Quoth the Maven

Is it real science, or a cheap imitation? Judges need to learn the difference.

Today judges across the country confront a terrifying new threat. They recognize that "deep anxieties and considerable hysteria" afflict a small community in Florida and a large one in New York. The fears are undoubtedly real, but most judges refuse to be swayed by "theoretical risk," "future theoretical harm," "pure speculation," or "irrational and unsupported belief." Almost without exception, judges brush aside popular fears and stake everything on real science. They empathize with those who are frightened but find themselves "duty bound to objectively evaluate the issue…and not be influenced by unsubstantiated fears of catastrophe."

"Little in science can be proved with complete certainty," a California judge concedes, but decisions cannot turn on some tiny, unerasable shadow of doubt. Judges rely instead on the "overwhelming weight of scientific evidence," the "clear weight of the expert medical evidence and opinion," "reasonable medical knowledge," and the consensus views of expert agencies and public-health authorities.

What could possibly elicit so much good sense in so many different quarters? Certainly not "chemical AIDS," the junk science alternative to the dread disease. The claims of clinical ecology—which alleges that a host of maladies, including suppression of the immune system, are caused by low-level exposure to industrial chemicals—have met with credulity and multimillion-dollar judgments in court. (See "Quack Attack," October.) No, at issue is AIDS itself, the new leprosy.

More specifically, the issue is whether AIDS can be transmitted by casual contact. The argument is always the same: Since we don't know absolutely for sure, shouldn't we err on the side of safety? It is the same argument that has been raised in connection with tuberculosis, electronic cruise controls, morning sickness drugs, contraceptives, vaccines, and industrial chemicals: Better to banish the small risk than to hazard the large harm.

But in answering this tired and superficial argument in the context of AIDS, courageous judges call the science straight. They side with doctors against a New York landlord who does not want them to treat AIDS patients on his premises, with AIDS children who want to attend schools in New York City and Florida, and with young Ryan Thomas, who begs leave to attend kindergarten in California. Predictably, the California school district does manage to dig up a Dr. Steven Armentrout, who opines—correctly, no doubt—that there is a tiny possibility that AIDS could be transmitted through as-yet-undiscovered vectors. But five experts, backed up by prestigious medical journals, present strong evidence that AIDS is not transmitted by casual contact, and that is enough.

And now a strange thing begins to happen: As one judge after another affirms the solid science and rejects the paranoid speculation, the courts help to educate the public and allay some of its most acute but least well-grounded fears. By refusing to take the junk science of AIDS seriously, wise judges help put a stop to it.

Why have courts been willing to rein in destructive legal speculation on genuine AIDS, when they are so reluctant to do the same when it comes to "chemical AIDS" or the other bogeys of junk science. One reason, no doubt, is that a young Ryan Thomas, with his heart set on kindergarten is a much more visible and sympathetic victim of junk science than a faceless, deep pocket chemical company. But giving in to scientific nonsense entails costs in either case. A rational court system can and should vindicate good science not just when the stakes are specially high and poignant but as an everyday matter for every kind of litigant.

But what is good science? How can we identify it? The trial lawyer and his acolytes will never tire of telling stories about how high priests of science have been proven badly wrong by "cranks" and "mavericks" in times past. Galileo, the patron saint of all heretics, figures often in such stories. Let's not ostracize the "mini Galileo," pleads a plaintiff's lawyer; the legal system must be "capable of advancing." Honor the expert "at the edge of the bell curve," advises the chemical-AIDs maven Alan Levin, "as was Galileo and as are other people at the frontiers of medicine or science."

No doubt about it: The views of the establishment are sometimes wrong, in science and medicine as in law. It's equally true that the isolated scientist, the iconoclast, the maverick, crank, or congenital rebel has sometimes been proved right. When the great Hungarian clinician Ignaz Philipp Semmelweis discovered the antiseptic properties of chlorinated lime in 1850, his findings were met with deep skepticism from many in the medical establishment.

But science has changed profoundly since the days of Galileo and Semmelweis. This is most particularly true of medical science. Until the late 17th century, as the historian James Burke recounts, a medical career "flourished or foundered according to the relationship the doctor managed to strike up at the bedside." The doctor would emphasize his "heroic and secret" insights into disease and its cure. Each individual's illness was thought to be a unique condition. Each doctor "would claim that all other doctors were quacks and their remedies ill-advised or dangerous." Doctors believed that every disease could exhibit every symptom. Therapies were correspondingly quirky.

In 1800, the French surgeon Xavier Bichat demonstrated that disease is a specific phenomenon peculiar to certain lesions or tissues. Doctors began to recognize that disease itself presents a specific and concrete target that transcends individual patients. As medicine raised its sights from the idiosyncratic and particular to the regular and general, it converged with statistics, a new branch of mathematics that was evolving during the same period. The center of medical learning shifted to the hospital, where patients could be studied in still larger numbers.

When cholera struck Europe in 1829, the focus shifted from the hospital to the city. William Farr, appointed controller of the General Register Office, set out to conquer cholera with a radically new medical instrument: the biometer. The biometer was pencil and paper—a life-table that insurance company actuaries had been using for years. Farr systematically analyzed who was dying and where. The most important things he discovered were negative. Wealth didn't protect you from cholera. Nor did occupation or residential proximity to the sea. What mattered was how high above the Thames you lived. Farr concluded that cholera was caused by the river's awesome stench.

He was wrong, but only in this single, last step of the analysis. It was left to another English physician, John Snow, to make the right connection in 1853. The key was not dirty air but dirty water; the London sewers emptied into the Thames, so the farther down-sewer you lived, the more likely you were to drink foul water. A few years later Parliament passed legislation to rebuild the sewers, and cholera disappeared from the city forever.

The story of cholera is the story of how medicine was transformed from black art into science, from a pseudoscience of the individual into a science of groups. The difference between the clinician who cures and the clinician who quacks is the difference between the intellectual hermit and the member in good standing of a community of scientists. For the one, medicine is shaped by an endless series of peculiar and individual cases; for the other, by broad perspective and consensus conclusions. One espouses fictions as changeable as the individual patient and doctor; the other, truths that apply to many people, not just to one. The rise of modern medical science, with its astonishing capacity to diagnose and cure, can be traced to the decline of individual eccentricity on both sides of the stethoscope.

Modern science is the study of facts that are regular, of things that recur in patterns. A courtroom trial is quintessentially singular. Science depends on placing facts in an orderly context, but a trial frames facts in isolation. Good science transcends the here and now, the individual and idiosyncratic, the single laboratory, the single nation, the single planet, even the single galaxy, but a trial typically examines a singular datum and demands that scientific truths be rediscovered anew every time. Scientific facts emerge from many isolated observations, as data are accumulated, vetted for error, tested for significance, correlated, regressed, and reanalyzed, but trials are conducted retail. Good science is open, collegial, and cumulative, but the courtroom setting is discrete, insular, and closed—a one-shot decision.

The methods of science are so fundamentally different from those of litigation that scientific anarchy in court is inevitable if rules of evidence are not strictly maintained. Scientific facts remain perpetually in play. Each patient, each injury, each illness becomes unique once again—or so says the 18th-century doctor, on the payroll of the 20th-century lawyer. Trials are not connected; the same question about the morning-sickness drug Bendectin, "sudden acceleration" in the Audi, or clinical ecology can be litigated again and again.

In the worst cases, courts drift through the degenerative sequence described by the historian Jerome Ravetz and thereafter elaborated by W.C. Clark. Tentative outlooks are often suppressed, views are quickly polarized, and a "great confidence game" replaces serious science. Recognition and money, Ravetz noted, flow "to those making the first, loudest, and most frightening noises." The careful skeptic is rewarded with "accusations of corruption, cowardice, or insensitivity." There will be "an accretion of cranks and congenital rebels whose reforming zeal is not matched by their scientific skill."

An accretion of cranks in court follows inevitably from the great paradox of modern liability science: In attempting to control quackery outside the courtroom, we invite quacks to the witness stand. If this degenerative process is to be halted—or, better still, reversed—judges must rediscover rules of evidence consonant with the essential collegiality of modem science. Such rules are not self-evident, nor can they be implemented mechanically, nor will they work their intended effect in the hands of jurists who hold science itself in no real respect. But rules can be formulated, and even modest rules, if enforced with evenhanded conviction and some measure of faith in the scientific method, will make a positive difference.

"Training is everything," Mark Twain once suggested. "Cauliflower is nothing but cabbage with a college education." Many, like Twain, will suppose that the cure to junk-science litigation is to have judges scrutinize professional credentials more carefully. It isn't; Twain was only half right. Strings of letters appended to last names do provide a useful initial screen against professional incompetence, but only a very coarse one. Even yesterday's stellar achievement offers little assurance that today's opinion is correct: Many a great scientist takes off sooner or later on some foolish frolic.

So while a résumé may be a necessary condition of expert competence, it is never a sufficient one. Twain's views notwithstanding, what defines a cauliflower is not its résumé but the views it shares with other cauliflowers. A cabbage with an M.D. is still a cabbage. Science is likewise defined by a community, not by the individual, still less by a résumé. Lawyers already know this. Credentials are all but irrelevant when a doctor sits at the defense table rather than in the witness box, and so they should be. Cucullus non facit monachum: The cowl does not make a monk.

This was the key insight in the Frye rule of evidence, now fallen into disfavor. Based on a 1923 federal appellate court decision, Frye required that expert testimony be founded on theories, methods, and procedures "generally accepted" as valid among other scientists in the same field. Frye directed the focus away from the individual, whatever his credentials might be, and toward the scientific consensus. Define the relevant community whose consensus views should prevail. Then require expert witnesses to report not their personal views but the consensus views of that community.

Applying the test is not always simple; there will always be room for quibbling. Any definition of "the relevant scientific community" will be somewhat arbitrary. But despite what some lawyers maintain, it isn't terribly difficult to decide which community of scientists to consult on Bendectin, cerebral palsy, or sudden acceleration. Lawyers in fact define similar communities all the time. A long-standing principle of negligence law is that doctors are held to the standards of the medical community in which they practice; one standard for an urban specialist, another for a rural generalist. Most of the time, common sense serves quite adequately to identify a relevant scientific community, far better than it does in ascertaining the science itself.

The second step in applying Frye—determining just where the mainstream scientific consensus lies—is usually not all that hard either. Careful reviews of current learning on one subject or another are published in top-notch scientific journals all the time. Such journals have long track records of accuracy and insight. They are backed by established scientific institutions. What they publish is reviewed by other scientists. A judge need not know the slightest thing about pharmacology or electronic cruise controls to make sensible calls about who speaks for mainstream science on such issues and who does not. Even a person who knows nothing about hydrology can distinguish the mainstream of the Mississippi from stagnant pools near its banks.

As the legal scholar Bert Black has lucidly discussed, a sophisticated, modern application of Frye looks to the methods behind a scientific report, not to its finely detailed conclusions. An epidemiological study will easily survive Frye even if it is the very first to report, for example, a link between Bendectin and birth defects, so long as standard protocols for conducting such studies have been observed and the data are reported with error bands, significance tests, and similar statements of caution suitable for a refereed professional journal. What should not survive, however, is a crude imitation of science—the unpublished hunch, the letter to the editor, the impressionistic "mosaic theory."

Lawyers should be the last to suggest that any of these ideas is radical or unreasonable, because lawyers apply every one of them to defendants, though not to witnesses. If the laws of negligence and strict liability can condemn doctors, chemists, pharmacologists, and car manufacturers for their ineptitude, it is because we believe that there is such a thing as ineptitude, that competence is ascertainably different from incompetence, that there are objective standards worth enforcing. But if people who really design cars or deliver babies are to be judged by professional standards in court, those who accuse them must be held to similar account. If the law is capable of holding defendants to professional standards, it is capable of holding witnesses to the same.

The scientific community supplies stopping points in abundance for those who care to find them. An authoritative scientific pronouncement on Bendectin by the Food and Drug Administration might be one. Or a report by the National Institutes of Health on electronic fetal monitors. Or one by the Centers for Disease Control on the pertussis vaccine or the causes of pelvic infection. Such institutions, established and funded to make difficult scientific calls, draw on the best and broadest scientific resources.

This is not to suggest that they are infallible; of course they aren't. They are just less fallible—much less fallible than a thousand juries scattered across the country grappling with the complexities of immune-system impairment after being educated by the likes of clinical ecologists Bertram Carnow and Arthur Zahalsky. Judges therefore have abundant reason to promote the former and to be far more cautious about admitting the latter.

Half the time, ironically, that is precisely what judges already do. If the FDA declares that thalidomide causes limb defects in the womb, there will not be a very long or complex trial if a drug company nonetheless sells the product and a child is born without arms. The case is easy not because the science is easy—it isn't—but because a much larger community has already thrashed out the questions and reached some consensus.

But there is consensus with Bendectin too, though in the opposite direction: There is no good evidence that it causes birth defects. Yet Bendectin trials have lumbered forward, one by one. If the FDA says the drug is a teratogen, liability will follow almost automatically. If the FDA says it isn't, ask a jury. And then another, and another, as many times in succession as the trial bar may deem to be justified by either visceral conviction or speculative greed. We find, once again, that our modem liability system is all accelerator and no brake.

When definitive pronouncements of the FDA, CDC, or surgeon general are not at hand, the next best place to look for the consensus views of mainstream science is in the peer-reviewed scientific literature. There is, indeed, a straightforward test for judges to determine which methods, procedures, and theories have not been "generally accepted" by other scientists: peer-reviewed publication.

A witness whose views have survived peer review in a professional journal will already have been forced into a candid disclosure of cautions and qualifications; good journals won't publish without them. If the published claim is of any importance, publication will also mobilize other scientists to repeat, verify, contradict, or confirm. By requiring professional publication as a basis for expert opinion, judges will help line up the larger community of scientists to shadow the necessarily smaller community of expert witnesses.

One might expect judges to be very comfortable with such a write-it-down rule of scientific evidence, for they apply similar rules elsewhere. The "parol evidence" rule, for example, declares that a written contract trumps the verbal discussions that lead up to it; writing is likewise required by the law to make binding a will or a contract for the sale of land. If the law already recognizes that only a written document can be trusted to determine who gets Aunt Agatha's prized collection of china dogs, it is hardly excessive to require formal writing by anyone who claims to have identified the causes of cerebral palsy or chemically induced AIDS.

There is one final test of expert competence, one more difficult to articulate but of great importance nonetheless. Scientific study of the causes of injury and disease, like scientific study of anything else, looks for regularities, patterns, and recurrences. Modern science uses rigorous, systematic methods to locate such regularities, methods that are rooted in statistics and significance tests, blind trials, and especially epidemiology. Each human spirit may be individual and unique, but the frailties of human flesh are shared by many. Medical science has recognized this fundamental fact for the better part of two centuries. Thus, any truly scientific claim about the causes of disease will be based on systematic observation of many patients or test subjects, not on off-the-cuff impressions developed in the course of clinical treatment. The considerable skills required to diagnose cerebral palsy, or perform infant surgery, or treat leukemia, are not the same skills required to determine the causes of those afflictions. The difference between the clinician and the scientist is one that courts must learn to understand and affirm.

To hold experts to serious scientific standards is not to abandon venerable legal principle but to reaffirm it. The expert witness is the only kind of witness who is permitted to reflect, opine, and pontificate, in language as conclusory as he may wish. Once we recognize the expert witness for what he is, an unusually privileged interloper, it becomes apparent why we must limit just how far the interloping may go. The expert whose testimony is not firmly anchored in some broader body of objective learning is just another lawyer, masquerading as a pundit.

It is heartening to record that at least some judges have arrived at this conclusion in recent years. Something of a turning point in judicial attitudes came in 1986, in a much cited opinion by federal appellate Judge Patrick Higginbotham. "Our message to our able trial colleagues: it is time to take hold of expert testimony in federal trials," he wrote. A slowly growing number of able colleagues have taken these sentiments to heart. Some have begun to emphasize that radical London novelty of 1660—the professional society and that radical Paris novelty of the 1820s—the professional journal and peer review. Others have begun to emphasize the importance of solid epidemiological data and to insist that scientists confine testimony to their specialties. But many judges still reject any such limits, or equivocate so much that defendants settle baseless claims rather than risk going to trial.

If judges will not screen witnesses retail, state legislatures can screen wholesale; a few have recently done so. In 1987, for example, Alabama passed a law requiring expert witnesses to have practiced recently in the same specialty as the doctor they charge with medical malpractice. Colorado passed a law in 1988 restricting malpractice expert testimony to licensed physicians who can demonstrate "substantial familiarity" with the applicable standard of care and the procedure being litigated. A recent Maryland law bars testimony from any malpractice expert who spends more than 20 percent of his time in court. Kansas, Michigan, Maryland, Rhode Island, and West Virginia have developed similar requirements. Most of these states also bar from the witness stand academics who do not practice at all.

The strongest antidote to bad science in court remains one that most American judges are still regrettably reluctant to use. European judges routinely summon their own experts. Our judges have similar powers, but few choose to exercise them. Most trial lawyers vehemently oppose court-appointed experts, perceiving (correctly, no doubt) that consensus cannot be good for a conflict-centered livelihood. Lawyers will therefore assure you that there is no such thing as a neutral expert. But it is obviously possible to find knowledgeable scientists of high principle, and having a nonpartisan judge do the finding considerably improves the prospect of locating a less partisan expert.

None of these ideas is the least bit radical. What they all come down to is biblical wisdom on the punishment of harlots. If not certifiably free of sin, the expert witness who casts the first stone should—at the very least—not be a notorious patron of the local scientific bordello.

Real science, science outside the courtroom, is an evenly balanced process of proposal and disposal. Science does, of course, require a steady supply of the new and different, the bold, the shocking, even the outrageous. But it requires even more a steady supply of replication, verification, and peer review, the patient development of consensus, the systematic weeding, pruning, and uprooting of spurious data and erroneous theory. Starting promising new lines of inquiry is important; no less important is stopping unpromising old ones. Beginning speculation may be more exciting than ending it. But good science depends quite as much on the patient, plodding rejection and elimination of bad data and mistaken theory.

Judges have understood this well enough when addressing unfounded prejudice against victims of AIDS. But when the targets of junk science are less sympathetic, and their pockets deeper, the legal record has been dismal. Let-it-all-in rules of scientific evidence have made it trivially easy to begin pseudoscientific speculation in court and almost impossible to end it. Courtroom science has come to revolve around the opinionated eccentric, the go-it-alone maverick. It is heavily biased in favor of the impresario who begins speculation, and against the plodders who end it.

The vindication of good science in court requires precisely the opposite. Judges, like scientists themselves, will never pen a final opinion on the laws of nature, nor should they try. What they can do, however, is realign courtroom science with the science of scientists. This means giving much less attention to the self-proclaimed new Galileos, and far more to the reticent stalwarts of the mainstream scientific community. No doubt, trials will always depend on individual witnesses and personal credibility. But just who those witnesses are and what they testify to can be controlled.

Lawyers and judges who claim so much aptitude in deterring incompetence and preventing accidents everywhere else can surely find the rules to deter and prevent them in court—if they can ever find the will. Until they do, courts will maintain their current renown for quixotic, pseudoscientific crusades, begun in haste and repeated at leisure. Until they do, what passes for science in court won't be.

Peter W. Huber is a senior fellow at the Manhattan Institute and a columnist for Forbes.