Seminar Series

We use an innovative strategy for linking parents to their adult children in the United States census to produce estimates of the intergenerational transmission of socioeconomic status from 1850 to 1940. We begin with data from a large, online, crowd- sourced genealogy platform (familysearch.org), which includes millions of users who personally link records to the profiles of their family members. We include the links created by these users in our data set, but also use information from the links they create to inform other supervised and unsupervised matching methods. Our completed data set, which we call the Census Tree, contains hundreds of millions of links among the 1850-1940 censuses. This data set is beyond the current frontier in terms of the precision, recall, and representativeness of the included links. We use these data to produce estimates of the intergenerational transmission of characteristics including occupation score, literacy, and fertility. Because family members do the linking and often know the maiden names of women in their family, we are able to include women in our analysis where previous research has omitted them.

Despite the centrality of structural explanations for understanding racialized inequality, less than one percent of studies on the link between race and health have focused on structural racism. Moreover, the conceptualization of structural racism in the race theory literature has often differed from the measurement strategies used in population health research.  This study advances the field by 1) distilling central tenets of theories of structural racism into concrete measures of structural racism, 2) conceptualizing U.S. states as racialized institutional actors shaping health,  3) developing a novel latent measure of structural racism in states across multiple domains, including political participation, education, economics, housing, and the judicial system, 4) mapping structural racism across states, and 5) quantifying the effects of structural racism on six individual-level health outcomes and state-level COVID-19 mortality among Black and white adults

Short birth spacing, a birth-to-conception interval of under 18 months, is common in the United States, with over one third of all second or higher order children conceived under 18 months after the birth of their older sibling. While associations between short birth spacing and adverse health outcomes immediately post-birth are well-documented, the linkages between short birth intervals and long-term child well-being remain understudied. In addition, studies linking birth spacing to long-term child health rarely account for birth order. In this talk, I outline a program of research to fill this research gap and examine connections between birth spacing and order and child health, focusing on the risk of abuse and neglect in early childhood. Using birth records data in North Carolina for the past 25 years linked to multiple administrative data sources, I estimate the probability of child maltreatment investigations within the first five years of life by family size and birth-to-conception interval. I also evaluate whether current recommendations for birth spacing are relevant for child maltreatment prevention and identify sub-populations at the highest risk of child maltreatment following short birth spacing. My findings offer a nuanced view of how family size and birth intervals can affect child well-being as well as a critical consideration of the role that poverty plays in driving such associations.

In the U.S., early childhood investments such as breastfeeding and daily reading are strongly promoted by pediatricians and public health campaigns as critical investments in children’s health and cognitive development. Qualitative research on gender, work, and family shows that women unambiguously find these investments difficult to combine with paid work. This study uses a nationally representative sample and an event study design to examine how breastfeeding intensity and reading daily to young children shape mothers’ labor supply, wages, and job characteristics over the short and long term in the U.S.

Studies on unemployment typically assess its costs on the individual level. However, when unemployment occurs, individuals, their families, and their kin all lose. Close kin provide the majority of social support for unemployed adults. Applying demographic and statistical techniques to official statistics and multiple survey datasets, we assess the prevalence of and exposure to unemployment in the United States from a kinship perspective. The results indicate dramatic racial disparities in the number of unemployed kin and the number of kin who would be affected by an unemployed person. Specifically, during the pandemic-induced recession, black Americans have 1.7unemployed people in their extended family compared to 1.2 among whites. Further, every job loss in a black extended family affects approximately 23 related members of the family through kinship ties, but this number among whites is only about 20. The findings of this study draw attention to the need for an understanding of unemployment and its demographic implications, which are stratified by race.

The United States is undergoing major demographic changes including an increase in life expectancy and a rapidly growing aging population. These demographic shifts are expected to strain social assistance programs and families. While the extended family plays a significant role as a source of caregiving and financial support to its members, decision-making within the extended family is relatively understudied due to lack of adequate data. In this project, I employ rich data from the Panel Study of Income Dynamics (PSID) and extend economic household decision-making models to examine how extended families allocate resources. I find that families do not pool their resources or allocate them in an efficient manner, leaving welfare gains on the table. I then explore why efficiency may not prevail in extended families by examining information asymmetry. I present descriptive evidence for the existence and extent of information asymmetry in the family, and I find that families with better information allocate their resources more efficiently. These findings provide suggestive evidence that information asymmetry may hinder efficient resource allocations in the family.

The life course perspective has long theorized that adversity in childhood, a sensitive period for mental, physical, and emotional development, can have long-lasting impacts on health and wellbeing. However, research on the long-term impact of childhood adversity has been disproportionally focused on studying a single adversity, or studying cumulative adversity operationalized as the sum of dichotomous (yes/no) indicators reflecting exposure to negative events. Although informative, these approaches mask how specific types of adversities co-occur, and how unique configurations of adversities relate to outcomes of interest. Using nationally representative data from the National Longitudinal Study of Adolescent to Adult Health (Add Health; Wave I-IV), I estimate cumulative adversity using latent class analyses. As opposed to a summative score, these classes reflect both the type and number of adverse events that may co-occur in childhood. I then investigate how these latent classes of adversity predict depressive symptoms from adolescence into early adulthood, clarifying the long-term mental health risks of early life adverse events. Throughout this study, I discuss the methodological benefits and challenges to estimating cumulative adversity using a latent class approach.

The US has a long history of housing discrimination.  Since the passage of the Fair Housing Act, HUD has sought to measure the extent of that discrimination with a series of audit studies.  Critiques of audit studies have questioned the ability of these and other similar experimental methods to measure discrimination on the margins where individuals are affected in actual market contexts.  We implement the largest correspondence study of discrimination in rental housing markets, and show how combining the results with actual market outcomes and with structural estimation techniques can confirm that discrimination indeed has market consequences, and that sorting behavior on the part of market participants can actually make the consequences of that discrimination worse.

The study of social inequality and stratification (e.g. ethnoracial and gender) has long been at the core of sociology and the social sciences. I argue that certain tendencies have become entrenched in our dominant paradigm that leaves many researchers pursuing coarse-grained analyses of how difference relates to inequality. Centrally, despite the importance of categories and categorization for how researchers study social inequality, contemporary theories of categories are poorly integrated into conventional research. I argue that the widespread and often unquestioned use of State categories as categories of analysis reinforces these tendencies. Using research on skin tone stratification (or colorism) as an inspiration, I highlight several components of an alternative model that better integrates contemporary theories of categories into measuring the social difference. Above all, this model proposes an analytic shift in focus from membership in categories to the cues of categories, membership in sub-categories, and perceived typicality.

Coercive labor (adult labor trafficking and child labor) is astonishingly prevalent worldwide. Poverty creates vulnerability to labor coercion, but quantitative evidence on how anti-poverty programs mitigate this vulnerability is scarce – particularly for adult labor trafficking. This paper provides new evidence on how conditional cash transfer programs, cornerstones of anti-poverty policy in lower-income countries, influence coercive labor risk, focusing on Brazil’s Bolsa Familia program. Using multiple regression discontinuity designs, we find evidence of limited effects on adult labor trafficking risk. By contrast, we also find that the program does substantially reduce child labor among the poor – but not among those classified as living in extreme poverty. Our results suggest that for adults, income gains alone may be insufficient to reduce labor trafficking risk -- and complementary action against criminal recruiters may be simultaneously required. Coercive labor (adult labor trafficking and child labor) is astonishingly prevalent worldwide.  Poverty creates vulnerability to labor coercion, but quantitative evidence on how anti-poverty programs mitigate this vulnerability is scarce – particularly for adult labor trafficking. This paper provides new evidence on how conditional cash transfer programs, cornerstones of anti-poverty policy in lower-income countries, influence coercive labor risk, focusing on Brazil’s Bolsa Familia program.  Using multiple regression discontinuity designs, we find evidence of limited effects on adult labor trafficking risk.  By contrast, we also find that the program does substantially reduce child labor among the poor – but not among those classified as living in extreme poverty.  Our results suggest that for adults, income gains alone may be insufficient to reduce labor trafficking risk -- and complementary action against criminal recruiters may be simultaneously required.