Ted Mouw
University of North Carolina, Chapel Hill

Presented at the NICHD panel,
"Visions of the Future: A Town Meeting on New Directions in Population Research"
Annual Meeting of the Population Association of America
March, 2000

In this presentation, I want to address two questions. First, what is demography? Second, given a working definition of demography, what are some promising fields of demographic research?

What is demography?

The reason I think this question is relevant to a discussion of new directions in demographic research is that I wonder what happens as we follow these "new directions". If these new demographic research areas lead us outside of traditional demographic areas, are we still demographers? To begin with, the very fact that we are having this town meeting suggests to me both a sense of optimism in the field about the expanding possibilities and new directions of demographic research and a sense of unease about what the identity of the field will be as a result of those developments. In a sense, the question of the identity of demography is already pertinent. It is impossible, for instance, to read the titles of the paper presentations for this year’s PAA and not wonder about the authors’ common connection as demographers. As a sociologist, I have always dealt with this question of disciplinary identity in relative terms---the thought of demography having an identity problem has always paled (in my mind) next to the perennial crisis of identity, ideological rifts, schisms, and fragmentation among sociologists. To claim to be a demographer at a cocktail party has often been easier and less anxiety provoking for me than to claim to be a sociologist. However, let me try to be more specific.

I will frame my answer to this question of demography’s identity as a response to a comment in the Summer 1999 newsletter of the American Sociological Association’s Sociology of Population Section. In this article, Dan Lichter claims that there is a generational shift going on in population research and training. Simply put, he argues there is a division between pre-1980 trained demographers who worry that the field is losing its intellectual core and post-1990 trained social demographers who are less interested in the traditional subjects of demography—fertility, mortality, and migration—than in a diverse set of population related issues (of which he mentions a few: welfare reform, aging, child development, teen pregnancy, economic inequality, and the spatial concentration of poor minorities). His focus on demographic training programs suggests the image of "barbarians at the gates" – young graduate students who would invigorate inter-disciplinary research topics but who would risk blurring or even losing the notion of demography as a distinct field of inquiry. Lichter asks: what is demography? "Is it simply a bag of tools? Or is there an intellectual core that must be reinforced in the current crop of students, lest it is lost forever? Or is demography defined simply by what (self-identified) demographers do?"

Personally, as a 1999-trained social demographer, I can say honestly that I do feel like a barbarian who made it through the gates. What attracted me to demography was not a formal training in fertility, mortality, and migration, and while in graduate school I did my best to evade being trained in these areas. However, at the conclusion of my graduate training I did feel like a demographer. But it was not a formal training in the core areas of traditional demography that made me feel that way. In addition, and forgive me for being irreverent, I do not believe that there is a common intellectual core to demography that connects all of the various corners of the field. However, I also believe that demography is more than a bag of tools or whatever demographers happen to be doing. Instead, I think that what connects us as demographers is a "demographic perspective" to doing social research. Let me call it a "reflex" to emphasize the distinction from a conscious intellectual commonality. I do not claim to be original here, but let me list the aspects of this "reflex":

When we look at the various sub-fields of demography supported by Demographic and Behavioral Sciences Branch of the NICHD (family and household demography, fertility, HIV and STD’s, mortality and health, population composition and change, and population movement and distribution), the difficulty of defining demography on the basis of either a shared intellectual core or "just what demographers do" is evident—aside from the fact that all the research is population related. Even within the final subfield, Population Movement and Distribution, the connection between them, as listed on the DBSB web page, is rather broad: "the determinants and consequences of international and internal migration, economic and social mobility, spatial demography, and the relationship of population change to the physical environment." What connects these fields to each other is not a shared set of intellectual interests but a common way of approaching research questions: begin with an empirical description of the population/s being studied and end with a careful analysis that hesitates to elide complexity of the data in the pursuit of theoretical simplicity.

Promising directions in population research

I will now move on to talk about a few promising directions in population research. In each of these areas, an important question is whether or not the best research on the topic—in other words the research that will answer the big questions—will be considered "demographic." In each of these topics I think there is a tension between demography as an intellectual core or a set of clearly defined subject areas, and demography as a way (or reflex) of doing social research.

We can divide promising future research topics into 2 categories: those that are important and interesting because they represent emerging social trends that are in need of an explanation and those questions that become empirically interesting as the result of changes in research technologies or methodological innovations. I will not try to be comprehensive by pretending to be scholarly on things I am just guessing about, but I will touch on four areas that have some experience or interest in.

Population-related problems in need of explanation.

1) Population size & the environment. As world population growth rates have fallen, it is tempting to think of the population bomb—the primary motivation behind the expansion of demographic research 3 decades ago—as having been defused. Does the fall in growth rates really mean that demography has entered into a new post-Malthusian phase? It is tempting, as researchers, to hedge our bets and to turn our attention to new "crisis" topics such as population aging and below-replacement fertility. (After all, the population-environment problem could turn out to be the 21^st century’s Y2K bug). Nonetheless, even medium range population projections suggest that the world’s population will come close to doubling again before it stabilizes, and the inherent uncertainty of these projections means we must put very wide confidence intervals on these estimates—tiny changes in the trajectory of total fertility rates result in large fluctuations around these projections. I would argue that the impact of large populations on local and global ecosystems is an important demographic topic if we base our interest on some sort of utilitarian "social problems" framework. What can demographers contribute to this topic?

As opposed to the question of population growth rates, the question of the environmental impact of population size is an interdisciplinary question. PhD-trained demographers feel out of place in the biological and physical scientific literature on the environment. It is tempting to leave the effects of population size to others, and concentrate on refining estimates of the confidence intervals on population projections. Moreover, reading the literature on global warming, erosion, acid rain, desertification, food production etc., one often gets the impression that population projections—with all of their lack of accuracy—may just be the most predictable aspect of the whole debate on global environmental change. Would a demographer with the necessary interdisciplinary training and interest to understand this problem still be considered a demographer?

2) Rising income inequality. Twenty-five years of rising income inequality in the United States is another population topic that is in need of an explanation. The importance of the demographic reflex in this case is that one should begin a conversation about inequality with the facts first: inequality has increased, but while this is worrisome, the overall impact may be similar to previous 20^th century fluctuations in the wage structure (i.e., wealth and income inequality were probably higher in the first three decades of the 20^th century than they are now¹). We should be careful about projecting current trends into the future until we understand the underlying causes at work. Demographers have already played key roles in documenting rising inequality and attempting to determine whether it can be explained by demographic factors (changes in family structure, rising female labor force participation, marriage markets etc.). However, after these trends in inequality have been carefully documented and population-level explanations have been dismissed, the question of why inequality is rising (which is important in the same sort of utilitarian way as population and environment) becomes more complicated as a demographic research agenda. What are the underlying causes? I would argue that—if trends in inequality continue—a central question in this literature should be whether or not rising inequality is due to the effects of technological change. At the moment, technological change is the default explanation in the economics literature, not because it has been empirically tested, but only because the empirically verifiable explanations have come up short. The usual data sets demographers use to measure changes in inequality—i.e. the Current Population Survey—provide a good means of tracking changes in inequality, but cannot really get at the question of why. I do not have perfect insight here, but my intuition is that actually getting a handle on the impact of technological change on the wage distribution would involve a return to some of the methods of classic industrial sociology—firm-level observation of the social dynamics of workplace interaction. Just like the question of the impact of population size on the environment, this poses the need for interdisciplinary research and also raises the question of whether the resulting research still falls under the rubric of demography.

Research innovations.

Let me now turn to two areas for demographic research that are promising not so much because of salient trends but because of innovations in research methods or data.

3) Geographic Information Systems. The growth of GIS software now makes it possible to easily map demographic data and look at spatial relationships. GIS software allows us to exercise our demographic reflex by looking at descriptive population statistics spatially. Mapping descriptive population characteristics—quickly and easily with the new GIS software—demonstrates the importance of physical space in local social systems—such as urban areas. For example, looking at a map of the percentage of black or white households by census tracts provides insight into patterns of residential segregation that you cannot get otherwise. This may lead to new innovations in the measurement of residential segregation that explicitly incorporate spatial patterns not readily visible with unmapped, tabular data. However, while GIS is an important descriptive tool in demographic research, it is more difficult to actually take the step of incorporating space into the analysis. This is for two reasons: First, the lack of population level behavioral models in which space plays a causal role means that in many cases it is difficult to argue why space should be an explanatory factor—despite descriptive evidence of spatial heterogeneity. Second, even in cases where it seems like geography should matter, the voluntary sorting of households into different neighborhoods across space makes it difficult to argue that location itself is causing geographic differences in the outcome variable. For example, in the literature on neighborhood effects—the effect of living in a good or bad neighborhood on socioeconomic outcomes—it is always reasonable to ask whether living in a good neighborhood helped you get a good job or whether having a good job helped you live in a good neighborhood. This sort of question underscores the difficulty of going from population maps as a descriptive tool for demographers, and geographic variation as an explanation for demographic outcomes. So again we must ask ourselves what is the aim of demographic research. Is it to provide numerical descriptions of population patterns? If so, then it is easy to recognize the power of GIS to facilitate demographic research. However, if the purpose is to understand causal processes, then I would argue that, at the moment anyway, GIS poses as many questions as it answers. Let me ask this question about GIS research: As in the case of income inequality, after the interesting descriptive questions have been carefully answered and the demographer turns towards the more difficult causal and analytical questions, is he/she still doing demographic research?

4) Social Networks. Research using social network data is similar to GIS in the promise and difficulties it brings to the field. Again, like GIS, social network research is not so much a separate research topic as it is a tool for adding to the realism of the data that we use—only in this case incorporating social structure (i.e. "who knows who" represented by network ties) rather than physical space. In descriptive analyses, social network data seems easy to incorporate within demographic research. For example, social network data can be combined with residential location data to look at the relationship between geographic segregation by race and social segregation by race. For the demographer, this provides insight on another dimension of segregation and at the same time suggests a potential critique of using census-tract defined neighborhoods as a meaningful social category. If the purpose of network data is to provide richer descriptions of populations, then, like GIS, its importance is clear. On the other hand, however, when demographers incorporate social network data into statistical models as independent variables, it difficult to argue that the coefficient on network variables represents a causal relationship. Network data may be systematically associated with unobserved individual characteristics. After all, we all acquire friendships in very non-random ways—this is the notion of social homophily (we acquire friends that are like ourselves). The cautiousness of the demographer in making causal statements makes it difficult to incorporate cross-sectional network data into statistical models. If we wish to go beyond this, then the type of data that we would need to collect (quasi-experimental) and the interdisciplinary nature of the resultant research brings up the question of whether the project would still be demographic.

In conclusion, I would like to stress that despite my criticisms of the new directions for research that I myself have proposed, I am very optimistic about the future of demography. However, I do feel that there is a tension (perhaps not really a crisis) about what demography really is, and whether the promising research areas of the future will reinforce or dilute our common identity as demographers. Ultimately, I believe (and I do not expect everyone to agree with me) that if we think of demography as a way of doing social research rather than a fixed list of topics, we have no reason to worry about the future health or vitality of the field as a whole.

Footnotes

¹ Plotnick, Smolensky, Evenhouse, and Reilly. 1998. "The Twentieth Century Record of Inequality and Poverty in the United States." Institute for Research on Poverty Discussion Paper no. 1166-98. University of Wisconsin.