Demography’s Quest for Learning Truth from Facts
Demographers are far off at the inductive end of the social sciences. I like to think that what we do is close to the Chinese injunction “learn truth from facts.” My own career has been heavily focused on questions of health. Social scientists have made big contributions to this field by focusing on the health of populations, a broader canvas than one finds in schools of medicine or public health. My dissertation provided an abrupt introduction to public discourse. It described how cigarette smoking by men was distorting mortality patterns in 17 countries, using demographic rather than epidemiologic data. When it was published as a monograph, it was roundly attacked by the Tobacco Institute, an organization funded by tobacco companies. So when a newspaper did a piece on my book, the Tobacco Institute would write a letter to the editor. Their claim at the time, believe it or not, was that the lungs of smokers were genetically different from the lungs of nonsmokers; smoker’s peculiar lungs caused them both to smoke and to die of lung cancer. I pointed out that genetically similar populations with different smoking patterns had very different death rates from lung cancer.Not surprisingly, the Surgeon General sided with me and cited the study in the introduction to the second edition of the famous Surgeon General’s report on the health hazards of cigarette smoking.
Shortly after the dissertation, as a very naïve assistant professor, I decided that it would be worthwhile to assemble all the national data that had ever been produced by statistical offices on the distribution of causes of death by age and sex. This is something that an aging full professor would find a million reasons not to do, but I wrote to statistical offices around the world, visited many libraries, developed comparability codes for the seven international classifications of disease that had then been used, supported by a grant for $26,000. The result was a 760-page book of tables and a summary monograph called “Mortality Patterns in National Populations.”One chapter in the monograph, also published separately, used historical data on mortality and per capita income to demonstrate that the relationship between the two had shifted upwards in the course of the twentieth century. Countries were achieving higher and higher levels of life expectancy at a particular income level as the century advanced. As a result, only about a fifth of the global advance in life expectancy could be attributed to income growth.The rest was attributable to more narrowly technical improvements in medicine and public health. This demonstration was at odds with the popular but rather loose arguments of a powerful British epidemiologist and medical historian named Thomas McKeown, who focused on nutritional factors. My chapter was reprinted last summer as a health classic by the International Journal of Epidemiology with commentary by six scholars, and the reactions were mixed. The former head of the Harvard Institute for Development Economics said (and here I indulge very heavily in self-promotion) “Many discussions and insights that are at the heart of economic and human development would not have arisen, or would have arisen much later, without Preston’s paper. Indeed, this paper demonstrates well the aptness of Oliver Wendell Holmes’ famous line, ‘One’s mind, once stretched by a new idea, never regains its original dimensions.’” Wow; I was very, very impressed with this. But apparently you had to have a particular mindset to begin with, so an epidemiologist who commented on the paper said, “Yes, this paper was pretty good but it’s had little impact. McKeown’s interpretation remains the dominant one in epidemiology,” he said, and he is right. The article has proven much more influential in the social sciences than in the health sciences, which is paradoxical because I was a social scientist saying that economic development did not matter, that the explanation was in the health domain, whereas McKeown was a physician and epidemiologist basically saying the explanation lay in economics rather than public health and medicine. I do not have an explanation for this, but I am afraid that it reflects in part the fact that we are still not beyond the point where we listen more carefully to people who are wearing our own disciplinary badge than we do to others.
Later I taught a graduate course at Penn in which the students and I analyzed large amounts of data that were becoming available around the world on child mortality. The data were drawn from censuses and surveys asking women how many children they had ever borne, and how many of those children were still alive. There were 15 students in the class, so we analyzed 15 countries and 15 variables. The result was a monograph published by the United Nations entitled “Socioeconomic Differentials in Child Mortality.” This monograph, I think, contained an important discovery. In country after country, child mortality was more responsive to the mother’s years of schooling than it was to the father’s years of schooling. This finding was at odds with the prevailing economic determinism of the time, since household income was more closely associated with father’s schooling than with mother’s. So this was another blow to economic determinism. Our work was followed up by many others and the generalization held. This relationship became one of the observations supporting a reorientation of development efforts away from steel mills and towards human capital, particularly toward the conditions and circumstances of women, who were finally being recognized for their critical roles as workers and household managers. In my files is a copy of a memo duplicated and sent to me by a thoughtful staff member at the World Bank. The memo was from James Grant, then head of UNICEF, to Robert McNamara, then head of the World Bank. It was a summary of our work on which McNamara had scribbled, “Fascinating, Jim;” a nice little trophy of mine in my files.
I should have stayed in the health area, but in the mid-1980s, I gave a Presidential address to the Population Association of America entitled “Children and the Elderly: Divergent Paths for America’s Dependence.” The article pointed out that on many indicators such as poverty and violent death, older people had been doing much better over time, and children had been doing much worse. It offered some explanations in terms of changing political constituencies and the dramatic changes that had occurred in the American family. A condensed version of the article was published in Scientific American and it set off a very vigorous reaction. Many advocates for the elderly resented the article because it undercut the image of older Americans as especially vulnerable. AARP’s journal Modern Maturity wrote an editorial calling me America’s leading crusader against the elderly. Within 24 hours I had calls both from my parents and my wife’s parents, asking me to defend myself. On the other hand, advocates for children, such as Pat Moynihan, found much in the article to support expanded programs for children. We began an extended correspondence and he cited my work in an article that appeared on the front page of The New York Times and in the first paragraph of his book “Family and Nation.” I have several file drawers full of reactions to this article, which was said by some to have set off a decade of generational warfare, including the formation of a very aggressive organization called Americans for Generational Equity (AGE). I tried to defend myself in Congressional testimony and before AARP tribunals and elsewhere, and my principal defense was the virtues of comparative analysis. We learn more by looking at these two groups together than we would by looking at either of them separately. This was not a convincing rationale to those who believe that everything is political. But everything is not political and I think Pat Moynihan, as a politician, understood that perhaps better than anyone. Social scientists with their broad perspectives and rigorous methods can help set the terms of discussion that are as much cultural as they are political. So, almost no one wants to set off a war, but in the end I think my article contributed to a national dialogue that had to occur at some point. There seems to be more realism now about the expansion of programs for the elderly, brought about undoubtedly by changing demography more than by demographers. On the other hand, despite the frequency of allusions to children in political campaigns, we have privatized childbearing to an even greater extent than earlier, and significant factions of the private sector are more poorly equipped than they were earlier to shoulder the burden. This remains, I think, one of the major dilemmas of American society.
Samuel H. Preston is Frederick J. Warren Professor of Demography at the University of Pennsylvania.
Preston is the the 2008 AAPSS Samuel A. Stouffer Fellow. These remarks were delivered upon accepting his Fellowship on May 8, 2008.
Press play to hear recording.