In scientific publishing, when scientists make a mistake, they publish an erratum or a corrigendum. If the mistake vitiates the study, then the erring scientists retract their article. To be sure, sometimes the retraction comes after an obscene delay, with the authors kicking and screaming.[1] Sometimes the retraction comes at the request of the authors, better late than never.[2]
Retractions in the biomedical journals, whether voluntary or not, are on the rise.[3] The process and procedures for retraction of articles often lack transparency. Many articles are retracted without explanation or disclosure of specific problems about the data or the analysis.[4] Sadly, however, misconduct in the form of plagiarism and data falsification is a frequent reason for retractions.[5] The lack of transparency for retractions, and sloppy scholarship, combine to create Zombie papers, which are retracted but continue to be cited in subsequent publications.[6]
LEGAL RETRACTIONS
The law treats errors very differently. Being a judge usually means that you never have to say you are sorry. Judge Andrew Hurwitz has argued that that our legal system would be better served if judges could and did “freely acknowledged and transparently corrected the occasional ‘goof’.”[7] Alas, as Judge Hurwitz notes, very few published decisions acknowledge mistakes.[8]
In the world of scientific jurisprudence, the judicial reticence to acknowledge mistakes is particularly dangerous, and it leads directly to the proliferation of citations to cases that make egregious mistakes. In the niche area of judicial assessment of scientific and statistical evidence, the proliferation of erroneous statements is especially harmful because it interferes with thinking clearly about the issues before courts. Judges believe that they have argued persuasively for a result, not by correctly marshaling statistical and scientific concepts, but by relying upon precedents erroneously arrived at by other judges in earlier cases. Regardless of how many cases are cited (and there are many possible “precedents”), the true parameter does not have a 95% probability of lying within the interval given by a given 95% confidence interval.[9] Similarly, as much as judges would like p-values and confidence intervals to eliminate the need to worry about systematic error, their saying so cannot make it so.[10] Even a mighty federal judge cannot make the p-value probability, or its complement, substitute for the posterior probability of a causal claim.[11]
Some cases in the books are so egregiously decided that it is truly remarkable that they would be cited for any proposition. I call these scientific Dred Scott cases, which illustrate that sometimes science has no criteria of validity that the law is bound to respect. One such Dred Scott case was the result of a bench trial in a federal district court in Atlanta, in Wells v. Ortho Pharmaceutical Corporation.[12]
Wells was notorious for its poor assessment of all the determinants of scientific causation.[13] The decision was met with a storm of opprobrium from the legal and medical community.[14] No scientists or legal scholars offered a serious defense of Wells on the scientific merits. Even the notorious plaintiffs’ expert witness, Carl Cranor, could muster only a distanced agnosticism:
“In Wells v. Ortho Pharmaceutical Corp., which involved a claim that birth defects were caused by a spermicidal jelly, the U.S. Court of Appeals for the 11th Circuit followed the principles of Ferebee and affirmed a plaintiff’s verdict for about five million dollars. However, some members of the medical community chastised the legal system essentially for ignoring a well-established scientific consensus that spermicides are not teratogenic. We are not in a position to judge this particular issue, but the possibility of such results exists.”[15]
Cranor apparently could not bring himself to note that it was not just scientific consensus that was ignored; the Wells case ignored the basic scientific process of examining relevant studies for both internal and external validity.
Notwithstanding this scholarly consensus and condemnation, we have witnessed the repeated recrudescence of the Wells decision. In Matrixx Initiatives, Inc. v. Siracusano,[16] in 2011, the Supreme Court, speaking through Justice Sotomayor, wandered into a discussion, irrelevant to its holding, whether statistical significance was necessary for a determination of the causality of an association:
“We note that courts frequently permit expert testimony on causation based on evidence other than statistical significance. See, e.g., Best v. Lowe’s Home Centers, Inc., 563 F. 3d 171, 178 (6th Cir 2009); Westberry v. Gislaved Gummi AB, 178 F. 3d 257, 263–264 (4th Cir. 1999) (citing cases); Wells v. Ortho Pharmaceutical Corp., 788 F. 2d 741, 744–745 (11th Cir. 1986). We need not consider whether the expert testimony was properly admitted in those cases, and we do not attempt to define here what constitutes reliable evidence of causation.”[17]
The quoted language is remarkable for two reasons. First, the Best and Westberry cases did not involve statistics at all. They addressed specific causation inferences from what is generally known as differential etiology. Second, the citation to Wells was noteworthy because the case has nothing to do with adverse event reports or the lack of statistical significance.
Wells involved a claim of birth defects caused by the use of spermicidal jelly contraceptive, which had been the subject of several studies, one of which at least yielded a nominally statistically significant increase in detected birth defects over what was expected.
Wells could thus hardly be an example of a case in which there was a judgment of causation based upon a scientific study that lacked statistical significance in its findings. Of course, finding statistical significance is just the beginning of assessing the causality of an association. The most remarkable and disturbing aspect of the citation to Wells, however, was that the Court was unaware of, or ignored, the case’s notoriety, and the scholarly and scientific consensus that criticized the decision for its failure to evaluate the entire evidentiary display, as well as for its failure to rule out bias and confounding in the studies relied upon by the plaintiff.
Justice Sotomayor’s decision for a unanimous Court is not alone in its failure of scholarship and analysis in embracing the dubious precedent of Wells. Many other courts have done much the same, both in state[18] and in federal courts,[19] and both before and after the Supreme Court decided Daubert, and even after Rule 702 was amended in 2000.[20] Perhaps even more disturbing is that the current edition of the Reference Manual on Scientific Evidence glibly cites to the Wells case, for the dubious proposition that
“Generally, researchers are conservative when it comes to assessing causal relationships, often calling for stronger evidence and more research before a conclusion of causation is drawn.”[21]
We are coming up on the 40th anniversary of the Wells judgment. It is long past time to stop citing the case. Perhaps we have reached the stage of dealing with scientific evidence at which errant and aberrant cases should be retracted, and clearly marked as retracted in the official reporters, and in the electronic legal databases. Certainly the technology exists to link the scholarly criticism with a case citation, just as we link subsequent judicial treatment by overruling, limiting, and criticizing.
[1] Laura Eggertson, “Lancet retracts 12-year-old article linking autism to MMR vaccines,” 182 Canadian Med. Ass’n J. E199 (2010).
[2] Notice of retraction for Teng Zeng & William Mitch, “Oral intake of ranitidine increases urinary excretion of N-nitrosodimethylamine,” 37 Carcinogenesis 625 (2016), published online (May 4, 2021) (retraction requested by authors with an acknowledgement that they had used incorrect analytical methods for their study).
[3] Tianwei He, “Retraction of global scientific publications from 2001 to 2010,” 96 Scientometrics 555 (2013); Bhumika Bhatt, “A multi-perspective analysis of retractions in life sciences,” 126 Scientometrics 4039 (2021); Raoul R.Wadhwa, Chandruganesh Rasendran, Zoran B. Popovic, Steven E. Nissen, and Milind Y. Desai, “Temporal Trends, Characteristics, and Citations of Retracted Articles in Cardiovascular Medicine,” 4 JAMA Network Open e2118263 (2021); Mario Gaudino, N. Bryce Robinson, Katia Audisio, Mohamed Rahouma, Umberto Benedetto, Paul Kurlansky, Stephen E. Fremes, “Trends and Characteristics of Retracted Articles in the Biomedical Literature, 1971 to 2020,” 181 J. Am. Med. Ass’n Internal Med. 1118 (2021); Nicole Shu Ling Yeo-Teh & Bor Luen Tang, “Sustained Rise in Retractions in the Life Sciences Literature during the Pandemic Years 2020 and 2021,” 10 Publications 29 (2022).
[4] Elizabeth Wager & Peter Williams, “Why and how do journals retract articles? An analysis of Medline retractions 1988-2008,” 37 J. Med. Ethics 567 (2011).
[5] Ferric C. Fanga, R. Grant Steen, and Arturo Casadevall, “Misconduct accounts for the majority of retracted scientific publications,” 109 Proc. Nat’l Acad. Sci. 17028 (2012); L.M. Chambers, C.M. Michener, and T. Falcone, “Plagiarism and data falsification are the most common reasons for retracted publications in obstetrics and gynaecology,” 126 Br. J. Obstetrics & Gyn. 1134 (2019); M.S. Marsh, “Separating the good guys and gals from the bad,” 126 Br. J. Obstetrics & Gyn. 1140 (2019).
[6] Tzu-Kun Hsiao and Jodi Schneider, “Continued use of retracted papers: Temporal trends in citations and (lack of) awareness of retractions shown in citation contexts in biomedicine,” 2 Quantitative Science Studies 1144 (2021).
[7] Andrew D. Hurwitz, “When Judges Err: Is Confession Good for the Soul?” 56 Ariz. L. Rev. 343, 343 (2014).
[8] See id. at 343-44 (quoting Justice Story who dealt with the need to contradict a previously published opinion, and who wrote “[m]y own error, however, can furnish no ground for its being adopted by this Court.” U.S. v. Gooding, 25 U.S. 460, 478 (1827)).
[9] See, e.g., DeLuca v. Merrell Dow Pharms., Inc., 791 F. Supp. 1042, 1046 (D.N.J. 1992) (”A 95% confidence interval means that there is a 95% probability that the ‘true’ relative risk falls within the interval”) , aff’d, 6 F.3d 778 (3d Cir. 1993); In re Silicone Gel Breast Implants Prods. Liab. Litig, 318 F.Supp.2d 879, 897 (C.D. Cal. 2004); Eli Lilly & Co. v. Teva Pharms, USA, 2008 WL 2410420, *24 (S.D.Ind. 2008) (stating incorrectly that “95% percent of the time, the true mean value will be contained within the lower and upper limits of the confidence interval range”). See also “Confidence in Intervals and Diffidence in the Courts” (Mar. 4, 2012).
[10] See, e.g., Brock v. Merrill Dow Pharmaceuticals, Inc., 874 F.2d 307, 311-12 (5th Cir. 1989) (“Fortunately, we do not have to resolve any of the above questions [as to bias and confounding], since the studies presented to us incorporate the possibility of these factors by the use of a confidence interval.”). This howler has been widely acknowledged in the scholarly literature. See David Kaye, David Bernstein, and Jennifer Mnookin, The New Wigmore – A Treatise on Evidence: Expert Evidence § 12.6.4, at 546 (2d ed. 2011); Michael O. Finkelstein, Basic Concepts of Probability and Statistics in the Law 86-87 (2009) (criticizing the blatantly incorrect interpretation of confidence intervals by the Brock court).
[11] In re Ephedra Prods. Liab. Litig., 393 F.Supp. 2d 181, 191 (S.D.N.Y. 2005) (Rakoff, J.) (“Generally accepted scientific convention treats a result as statistically significant if the P-value is not greater than .05. The expression ‘P=.05’ means that there is one chance in twenty that a result showing increased risk was caused by a sampling error — i.e., that the randomly selected sample accidentally turned out to be so unrepresentative that it falsely indicates an elevated risk.”); see also In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 289 F.Supp. 2d 1230, 1236 n.1 (W.D. Wash. 2003) (“P-values measure the probability that the reported association was due to chance… .”). Although the erroneous Ephedra opinion continues to be cited, it has been debunked in the scholarly literature. See Michael O. Finkelstein, Basic Concepts of Probability and Statistics in the Law 65 (2009); Nathan A. Schachtman, “Statistical Evidence in Products Liability Litigation,” at 28-13, chap. 28, in Stephanie A. Scharf, George D. Sax, & Sarah R. Marmor, eds., Product Liability Litigation: Current Law, Strategies and Best Practices (2d ed. 2021).
[12] Wells v. Ortho Pharm. Corp., 615 F. Supp. 262 (N.D. Ga.1985), aff’d & modified in part, remanded, 788 F.2d 741 (11th Cir.), cert. denied, 479 U.S. 950 (1986).
[13] I have discussed the Wells case in a series of posts, “Wells v. Ortho Pharm. Corp., Reconsidered,” (2012), part one, two, three, four, five, and six.
[14] See, e.g., James L. Mills and Duane Alexander, “Teratogens and ‘Litogens’,” 15 New Engl. J. Med. 1234 (1986); Samuel R. Gross, “Expert Evidence,” 1991 Wis. L. Rev. 1113, 1121-24 (1991) (“Unfortunately, Judge Shoob’s decision is absolutely wrong. There is no scientifically credible evidence that Ortho-Gynol Contraceptive Jelly ever causes birth defects.”). See also Editorial, “Federal Judges v. Science,” N.Y. Times, December 27, 1986, at A22 (unsigned editorial) (“That Judge Shoob and the appellate judges ignored the best scientific evidence is an intellectual embarrassment.”); David E. Bernstein, “Junk Science in the Courtroom,” Wall St. J. at A 15 (Mar. 24,1993) (pointing to Wells as a prominent example of how the federal judiciary had embarrassed American judicial system with its careless, non-evidence based approach to scientific evidence); Bert Black, Francisco J. Ayala & Carol Saffran-Brinks, “Science and the Law in the Wake of Daubert: A New Search for Scientific Knowledge,” 72 Texas L. Rev. 715, 733-34 (1994) (lawyers and leading scientist noting that the district judge “found that the scientific studies relied upon by the plaintiffs’ expert were inconclusive, but nonetheless held his testimony sufficient to support a plaintiffs’ verdict. *** [T]he court explicitly based its decision on the demeanor, tone, motives, biases, and interests that might have influenced each expert’s opinion. Scientific validity apparently did not matter at all.”) (internal citations omitted); Bert Black, “A Unified Theory of Scientific Evidence,” 56 Fordham L. Rev. 595, 672-74 (1988); Paul F. Strain & Bert Black, “Dare We Trust the Jury – No,” 18 Brief 7 (1988); Bert Black, “Evolving Legal Standards for the Admissibility of Scientific Evidence,” 239 Science 1508, 1511 (1988); Diana K. Sheiness, “Out of the Twilight Zone: The Implications of Daubert v. Merrill Dow Pharmaceuticals, Inc.,” 69 Wash. L. Rev. 481, 493 (1994); David E. Bernstein, “The Admissibility of Scientific Evidence after Daubert v. Merrell Dow Pharmacueticals, Inc.,” 15 Cardozo L. Rev. 2139, 2140 (1993) (embarrassing decision); Troyen A. Brennan, “Untangling Causation Issues in Law and Medicine: Hazardous Substance Litigation,” 107 Ann. Intern. Med. 741, 744-45 (1987) (describing the result in Wells as arising from the difficulties created by the Ferebee case; “[t]he Wells case can be characterized as the court embracing the hypothesis when the epidemiologic study fails to show any effect”); Troyen A. Brennan, “Causal Chains and Statistical Links: Some Thoughts on the Role of Scientific Uncertainty in Hazardous Substance Litigation,” 73 Cornell L. Rev. 469, 496-500 (1988); David B. Brushwood, “Drug induced birth defects: difficult decisions and shared responsibilities,” 91 W. Va. L. Rev. 51, 74 (1988); Kenneth R. Foster, David E. Bernstein, and Peter W. Huber, eds., Phantom Risk: Scientific Inference and the Law 28-29, 138-39 (1993) (criticizing Wells decision); Peter Huber, “Medical Experts and the Ghost of Galileo,” 54 Law & Contemp. Problems 119, 158 (1991); Edward W. Kirsch, “Daubert v. Merrell Dow Pharmaceuticals: Active Judicial Scrutiny of Scientific Evidence,” 50 Food & Drug L.J. 213 (1995) (“a case in which a court completely ignored the overwhelming consensus of the scientific community”); Hans Zeisel & David Kaye, Prove It With Figures: Empirical Methods in Law and Litigation § 6.5, at 93(1997) (noting the multiple comparisons in studies of birth defects among women who used spermicides, based upon the many reported categories of birth malformations, and the large potential for even more unreported categories); id. at § 6.5 n.3, at 271 (characterizing Wells as “notorious,” and noting that the case became a “lightning rod for the legal system’s ability to handle expert evidence.”); Edward K. Cheng , “Independent Judicial Research in the ‘Daubert’ Age,” 56 Duke L. J. 1263 (2007) (“notoriously concluded”); Edward K. Cheng, “Same Old, Same Old: Scientific Evidence Past and Present,” 104 Michigan L. Rev. 1387, 1391 (2006) (“judge was fooled”); Harold P. Green, “The Law-Science Interface in Public Policy Decisionmaking,” 51 Ohio St. L.J. 375, 390 (1990); Stephen L. Isaacs & Renee Holt, “Drug regulation, product liability, and the contraceptive crunch: Choices are dwindling,” 8 J. Legal Med. 533 (1987); Neil Vidmar & Shari S. Diamond, “Juries and Expert Evidence,” 66 Brook. L. Rev. 1121, 1169-1170 (2001); Adil E. Shamoo, “Scientific evidence and the judicial system,” 4 Accountability in Research 21, 27 (1995); Michael S. Davidson, “The limitations of scientific testimony in chronic disease litigation,” 10 J. Am. Coll. Toxicol. 431, 435 (1991); Charles R. Nesson & Yochai Benkler, “Constitutional Hearsay: Requiring Foundational Testing and Corroboration under the Confrontation Clause,” 81 Virginia L. Rev. 149, 155 (1995); Stephen D. Sugarman, “The Need to Reform Personal Injury Law Leaving Scientific Disputes to Scientists,” 248 Science 823, 824 (1990); Jay P. Kesan, “A Critical Examination of the Post-Daubert Scientific Evidence Landscape,” 52 Food & Drug L. J. 225, 225 (1997); Ora Fred Harris, Jr., “Communicating the Hazards of Toxic Substance Exposure,” 39 J. Legal Ed. 97, 99 (1989) (“some seemingly horrendous decisions”); Ora Fred Harris, Jr., “Complex Product Design Litigation: A Need for More Capable Fact-Finders,” 79 Kentucky L. J. 510 & n.194 (1991) (“uninformed judicial decision”); Barry L. Shapiro & Marc S. Klein, “Epidemiology in the Courtroom: Anatomy of an Intellectual Embarrassment,” in Stanley A. Edlavitch, ed., Pharmacoepidemiology 87 (1989); Marc S. Klein, “Expert Testimony in Pharmaceutical Product Liability Actions,” 45 Food, Drug, Cosmetic L. J. 393, 410 (1990); Michael S. Lehv, “Medical Product Liability,” Ch. 39, in Sandy M. Sanbar & Marvin H. Firestone, eds., Legal Medicine 397, 397 (7th ed. 2007); R. Ryan Stoll, “A Question of Competence – Judicial Role in Regulation of Pharmaceuticals,” 45 Food, Drug, Cosmetic L. J. 279, 287 (1990); Note, “A Question of Competence: The Judicial Role in the Regulation of Pharmaceuticals,” Harvard L. Rev. 773, 781 (1990); Peter H. Schuck, “Multi-Culturalism Redux: Science, Law, and Politics,” 11 Yale L. & Policy Rev. 1, 13 (1993); Howard A. Denemark, “Improving Litigation Against Drug Manufacturers for Failure to Warn Against Possible Side Effects: Keeping Dubious Lawsuits from Driving Good Drugs off the Market,” 40 Case Western Reserve L. Rev. 413, 438-50 (1989-90); Howard A. Denemark, “The Search for Scientific Knowledge in Federal Courts in the Post-Frye Era: Refuting the Assertion that Law Seeks Justice While Science Seeks Truth,” 8 High Technology L. J. 235 (1993)
[15] Carl Cranor & Kurt Nutting, “Scientific and Legal Standards of Statistical Evidence in Toxic Tort and Discrimination Suits,” 9 Law & Philosophy 115, 123 (1990) (internal citations omitted).
[16] 131 S.Ct. 1309 (2011) [Matrixx]
[17] Id. at 1319.
[18] Baroldy v. Ortho Pharmaceutical Corp., 157 Ariz. 574, 583, 760 P.2d 574 (Ct. App. 1988); Earl v. Cryovac, A Div. of WR Grace, 115 Idaho 1087, 772 P. 2d 725, 733 (Ct. App. 1989); Rubanick v. Witco Chemical Corp., 242 N.J. Super. 36, 54, 576 A. 2d 4 (App. Div. 1990), aff’d in part, 125 N.J. 421, 442, 593 A. 2d 733 (1991); Minnesota Min. & Mfg. Co. v. Atterbury, 978 S.W. 2d 183, 193 n.7 (Tex. App. 1998); E.I. Dupont de Nemours v. Castillo ex rel. Castillo, 748 So. 2d 1108, 1120 (Fla. Dist. Ct. App. 2000); Bell v. Lollar, 791 N.E.2d 849, 854 (Ind. App. 2003; King v. Burlington Northern & Santa Fe Ry, 277 Neb. 203, 762 N.W.2d 24, 35 & n.16 (2009).
[19] City of Greenville v. WR Grace & Co., 827 F. 2d 975, 984 (4th Cir. 1987); American Home Products Corp. v. Johnson & Johnson, 672 F. Supp. 135, 142 (S.D.N.Y. 1987); Longmore v. Merrell Dow Pharms., Inc., 737 F. Supp. 1117, 1119 (D. Idaho 1990); Conde v. Velsicol Chemical Corp., 804 F. Supp. 972, 1019 (S.D. Ohio 1992); Joiner v. General Elec. Co., 864 F. Supp. 1310, 1322 (N.D. Ga. 1994) (which case ultimately ended up in the Supreme Court); Bowers v. Northern Telecom, Inc., 905 F. Supp. 1004, 1010 (N.D. Fla. 1995); Pick v. American Medical Systems, 958 F. Supp. 1151, 1158 (E.D. La. 1997); Baker v. Danek Medical, 35 F. Supp. 2d 875, 880 (N.D. Fla. 1998).
[20] Rider v. Sandoz Pharms. Corp., 295 F. 3d 1194, 1199 (11th Cir. 2002); Kilpatrick v. Breg, Inc., 613 F. 3d 1329, 1337 (11th Cir. 2010); Siharath v. Sandoz Pharms. Corp., 131 F. Supp. 2d 1347, 1359 (N.D. Ga. 2001); In re Meridia Prods. Liab. Litig., Case No. 5:02-CV-8000 (N.D. Ohio 2004); Henricksen v. ConocoPhillips Co., 605 F. Supp. 2d 1142, 1177 (E.D. Wash. 2009); Doe v. Northwestern Mutual Life Ins. Co., (D. S.C. 2012); In re Chantix (Varenicline) Prods. Liab. Litig., 889 F. Supp. 2d 1272, 1286, 1288, 1290 (N.D. Ala. 2012); Farmer v. Air & Liquid Systems Corp. at n.11 (M.D. Ga. 2018); In re Abilify Prods. Liab. Litig., 299 F. Supp. 3d 1291, 1306 (N.D. Fla. 2018).
[21] Michael D. Green, D. Michal Freedman & Leon Gordis, “Reference Guide on Epidemiology,” 549, 599 n.143, in Federal Judicial Center, National Research Council, Reference Manual on Scientific Evidence (3d ed. 2011).
