within their relevant scientific communities" but recently called "into varying degrees of question." Thus, Williams is an invitation for defense counsel to challenge the admissibility of such evidence. To borrow the title of an article that Judge Fahey cites, the court now recognizes that "dubious science can spread like a virus" if not put to the Frye test.
• The court's invitation to apply Frye to more than just novel scientific evidence could change the outcome of cases like People v. Flores-Estrada, 55 Misc.3d 1015, decided in Kings County in 2017. There, the defendant moved for a Frye hearing to challenge the admissibility of shaken baby syndrome evidence—expert testimony that shaking a baby can result in permanent brain damage or death. Although recognizing that "a significant and legitimate debate in the medical community about the issue has developed in the past ten years," the trial court found dispositive the fact that such evidence has been admitted in New York courts "for over two decades." As the court saw it, disagreement in the scientific community, "raise[d] a question of fact, which [was] a matter for the jury"; it did not require a hearing. Williams calls the reasoning of Flores-Estrada into serious question. "Past is prologue" Shakespeare wrote, but Williams teaches that past admissibility does not guarantee present admissibility when a significant portion of the expert community develops significant doubts. (I thank Prof. Vincent Alexander for bringing Flores-Estrada to my attention; the issue is also discussed in a thoughtful amicus brief submitted by several defender organizations in Williams.)
• Now that Frye is not limited to novel scientific evidence, a question arises: What is the difference between Frye and Daubert, the federal standard? Daubert v. Merrell Dow Pharm., 509 U.S. 579 (1993). The answer, it seems, is that Frye makes scientists the gatekeepers for the admissibility of scientific evidence, whereas Daubert assigns that role to judges. Frye counts the votes of scientists while Daubert reposes responsibility in judges to hear the testimony of competing scientists and decide if the proffered testimony is reliable. See People v. Wesley, 83 N.Y.2d 417, 439 (1994) (Kaye, C.J., concurring) ("[t]he Frye test emphasizes counting scientists' votes rather than on verifying the soundness of scientific conclusions"). A federal judge can side with the minority of scientists and admit evidence if she concludes that the minority has it right. A state judge cannot. Given how much bad science has been admitted in criminal cases and how many wrongful convictions have resulted therefrom, Frye's more conservative test, if not limited to novel scientific evidence, may be what is needed.
• Two additional notes: First, Judge Fahey's opinion includes a footnote written as an "epilogue to this case." It notes that, since the trial in Williams, the OCME "has abandoned LCN testing," and it questions whether the "People would defend an application seeking a Frye hearing on that issue today." (The abandonment is discussed in a September 2017 New York Times article titled "Traces of Crime: How New York's DNA Techniques Became Tainted.") Judge Fahey's question is a fair one and begs another: If the OCME abandoned LCN testing in 2017, why did the People seek to defend it as generally accepted and reliable in 2019 in Williams?
• Evenhandedness requires a jab at the defense. In seeking to avoid a harmless error ruling, Williams' lawyers wrote this to the Court of Appeals: "Had Mr. Williams not been forced to fashion a defense that accommodated the evidence showing that he likely handled the murder weapon, he might have declined to take the stand and assert a justification defense, and may have chosen [instead] to defend on the grounds that he had not held the gun or fired the shot that killed Sackey." Really? Should a defense lawyer tell the Court of Appeals that his client gave perjurious testimony at trial because the People's evidence left him no choice?

In sum, Williams is good reading for those sheltering in place.

Paul Shechtman is a partner at Bracewell in New York and teaches evidence and criminal procedure at Columbia Law School.