In “Cheng’s Proposed Consensus Rule for Expert Witnesses,”[1] I discussed a recent law review article by Professor Edward K. Cheng,[2] who has proposed dispensing with expert witness testimony as we know it in favor of having witnesses tell juries what the scientific consensus is on any subject. Cheng’s project is fraught with difficulties and contradictions; and it has clearly anticipatable bad outcomes. Four Supreme Court cases (Daubert, Joiner, Kumho Tire, and Weisgram), and a major revision in Rule 702, ratified by Congress, all embraced the importance of judicial gatekeeping of expert witness opinion testimony to the fact-finding function of trials. Professor Cheng now wants to ditch the entire notion of gatekeeping, as well as the epistemic basis – sufficient facts and data – for expert witnesses’ opinions in favor of reportage of which way the herd is going. Cheng’s proposal is perhaps the most radical attack, in recent times, on the nature of legal factfinding, whether by judges or juries, in the common law world.
Still, there are two claims within his proposal, which although overstated, are worth further discussion and debate. The first is that the gatekeeping role does not sit well with many judges. We see judges ill at ease in their many avoidance tactics, by which they treat serious methodological challenges to expert witness testimony as “merely going to the weight of the conclusion.” The second is that many judges, and especially juries, are completely at sea in the technical knowledge needed to evaluate the scientific issues in many modern day trials.
With respect to the claimed epistemic incompetence, the simpler remedy is to get rid of incompetent judges. We have commercial courts, vaccine courts, and patent courts. Why are litigants disputing a contract or a commercial practice entitled to epistemically competent judges, but litigants in health claim cases are not? Surely, the time has come to have courts with judges that have background and training in the health and statistical sciences. The time for “blue ribbon” juries of properly trained fact finders seems overdue. Somehow we must reconcile the seventh amendment right to a jury with the requirement of “due process” of law. The commitment to jury trials for causes of action known to the common law in 1787, or 1791, is stretched beyond belief for the sorts of technical and complex claims now seen in federal courts and state courts of general jurisdiction.[3]
Several courts have challenged the belief that the seventh amendment right to a jury applies in the face of complex litigation. The United States Court of Appeals explained its understanding of complexity that should remove a case from the province of the seventh amendment:
“A suit is too complex for a jury when circumstances render the jury unable to decide in a proper manner. The law presumes that a jury will find facts and reach a verdict by rational means. It does not contemplate scientific precision but does contemplate a resolution of each issue on the basis of a fair and reasonable assessment of the evidence and a fair and reasonable application of the relevant legal rules. See Schulz v. Pennsylvania RR, 350 U.S. 523, 526 (1956). A suit might be excessively complex as a result of any set of circumstances which singly or in combination render a jury unable to decide in the foregoing rational manner. Examples of such circumstances are an exceptionally long trial period and conceptually difficult factual issues.”[4]
The Circuit’s description of complexity certainly seems to apply to many contemporary claims of health effects.
We should recognize that Professor Cheng’s indictment, and conviction, of judicial gatekeeping and jury decision making as epistemically incompetent directly implies that the judicial process has no epistemic, truth finding function in technical cases of claimed health effects. Cheng’s proposed solution does not substantially ameliorate this implication, because consensus statements are frequently absent, and even when present, are plagued with their own epistemic weaknesses.
Consider for instance, the 1997 pronouncement of the International Agency for Research on Cancer that crystalline silica is a “known” human carcinogen.[5] One of the members of the working group responsible for the pronouncement explained:
“It is hardly surprising that the Working Group had considerable difficulty in reaching a decision, did not do so unanimously and would probably not have done so at all, had it not been explained that we should be concerned with hazard identification, not risk.”[6]
And yet, within months of the IARC pronouncement, state and federal regulatory agencies formed a chorus of assent to the lung cancer “risk” of crystalline silica. Nothing in the scientific record had changed except the permission of the IARC to stop thinking critically about the causation issue. Another consensus group came out, a few years after the IARC pronouncement, with a devastating critical assessment of the IARC review:
“The present authors believe that the results of these studies [cited by IARC] are inconsistent and, when positive, only weakly positive. Other, methodologically strong, negative studies have not been considered, and several studies viewed as providing evidence supporting the carcinogenicity of silica have significant methodological weaknesses. Silica is not directly genotoxic and is a pulmonary carcinogen only in the rat, a species that seems to be inappropriate for assessing particulate carcinogenesis in humans. Data on humans demonstrate a lack of association between lung cancer and exposure to crystalline silica. Exposure-response relationships have generally not been found. Studies in which silicotic patients were not identified from compensation registries and in which enumeration was complete did not support a causal association between silicosis and lung cancer, which further argues against the carcinogenicity of crystalline silica.”[7]
Cheng’s proposal would seem to suppress legitimate courtroom criticism of an apparent consensus statement, which was based upon a narrow majority of a working group, on a controversial dataset, with no examination of the facts and data upon which the putative consensus statement was itself based.
The Avandia litigation tells a cautionary tale of how fragile and ephemeral consensuses can be. A dubious meta-analysis by a well-known author received lead article billing in an issue of the New England Journal of Medicine, in 2007, and litigation claims started to roll in within hours.[8] In face of this meta-analysis, an FDA advisory committee recommended heightened warnings, and a trial court declined to take a careful look at the methodological flaws in the inciting meta-analytic study.[9] Ultimately, a large clinical trial exculpated the medication, but by then the harm had been done, and there was no revisiting of the gatekeeping decision to allow the claims to proceed.[10] The point should be obvious. In 2007, there appeared to be a consensus, with led to an FDA label change, despite the absence of sufficient facts and data to support the litigation claims. Even if plaintiffs’ claims passed through the gate in 2008, they were highly vulnerable to courtroom challenges to the original meta-analysis. Cheng’s proposal, however, would truncate the litigation process into an exploration whether or not there was a “consensus.”
Deviation from Experts’ Standards of Care
The crux of many Rule 702 challenges to an expert witness is that the witness has committed malpractice in his discipline. The challenger must identify a standard of care, and the challenged witness’s deviation(s) from that standard. The identification of the relevant standard of care will, indeed, sometimes involve a consensus, evidenced by texts, articles, professional society statements, or simply implicit in relevant works of scholarship or scientific studies. Consensuses about standards of care are, of course, about methodology. Consensuses about conclusions, however, may also be relevant because if a litigant’s expert witness proffers a conclusion at odds with consensus conclusions, the deviant conclusion implies deviant methodology.
Cheng’s treatment of statistical significance is instructive for how his proposal would create mischief in many different types of adjudications, but especially of claimed health effects. First, Cheng’s misrepresentation of consensus among statisticians is telling for the validity of his project. After all, he holds an advanced degree in statistics, and yet, he is willing write that that:
“[w]hile historically used as a rule of thumb, statisticians have now concluded that using the 0.05 [p-value] threshold is more distortive than helpful.”[11]
Statisticians, without qualification! And as was shown, Cheng is demonstrably wrong in his use of the cited source to support his representation of what certainly seems like a consensus paper. His précis is not even remotely close to the language of the paper, but the consensus paper is hearsay and can only be used by an expert witness in support of an opinion. Presumably, another expert witness might contradict the quoted opinion about what “statisticians” have concluded, but it is unclear whether a court could review the underlying A.S.A. paper, take judicial notice of the incorrectness of the proffered opinion, and then exclude the expert witness opinion.
After the 2016 publication of the A.S.A.’s consensus statement, some statisticians did indeed publish editorials claiming it was time to move beyond statistical significance testing. At least one editorial, by an A.S.A. officer was cited as representing an A.S.A. position, which led the A.S.A. President to appoint a task force to consider the call for an across-the-board rejection of significance testing. In 2021, that task force clearly endorsed significance testing as having a continued role in statistical practice.[12]
Where would this situation leave a gatekeeping court or a factfinding jury? Some obscure psychology journals have abandoned the use of significance testing, but the New England Journal of Medicine has retained the practice, while introducing stronger controls for claims of “significance” when the study at hands has engaged in multiple comparisons.
But Cheng, qua law professor and statistician (and would-be expert witness) claims “statisticians have now concluded that using the 0.05 [p-value] threshold is more distortive than helpful,” and the trial must chase not the validity of the inference of claimed causation but whether there is, or is not, a census about the use of a pre-specified threshold for p-values or confidence intervals. Cheng’s proposal about consensuses would turn trials into disputes about whether consensuses exist, and the scope of the purported agreement, not about truth.
In some instances, there might be a clear consensus, fully supported, on a general causation issue. Consider for instance, the known causal relationship between industrial benzene exposure and acute myelogenous leukemia (AML). This consensus turns out to be rather unhelpful when considering whether minute contamination of carbonated water can cause cancer,[13] or even whether occupational exposure to gasoline, with its low-level benzene (~1%) content, can cause AML.[14]
Frequently, there is also a deep asymmetry in consensus statements. When the evidence for a causal conclusion is very clear, professional societies may weigh in to express their confident conclusions about the existence of causation. Such societies typically do not issue statements that explicitly reject causal claims. The absence of a consensus statement, however, often can be taken to represent a consensus that professional societies do not endorse causal claims, and consider the evidence, at best, equivocal. Those dogs that have not barked can be, and have been, important considerations in gatekeeping.
Contrary to Cheng’s complete dismissal of judges’ epistemic competence, judges can, in many instances, render reasonable gatekeeping decisions by closely considering the absence of consensus statements, or systematic reviews, favoring the litigation claims.[15] At least in this respect, Professor Cheng is right to emphasize the importance of consensus, but he fails to note the importance of its absence, and the ability of litigants and their expert witnesses to inform gatekeeping judges of the relevance of consensus statements or their absence to the epistemic assessment of proferred expert witness opinion testimony.
[1] “Cheng’s Proposed Consensus Rule for Expert Witnesses,” (Sept. 15, 2022).
[2] Edward K. Cheng, “The Consensus Rule: A New Approach to Scientific Evidence,” 75 Vanderbilt L. Rev. 407 (2022) [Consensus Rule]
[3] There is an extensive discussion and debate of viability and the validity of asserting rights to trial by jury for many complex civil actions in the modern era. See, e.g., Stephan Landsman & James F. Holderman, “The Evolution of the Jury Trial in America,” 37 Litigation 32 (2010); Robert A. Clifford, “Deselecting the Jury in a Civil Case,” 30 Litigation 8 (Winter 2004); Hugh H. Bownes, “Should Trial by Jury Be Eliminated in Complex Cases,” 1 Risk 75 (1990); Douglas King, “Complex Civil Litigation and the Seventh Amendment Right to a Jury Trial,” 51 Univ. Chi. L. Rev. 581 (1984); Alvin B. Rubin, “Trial by Jury in Complex Civil Cases: Voice of Liberty or Verdict by Confusion?” 462 Ann. Am. Acad. Political & Social Sci. 87 (1982); William V. Luneburg & Mark A. Nordenberg, “Specially Qualified Juries and Expert Nonjury Tribunals: Alternatives for Coping with the Complexities of Modern Civil Litigation,” 67 Virginia L. Rev. 887 (1981); Richard O. Lempert, “Civil Juries and Complex Cases: Let’s Not Rush to Judgment,” 80 Mich. L. Rev. 68 (1981); Comment, “The Case for Special Juries in Complex Civil Litigation,” 89 Yale L. J. 1155 (1980); James S. Campbell & Nicholas Le Poidevin, “Complex Cases and Jury Trials: A Reply to Professor Arnold,” 128 Univ. Penn. L. Rev. 965 (1980); Barry E. Ungar & Theodore R. Mann, “The Jury and the Complex Civil Case,” 6 Litigation 3 (Spring 1980); Morris S. Arnold, “A Historical Inquiry into the Right to Trial by Jury in Complex Civil Litigation,”128 Univ. Penn. L. Rev. 829 (1980); Daniel H. Margolis & Evan M. Slavitt, “The Case Against Trial by Jury in Complex Civil Litigation,” 7 Litigation 19 (1980); Montgomery Kersten, “Preserving the Right to Jury Trial in Complex Civil Cases,” 32 Stanford L. Rev. 99 (1979); Maralynne Flehner, “Jury Trials in Complex Litigation,” 4 St. John’s Law Rev. 751 (1979); Comment, “The Right to a Jury Trial in Complex Civil Litigation,” 92 Harvard L. Rev. 898 (1979); Kathy E. Davidson, “The Right to Trial by Jury in Complex Litigation,” 20 Wm. & Mary L. Rev. 329 (1978); David L. Shapiro & Daniel R. Coquillette, “The Fetish of Jury Trial in Civil Cases: A Comment on Rachal v. Hill,” 85 Harvard L. Rev. 442 (1971); Comment, “English Judge May Not Order Jury Trial in Civil Case in Absence of Special Circumstances. Sims v. William Howard & Son Ltd. (C. A. 1964),” 78 Harv. L. Rev. 676 (1965); Fleming James, Jr., “Right to a Jury Trial in Civil Actions,” 72 Yale L. J. 655 (1963).
[4] In re Japanese Elec. Prods. Antitrust Litig., 63` F.2d 1069, 1079 (3d Cir 1980). See In re Boise Cascade Sec. Litig., 420 F. Supp. 99, 103 (W.D. Wash. 1976) (“In sum, it appears to this Court that the scope of the problems presented by this case is immense. The factual issues, the complexity of the evidence that will be required to explore those issues, and the time required to do so leads to the conclusion that a jury would not be a rational and capable fact finder.”). See also Ross v. Bernhard, 396 U.S. 532, 538 & n.10, 90 S. Ct. 733 (1970) (discussing the “legal” versus equitable nature of an action that might give rise to a right to trial by jury). Of course, the statistical and scientific complexity of claims was absent from cases tried in common law courts in 1791, at the time of the adoption of the seventh amendment.
[5] IARC Monograph on the Evaluation of Carcinogenic Risks to Humans of Silica, Some Silicates, Coal Dust and para-Aramid Fibrils, vol. 68 (1997).
[6] Corbett McDonald & Nicola Cherry, “Crystalline Silica and Lung Cancer: The Problem of Conflicting Evidence,” 8 Indoor Built Env’t 121, 121 (1999).
[7] Patrick A. Hessel, John F. Gamble, J. Bernard L. Gee, Graham Gibbs, Francis H.Y. Green, W. Keith C. Morgan, and Brooke T. Mossman, “Silica, Silicosis, and Lung Cancer: A Response to a Recent Working Group Report,” 42 J. Occup & Envt’l Med. 704, 704 (2000).
[8] Steven Nissen & K. Wolski, “Effect of Rosiglitazone on the Risk of Myocardial Infarction and Death from Cardiovascular Causes,” 356 New Engl. J. Med. 2457 (2007); Erratum, 357 New Engl. J. Med. 100 (2007).
[9] In re Avandia Mktg., Sales Practices & Prods. Liab. Litig., 2011 WL 13576 (E.D. Pa. Jan. 4, 2011).
[10] Philip D. Home, Stuart J Pocock, et al., “Rosiglitazone Evaluated for Cardiovascular Outcomes in Oral Agent Combination Therapy for Type 2 Diabetes (RECORD),” 373 Lancet 2125 (2009). The hazard ratios for cardiovascular death was 0.84 (95% C.I., 0·59–1·18), and for myocardial infarction, 1·14 (95% C.I., 0·80–1·63).
[11] Consenus Rule at 424 (emphasis added) (citing Ronald L. Wasserstein & Nicole A. Lazar, “The ASA Statement on p-Values: Context, Process, and Purpose,” 70 Am. Statistician 129, 131 (2016)).
[12] Yoav Benjamini, Richard D. DeVeaux, Bradly Efron, Scott Evans, Mark Glickman, Barry Braubard, Xuming He, Xiao Li Meng, Nancy Reid, Stephen M. Stigler, Stephen B. Vardeman, Christopher K. Wikle, Tommy Wright, Linda J. Young, and Karen Kafadar, “The ASA President’s Task Force Statement on Statistical Significance and Replicability,” 15 Annals of Applied Statistics 1084 (2021); see also “A Proclamation from the Task Force on Statistical Significance” (June 21, 2021).
[13] Sutera v. Perrier Group of America, Inc., 986 F. Supp. 655, 664-65 (D. Mass. 1997).
[14] Burst v. Shell Oil Co., 2015 WL 3755953, at *9 (E.D. La. June 16, 2015), aff’d, 650 F. App’x 170 (5th Cir. 2016). cert. denied. 137 S. Ct. 312 (2016); Henricksen v. ConocoPhillips Co., 605 F. Supp. 2d 1142, 1156 (E.D. Wa. 2009).
[15] In re Mirena Ius Levonorgestrel-Related Prod. Liab. Litig. (No. II), 341 F. Supp. 3d 213 (S.D.N.Y. 2018), aff’d, 982 F.3d 113 (2d Cir. 2020); In re Lipitor (Atorvastatin Calcium) Mktg., Sales Pracs. & Prods. Liab. Litig., 227 F. Supp. 3d 452 (D.S.C. 2017), aff’d, 892 F.3d 624 (4th Cir. 2018); In re: Zoloft (Sertraline Hydrocloride) Prod. Liab. Litig., No. 12-MD-2342, 2015 WL 7776911, at *1 (E.D. Pa. Dec. 2, 2015), aff’d, 858 F.3d 787 (3d Cir. 2017); In re Incretin-Based Therapies Prods. Liab. Litig., 524 F. Supp. 3d. 1007 (S.D. Cal. 2021); In re Viagra (Sildenafil Citrate) & Cialis (Tadalafil) Prod. Liab. Litig., 424 F. Supp. 3d 781, 798–99 (N.D. Cal. 2020).
