Research on Children and Families

The demography of children and families has become an increasing concern of population research. Dimensions of interest include various aspects of child and family well-being, as well as the individual, family-, and community-level processes and domains of decision-making that influence well-being. Relevant aspects of well-being include physical and mental health, socio-economic status, and access to and utilization of resources such as health care, early care and education, and schooling. Areas of decision-making include family formation and dissolution, consumption and savings, intergenerational exchanges, and investments in human capital through schooling, health, residential mobility, and migration.

Research on children and families must confront a number of methodological and analytical challenges. First, there are multiple dimensions of well-being and multiple-levels of factors that influence decision-making. This poses a challenge in terms of having appropriate data to capture the outcomes of interest and the determinants at the individual, family, or community level. Second, the interrelationships between behaviors are complex, and, often, causality runs in both directions. Decision-making takes place over the life-course; thus, outcomes may be best understood in a dynamic context. Thus, researchers must develop appropriate models to account for simultaneity and dynamic processes.

Third, there are often measurement challenges associated with key dimensions of well-being such as health, income, or wealth. A related issue is that some population subgroups may not be captured in large enough numbers to be studied in depth.

Contributions by PRC Staff

These complex issues are best understood from a multidisciplinary perspective that brings together methods from demography, economics, psychology, sociology, and other fields. Taking this approach and using data from both the United States and overseas, the RAND PRC scholars have made recent contributions and continue to add to our knowledge base in these areas.

Family Formation

Loughran (2002a) examines the linkage between rising inequality and increases in the age at first marriage and provides evidence that increasing male wage inequality has affected family formation within geographically and demographically defined marriage markets. Loughran and Zissimopoulos (2004) study the impact of age at first marriage on career development and wages finding that the wages of women in particular benefit from delayed marriage. They also find that women who marry early tend to marry husbands with relatively high wages. In ongoing work, Loughran, Zissimopoulos and Lakdawalla are exploring how well women predict future fertility and whether delayed childbearing is more related to choice or chance events. Rendall has been developing methods for combining survey and population data to improve estimates of fertility and family behavior, with applications to populations in the United States and Europe (Handcock, Huovilainen, and Rendall, 2000; Hancock, Rendall, and Cheadle, 2003). An important application of this work is to develop unbiased estimates of men’s family formation and dynamics (Rendall et al., 1999, 2001; Rendall, 2003). Bitler has also studied the impact of the Medicaid expansions of the early 1990s on overall fertility and birth spacing (Bitler and Zavodny, 2002c).

Back to Top

Family Maintenance and Dissolution

Karney and his collaborators have been using repeated-measures longitudinal designs to examine the effects of environmental stress on stability, change, and divorce in the early years of marriage. This work has provided direct evidence of the constraints that environmental stressors place on relationship maintenance processes (Neff & Karney, 2004) and has shown how chronic and acute stressors interact to affect long-term marital outcomes (Karney, Story, & Bradbury, in press). Karney also directed survey research commissioned by the state of Florida that compared the correlates of stable families across high, middle, and low-income groups (Karney, Garvan, & Thomas, 2003). This work is being used to support critical analyses of proposals to promote and strengthen marriage among low-income populations (Bradbury & Karney, 2004; Karney & Springer, 2004).

Back to Top

Income, Wealth, and Consumption

Smith has studied the determinants of wealth and savings among couples and demonstrates that falling rates of marriage can explain a large share of the decline in household savings and wealth levels (Juster, Lupton, Stafford and Smith, forthcoming), while Karoly and Klerman have studied the impact of welfare reform on income and broader measures of well-being (Grogger, Karoly and Klerman, 2002a, 2002b). Hurd and Rohwedder are currently studying the material well-being of the elderly using new data on total spending as an alternative to traditional poverty measures based on income. Hurd and Smith have used data from HRS and AHEAD to study the magnitude and distribution of bequests (Hurd and Smith, 2001). They find that actual bequests go overwhelmingly to family members and that among families with children, bequests are typically equally divided. Data on anticipated bequests suggest that the typical or median bequest to the baby boom generation will not be large, contrary to impressions created in the popular press. However, Rendall has shown that bequests to the baby boom generation will operate to increase racial economic inequality (Avery and Rendall, 2002).

Back to Top

Immigration and Assimilation of Immigrants

Smith and his collaborators have contributed to our understanding of the economic progress of legal immigrants with estimates from new nationally representative survey data that both male and female legal immigrants experience a substantial economic gain (over 60 percent increase in earnings) in their first job in the United States compared with their last job abroad (Jasso et al., 2000a). Meanwhile, Rendall’s recent research indicates that the fertility contribution of recent decades’ waves of Mexican immigrants is likely to be much greater than previously projected (Jonnson and Rendall, 2004). Internationally, Rendall and colleagues also explored the use of household survey data to measure and analyze the country origins of European migration flows (Rendall, Tomassini, and Elliot 2003).

Back to Top

Family Planning

DaVanzo, using experimental data from the Matlab region of Bangladesh, finds that the better services available in the treatment area reduced rates of unintended pregnancy and led to lower rates of abortion (Rahman, DaVanzo, and Razzaque, 2001). The research explains how it is possible for abortion rates to rise during the fertility transition. In particular, as women in Bangladesh became more intent on controlling their fertility and not exceeding their family size goals, the likelihood that intended pregnancies would be aborted increased. In the comparison area of Matlab, this increase was large enough to more than offset a decline in the incidence of unintended pregnancy. DaVanzo has also studied the impact of the economic crisis on child schooling in Cameroon (Eloundou-Enyegue and DaVanzo, 2003).
At the international level, DaVanzo has participated in a multicountry study of family policy regimes and their effects on overall fertility levels, generally measured by the total fertility rate. Rendall’s work includes both two-country (Ekert et al. 2002, Rendall et al. 2004) and multicountry (Rendall et al., 2003) comparisons of fertility that include detailed analysis of fertility timing and parity progression by education and occupation, as well as analysis of the differences in family-policy regimes associated with different life-course fertility patterns. Among their findings is an increasing divergence in age at first childbearing in the UK along occupational lines, contrasted with a universal shift towards later childbearing in France. They interpret these findings in the context of the UK’s means-tested family policy and more comprehensive, “universalist” family policy of France.

Back to Top

Role of Neighborhoods and Communities

PRC staff have been examining concepts for defining neighborhoods and the impact of neighborhood characteristics on child and adult outcomes. For example, Sastry and Pebley are using data from the L.A.FANS to investigate how adults and children define their own neighborhoods and the importance of neighborhoods to daily life (Sastry, Pebley and Zonta, 2002). The same data are shedding light on the contribution of neighborhood factors to adult and child health (Pebley and Sastry, 2002; Sastry and Pebley, 2003) and to children's reading and mathematics achievement (Pebley and Sastry, 2003). Finch has also examined neighborhood impacts, specifically on health and risky behaviors, showing geographic clustering in such outcomes as adult substance use and the impact of neighborhood characteristics on substance use (Finch, 2001; Finch, Kolody, and Vega, 1999; Finch, Vega and Kolody, 2001). Focusing on a lower-income setting, Sastry is using data from Brazil to study the role that changes in community characteristics, such as public health measures, have played in shaping the improvement and the pattern of differentials in child health and survival between the early 1970s and the present day (Sastry and Burgard, 2001; Sastry and Ferreira, 2001; Sastry, 2004).

RAND’s current and future work on these issues draws on numerous data sources, including three innovative data collection projects (IFLS, L.A.FANS, NIS) mentioned above.

Back to Top

Future Directions and Scientific Objectives

Through these funded and pending projects and future projects that are still in the planning stages or further on the horizon, our future research in this area has two primary scientific objectives:

  • To improve our understanding of key dimensions of child and family well-being and their determinants
  • To add substantially to what is known about the role of neighborhoods and communities in determining child and family well-being and in enhancing human capital investments in diverse and changing environments

Examples of these new directions include:

The Impact of Immigration and the Economic Progress of Immigrants

The NIS will be an incomparable resource for evaluating the assimilation and experiences of immigrant adults and children, the impact of immigration on the United States, and the impact of U.S. immigration law. The data will support analyses of the health and skill composition of immigrant entry cohorts; the transitions between legal and illegal statuses; the geographic mobility of migrants, including return migration; the process of family formation and the role of the family and social capital in the immigration process; and the contribution of immigrants to the U.S. economy. The survey will provide rich information on health, on economic status, on schooling, and on children’s well-being from a population heterogeneous in English language abilities and in native languages. The NIS will allow researchers to overcome biases inherent in cross-sectional datasets typically used to study immigrant populations (e.g., Census, CPS), thereby providing more reliable assessments of the dynamic nature of the immigration process and immigrant experiences in the United States. Smith will lead RAND’s future research on these issues, drawing in other researchers over time with interests in this area.

Role of Neighborhoods and Communities

Although characteristics of communities are often hypothesized to affect individual behaviors and outcomes, rarely do we have longitudinal data on both households and on the communities in which they are located. At the same time, given that families have some degree of choice over where they live, efforts to quantify the causal contribution of community-level factors are confounded by the selectivity of residential choice. The rich longitudinal data being collected by the IFLS and L.A.FANS are ideally suited for addressing these issues and overcoming methodological problems in prior studies.

Using these data and complementary surveys in the United States and overseas, PRC staff will investigate the contribution of neighborhood resources and characteristics to child academic achievement, child health, youth schooling attainments, and other outcomes. The research will delve into the most salient definition of neighborhoods and communities and the characteristics that appear to have the most impact on child and family outcomes. The relationships between neighborhoods and social networks will be another area of future inquiry. The research will exploit spatially defined data and use appropriate tools for analyzing such data. Methodological advances in approaches for defining neighborhoods, for measuring characteristics at the local level, and for using multi-level modeling techniques with panel data will also advance research in this area. These analyses will be conducted by a number of PRC staff, including Finch, Klerman, Sastry, and Strauss.

In investigating many of these issues, PRC researchers are in a unique position of both designing and collecting the data for the IFLS, L.A.FANS, and NIS and conducting analyses using these data. In this way, the surveys reflect the scientific objectives of the projects. At the same time, findings from the research projects based on early waves of the data will help to inform subsequent waves. Ultimately, we are able to collect better and more appropriate data because the same research teams both collect and analyze these unique surveys.

In addition to advancing work in these areas, other projects will contribute to our scientific objectives by:

  • Analyzing delayed childbearing and childlessness to understand to what extent these outcomes are determined by choice and by chance (Lakdawalla, Loughran, Zissimopoulos)
  • Analyzing family formation and childbearing to better understand the determinants of the rising age at first marriage and the changes in teen and out-of-wedlock childbearing, with follow-on work to determine the consequences for child well-being (Klerman, Loughran)
  • Further developing new methods for joint analysis of population and survey data, with particular applications to fertility and family structure (Rendall)
  • Examining trends in demographic outcomes, their social consequences, and the implications for policy in countries in the European Union (Davanzo and Rendall)
  • Linking data on bequests across several generations to predict the impact of future bequests on the economic status of the grandchildren of the deceased (Hurd and Smith)
  • Using data from the HRS to study intergenerational exchanges of money and time (Zissimopoulos)
  • Studying especially vulnerable groups and specialized populations such as children of HIV-infected adults (Beckett)

These projects and others expected to be active over the five-year funding cycle share a common interest in understanding the diversity of experiences across different demographic groups defined by race, ethnicity, sex, family structure, and other characteristics. Both disadvantaged and advantaged families and children will be studied to further shed light on the full range of outcomes and their determinants. The projects recognize that child and family well-being is multidimensional, so they explore a rich array of subjective and objective measures available through high-quality survey data, including data collected by RAND. The relationships are also viewed as multilevel, so demographic data at the individual and family levels are combined with spatial information to investigate the contributions of factors at different levels of aggregation. Where possible, the future work will exploit longitudinal data to assess the dynamics of well-being, along with differences at a point in time.

Repeat observations on the same individuals over time, or on other family members in a single cross-section, provide an opportunity for many of these projects to control for both observed and unobserved individual- and/or family-level determinants of well-being, thereby strengthening inferences about causal relationships. The ability to make causal inferences is further enhanced by the fact that some of the data sets span periods of large exogenous changes in the environment. Future work will also seek to exploit cross-national data to examine differences in child and family well-being in both developed and developing countries.