IPUMS30 Research Showcase Abstracts

Panel Session 1: Examining Marriage and Families

A Dream Come True: IPUMS International and the Study of the Family on a Global Scale. Albert Esteve

IPUMS emerged to revolutionize access to US census microdata but ultimately transformed global access. IPUMS International is the result of this remarkable global achievement, now covering 100+ countries and expanding to diverse demographic data sources (e.g. Demographic Health Surveys). In this presentation, I'll illustrate with my own work how IPUMS International empowers global family research, discussing the spread of non-marital cohabitation in Latin America, the impact of the gender-gap reversal in education on marriage markets, and the worldwide transformations in living arrangements.

Young Children and Parents' Labor Supply during COVID-19. Joanne Song McLaughlin

We study the effect of childcare needs during the COVID-19 pandemic on parents’ labor supply. Using monthly Current Population Survey data and following a pre-analysis plan, we implement three variations of an event-study comparing workers with and without (or less) childcare responsibilities. The first compares parents with young children to adults without young children, while the second and third rely on the presence of someone who could provide childcare in the household: a teenager in one and a grandparent in the other. Across these approaches, we find childcare needs did not negatively affect parents’ labor supply during the pandemic. We also do not find any difference in the likelihood of working between men and women. At the onset of the pandemic, many employers adopted flexible working arrangements. We provide evidence suggesting the ability to work remotely may have helped many parents avoid labor supply decreases.

Trends in marriage and health among sexual minorities in the US. Gilbert Gonzales

Community characteristics and armed conflict: Comparative research with IPUMS DHS. Elizabeth Heger Boyle

The Demographic and Health Surveys (DHS) are uniquely suited for contextual research because the data include longitude and latitude coordinates of sampling clusters, slightly displaced. Recent NICHD supplemental funding allows IPUMS DHS to begin disseminating these coordinates. This study links IPUMS DHS data to armed conflict and other contextual data for an analysis of whether and how local demographic and social characteristics can be useful in predicting the outbreak of armed conflict. Previous research on the outbreak of armed conflict has tended to focus solely on political and geographic (terrain) indicators; IPUMS DHS' geocoded and harmonized data make it possible to broaden the scope of these analyses.

Panel Session 2: Reconsidering Community and Geographic Contexts

Policy Backlash or Structural Inertia? A Micro-Macro Analysis of White Residential Decline from Desegregating City School Districts Between 1970 and 1990. Peter Rich

School desegregation plans are credited for exacerbating White suburbanization in the 1970s and 1980s, but this understanding draws from population-level causal estimates prone to ecological fallacy. This is the first study to systematically investigate whether White families were more likely to exit and more likely to avoid moving into city school districts because of mandated desegregation. Using annual family residential addresses from the Panel Study of Income Dynamics, geocoded by block and linked to school districts, I find that White residential exit did increase slightly in response to desegregation policy. In a micro-macro simulation model, however, I show that this small effect paled in comparison to a) the influence of White avoidance–both pre and post desegregation implementation, and b) demographic differences in the fertility and age profiles of suburban and urban residents established in prior decades. White population loss from city school districts was the result of structural inertia rather than an unintended consequence of desegregation policy. This conclusion has implications for the contemporary politics of structural reform, where fear of perverse incentives looms large. The tendency to analyze either micro-level policy responses or population-level trends, without linking the two, can reify assumptions and conceal complex sociological forces undertow.

Climatic Variability and Internal Migration in Asia: Evidence from Big Microdata. Brian Thiede

The potential effects of climate change on human migration have received widespread attention, driven in part by concerns about large-scale population displacements. Recent studies demonstrate that climate-migration linkages are often complex, and climatic variability may increase, decrease, or have null effects on migration. However, the use of non-comparable analytic strategies across studies makes it difficult to disentangle substantive variation in climate effects from methodological artifacts.  We address this gap by using census and survey micro-data from six Asian countries (n=54,987,838) to measure climate effects on interprovincial migration, overall and among sub-populations defined by age, sex, education, and country. We also evaluate whether climate effects differ according to the distance and type of move. We find non-linear precipitation effects across the sample, with exposure to precipitation deficits leading to substantively large reductions in out-migration. Both precipitation and temperature effects vary among focal sub-populations. Precipitation deficits reduce internal migration to both adjacent and non-adjacent provinces and also reduce the probability of work-related moves in the countries that the reason for migration is measured. Temperature anomalies reduce work-, education-, and family-related moves. Our findings provide evidence of climate-related reductions in migration and suggest these effects are driven largely by economic factors.

(Infra)Structural Racism: Examining Social and Spatial Division in Road Networks. Elizabeth Roberto

Racism, discriminatory practices, institutional bias, and systematic exclusion can take lasting physical form in the built environment. There has been growing attention to the long-term consequences of housing policies and practices on social and economic outcomes, including racial segregation, but comparatively less attention to other aspects of the built environment, such as road networks and spatial connectivity. In this paper, we use a newly developed method that identifies missing road segments that we would expect to exist given the surrounding infrastructure. We find that unexpected disconnectivity in a city’s road network is associated with racial differences in nearby areas and contributes to higher levels of segregation at the local and city level. Our findings emphasize the power of the built environment and suggest that road net- works warrant more attention as a factor that may contribute to the persistence of segregation.

Panel Session 3: Measuring and Understanding Daily Life

Did the COVID-19 Pandemic Make it Worse? Working from Home and Affective Wellbeing at the Intersections of Parental Status and Occupation. Yue Qian and Wen Fan

The COVID-19 pandemic led to an unprecedented expansion of working from home. As millions of workers moved to remote work overnight, what happened to their affective well-being? Using data from the 2003–2021 American Time Use Survey, 2021 American Community Survey, 2021 Current Population Survey, and Oxford COVID-19 Government Response Tracker, we examine how workers’ affective well-being changed from pre-pandemic to the pandemic era and how this change varied at the intersection of work location, parental status, and occupational telecommutability. We find that the pandemic exacerbated negative affect for parents who were in less telecommutable occupations but found themselves working from home during the pandemic. This pandemic impact remained even after accounting for a wide array of sociodemographic, health, family, time use, and occupation-level characteristics. Additional analysis shows that the pandemic impact was more pronounced when school closing policies were more stringent, suggesting that rising difficulty balancing work and family demands during the pandemic was an important factor precipitating higher negative affect felt by remote-working parents in less telecommutable occupations. This study reveals the heterogeneous impacts of working from home on affective well-being. The results also highlight the negative implications of weak care infrastructures and workplace support for parental well-being.

From Data to Dreams: IPUMS and the Study of Rapidly Changing Sleep Trends. Connor Sheehan

Sleep is increasingly understood to shape human behaviors, well-being, and longevity.  However, sleeping patterns are also influenced by external social factors. Accordingly, as society changes, it is important to systematically analyze alterations in population-level sleeping patterns. Using data from the 2004-2018 National Health Interview Survey, harmonized and prepared by IPUMS, I delineate to two concerning trends regarding the sleeping patterns of American adults. First, American adults are reporting worse sleeping patterns, a trend that began abruptly in 2013. Notably, this decline in sleep was especially concentrated among People of Color. Second, my findings reveal that cohorts now entering midlife, when compared to their older counterparts, are reporting both shorter sleep durations and poorer sleep quality. In the face of ongoing societal transformations, researchers leveraging the capabilities of IPUMS will be well-prepared to comprehensively chronicle secular changes in sleep patterns, and their increasingly complex explanations.

The Mismeasurement of Work Time and Its Implications for Wage Discrimination and Inequality. Daniel S. Hamermesh and George Borjas

A comparison of the measures of work time available in the CPS-ASEC data file (based on recall) with contemporaneous measures reveals many logical inconsistencies and probable errors. About 8 percent of ASEC respondents report weeks worked last year that are inconsistent with their work histories in the Basic monthly interviews, with the error rate rising to over 50 percent for those who move in and out of the workforce across the monthly interviews. Similarly, 22 percent give contradictory information about whether they usually work a full-time weekly schedule (35 or more hours per week). The errors are not random: They differ by gender and race, and group error rates change differently over time. The measurement issues suggest rethinking how wages changed from 1978 to 2018. After adjusting for the measurement errors, we find that the gender and male black-white wage gaps among all workers narrowed by about 4 log points more than is commonly reported, and that inequality decreased by 4 log points more. The biases produced by recall error also exist in measures of the gender wage gap and residual inequality among full-time year-round workers. Using a more carefully defined sample of workers with this employment status shows that gender wage differentials have been slightly smaller than previously estimated, as has residual wage inequality.

Toxified to the Bone: Early-Life and Childhood Exposure to Lead and Men’s Old-Age Mortality. Jason Fletcher and Hamid Noghanibehambari

Several strands of research document the life-cycle impacts of lead exposure during the critical period of children’s development. Yet little is known about long-run effects of lead exposure during early-life on old-age mortality outcomes. This study exploits the staggered installation of water systems across 761 cities in the US over the first decades of the 20th century combined with cross-city differences in materials used in water pipelines to identify lead and non-lead cities. An event-study analysis suggests that the impacts are more concentrated on children exposed during in-utero up to age 2. The results of difference-in-difference analysis suggests an intent-to-treat effect of 1.6 months reduction in old-age longevity for fully exposed cohorts. A heterogeneity analysis reveals effects that are 2.8 times larger among nonwhite subpopulation. We also find reductions in education and socioeconomic standing during early adulthood as candidate mechanism. Finally, we employ WWII enlistment data and observe reductions in height-for-age among lead-exposed cohorts.

Panel Session 4: Disentangling Discrimination

Ethnic Identity and Anti-Immigrant Sentiment: Evidence from Proposition 187. Francisca Antman and Brian Duncan

In recent years, many elections have centered on stoking racial and ethnic divisions, thus raising the possibility that individuals’ self-reported racial and ethnic identities may have responded to the increasingly tense political climate. The expected impacts, however, are unclear, as individuals from ethnic or racial minorities might be less likely to identify with their minority groups if they fear the political discourse may have negative consequences for them, or they may be more likely to identify if the heightened saliency of race and ethnicity has raised their consciousness surrounding this aspect of their identity. We use data surrounding one of the first and most well-known ballot measures widely seen to be anti-immigrant, Proposition 187, to address this question. Using data from the US Census and Current Population Survey, as well as county-level vote share information on the passage of Prop 187, we ask whether individuals with documented Hispanic ancestry or parentage are more likely to identify ethnically as Hispanic/Latino, or with any subnational Hispanic/Latino group, after the passage of Prop 187 and in areas with higher support for the initiative. To our knowledge, this is the first investigation into this question and the first to document a connection between political discourse, participation, and endogenous ethnic identity.

Racial residential segregation and child mortality in the southern United States at the turn of the 20th century. J'Mag Karbeah and J. David Hacker

A growing body of research considers racial residential segregation to be a form of systemic racism and a fundamental cause of persistent racial disparities in health and mortality. Historical research examining the impact of segregation on health and mortality, however, is limited to a few studies with poor data and inconsistent results. In this study, we examine the association between racial residential segregation and child mortality in the South at the turn of the 20th century. We rely on the new IPUMS 1900 and 1910 complete-count databases to estimate child mortality in the 5 years before each census and construct segregation measures at the census enumeration district (ED), the lowest level of geography consistently available in the census. We calculate the proportion of households headed by Black individuals in each ED, and the Sequence Index of Segregation (SIS), which is based on the racial sequencing of household heads within each district. We construct models of child mortality for rural and urban areas, controlling for a wide variety of demographic and socioeconomic variables. The results indicate that proportion Black and SIS were strongly and positively associated with the mortality of Black children in most models and in both rural and urban areas. Proportion Black was also positively but more moderately correlated with the mortality of White children, while SIS was not correlated or negatively correlated. These results suggest that racial segregation was a long-standing fundamental cause of race disparities in health and mortality in the United States.

Ethnic Identificational Change and Educational Mobility Among Mexican Americans. Jennifer Van Hook, James D. Bachmeier, Kendal Lowrey, and Cheyenne Lonobile

Mobility scholars have demonstrated that disparities in socioeconomic status are rooted in historical inequities in intergenerational mobility and reproduction (Song 2020). Building on earlier research (Duncan and Trejo 2011), we theorize how a third factor, ethnic change, may impact educational disparities, and we quantify the degree this factor helps explain the large and persistent educational gaps between Mexican Americans and Whites. Using linked Census data across 80 years and 3 generations, we find that Mexican ethnic attrition may be even more widespread than earlier estimates suggest, and that the share who identify as Mexican declines at higher levels of parental and own educational attainment. Additionally, we find that ethnic change accounts for between 30 and 50 percentage of the gap in postsecondary and college attainment between 3rd generation Mexican Americans and their contemporary white counterparts.  The results complicate simplistic understandings of ethnic disparities in educational attainment and their persistence in the contemporary era. Not only are they a consequence of historical inequities in mobility and reproduction, they are also related to the tendency for those with higher levels of educational attainment to no longer identify with the ethnic group.