In recent years, federal courts have faced a wave of challenges to new and increasingly onerous state restrictions on abortion. Last July, in MKB Management Corp. v. Stenehjem,1×1. 795 F.3d 768 (8th Cir. 2015), cert. denied, 2016 WL 280836 (U.S. Jan. 25, 2016). the Eighth Circuit invalidated a North Dakota law banning abortion once a fetal heartbeat has been detected2×2. Id. at 770 (describing the North Dakota law). as incompatible with Supreme Court precedent established by Roe v. Wade.3×3. 410 U.S. 113 (1973). The court spent the bulk of its opinion, however, laying out what it saw as the many reasons to reevaluate the current standard governing the constitutionality of abortion restrictions in the United States. In doing so, the Eighth Circuit suggested — unnecessarily and on the basis of an inadequate record — that abortions may cause adverse consequences for women’s health and well-being. Though it is not unusual for appellate courts to opine on such generalized questions of fact, a more restrained approach would foster confidence in the judicial system by avoiding opinions that draw conclusions based on weak factual foundations.
In 1973, the Supreme Court’s landmark decision in Roe v. Wade recognized a fundamental right to abortion in the United States.4×4. Id. at 152–53. The Roe Court held that state restrictions on first-trimester abortions are unconstitutional deprivations of a woman’s liberty (specifically, her right of privacy) without due process of law under the Fourteenth Amendment.5×5. Id. at 153. The Roe Court acknowledged that the right might be alternatively grounded in the Ninth Amendment’s reservation of rights to the people. Id. Roe recognized that the state also has an interest “in protecting the potentiality of human life,”6×6. Id. at 162. and that this interest becomes “compelling” as a pregnancy progresses.7×7. Id. at 163. In certain circumstances, therefore, the state may regulate, and even proscribe, abortion when the regulation is narrowly tailored to serve a compelling government interest.8×8. Id. at 155. In Planned Parenthood of Southeastern Pennsylvania v. Casey,9×9. 505 U.S. 833 (1992). the Court affirmed the central holding of Roe that the point in a woman’s pregnancy at which the state’s interest in protecting the potentiality of human life becomes compelling is when the fetus becomes “viable.”10×10. Id. at 860. The Court has defined “viable” to mean capable of prolonged life outside the mother’s womb, “with or without artificial support.”11×11. Colautti v. Franklin, 439 U.S. 379, 388 (1979). Expert testimony in Stenehjem indicated that the current medical point of viability generally occurs twenty-four weeks into pregnancy.12×12. Stenehjem, 795 F.3d at 773.
For a number of years, North Dakota’s laws adhered to the standards articulated in Roe and Casey.13×13. See N.D. Cent. Code § 14-02.1-04(3) (2009) (prohibiting abortion “[a]fter the point in pregnancy when the fetus may reasonably be expected to have reached viability,” except when necessary to preserve the life or health of the mother). North Dakota defines viability as “the ability of a fetus to live outside the mother’s womb, albeit with artificial aid.” Id. § 14-02.1-02(10). In 2013, however, the North Dakota legislature passed House Bill 1456, a bill that would have banned abortion once a fetal heartbeat has been detected.14×14. Stenehjem, 795 F.3d at 770. This detection often occurs as early as six weeks into pregnancy.15×15. Id. at 771. Before House Bill 1456 took effect, Red River Women’s Clinic, the sole abortion provider in North Dakota, and its medical director, Dr. Kathryn Eggleston, brought suit in district court, challenging the law’s constitutionality and seeking injunctive relief from its enforcement.16×16. Id. at 770. The district court granted a preliminary injunction enjoining the implementation of the law.17×17. Id. The plaintiffs then moved for summary judgment, arguing that the law unconstitutionally infringed on a woman’s fundamental right to an abortion.18×18. Id. at 771. The State’s expert, Dr. Jerry Obritsch, argued that, contrary to the prevailing definition, fetal viability actually occurs at conception because in vitro fertilization (IVF) techniques “allow an embryonic unborn child to live outside the [womb] for 2–6 days after conception.”19×19. Id. If the fetus is technically “viable” at conception, the state would be permitted to ban abortion at virtually any point during pregnancy, and North Dakota’s heartbeat bill would clearly be constitutional.
The district court found that Dr. Obritsch’s argument did not raise a genuine issue of material fact because his proposed definition of “viability” differed from “the one used by either the United States Supreme Court or the medical community generally,” and was unsupported by “any medical literature.”20×20. MKB Mgmt. Corp. v. Burdick, 16 F. Supp. 3d 1059, 1073 (D.N.D. 2014). Since the bill “clearly prohibits pre-viability abortions in a very significant percentage of cases in North Dakota,”21×21. Id. at 1074. the district court granted summary judgment to the plaintiffs, permanently enjoining House Bill 1456.22×22. Id. at 1075. The State appealed.
The Eighth Circuit affirmed.23×23. Judge Shepherd wrote the opinion for a unanimous panel, which also included Judges Smith and Benton. As an initial matter, the court agreed with the State that, since Roe, the Supreme Court’s decisions on the subject of abortion have increasingly recognized “states’ profound interest in protecting unborn children.”24×24. Stenehjem, 795 F.3d at 771; see also id. at 771–72 (characterizing the Supreme Court’s doctrinal development from Roe to Casey to Gonzales v. Carhart, 550 U.S. 124 (2007), as moving toward giving states greater leeway in regulating abortion). The court observed that the Supreme Court’s decision in Gonzales v. Carhart,25×25. 550 U.S. 124 (2007). in which the majority chose to merely “assume” that Casey was correctly decided,26×26. Id. at 146. “may . . . signal the Court’s willingness to reevaluate its abortion jurisprudence.”27×27. Stenehjem, 795 F.3d at 772. In fact, this past November, the Court agreed to hear a case considering the constitutionality of a Texas abortion law. Whole Woman’s Health v. Cole, No. 15-274, 2015 WL 5176368, at *1 (U.S. Nov. 13, 2015). But since the Supreme Court “has yet to overrule” Roe,28×28. Stenehjem, 795 F.3d at 772. the court found itself tied to the Supreme Court’s current definition of viability and “bound by Supreme Court precedent holding that states may not prohibit pre-viability abortions.”29×29. Id. at 773.
While the Stenehjem court accepted that Supreme Court precedent dictated the outcome of the case, it went on to state that “good reasons exist for the [Supreme] Court to reevaluate its jurisprudence.”30×30. Id. First, the court claimed that the current fetal-viability standard “has proven unsatisfactory because it gives too little consideration to the ‘substantial state interest in potential life throughout pregnancy.’”31×31. Id. at 774 (quoting Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 876 (1992) (plurality opinion)). In the court’s view, the Supreme Court’s viability standard undermines states’ “interest in protecting unborn children” because it “removed the states’ ability to account for ‘advances in medical and scientific technology [that] have greatly expanded our knowledge of prenatal life.’”32×32. Id. (alteration in original) (quoting Hamilton v. Scott, 97 So. 3d 728, 746 (Ala. 2012) (Parker, J., concurring specially)).
The court next stated that “[a]nother reason for the [Supreme] Court to reevaluate its jurisprudence is that the facts underlying Roe and Casey may have changed.”33×33. Id. at 775. First, based on the declarations of several women who had undergone abortions, the court asserted that Roe’s assumption that the decision to have an abortion would be made in close consultation with a doctor was not necessarily true.34×34. Id. Second, the court cited the declaration of an obstetrician who claimed that coercion or pressure to have an abortion “occurs with frequency,” and the court recounted the story of one woman who faced being kicked out of her house by her husband if she did not have an abortion.35×35. Id. Third, the court noted that “abortions may cause adverse consequences for the woman’s health and well-being,” citing declarations from three different women to that effect.36×36. Id. The court also gave weight to a statement by Dr. Obritsch, who claimed that some studies support a connection between abortion and breast cancer.37×37. Id. Fourth, the court found it significant that the plaintiffs in two of the Supreme Court’s foundational abortion cases — Norma McCorvey, the “Roe” of Roe v. Wade, and Sandra Cano, the “Doe” of Doe v. Bolton38×38. 410 U.S. 179 (1973). — later advocated against those decisions.39×39. Stenehjem, 795 F.3d at 775–76. And finally, the court observed that a North Dakota law that permits parents to abandon unwanted children at hospitals without consequence has reduced the burdens associated with childcare identified in Roe.40×40. Id. at 776.
For these reasons, the court ultimately concluded that the current standard “discounts the legislative branch’s recognized interest in protecting unborn children,” and urged the Supreme Court to reconsider its jurisprudence on the issue.41×41. Id. In so doing, the Eighth Circuit weighed in on a generalized factual issue — asserting that “abortions may cause adverse consequences for [a] woman’s health and well-being”42×42. Id. at 775. — unnecessarily and on the basis of an inadequate record. While this practice is not uncommon, judicial efforts to avoid opinions that rest on weak factual foundations, even in dicta, would reduce the danger of such unreliable information misleading both other courts and the wider public.
The United States’ commitment to adversarial justice is a defining feature of its legal system.43×43. See Brianne J. Gorod, The Adversarial Myth: Appellate Court Extra-Record Factfinding, 61 Duke L.J. 1, 3 (2011) (“[T]he U.S. legal system’s commitment to adversarial justice derives from the belief that adversarial testing is the surest route to truth.”). The adversarial system was crafted for “adjudicative facts,” that is, facts about the particular circumstances of the case.44×44. Id. at 38. Yet court decisions often turn on “legislative facts” — general factual questions that are not limited to any specific case.45×45. Allison Orr Larsen, Confronting Supreme Court Fact Finding, 98 Va. L. Rev. 1255, 1255–57 (2012). In cases that turn on legislative facts, there are very few procedures “designed to ensure the quality of the sources on which the courts rely or to ensure that all views are adequately tested.”46×46. Gorod, supra note 43, at 45. As one scholar observes, “there is a real danger when judges, inexperienced in making empirical judgments and unrestrained in how they do so, are forced to make factual determinations that are highly contestable and ideologically laden without any guidelines.”47×47. Id. at 49–50. For instance, in Stenehjem, had the court accepted North Dakota’s invitation to reconsider Roe and Casey, one factual question relevant to the case would have been whether abortions cause adverse consequences for women’s health and well-being — precisely the type of factfinding that judges may be underequipped to perform.
Within the adversarial process, expert testimony has long played an important role in establishing the foundation of legislative facts.48×48. See id. at 59 n.257; see also Fed. R. Evid. 702 advisory committee’s note (1972 Proposed Rules) (“An intelligent evaluation of facts is often difficult or impossible without the application of some scientific, technical, or other specialized knowledge. The most common source of this knowledge is the expert witness . . . .”). See generally Learned Hand, Historical and Practical Considerations Regarding Expert Testimony, 15 Harv. L. Rev. 40 (1901) (discussing development of the use of expert witnesses). As Justice Scalia once pointed out, “[a]n adversarial process in the trial courts can identify flaws in the methodology of the studies that the parties put forward.”49×49. Sykes v. United States, 131 S. Ct. 2267, 2286 (2011) (Scalia, J., dissenting), overruled by Johnson v. United States, 135 S. Ct. 2551 (2015). However, concerns about “junk science” and fears that jurors would be unable to appropriately weigh such evidence highlighted a need for a reliability standard addressing the admissibility of expert testimony.50×50. See Cassandra H. Welch, Note, Flexible Standards, Deferential Review: Daubert’s Legacy of Confusion, 29 Harv. J.L. & Pub. Pol’y 1085, 1085 (2006); see also Kenneth R. Foster & Peter W. Huber, Judging Science 16–19 (1997) (exploring situations where scientific evidence may be deemed “reliable”). In 1993, the Supreme Court responded to these concerns in Daubert v. Merrell Dow Pharmaceuticals, Inc.,51×51. 509 U.S. 579 (1993). which envisioned a “gatekeeping role” for federal judges in order to screen out unreliable expert testimony.52×52. Id. at 597; see also Foster & Huber, supra note 50, at 15. The Court found that scientific evidence must have “a grounding in the methods and procedures of science” and must be “more than subjective belief or unsupported speculation.”53×53. Daubert, 509 U.S. at 590. The holding of Daubert was later incorporated into Rule 702 of the Federal Rules of Evidence. This rule now states that witnesses may testify to scientific evidence only if “the testimony is based on sufficient facts or data; the testimony is the product of reliable principles and methods; and the [witness] has reliably applied the principles and methods to the facts of the case.”54×54. Fed. R. Evid. 702. These restrictions on admitted evidence serve to protect parties from arbitrary judgments based on unfounded, discredited, or unscientific evidence.
Because Stenehjem was decided on summary judgment after only limited discovery, the record on the question of abortions’ effects on women’s health had not been fully developed.55×55. See MKB Mgmt. Corp. v. Burdick, 16 F. Supp. 3d 1059, 1061 (D.N.D. 2014). Nonetheless, even though it had resolved the straightforward legal question before it for the plaintiffs, the Eighth Circuit took the opportunity, in dicta, to weigh in on that generalized factual question. In doing so, the court included and relied on testimony from the record: Dr. Obritsch’s testimony about the link between abortion and breast cancer and anecdotal testimony from women who had had abortions.56×56. See Stenehjem, 795 F.3d at 775. The court should have refrained from opining on a generalized factual question on the basis of such incomplete sources.
First, the court made its factual assertions on the basis of a scant record that had not been developed through the adversarial process. The lower court found in favor of the plaintiffs on summary judgment, before the case had gone to trial.57×57. Burdick, 16 F. Supp. 3d at 1060. Because the plaintiffs could win on summary judgment simply by showing that North Dakota’s expert witness had used a definition of viability that was not consistent with the Supreme Court’s definition, they had no reason to rebut North Dakota’s assertions about the adverse health effects of abortion. Moreover, courts typically do not make Daubert rulings on expert testimony until trial.58×58. See Cortés-Irizarry v. Corporación Insular de Seguros, 111 F.3d 184, 188–89 (1st Cir. 1997) (arguing that the Daubert inquiry is “best performed by trial judges,” id. at 189). Thus, the Eighth Circuit used an undeveloped record and potentially unreliable testimony to call for reconsideration of a forty-year-old precedent.
Second, although the court’s comment on what it saw as the changed circumstances since Roe was only in dicta, dicta have a tendency to leak into subsequent briefs and judicial opinions as though they were good law.59×59. See Judith M. Stinson, Why Dicta Becomes Holding and Why It Matters, 76 Brook. L. Rev. 219, 221 (2010) (“[T]oo often lawyers argue for, and judges treat, extraneous statements made in a prior case — that is, dicta — as holding.”). If dicta are treated as the court’s holding, there is a danger that other courts, lawyers, and the public will treat the “scientific facts” asserted in dicta similarly to facts that, having been subjected to evidentiary safeguards, can be assumed to be more reliable. Moreover, on issues of great social interest, “dicta are usually the most quoted language by popular media outlets and non-legal sources, are often the most remembered part of the case, and are thematically opportunistic in guiding the debate and framing the narrative for use in future cases.”60×60. Stacy A. Scaldo, Deadly Dicta: Roe’s “Unwanted Motherhood,” Carhart II’s “Women’s Regret,” and the Shifting Narrative of Abortion Jurisprudence, 6 Drexel L. Rev. 87, 90 (2013). Indeed, most of the media coverage of Stenehjem has focused on the dicta rather than the holding, in some instances without questioning its science.61×61. See, e.g., Jonathan H. Adler, Appeals Court Urges Supreme Court to Revisit Constitutional Limits on State Abortion Restrictions, Wash. Post: Volokh Conspiracy (July 25, 2015), http://www.washingtonpost.com/news/volokh-conspiracy/wp/2015/07/25/appeals-court-urges-supreme-court-to-revisit-constitutional-limits-on-state-abortion-restrictions [http://perma.cc/RUF2-2YPC]; Ed Whelan, Eighth Circuit Calls for Supreme Court to Reconsider Abortion Precedents, Nat’l Rev. Online: Bench Memos (July 23, 2015, 9:56 AM), http://www.nationalreview.com/bench-memos/421515/eighth-circuit-calls-supreme-court-reconsider-abortion-precedents-ed-whelan [http://perma.cc/S6D2-SNNJ]. Ultimately, dicta can impact public discourse and future judicial opinions, despite not reflecting the holding itself or being subject to the same evidentiary standards.
The Eighth Circuit’s speculations in Stenehjem — namely, its assertions and statements of fact in dicta based on an undeveloped and potentially unreliable record — are not an uncommon practice among appellate courts. But the Eighth Circuit’s reasoning here is particularly concerning for two reasons. First, the court cited the record selectively, with no mention of information that cut the other way. For example, the Stenehjem court referred to Dr. Obritsch’s claim that some studies support a connection between abortion and breast cancer without acknowledging the contradictory evidence from the amicus brief submitted by the American College of Obstetricians and Gynecologists and Physicians for Reproductive Health.62×62. Compare Stenehjem, 795 F.3d at 775, with Amici Curiae Brief of American College of Obstetricians and Gynecologists and Physicians for Reproductive Health in Support of Plaintiffs-Appellees, Stenehjem, 795 F.3d 768 (No. 14-2128) [hereinafter ACOG Amici Brief]. This amicus brief argued that the majority of the studies suggesting a link between abortion and breast cancer have “been rejected due to methodological problems . . . . In contrast, exhaustive review by panels convened by the American and British governments, including one involving more than 100 of the world’s leading experts at the U.S. National Cancer Institute’s 2003 workshop, have consistently found no association between abortion and breast cancer.”63×63. ACOG Amici Brief, supra note 62, at 22–23 (citations omitted). Yet rather than provide a complete survey of the pertinent information in the record, the court cited only to Dr. Obritsch’s opinion, “[d]espite overwhelming medical evidence to the contrary.”64×64. Brief of Amicus Curiae Program for the Study of Reproductive Justice — Information Society Project at the Yale Law School at 2, Stenehjem, 795 F.3d 768 (No. 14-2128).
Second, the court supported its sweeping statement about the possible adverse health consequences of abortion with anecdotes — declarations from three women who claimed to have experienced adverse side effects.65×65. Stenehjem, 795 F.3d at 775. In the absence of any scientific medical study, it is impossible to say that the medical conditions the women experienced were causally linked to their abortions. In this case, the nature of the proceeding allowed the court to present these anecdotes as scientific fact without the benefit of an adversarial challenge.
Other courts have also used this troubling practice in abortion cases. In Gonzales, the Supreme Court case addressing the constitutionality of a ban on partial-birth abortion, the Court drew conclusions from one amicus brief to assert that women who have abortions consequently suffer from “[s]evere depression and loss of esteem.”66×66. 550 U.S. 124, 159 (2007). Yet as one scholar points out, “[t]he Court, of course, did not actually have the opportunity to hear these women testify, nor was there adequate opportunity for the other side to give the Court a sense of how representative the women who had signed that amicus brief were.”67×67. Gorod, supra note 43, at 32–33. In her dissent, Justice Ginsburg argued that the majority had no reliable evidence for their claim.68×68. Gonzales, 550 U.S. at 183 & n.7 (Ginsburg, J., dissenting). To back up her own assertion, Justice Ginsburg cited ten different “fact-based authorities — from medical journals to New York Times articles to briefs from the American Psychological Association.”69×69. Allison Orr Larsen, The Trouble with Amicus Facts, 100 Va. L. Rev. 1757, 1777 (2014) (citing Gonzales, 550 U.S. at 183 n.7 (Ginsburg, J., dissenting)). The Eighth Circuit’s unscientific discussion repeats the Gonzales majority’s similarly troubling use of anecdotes in lieu of reliable scientific fact.
In Stenehjem, the court swiftly and efficiently resolved the legal issue at hand by holding that North Dakota’s “heartbeat bill” was clearly unconstitutional in light of Roe v. Wade. However, the nature of the proceeding allowed the Eighth Circuit to weigh in on a “legislative fact” unnecessarily and on the basis of an inadequate record. Though this troubling practice is certainly not unusual, a more restrained approach by appellate courts to offering dicta based on general, unlitigated facts would foster greater confidence in the judicial system by avoiding opinions that rest on weak and untested factual foundations.