Jeffrey Bossert Clark*
The Federalist Society's three E.L. Wiegand Practice Groups in
Administrative Law & Regulation, Environmental Law & Property
Rights, and Litigation held a colloquium entitled "Junk Science,
the Courts, and the Regulatory State" on July 10, 1997 at the
University Club in Washington, D.C. The five participants in the
morning session of the colloquium focused on the interplay between
risk regulation in administrative agencies and risk regulation through
the tort system, while the two participants in the afternoon session
examined the question of who should decide scientific questions
in the toxic-tort context -- judges, juries, or expert panels. Peter
Huber, one of the pioneers of the attack on the use of "junk
science" in the courtroom and a partner in the law firm of
Kellogg, Huber, as well as a Senior Fellow of the Manhattan Institute,
delivered a luncheon address exploring the meaning of the evidentiary
tests for true "science" announced by the Supreme Court
in its landmark case, Daubert v. Merrell Dow Pharm., Inc., 509 U.S.
Panelists in both sessions, including a representative of the plaintiffs'
bar, seemed to agree that "junk science" was to some degree
a problem. In fact, Peter Huber went so far as to claim that the
battle to establish the need to eradicate "junk science"
from the courts had been won in Daubert and that the principal task
now remaining undone was to work out how to apply Daubert properly.
Most panelists, however, coalesced around the idea that much more
needed to be done to solve the vexing problems posed by "junk
science." Edward Warren, a partner in the law firm of Kirkland
& Ellis and a participant in the morning session, captured this
view when he noted that Daubert seemed a second-best solution to
a recurring legal and policy problem demanding bolder action. Panelists
diverged on exactly how the regime of legal rules governing the
use of science in the courtroom and in the halls of administrative
agencies should be improved, although a number of potentially constructive
solutions were advanced. These potential solutions ranged from toughening
judicial review of agency decisionmaking by cross-applying Daubert
in the administrative context, to reforming the rules in other problematic
areas of the law such as class actions, punitive damages, and discovery,
to taking the scientific fact-finding powers away from juries almost
entirely by instructing juries in scientific facts as they are now
instructed in the law.
The four-hour colloquium was brisk, intellectually rigorous, and
even when there were disagreements, conducted in an atmosphere of
collegiality. As a colleague sitting near me throughout the colloquium
remarked, the "junk science" program was conducted on
too high a plane and was far too much fun to warrant continuing-legal-education
(CLE) credit. But it should come as no surprise to Federalist Society
members that its programs are of a consistently high quality and
are much more provocative than standard bar fare.
What follows is Part I of a two-part series of articles.
The morning session began with a panel entitled "Junk Science:
The Interplay Between Risk Regulation and the Torts System."
Panelists included: Arthur Bryant, Executive Director, Trial Lawyers
for Public Justice; Dr. George Ehrlich, University of Pennsylvania
Medical School; Alan Raul, Sidley & Austin, Chairman of the
E.L. Wiegand Practice Group in Environmental Law & Property
Rights; Edward Warren, Kirkland & Ellis, Chairman of the E.L.
Wiegand Practice Group in Administrative Law & Regulation; James
Gauch, as moderator, Jones, Day, Reavis & Pogue, Vice Chairman
of Programs for the E.L. Wiegand Practice Group in Administrative
Law & Regulation.
James Gauch introduced the morning's panelists and noted that it
was his intention to place the spotlight on neglected topics --
should courts exercise the same gatekeeping function over the uses
of science by administrative agencies that they now serve over science
in the courtroom, and is science consumed by agencies in the same
way it is consumed by the courts? Gauch began by introducing Edward
Warren and commending to the audience Warren's article bearing on
these questions -- Judge Leventhal's Revenge: The Courts as "Gatekeepers"
of "Good Science" After Daubert, 1994 Pub. Int. L. Rev.
93 (1994) (arguing that Daubert was similar to former D.C. Circuit
Judge Harold Leventhal's "hard-look" brand of judicial
review of technical and scientific agency decisionmaking). Gauch
then noted that Dr. George Ehrlich has been an advisor to the FDA
and consultant to the manufacturers of breast implants. Gauch said
Alan Raul planned to focus on relevant lessons from tobacco-related
regulation. Lastly, Gauch introduced Arthur Bryant as a champion
of the plaintiffs' bar who would do his best to rebut what other
panelists would say.
Alan Raul, partner at Sidley & Austin and Chairman of the practice
group in Environmental Law & Property Rights, made the first
presentation. Holding up his pocket computer, Raul joked that it
was only a matter of time before he became a member of a class action
involving radio emissions or carpal-tunnel syndrome. Throughout
his presentation, Raul peppered his analysis with examples drawn
from his experience with the regulatory treatment of and science
surrounding so-called "environmental tobacco smoke" (ETS)
or "second-hand smoke." Raul's thesis was that Daubert
should be applied "actively and aggressively" by federal
judges reviewing the scientific underpinnings of agency action.
Acknowledging the "[anti-]democratic risk" of his approach,
Raul cautioned that he was not advocating the substitution by judges
of their own preferences for those of Congress or for the preferences
of agencies legitimately delegated lawmaking authority. What he
was advocating and what deeply troubled him was the tendency he
perceived for agencies to make rules and other decisions based on
various assumptions, default principles, and "fudge factors"
that often remain undisclosed to the regulated public. According
to Raul, federal judges should follow the example of Daubert in
the tort context by excising such unexamined and unscientific regulatory
assumptions with the scalpel of searching judicial review.
Raul recommended that the audience read Wendy E. Wagner's article,
The Science Charade in Toxic Risk Regulation, 95 Colum. L. Rev.
1613 (1995). Though it approached the issues posed by regulatory
science from a liberal perspective, Raul argued that the article's
observation that agencies were deliberately obfuscating their modus
operandi by calling it "science" rather than naked "policymaking"
was correct and should be taken to heart. Raul apparently parts
company with Wagner's follow-up argument that agencies shouldn't
have to pretend and thus should be allowed to make policy openly.
In Raul's view, the problem in the regulatory context in most cases
isn't "junk science," it's "junk policy." The
science is fine; it's what the agency does with the science that's
invalid. In the ETS context, for example, perfectly valid epidemiologic
studies establish that ETS increases the risk of certain diseases
by a factor of 1.19. According to Raul, however, epidemiologists
are in general agreement that such a factor would have to exceed
at least 2 and possibly 3 before there was any cause for alarm,
yet ETS studies far below that threshold are currently being cited
by regulators to justify administrative action.
In Raul's view, agencies regulating on the basis of unarticulated
assumptions become essentially a "farm team" for the plaintiffs'
bar. Whenever an agency takes action or does a study suggesting
that a product or service causes harm and there is a "deep
pocket" anywhere in the vicinity, litigation is inevitable.
For support, Raul ticked off the examples of Love Canal, the "junk
science" behind banning asbestos in building materials, the
baselessness of breast-implant litigation, and the dioxin scare.
To his recommended reading list Raul added a publication by the
American Council on Science and Health, Facts vs. Fears, which reviews
the twenty greatest modern health scares perpetrated in our country.
To solve the problems he identified, Raul advocated using the Daubert-like
approach to judicial review of regulation described above, an approach
that Raul believes Judge Leventhal would have applauded. Unfortunately,
Raul noted that the only court to address this question explicitly,
the Seventh Circuit, has rejected a similar argument. See Sierra
Club v. Marita, 46 F.3d 606, 621-22 (7th Cir. 1995) ("While
such a proposal might assure better documentation of an agency's
scientific decisions, we think that forcing an agency to make such
a showing as a general rule is intrusive, undeferential, and not
The next to speak was Dr. George Ehrlich from the University of
Pennsylvania Medical School. Dr. Ehrlich explained that he was a
man of "strong opinions," as reflected in his television
appearances concerning breast-implant litigation and science. He
noted that it is inevitable that as a new medical product or service
becomes more widely used in society, the segment of the population
using the new product or service will begin to show some incidence
of the rare diseases that manifest themselves in the population
at large. Breast implants are only one example of this phenomenon.
Dr. Ehrlich suggested that the recent uproar over the weight-loss
drug fen-phen was another.
Given this rather obvious statistical fact, Dr. Ehrlich posed the
question of how it is that widespread claims of rare diseases being
caused by medical products or services are taken seriously in the
courts and by doctors, even though the science supporting such claims
is flimsy or nonexistent. He believes the answer lies in "Bergson's
fallacy." Dr. Ehrlich explained that Bergson was a statistician
at the Mayo Clinic who was often being asked by his doctor colleagues
how it could be that they were seeing more and more patients with
rare diseases unless the incidence of such diseases was truly increasing.
The answer, Bergson explained, was that the doctors at the Mayo
Clinic were super-specialists. Patients with rare diseases were
concentrated at the Mayo Clinic because of its world-renowned reputation.
(Psychologists Kahneman and Tversky have labeled the phenomenon
also identified by Bergson the "availability heuristic"
-- people tend to generalize inappropriately from what is common
or rare in their own experience (including media reports) to conclusions
about what is common or rare in the world as a whole. See Amos Tversky
& Daniel Kahneman, Availability: A Heuristic for Judging Frequency
and Probability, in Daniel Kahneman, et al., eds., Judgment under
Uncertainty: Heuristics and Biases 166 (Cambridge 1982).) As an
example, Dr. Ehrlich pointed out that many doctors practicing on
the West Coast of Florida became convinced that breast implants
were causing various health problems because they were seeing a
lot of women with breast implants who had such problems. Of course,
those doctors had fallen into Bergson's fallacy by failing to consider
that breast implants were especially common in that part of the
country and that women conscious of their appearances enough to
obtain implants were more likely to search out doctors for any health
problems they were experiencing.
Further elaborating on the example of breast implants, Dr. Ehrlich
explained that the mine run of epidemiologic studies demonstrate
consistently that breast implants do not cause diseases such as
scleroderma or rheumatoid arthritis. (Dr. Ehrlich is a rheumatologist.)
Thus, breast-implant plaintiffs have resorted to arguing that they
suffer from what Dr. Ehrlich called "fake" illnesses,
such as "atypical connective-tissue disease." The problem
with these diseases is that they are non-falsifiable because they
rely on subjective expressions of pain -- that is why the list of
symptoms for these kinds of "diseases" is at 150 and growing.
In the same category Dr. Ehrlich put other so-called diseases such
as "fibromyalgia," "chemical sensitivity syndrome,"
and "repetitive-strain syndrome."
In contrast to Raul's observation that regulatory action tends
to spur litigation, Dr. Ehrlich seemed more concerned that agency
mandates have expanded because of action by an aggressive plaintiffs'
bar. He described how the Food and Drug Administration's (FDA's)
powers to regulate have expanded from the power to ensure safety
to include powers to ensure efficacy. In Dr. Ehrlich's view, these
new powers give regulators the ability to impose the impossible
burden on manufacturers to prove scientifically that their products
are absolutely safe.
Dr. Ehrlich also argued that the proliferation of spurious science
is not entirely the fault of plaintiffs themselves. Dr. Ehrlich
lays blame at the feet of both plaintiffs' attorneys and doctors.
Many of the plaintiffs in breast-implant cases are proceeding in
good faith, according to Dr. Ehrlich -- it's simply that their doctors
and lawyers have convinced them that inside their breasts wait ticking
time bombs and therefore that they should sue now before the inevitable
Moderator James Gauch next introduced Edward Warren, a partner
at Kirkland & Ellis and Chairman of the Administrative Law &
Regulation practice group. Borrowing the thesis of an obscure article
written at the turn of the century in the Albany Medical Journal,
Warren argued that Daubert was a second-best solution to a very
old legal problem. Turns out that the article was written by none
other than the famed jurist Learned Hand and reprinted shortly thereafter
in the Harvard Law Review. See Historical and Practical Considerations
Regarding Expert Testimony, 15 Harv. L. Rev. 40 (1902). In Warren's
view, Judge Hand's analysis of the problems posed by expert scientific
testimony were "prescient" and so fresh they "could
have been written yesterday."
In his article, Judge Hand analyzed a long line of common-law decisions
to make the point that it was an anomaly in Anglo-Saxon jurisprudence
for witnesses to be allowed to testify to opinions. The liberal
treatment of experts in this regard was thus an exception to that
general rule. Judge Hand thought that this exception was totally
unwarranted. The usual trope advanced to justify this exception
even in Judge Hand's day was that juries were incapable of applying
scientific learning to pure facts to draw valid inferences because
such learning was outside their experience or maybe beyond their
ken. But Judge Hand emphasized that this same problem also clearly
counseled against allowing juries to weigh the opinions of conflicting
experts at all. The solution to the problem of jury incompetence
in this area, according to Judge Hand, was to treat scientific knowledge
in the same way courts are accustomed to treating something else
universally acknowledged to be beyond the jury's powers -- the law.
Thus, juries should be instructed about scientific conclusions as
if they were law. In Judge Hand's schema, either judges or expert
panels of neutral scientists would be tasked with crafting the "science
instructions" in a particular case. Warren subscribes fully
to Judge Hand's view as the "first-best" approach that
is correct as a matter of logic and law.
Turning from the ideal world to the existing one dominated by Daubert's
schema, Warren noted that the lower federal courts have by and large
applied Daubert faithfully. The most glaring exception to that trend
has been the Eleventh Circuit's decision in Joiner v. General Elec.
Co., 78 F.3d 524 (11th Cir. 1996) (Barkett, J.), cert. granted,
117 S. Ct. 1243 (1997), a case in which the plaintiff alleged that
his lung cancer was caused by PCBs and the court reasoned, despite
a careful district court opinion to the contrary below, that two
mouse studies and the mere credentials of the plaintiff's experts
were enough to allow the case to go to a jury. The Supreme Court
has granted certiorari in Joiner, and in Warren's view, the Court
did not take that step with an eye to affirm. Therefore, Warren
expects that Daubert will be strengthened in some way or confirmed
in the Court's next term. Alternatively, the Court could choose
to use the case as a vehicle to explore the powers of the courts
of appeals to reverse evidentiary rulings excluding experts, but
at the very least that approach would leave Daubert intact.
Expanding on the analysis he advanced in his piece in the Public
Interest Law Review, Warren next explored the differences between
how science is used by agencies and how science is used by courts.
By contrast to Raul, Warren thinks that agencies should be given
a wider latitude in their use of science than courts. This is because
it is the task of agencies to regulate in a forward-looking, prophylactic
manner, while courts are tasked with deciding individual cases under
the backward-looking standard of whether a plaintiff can show that
his particular injury was more likely than not caused by a defendant's
actions. Warren then expressed his view that Judge Alex Kozinski
properly emphasized the point on remand in Daubert that the "more
likely than not" standard is a "pretty tough test."
See Daubert v. Merrell Dow Pharm., Inc., 43 F.3d 1311 (9th Cir.),
cert. denied, 116 S. Ct. 189 (1995). Because agencies use science
in a different way than courts, Warren concluded by arguing that
courts should not defer to agencies whenever agencies decide that
the risks warrant taking administrative action. On the other hand,
if an agency, with the broader scope of action entrusted to it,
decides not to regulate, then courts should in most cases defer
to the expert agency's determination and thus block lawsuits running
contrary to such an agency's effective determination that a product
or service is safe.
The last panelist to make a presentation in the morning was Arthur
Bryant, Executive Director of Trial Lawyers for Public Justice.
Bryant joked that he felt like the man invited to a barbecue only
to find out that he was the main course. Bryant predicted that the
audience would be shocked by his exclamation: "Ich bin ein
Federalist." Bryant chastised other panel members for advocating
an expanded role for federal judges and agencies to control "junk
science" as anti-Federalist and argued that because of the
more liberal treatment trial lawyers champion for expert testimony
in the law they are actually more in line with Federalist principles
than the Federalist Society.
The theme of Bryant's presentation was that there were no easy
answers -- each case had to be decided on its facts and each agency
possessed unique problems and capabilities that needed to be considered.
Bryant also argued that the problem of "junk science"
cuts both ways -- that corporate defendants often deploy "junk
science" in order to ward off valid claims. He pointed to a
Title IX case that he recently litigated against Brown University
in which Brown attempted to cut its female gymnastics and volleyball
programs costing about $60,000 annually. Brown paid $100,000, however,
to commission a study designed to show that men are generally more
interested in participating in college-sports programs than women.
In Bryant's view, that study was a prime example of "junk science."
Continuing with his "complexity" theme, Bryant asked whether
defendant tobacco manufacturers would agree with him that Daubert
bars them from presenting an expert to testify before a jury that
smoking does not cause lung cancer, despite the current scientific
consensus to the contrary.
According to Bryant, there are two reasons why plaintiffs are commonly
perceived as being more set back by Daubert than defendants: (1)
plaintiffs have the burden of proof, and if all scientific evidence
is excluded when scientific evidence is in fact necessary to establish
liability, then plaintiffs obviously lose; and (2) Peter Huber successfully
framed the issue this way in his book, Galileo's Revenge: Junk Science
in the Courtroom (Basic Books 1991). Likening Huber's method to
the Spanish Inquisition (proving only what it set out to prove),
Bryant accused Huber of ironically providing no scientific basis
for the claims he made in that book. For Bryant, Huber has claimed
the commanding heights on this issue only because of a vigorous
public-relations campaign to promote the book conducted by the Manhattan
Pushing his oversimplification thesis, Bryant argued that "eggshell"
plaintiffs deserve protection by our tort system, and that some
breast implants cause some diseases, but not others. He explained
his point that not all agencies are created equal by pointing to
the example of the FDA, which lacks subpoena powers and thus was
apprised of some of the evidence available to buttress claims of
breast-implant risk only by plaintiffs' lawyers, sometimes in violation
of judicial protective orders. In Bryant's view, truth isn't absolutely
knowable and unchangeable and thus, while cases must be decided
at a specific point in time, courts should never crystallize the
prevailing view of mainstream science in the law. To support this
claim he relied on the Supreme Court's rejection of the asbestos
class-action settlement in Amchem Prods., Inc. v. Windsor, No. 96-270,
1997 WL 345149 (June 25, 1997), where a district court was reversed
for binding future asbestos claimants to the settlement, despite
the possibility that their claims might be different from those
of current plaintiffs and that the ability of future plaintiffs
to prove their claims might similarly be different in light of subsequent
advances in science. Bryant also noted that many of the proposals
for reform advanced by other panelists relied on giving a freer
rein to judges, though in some cases the judiciary inspires less
confidence than juries. Summing up, Bryant cautioned everyone to
"be more skeptical" of attacks on "junk science"
because "science is about as an efficient search for truth
as the legal system is an efficient search for justice."
When initial morning presentations concluded, Gauch gave each of
the panelists a chance to react to their fellow panelists' arguments.
Raul began by turning around Bryant's point about public relations.
He argued that in reality the so-called public interest groups have
been far more effective in playing the public-relations game than
those of Peter Huber's persuasion. He pointed out that a single
person can place a call to the Larry King Live cable-television
program claiming that his wife died of brain cancer because she
frequently used a cellular telephone and soon there is a national
panic. Raul did give credit to agencies here in rejecting calls
to regulate cellular telephones on the ground their use causes physical
harm, however. On the whole, though, Raul thought that Bryant had
been quite reasonable, noting that both sides in the debate are
sometimes prone to oversimplification. In reality, subtle questions,
not easy questions, are involved in this issue.
Dr. Ehrlich agreed with Bryant's point that courts should never
freeze current scientific views into the law, quoting Captain Cook's
quip that "There are no black swans until you encounter your
first one." He disagreed with Bryant's point that there are
two sides to science, however. He argued that there is always only
one side that is currently supportable and, thus, the other side
must be presumed to engage in speculation. Case reports and the
like can provide useful signals that the current orthodoxy should
be changed, but case studies alone cannot be the basis for doing
so. He referenced the principles of bacteriology that maintain that
even epidemiologic evidence not be taken as conclusive until medical
science has established the causal links in a chain operating at
the cellular level.
Returning to the subject of breast implants, Dr. Ehrlich pointed
out that the types of silicone used in such implants are safe and
inert. In fact, silicone is used to coat needles and to make artificial
limbs, in pacemakers, in devices like Norplant, and even in anti-flatulence
drugs and in breakfast cereals. The only silicone ever shown to
cause health problems according to Dr. Ehrlich is a type of silicone
that Japanese prostitutes injected into their breasts in the aftermath
Responding to criticisms on federalism grounds of his recommendations
for reforming the problems associated with "junk science,"
especially to his arguments that courts should defer to agencies
that decide not to regulate, Warren indicated that he was not at
that point arguing for the preemption of state lawsuits, merely
that a flexible principle of deference should be voluntarily recognized
and applied. Warren also stated his opinion that "junk science"
is the symptom of a much larger problem and not the cause. In this
vein he argued that there is too much dual regulation between the
federal agencies and the tort system and thus that Congress should
explicitly preempt more tort law. (The implementation of Judge Hand's
solution would also require legislative action.) He applauded the
Supreme Court's recent ruling in Metro-North Commuter R.R. Co. v.
Buckley, No. 96-320, 1997 WL 338550, (June 23, 1997) (rejecting
a fear-of-cancer tort under the Federal Employers' Liability Act
(FELA)). Like Bryant, Warren also referred to the Supreme Court's
rejection of the asbestos class action in Windsor, but as an example
of the rampant abuse of the rules of civil procedure by plaintiffs.
Finally, Warren decried unbelievable punitive-damage awards and
the rise of a "discovery tort" used by the plaintiffs'
bar to shift the focus from the reality of cases of dubious scientific
merit to alleged misconduct by products-liability defendants, who
are often crushed by oppressive discovery requests in multiple fora.
Bryant responded to his critics by noting that there was little
disagreement with his "oversimplification" thesis. He
then recounted how the first breast implant case settled for a sizable
sum in exchange for a sealing of the record, suggesting that "there
must have been something there." Dr. Ehrlich couldn't help
but exclaim that it's often cheaper for defendants to settle than
to litigate. Bryant then moved to a different subject, agreeing
with Warren's claim that science is and should be used in different
ways in the regulatory and judicial contexts. In Bryant's words,
"the agencies work wholesale, while the courts work retail."
He took issue, however, with Dr. Ehrlich's point that there is only
one side to science. Bryant said the scientists he talks to tell
him that there is plenty of room for disagreement on many scientific
questions. Finally, Bryant attacked Warren's claim that courts should
defer to agencies when they decide not to regulate. He suggested
that embedded within any such argument is an ideological assumption
that the agencies always do their best to assert that a potentially
regulable product or service causes harm. He did not dispute that
deference was appropriate in some cases, but argued that plaintiffs
should be able to present evidence to a jury that an agency decision
not to regulate was caused by a lack of information or by political
The first question from the audience was put to Raul and focused
on whether institutional pressures creating a "flight from
science and reason" turned too many scientists into cowards.
Raul acknowledged that it is dangerous for a scientist to be caught
outside the mainstream -- that even scientists can fall prey to
"political correctness" because they fear losing the right
to compete on a level playing field for future grants. Raul was
considerably more sanguine than the questioner that good science
could win out, however, because the light of full disclosure is
a powerful medicine. He pointed to the example of the Congressional
Research Service's unmasking of the fact that EPA reduced the standards
for statistical significance when reviewing the studies on ETS.
Warren primarily fielded a question arguing that it was ironic
for Federalists to be advocating giving judges more power in order
to solve the problem of "junk science." Warren responded
that, as Judge Hand had recognized, allowing expert witnesses to
testify to opinions is a rule at war with our legal tradition. Therefore,
any qualms Federalists have with fixing the problems of "junk
science" by strengthening the role of the judiciary operates
from an incorrect legal baseline. The best solution to the problem
of "junk science" is for courts to impanel expert advisory
panels in Warren's view. (As additional support for Warren's argument
that Judge Hand's solution is not radical, consider the fact that
Lord Mansfield, to cite the practice of only one eminent common-law
judge, convened expert juries to address complex questions arising
under the commercial law. See 1 James Oldham, The Mansfield Manuscripts
and the Growth of English Law in the Eighteenth Century 93-99 (1992)).
Warren also agreed with Raul that while there is a price to be paid
for speaking out against "junk science" we should be optimistic
that scientists generally have enough courage to do so. In particular,
Warren argued that reputable scientists would participate in the
expert panels he recommended be utilized because this move would
free them from the taint associated with being labeled "hired
*Jeffrey Bossert Clark is an associate at the law firm of Kirkland