Jennifer Hirsch
Emory University

The DBSB could play an important role in supporting research that helps us learn about the contribution of cultural factors to demographic phenomena. This would not necessarily mean trying to squash culture into quantitative models, though it might certainly incorporate that task. I see three specific areas of opportunity:

• What do our dependent and independent variables mean to people, and-looking across social groups and across time-what is the diversity of meanings?
• In what cases is it possible and useful to construct ethnographically informed, quantitatively robust, measures that reflect the range of cultural meanings?
• How can we use these measures to explore the interconnections between culture, social structure, and health?

1. Culture as "Attitudes": Include a Structure and Strategy Perspective; Attend to Heterogeneity
   
   If you think about culture as what is meaningful to people, then clearly there are a set of questions here relevant to the DBSB's research mission (e.g., explorations of the socially constructed meanings of fertility, reproductive health, mortality and morbidity, migration, family and household structures). We could profitably explore the range of local meanings people give to our independent (e.g., gender ideologies, women's labor force participation) or dependent variables (population processes). There has already been a good bit of this descriptive cross-cultural research in population studies, particularly in terms of social structures and behaviors that are understood to be related to negative public health outcomes (e.g., work on the cultural constructions of sexuality among men who have sex with men, symbolic meanings of condom use, family structures among urban ethnic minorities in the United States, meanings of fertility in sub-Saharan Africa) or, in the case of migration, to politically problematic aspects of population processes (there is a large-and growing-anthropological literature on migration, some of which has been supported by the DBSB). Across all of the Branch's substantive areas, some niches have been relatively neglected. We know very little, for example, about heterosexuality as a cultural construct, or about symbolic meanings of contraceptive methods other than condoms, or about the meaning of single parenthood to the parents and children in these fragile family structures. For each of the substantive areas of opportunity, those most familiar with the research could certainly generate a list of dependent or independent variables for which it might be useful to learn more about local meanings.
   
   The DBSB could also work strategically to improve the theoretical and methodological quality of research on cultural factors within population studies. Much of this research fails to address the fact that "cultures" are not internally homogeneous, undifferentiated masses. For example, theoretical sampling-creating samples that differ across social strata-can help us understand how shared interpretations of demographic phenomena are differentially advantageous to actors, depending on social location, and thus might help people capture a less static, more "structure and strategy" approach to culture. Ensuring that those who employ qualitative methods are trained in a broader range of data collection techniques will also allow for the generation of more subtle understandings of the relationship between ideology and behavior.
   
   This raises the question of how the DBSB could best promote interdisciplinary research. Concretely, I see two ways to improve the methodological and theoretical quality of cultural research on the Branch's issues: either support better training for demographers interested in these approaches, or promote more collaboration between anthropologists and demographers. The Mellon Foundation is no longer supporting programs in anthropological demography, and although some of the joint-degree programs continue to turn out graduates who are very sharp methodologically and theoretically, there is the real possibility that the ongoing legacy of Mellon's support will be that demographers feel compelled to throw in a few focus groups, but are never provided with rigorous training in methodology and theory.
   
   

   Given these substantive and methodological considerations, the Branch might focus on several areas:

   • Pushing culturally-oriented researchers to consider how cultural meanings change over time and to explore the key social factors influencing those meanings. The qualitative research presented at PAA, for example, frequently depicts static, homogeneous portrayals of cultural factors. The DBSB ought to encourage more theoretically sophisticated work on culture as dynamic and historically variable. For example, power is completely invisible in much current research on cultural influences.
     
     
   • Encouraging researchers to explore the relationship between meaning and behavior by working comparatively. Working comparatively (across social classes or cultural groups, for example), allows you to test empirically hypotheses about the relationships between culture and behavior.
     
     
   • Exploring the role of culture at levels other than the individual. Most of the people on whom DBSB grantees conduct research are somewhere between moderately and desperately poor; their understanding of their experience is not irrelevant, but we could use this concept of "culture" to study the belief systems of much more powerful groups. Culturally-oriented research on reproductive health, for example, has looked at the cultures of pharmaceutical companies, policy-makers and health systems (including the cultures of institutions which train reproductive health providers); the relationships between providers and recipients, and the role of family planning programs in promoting ideologies of reproductive modernity.
     
     
   • The anthropology of demography (as opposed to anthropological demography). There are two ways to help demographers avoid mistaking emic and etic phenomena, either (as above) encourage exploration of the local symbolic meanings of the things we try to measure quantitatively, or else encourage a more reflexive demography, in which demographers articulate and critique our own culturally-constructed worldview. This is a special case of the example above, of studying the culture of powerful organizations.
     
     
   • Methodologically, the best way to capture these local-level systems of meaning is through with intensive work that collects in-depth data, usually from relatively small samples. The application of qualitative methods in demographic research- frequently as an add-on to larger survey research projects-has been characterized by more enthusiasm than rigor. The DBSB might take the lead on supporting research that employs a wider range of ethnographic methods (participant observation, discourse analysis, life histories, in-depth interviews) in addition to focus groups, that employs a more thoughtfully articulated approach to ethnographic sampling, and that relies on more theoretically informed analytic approaches, including addressing issues of representation.
     
     
2. Connecting Ethnographic and Demographic Research
   
   Small-scale ethnographic research may be the best way to generate information on local meanings, but most of what demographers do is based on the universal applicability of quantitative measures. So how to integrate the two? The use of qualitative research as a formative tool to increase question validity is well accepted and fairly routine, but it's very time consuming to create locally relevant measures, and, of course, if we measure reproductive health differently in each place we work, then we can't make cross-country comparisons. So one project here would be to articulate areas in demographic research where it might be particularly useful to have ethnographically informed measures as outcomes; in other words, to think about areas in which culture might matter the most. The Branch's currently supported work on pregnancy intentions is a good example of this, as is Bledsoe's work in the Gambia.
   
   Methodologically, there are three general strategies for combining ethnography and survey work:
   • Using ethnographic work to explore unexplained patterns in survey data
   • Using ethnographic research to develop hypotheses to be tested at the population level
   • via surveys
   • Simultaneously using both approaches to study a problem
     
     
   Each strategy has pluses and minuses, but the last is particularly under-utilized
   
   The real opportunity here, however, is that anthropologists are increasingly interested in pushing studies of cultural influences beyond the local level. Susan Greenhalgh signaled this direction a decade ago, with her article on A Political Economy of Fertility. (Without going into an intellectual history of the discipline, a short explanation for this might be that the influence of post-colonial studies and the seminal work of Sid Mintz and Eric Wolf, combined with current concerns about the importance of globalization, have made many anthropologists much more aware that work conducted only on the local level risks irrelevance, and thus that it is crucially important to focus on multi-level analysis.) In a clever shorthand for this desire to put the role of cultural influence in its proper social context, Paul Farmer cautions against confusing cultural difference with structural violence. This leads to my third suggestion.
   
   
3. Integrating Culture and Social Stratification
   
   Work on cultural influences, without any consideration of the role of social structure or inequality, represents a theoretical stance that most critical anthropologists abandoned two decades ago (though there is a whole industry of applied anthropologists churning out research on cultural barriers and cultural determinants of this or that under subcontracts to the Agency for International Development). While there is some role for this sort of descriptive cross-cultural research, we should see it as only a starting point for the two key ways that demographers can engage with research on culture, which are:
   1. To use quantitative methods to link ideational factors to individual strategies
      
      
   2. To explore how meaning intersects with social stratification (social class, gender, race) to create socially and culturally variable opportunity structures (again, Bledsoe's work is a good example of this)
      
      
   The way to answer important questions about cultural factors without abandoning the demographic strength of quantitative measurement is not to try to measure local meanings with surveys; rather, we can use ethnography to identify behaviors (or even attitudes) that map closely with certain cultural forms, and conduct survey research measure those. So that in engaging with "culture", we should let demographers do what they do best-measure the distribution of etic phenomena to help us understand how new cultural forms are differentially advantageous. Anthropologists have excelled at developing finely grained analysis of the range of local meanings, but the field has not done a very good job of exploring the distributions of these meanings across populations, or how these meanings intersect with socially stratified access to resources. The point here is to use the traditional strengths of population research to evaluate the role of ideas in demographic change.
   
   These questions, of course, about the relative role of culture and structure go back to the Princeton European Fertility Project (and beyond), but this is not an either-or game; rather, it is an argument for a culturally-informed political economy, in which locally-variable meanings are just one of the influences on individual choice. There are a number of anthropologists who have worked in this vein: Feldman-Savelsberg, Rapp, Scheper-Hughes are just a few, and for all of them the explanatory power of their work would be significantly expanded by large-scale survey work. This focus on linking culture and social inequality fits in with one of the Branch's strategic goals, "encourage research that examines the interrelationships among health, socioeconomic status, and demographic processes over the life course," and also helps link that third goal to the fourth (encourage research to improve understandings of social, interpersonal and cultural influences upon individuals' and couples' sexual behaviors). The intellectual exchange between anthropology and population studies has been pretty one-sided, with anthropologists much less interested in learning from demographers than the other way around; but if anthropologists want to really explore the relationship between social structure and behavior, they will need to address questions of distribution. In other words, the anthropological interest in political economy might open the door to more equitable collaboration. People doing research on cultural factors have traditionally focused on the less powerful, but you can't see really see the effect of poverty on health if you only work with the poor. Again, this goes back to the question of sampling for cultural research: without socially diverse samples, you run the research of overemphasizing the importance of culture because social stratification is invisible.
   
   
4. Examples of Areas in which This is Already Taking Place, and Where It Might Happen More
   
   The area of gender and reproductive health is a good example of this over-emphasis on cultural factors and relative neglect of social inequality. At the recent Population Council/International Planned Parenthood Federation meeting on Power in Sexual Relationships, for example, all of the intervention research presented there conceptualized gender as men's and women's individual beliefs about masculinity and femininity, rather than as men's and women's socially structured access to resources. We (as social scientists) know about the importance of social stratification but it seems as if when we conduct qualitative research, we forget all that we know theoretically and focus only on individual attitudes and beliefs. The Branch could play an important role in supporting more theoretically sophisticated work on gender and health.
   
   Sexuality is another substantive area in which we know a lot about what Richard Parker calls erotic maps or grammars (the local-level meanings of sexual identities, sexual relationships, sexual behaviors), and not enough about how these "grammars" intersect with social conditions. Beliefs are not unimportant to understanding HIV risk behavior, but what is really determinative about a commercial sex worker's HIV risk is not only what she believes about sexuality (and about HIV), but the social conditions that led her into sex work in the first place. Similarly, lot of energy been expended on things like men's beliefs about condoms and sexual pleasure-not unimportant-but much less on factors like migrant labor, which may shape men's access to multiple partnerships in the first place
   
   Some current DBSB research already integrates some of these questions: the Three Cities study, and the study of the role of institutions in the politics of sexuality in Chicago neighborhoods are both good examples (p. 51, Program Activities and Progress report). Suggestions for areas in which we might do more includes:
   
   
   • Culture and teen pregnancy. The Program Activities and Progress Report notes that teen pregnancy is "strongly influenced by poverty, school-failure, lack of opportunity and family disruption." So then here we obviously wouldn't want to encourage people to think only about what sex, relationships, and childbearing means to these kids, but rather to put those questions in the context of social structure.
   • Fatherhood. If we are interested in exploring "what roles...men play in the events and decisions leading to childbearing," we might think about connecting these roles both to men's ideas about masculinities and also to gendered opportunity structures.
   • Immigrant health. Much attention has been paid in the news (and by medical anthropologists) to the exotic health beliefs of immigrants, but DBSB research has indicated that poverty and access to insurance are also vitally important. How are these things intertwined?
   • Contraceptive choice. What is the relative role of factors such as symbolic aspects of contraceptive methods, and what is the relative role of things like provider preference (culture at a higher level) and insurance reimbursement?