It is the peculiar and perpetual error of the human understanding to be more moved and excited by affirmatives than by negatives.
The root of all superstition is that men observe when things hit but not when they miss; and commit to memory the one and forget and pass over the other.
—FRANCIS BACON
During World War II the Allies carried out a strategic bombing campaign against the German industrial heartland from airfields in Britain. The main workhorse of the campaign was the Lancaster four-engine bomber, which, owing to its weight and slow speed, suffered punishing losses from German night fighters. By one estimate, the chances of a crew reaching the end of a thirty-mission tour were about 25 percent. The British military called in experts, including the young Freeman Dyson, to determine how to reduce the staggering casualty rates. Owing to their heavy armor plating and gun turrets, the planes were forced to fly at a low altitude and were painted black to make them less visible during their night runs. Dyson tells of a vice air marshal, Sir Ralph Cochrane, who proposed ripping out the gun turrets and other dead weight from one of the Lancasters, painting it white, and flying it high over Germany. But the military command rejected this audacious experiment owing to what Dyson, following Daniel Kahneman, calls the “illusion of validity”—the deep-seated human need to believe that our actions are well-founded.
1All those involved in the air war believed in the tightly knit bomber crew, with the gunner playing a crucial role in defending the aircraft, and the pilot using his experience to take evasive actions. Dyson writes, “The illusion that experience would help them to survive was essential to their morale. After all, they could see in every squadron a few revered and experienced old-timer crews who had completed one tour and had volunteered to return for a second tour. It was obvious to everyone that the old-timers survived because they were more skillful. Nobody wanted to believe that the old-timers survived only because they were more lucky.”
When Dyson undertook a careful analysis of the correlation between the experience of the crews and their loss rates, taking into account the possible distorting effects of weather and geography, he found that experience had no effect on whether a plane returned home. “So far as I could tell, whether a crew lived or died was purely a matter of chance. Their belief in the life-saving effect of experience was an illusion.”
Dyson’s demonstration that experience had no effect on losses should have provided strong support for Cochrane’s idea of tearing out the gun turrets. But it did nothing of the sort. He tells us that “everyone at Bomber Command, from the commander in chief to the flying crews, continued to believe in the illusion. The crews continued to die, experienced and inexperienced alike, until Germany was overrun and the war was finally ended.”
It took another outsider to come up with a dazzling insight into the reasons for the heavy toll on British bombers. Abraham Wald was a Jewish mathematician from Eastern Europe who had come to the United States in the late 1930s to escape persecution. During the war he used his knowledge of statistics to analyze the problem of the aircraft losses. Analysts had proposed adding armor to those areas of the aircraft that showed the most damage. What Wald realized was that the damage sustained by the aircraft that returned safely represented areas that were not fatal to the plane’s survival. The fact that there were areas of the returning planes that showed no damage led him to surmise that these were the vulnerable spots that must have led to the loss of the planes to enemy fire. Thus it was these areas that needed to be reinforced.
2Making an inspired leap, Wald posited that there must be a crucial difference in the pattern of damage between those bombers that returned and those that did not. He saw that the missing data—the bombers that never made it back—provided the key to the problem, and he analyzed the pattern of nonfatal damage displayed by the returning bombers to intuit the pattern of fatal damage to the planes that did not return. What his analysis showed was that the planes’ engines were vulnerable and needed shielding.
Wald’s approach to estimating aircraft survivability was used during World War II, as well as by the U.S. Navy and Air Force during the Korean and Vietnam Wars. Today his analysis—which was carried out without computers—is considered a seminal contribution to statistics and specifically to the problem of “missing data.” Writing about Wald’s work on aircraft survivability in the leading statistics journal in 1984, two statisticians concluded that, “while the field of statistics has grown considerably since the early 1940’s, Wald’s work on this problem is difficult to improve upon…. By the sheer power of his intuition, [he] was… able to deal with both structural and inferential questions in a definitive way.”
3More broadly, Wald’s analysis provides an example of how crucial it is to consider the full range of relevant data, rather than confining oneself to a biased sample (i.e., the planes that returned safely) or to the usual categories. It is an inspired example of what we refer to (perhaps too lazily) today as “thinking outside the box.” It underscores the need to be open to new ways of seeing, going beyond the limits of our habitual thinking, and looking for answers in places where we might not immediately think to look.
4In fact, Dyson’s “illusion of validity” and Wald’s “negative thinking” represent two sides of a single coin. Taken together, the stories of Dyson and Wald provide inspired examples of overcoming the impediments to thinking afresh about a problem, divesting oneself of preconceptions and habitual ways of looking at things. We all tend to focus on certain salient aspects of a problem, and these can obscure other aspects, which may be essential to consider. Experts are not exempt from this tendency, which, it has been noted, is particularly in evidence among those who formulate policy.
5* * *
Since World War II, science has made remarkable progress in medicine, genetics, molecular biology, and epidemiology. And yet, in spite of this progress, our understanding of what causes many chronic diseases and how to prevent them is still humblingly limited. Furthermore, widespread confusion reigns about what are the real threats that are likely to affect our lives. For example, there are controversies raging within the scientific community or wider society regarding a wide range of issues, including radiofrequency radiation from cellular telephones and other wireless technology, “endocrine disrupting chemicals” including pesticides and other contaminants in our food and consumer products, what constitutes a healthy diet, vaccines, obesity, genetically modified foods, the use of hydraulic fracturing (“fracking”) to extract oil and gas, alternative and complementary medicine, and particulate air pollution—to name some of the more prominent topics.
These threats, which are so much in view, tap into reflexes that allowed our ancestors to survive hundreds of thousands of years ago in the African savannah. But the instinctual reaction that served us well when the task of not being eaten by a predator was paramount is less suited to the modern world, which is a much more complicated environment to navigate. It is not that we are wrong to be mistrustful and wary of our environment or to question information put out by the authorities, but when we adopt an extreme position—embracing conspiracy theories and rejecting objective evidence that comes from impartial sources—we are apt to fall for the “illusion of validity” and fail to recognize other real dangers.
Similarly, when scientists become wedded to a particular hypothesis and resist considering contradictory evidence and alternative explanations, they narrow their field of vision and close off what may be more productive lines of inquiry.
This brings us to the two very different outcomes of scientific research in the area of health and health risks that are the focus of this book. At the outset, it needs to be said that the vast majority of research never attracts the attention of the media or the public. So the contrast I am setting up is one of extremes.
Research that succeeds in uncovering new knowledge involves the painstaking process of formulating a hypothesis, obtaining meaningful data, ruling out artifacts and overcoming biases, comparing results from different research groups, and considering and excluding alternative explanations at each step of the way. At the heart of this process is a tension between the researcher’s hypothesis and the evolving evidence bearing on it. It is only natural that a researcher can become deeply invested in a particular hypothesis. But, at the same time, he or she has to be the most relentless critic of the hypothesis and be willing to modify or reject it if it conflicts with the evidence. In pursuing an initial idea, a researcher will often be led to a more promising idea that was not envisaged at the outset. All this takes place out of the spotlight, for the simple reason that until one has followed the line of inquiry and obtained a solid result, there is no reason to get the media and the public stirred up about the possible significance of the work. (An added motivation for caution is that one doesn’t want to end up looking like a fool.)
Some hypotheses may be weak but may nevertheless merit study. If research does not provide support for the hypothesis, in due course it would normally be abandoned for other lines of research. However, in cases where a weak hypothesis touches on a topic that has the potential to galvanize public concern, what is at heart a scientific question can attract the attention of nonscientists, including regulators, funding agencies, advocates, journalists, and others. When such an issue is framed in a narrow way—is X a problem?—a way that restricts attention to the putative threat and fails to put it in perspective, it can take on a life of its own. Regulators may feel the need to consider the question. Funding agencies may decide to support further research. These actions, which attract news coverage and generate more concern in the public, keep the issue in the public eye. Some scientists may believe that there is evidence to support the hypothesis and may have a strong stake in it. Furthermore, because their findings speak to deep-seated fears relating to a publicized issue, the work of these scientists can be championed by advocates who believe that they are telling the truth, in opposition to “establishment scientists,” who are minimizing or suppressing a real problem. We will see examples of this second outcome of research in the stories of cell phones, endocrine-disrupting chemicals, and particularly BPA.
* * *
It is a striking paradox that the stories concerning the causes of chronic disease that get the most attention often involve findings that are questionable but that have the power to arouse anxiety, whereas stories involving painstaking, incremental work that, over time, leads to major, life-saving advances get little attention. Both types of stories are the product of “science,” and yet, depending on the research question and the social context surrounding that question, the prospects for uncovering new and important knowledge can be radically different. Strong scientific results speak for themselves—they lead to tangible and reproducible results. In contrast, when research fails to make solid progress in an area that arouses public concern, it is only too easy for researchers and advocates to offer up weak or erroneous results to the public as meaningful findings. Research that for various reasons goes off the rails and ends up misleading us rather than yielding useful knowledge has been variously referred to as “bad science,” “voodoo science,” “cargo cult science,” “pathological science,” and “parascience.”
6 Successful research exists in a separate realm, and there are many valuable accounts of the process of scientific discovery.
7 However, rarely have the two very different outcomes been forced to confront each other. What follows is an exploration of what distinguishes these two contrasting outcomes of scientific research.
Chapter 2 describes the fundamentals of observational studies in the area of public health. Studies reporting associations between an exposure and a disease vary greatly in import depending on prior knowledge regarding the association, the ability to measure the exposure and the disease condition accurately, and other methodological factors. Thus it is essential to realize that not all reported associations are created equal. However, because it is challenging to identify the important causes of complex chronic diseases that are multiply determined and that may take decades to develop, there is a tendency to latch onto findings that appear to point to a cause, even when the methodology is weak. If science in this area means anything, it means the uncompromisingly critical assessment of the relevant evidence on a question.
Chapter 3 describes how science in this area is embedded in a society that is highly attuned to the latest potential threat or breakthrough. Findings from rudimentary studies often are reported as if they were likely to be true when, in fact, most research findings are false or exaggerated, and the more dramatic the result, the less likely it is to be true. The public’s hunger for novel information about health threats and breakthroughs creates a fertile soil for biases that come into play in interpreting the results of observational studies and disseminating findings to the public. Reports of exaggerated findings can, in turn, give rise to “information cascades”—highly publicized campaigns that can sow needless alarm and lead to misguided regulation and policies.
The question of whether exposure to radiofrequency energy (RF) causes brain cancer arose over twenty years ago and is still a cause of controversy and confusion.
Chapter 4 examines what science has to say about the disturbing possibility that the worldwide adoption of a novel technology within a short time span could be causing a terrifying fatal disease. In fact, extensive research carried out over two decades provides no strong or consistent evidence to support this possibility.
Chapter 5 explores the main lines of the preoccupation with “endocrine disrupting chemicals” in the environment; how this question first arose; what we have learned from decades of research, including false ideas based on poor data that got enormous attention; and how to make sense of a bitter controversy that is currently raging in the scientific and regulatory communities in Europe and the United States.
Chapter 6 describes a little-known story linking a long-standing, enigmatic disease in the Balkans to dietary exposure to a toxic herb that has been used in traditional cultures throughout history, right up to the present. The mystery was fortuitously illuminated by a dramatic outbreak of kidney disease among women attending a weight-loss clinic in Brussels. Research on the potent toxin and carcinogen aristolochic acid contained in certain varieties of the herb
Aristolochia has led to new insights into the carcinogenic process, as well as highlighting the threat to public health posed by the woefully inadequate regulation of thousands of products marketed as “dietary supplements.”
Chapter 7 recounts how the long-standing question of what causes cervical cancer led, over a period of thirty years, to the identification of a small number of highly specific carcinogenic subtypes of the human papillomavirus (HPV) and to the understanding that persistent infection with one or more of these subtypes is necessary to cause the disease. This knowledge in turn has led to the development of vaccines that have the potential to virtually eliminate cervical cancer—a major cause of cancer death among women worldwide—as well as to fundamental new knowledge about how the virus evolved to cause cancer.
The conclusion emphasizes the need for a more nuanced and realistic view of science, which acknowledges the enormous challenges, promotes skepticism toward widely circulated but questionable ideas, and at the same time pays attention to what science can achieve at its best.