The so-called Barker hypothesis, published in 1989 by the British physician David Barker, is a prominent underpinning of the central conception of the authors: that a fetus deprived of nutrition may be predisposed to diseases that typically appear in adulthood. This view was formulated by retrospectively studying the medical records of adults in the United Kingdom and correlating birth weights with risks for disorders like heart attack and diabetes that occur later in life. The authors write:
Even if malnutrition in childhood is followed by improved economic circumstances in later life, so that there is then no lack of energy for current work and other activity, the earlier deprivation can have long-term consequences.
In The Changing Body, they further argue that increases in average height of men and women can be taken as a reliable measure of the success of different societies, specifically the productivity of their economies, the fair distribution of their resources, and wider access to scientific advances in health care and the workplace.
The voluminous data cited in The Changing Body to statistically support its theses come from observational surveys. These derive information on changes in the heights of military recruits, factory workers, and residents of geographic areas from the 1700s to the present from historical records and other documents, not scientific experiments. No one has conducted (or should) randomized controlled trials to more scientifically assess the impact of factors like dietary nutrients on health and productivity by, say, assigning one group of mothers to a robust diet and another group to limited food. But extreme care needs to be exercised to avoid overinterpreting statistical correlations from data like birth weights and later maladies as indicating causation.
This caveat about confusing correlation with causation is raised by the authors, but the temptation is nonetheless powerful enough for them to draw attention to research that stands on thin etiological grounds. And it is here that the book stumbles. For example, the authors cite a study that showed “that there is a negative association between height and certain causes of premature death, such as prostate cancer, lymphoma, and colorectal cancer.”6 But each of these malignancies is very different in its biology. Moreover, the ages of the patients at the time of diagnosis of each cancer are not similar, and the increased adoption over recent decades of screening tests such as PSA and colonoscopy have changed the incidence of those tumors, resulting in a so-called “lead time bias,” where survival appears to change but in fact only reflects an earlier time of detection.
Rarely are the findings from observational studies of such magnitude that the high degree of correlation in conjunction with solid biological research mitigates the danger of confusing correlation with causation. Cigarette smoking and its correlation with development of lung cancer or emphysema is arguably the best example. In that case, epidemiological studies, complemented by laboratory research on toxins in tobacco smoke, overwhelmingly showed a cause-and-effect relationship.7
By contrast, in 2000 Peter C. Austin, a medical statistician at the University of Toronto, and his colleagues conducted a study of all 10,674,945 residents of Ontario aged between eighteen and one hundred. Residents were randomly assigned to different groups, in which they were classified according to their astrological signs. The research team then searched through more than two hundred of the most common diagnoses of hospitalization until they identified two where patients under one astrological sign had a significantly higher probability of hospitalization compared to those born under the remaining signs combined: Leos had a higher probability of gastrointestinal hemorrhage while Sagittarians had a higher probability of fracture of the upper arm compared to all other signs combined.
It is thus relatively easy to generate statistically significant but spurious correlations when examining a very large data set and a similarly large number of potential variables. Of course, there is no biological mechanism whereby Leos might be predisposed to intestinal bleeding or Sagittarians to bone fracture, but Austin notes, “It is tempting to construct biologically plausible reasons for observed subgroup effects after having observed them.” Such an exercise is termed “data mining,” and Austin warns, “Our study therefore serves as a cautionary note regarding the interpretation of findings generated by data mining, and suggests that conclusions obtained from data mining should be use with a healthy degree of skepticism.”8
The authors’ model of “technophysio evolution” in The Changing Body is most convincing in the dramatic increases in life expectancy in developed nations. They then ask: How long might future generations live?
So far the maximal capacity of human longevity has been restricted by limited diet and lifetime health insults. But human beings may live up to 130 years if they are well fed and if there were no health insults over the lifetime.
The authors note that this estimate is based on the work of Leonard Hayflick, who demonstrated that cultured human cells have a limited capacity for replication, and so eventually enter a phase of senescence (the Hayflick Limit), in which they no longer can sustain life. The limit is presumed to occur because the structures at the end of chromosomes, called telomeres, reach a critical length. Cells in culture derived from normal human fetal tissue (those presumably protected from “health insults” such as ambient radiation) will divide between forty and sixty times before entering the senescent phase. Each mitosis or cell division shortens the telomeres on the chromosomes of the cell, which eventually makes cell division impossible. But again, extreme caution needs to be exercised in extrapolating from cell growth in a laboratory dish to the fate of human beings. And it is difficult to envision a life without any “health insults,” in a changing and sometimes noxious environment.
They are on somewhat firmer footing when addressing cognitive capacity and social factors. “It seems likely,” they write, that “poor nutrition affects labor productivity because it diminishes cognitive ability and the capacity to undertake and benefit from education.” They note that poorly nourished children often start school later, progress through school less rapidly, and do less well on cognitive achievement tests when older. Birth weight was significantly and positively associated with cognitive ability at age eight in a cohort of Britons born in 1946, even after the factors of gender, father’s social class, mother’s education, and birth order were discounted:
On the basis of all this evidence from the modern world, it seems highly likely that populations in the past, with high levels of malnutrition, suffered from low productivity not simply because of diminished physical strength but also because of diminished cognitive ability or intelligence. The relative contribution of these two potential causes of low productivity may be impossible to determine, but they are likely to have reinforced each other.
They recognize that this is “a potentially contentious statement—particularly when applied to differences in stature between and within modern populations.” Thus they emphasize their belief that “all human populations have equal potential which will be fulfilled under conditions of optimal nutritional status—even if that state has possibly never been achieved.”
While The Changing Body is written as a textbook with numerous graphs and equations, the intriguing subject of anthropometry has been examined in popular writing. Stephen S. Hall in his book Size Matters: How Height Affects the Health, Happiness, and Success of Boys—and the Men They Become9 devotes a chapter to his meeting with Richard Steckel, a student of Fogel’s when he wrote Time on the Cross. Steckel is now an economic historian at Ohio State University, and with Hall toured the Smithsonian Institution’s National Museum of the American Indian in Lower Manhattan. They lingered in front of a large watercolor and ink drawing of the Battle of Little Big Horn from the 1880s, painted by Standing Bear, who as a seventeen-year-old had participated in the battle. “While most museum goers (including me),” Hall writes,
predictably search for the likeness of George Armstrong Custer in the tableau, Steckel saw in it a crystallization of all the invisible factors that can contribute to a unique quality of life, and that quality revealed itself in a single characteristic of the figures peopling the painting: their height…. The individual Lakota Sioux warriors appeared to be as tall as, or taller, than their American adversaries.
Steckel attributes their robust form to several factors: the high-protein diet of the Plains tribes; their excellent dental health; the ways their use of horses reduced their work effort, and may also have allowed these tribes to disperse quickly at the first sign of epidemic disease. The Plains Indians also may have benefited nutritionally from their custom of long-term breast-feeding.
Hall points out that height data from the muster roles of Custer’s 7th Cavalry show that the soldiers averaged slightly more than five foot seven in height, an inch or two shorter than the average height of the tribes they fought, using height measures of Native Americans collected in the nineteenth century. Indeed,
the Plains tribes were on average among the tallest people in the world at the time—taller than native-born white Americans, taller than Europeans, taller than virtually any national group for whom reliable data exists.
Although the Native Americans were viewed as “poor,” Steckel and his colleague Joseph M. Prince of the University of Tennessee wrote that “height and health are known to be sensitive to inequality,” and suggested that egalitarian values, a steady food supply, and “social and economic fluidity” allowed these Native American nomads to build a stronger social safety net than their European American contemporaries.” The health and welfare of society, according to Steckel and Prince, may be best gauged not by monitoring numbers such as average income or gross domestic product, but by the communal commitment and collective ability to care for the young, resist disease, and maximize the salutary effects of good nutrition. All these factors help a given society to maximize the genetic potential of its people, and that maximized potential translates into a single salient characteristic: greater average height.
Another prolific scholar influenced by Fogel is the Hungarian-born and American-bred John Komlos, who has created a research center at the University of Munich focused on anthropometric studies. He is featured in Burkhard Bilger’s New Yorker article “The Height Gap,” in which Bilger wrote:
The Netherlands, as any European can tell you, has become a land of giants. In a century’s time, the Dutch have gone from being among the smallest people in Europe to the largest in the world. The men now average six feet one—seven inches taller than in Van Gogh’s day—and woman five feet eight…. From Rotterdam to Eindhoven, ceilings have had to be lifted, furniture redesigned, lintels raised to keep foreheads from smacking them.10
Bilger further observes that “for more than two centuries,” Americans “had been so healthy and so prosperous that they towered above the rest of the world—about four inches above the Dutch, for example, for most of the nineteenth century.” But now, the average American man is about five foot nine—four inches shorter than the average Dutch man.
6 George Davey Smith et al., "Height and Risk of Death Among Men and Women: Aetiological Implications of Associations with Cardiorespiratory Disease and Cancer Mortality," Journal of Epidemiology and Community Health, February 2000. See also George Davey Smith, Martin Shipley, and David A. Leon, "Height and Mortality from Cancer Among Men: Prospective Observational Study," British Medical Journal, November 14, 1998. ↩
7 The cardinal paper setting forth criteria to link correlation and etiology: Sir Austin Bradford Hill, "The Environment and Disease: Association or Causation?" Proceedings of the Royal Society of Medicine, Vol. 58 (1965). See also Jerome Groopman, "Birth Pangs," The New York Times, October 3, 2010, and Jerome Groopman, "The Plastic Panic," The New Yorker, May 31, 2010. ↩
8 Peter C. Austin et al., "Testing Multiple Statistical Hypotheses Resulted in Spurious Associations: A Study of Astrological Signs and Health," Journal of Clinical Epidemiology, Vol. 59 (2006). ↩
9 Stephen S. Hall, Size Matters: How Height Affects the Health, Happiness, and Success of boys—and the Men They Become (Houghton Mifflin, 2006). ↩
10 Burkhard Bilger, "The Height Gap: Why Europeans Are Getting Taller and Taller—and Americans Aren't," The New Yorker, April 5, 2004. ↩
George Davey Smith et al., “Height and Risk of Death Among Men and Women: Aetiological Implications of Associations with Cardiorespiratory Disease and Cancer Mortality,” Journal of Epidemiology and Community Health, February 2000. See also George Davey Smith, Martin Shipley, and David A. Leon, “Height and Mortality from Cancer Among Men: Prospective Observational Study,” British Medical Journal, November 14, 1998. ↩
The cardinal paper setting forth criteria to link correlation and etiology: Sir Austin Bradford Hill, “The Environment and Disease: Association or Causation?” Proceedings of the Royal Society of Medicine, Vol. 58 (1965). See also Jerome Groopman, “Birth Pangs,” The New York Times, October 3, 2010, and Jerome Groopman, “The Plastic Panic,” The New Yorker, May 31, 2010. ↩
Peter C. Austin et al., “Testing Multiple Statistical Hypotheses Resulted in Spurious Associations: A Study of Astrological Signs and Health,” Journal of Clinical Epidemiology, Vol. 59 (2006). ↩
Stephen S. Hall, Size Matters: How Height Affects the Health, Happiness, and Success of boys—and the Men They Become (Houghton Mifflin, 2006). ↩
Burkhard Bilger, “The Height Gap: Why Europeans Are Getting Taller and Taller—and Americans Aren’t,” The New Yorker, April 5, 2004. ↩