As part of the project Death and Burial Data: Ireland 1864–1922 (DBDIrl), a web application was created in DIME, a low-code web application development environment. DIME is based on the popular IDE Eclipse and utilizes three distinct graphical model types (data model, process model and GUI model) as Domain Specific Language (DSL). Web applications are defined by means of these models in the DIME integrated modelling environment in a simple way, and these models are used to generate the web application code that is then deployed and made available to the users. This paper provides an overview of how we use a web application created in DIME, a low-code application development tool to establish interoperability between Irish historical civil registration of death data and Irish cross-sectional decennial census data from 1901 and 1911. The former data collections are complex and require manual interventions like transcription and considerable cleaning to render them machine readable, the latter are fully transcribed. Here we present a case study concerning the analysis of deaths in old age in Dublin City and how they correlate with census returns. We then turn our attention to a use case of marital status and the algorithmic problems caused by irregular replies in census data. Our objective was to test how DIME could replicate the decisions taken by domain experts using old age and marital status as the primary linkage criteria.

I study the long-run effects of displacement and neighborhood division by looking at individuals affected by the construction of the Interstate Highway System. To do so, I track individuals over time by linking the 1940 census to administrative mortality records from 1995 to 2005. I find that displaced individuals die three months younger, are more likely to leave their neighborhoods, and reside in areas with lower socioe- conomic characteristics at the time of death. I also find highly localized spillovers: individuals living within 100 meters of a highway are more likely to leave their neigh- borhoods and relocate to lower socioeconomic areas, yet they do not experience in- creased mortality. The neighborhoods where individuals relocate after displacement explain 30% of the displacement-mortality effect. Accounting for the mortality ef- fects of displacement would have increased the cost of building the highway system by 10%. Together, these results enhance our understanding of the costs displacement imposes on individuals and their communities and provide new insights for the de- sign of future infrastructure projects.

The study focuses on the temporal differences (30 years on average) between ethnic groups migrating from the Austro-Hungarian Empire to the U.S. between 1850 and 1910. In our study, we argue that the main factor that led to differences in the timing of emigration was the differences in regional economic development of different ethnic groups living in different regions of the Empire. Migration costs: before the 1864 in- troduction of steam engine technology in transatlantic maritime transport, emigration costs were not affordable for Hungarians and Slovaks due to the sea and land voyage high ticket prices. Therefore, with more resources, Austrians and Czechs could afford to migrate earlier. However, after the introduction of steamship technology and the technological change in ship engines, travel became more affordable due to reduced ticket prices, faster voyages, and increased capacity. This allowed Hungarians and Slo- vaks from poorer regions to begin migrating in larger numbers as migration became economically feasible. In this study, we utilise a complete count of the U.S. Census records from 1900 and 1910 (Helgertz et al., 2023; Ruggles et al., 2021), which Inte- grated Public Use Microdata Series (IPUMS). On the other hand, we utilise economic indicators, which are regional daily wage, GDP per capita income and living standard data for the Austria-Hungary Empire from Cvrcek (2013) and Schulze (2000)

We study the impact of Federal alcohol Prohibition in 1919 on workers in the alcohol industry and their families using newly linked census records that allow us to follow spouses, sons and daughters. Immediately after Prohibition, men previously working in alcohol-related industries were less likely to be in the labor force, and when working, employed in lower skilled occupations. By 1940, 21 years after Prohibition, workers were still more likely to be in unskilled occupations, but they were more likely to be employed, consistent with delayed retirement. In the short run, sons are largely unaffected but in the long run, they complete slightly more schooling and earn more. Interestingly, daughters were more likely to

Understanding life course trajectories requires data with a long time horizon or sufficiently detailed retrospective recall measures; extending this to include intergenerational processes or mortality outcomes places further demands on data. Historical full count census data from IPUMS are a rich and underutilized source of information on early life conditions and can be harnessed to summarize millions life histories for individuals and trace families over multiple generations. IPUMS full count census data covers 1850-1950 and boast nearly 800 million individual records, which can be linked longitudinally across decennial censuses as well as to external data sources to answer questions about exposures and outcomes across the life course. This session will feature research presentations on the power of linked full count US census data to explore life course trajectories including substantive applications focused on racial equity in older-age mortality, negative economic shocks and long-term wellbeing, segregation and exposure to air pollution, and developing scores for estimating lifespan using commonly available socioeconomic and demographic measures. This session will highlight the value of historical census data to understanding current population and aging issues. The presentations will showcase research that links full count census data in multiple ways and demonstrates the flexibility and wealth of information available in these data.

Over the past five years, the IPUMS Multigenerational Longitudinal Panel (MLP) has constructed a massive longitudinal population panel by linking together American censuses, surveys, administrative sources, and vital records spanning the period from 1850 to the present. MLP is designed as a general-purpose resource for studying longitudinal processes and long-run social and economic change. By linking individuals across generations from birth through death using data from multiple sources, MLP enables reproducible research on shifting patterns of life course patterns and intergenerational processes. The large scale of the MLP database allows researchers both to study particular communities and small dispersed populations and to conduct big studies spanning many places and periods. Data that allow investigators to examine simultaneously the broad sweep of time and fine spatial detail will yield new insights into ongoing transformations of demographic behavior. The presentation will describe the creation of the IPUMS MLP database and the current MLP linking strategy. We then detail our plans for expansion and improvement of MLP over the next five years, including the incorporation of additional data sources, the development of a “linkage hub” to connect MLP with other major record linkage efforts, and the refinement of our technology and dissemination efforts. We will conclude by describing a few early examples of MLP-based research.

Secondary data from longstanding population-representative sources hold powerful potential for public health research. Their rich socioeconomic and demographic measures, relative continuity over time, and large sample sizes mean they can be used to study change, conduct comparative research, situate individuals within community and family contexts, or serve as an external benchmark for smaller studies. IPUMS (ipums.org) harmonizes valuable secondary data sources, applying a consistent coding scheme across time along with integrated documentation. Learn about research ready data from IPUMS that are available without charge to the research community.

We propose a new measure of cultural distance based on differences in the composition of first names and church denominations between locations. We use a gravity equation to estimate the elasticity of migration flows with respect to the two components of cultural distance as well as a standard measure of travel costs via the transportation network in the United States between 1850 and 1870. Our findings indicate a modest role for cultural distance relative to travel costs in explaining migration flows. We construct migration costs that reflect the distinct contributions of cultural distance and travel costs, and use an economic geography model of migration to quantify their effects. Travel costs are substantially more important than cultural distance for aggregate welfare. Nevertheless, we provide evidence that the components of cultural distance play a role in shaping of how many people move and their final destinations.

This study examines the long-run effects of radiation exposure from U.S. at- mospheric nuclear weapon tests on mortality. Using individual-level vital records and a differences-in-differences approach, I analyze county-level radiation fallout data from tests conducted at the Nevada Test Site between 1951 and 1962 to assess the impact on mortality in the continental U.S. I find significant negative effects of radiation exposure on longevity with heterogeneous impacts by age and gender. Those exposed before age 15 experience an average 16-month reduction in lifespan, while exposure after age 60 results in a 30-month reduction. The effects are larger for aged females compared to aged males. Prenatal radiation exposure, especially pre-conception, also has significant negative effects on mortality - a 1% increase in ground-level radiation fallout in the 10-12 months before birth reduces the likeli- hood of surviving to ages 1, 3, and 50 by 0.3%, 0.4%, and 0.5%, respectively. In contrast, radiation exposure during other prenatal months does not have signifi- cant effects on future mortality. These findings have important implications for understanding the long-term public health consequences of nuclear events and can inform policies related to nuclear energy adoption and the importance of avoiding nuclear war.

Academic researchers in the U.S. and Canada have partnered with genealogical organizations and statistical agencies to create massive new scientific data collections from census enumerations. These data are creating new opportunities for research across the social and health sciences. By linking individuals and families across censuses, analysts can create national longitudinal panels that trace the characteristics of individuals over their lives and families over multiple generations. IPUMS disseminates full count census enumerations for ten U.S. census years from 1850 to 1950. These full count data cover almost 800 million individual records and the IPUMS Multigenerational Longitudinal Panel (MLP) project links individuals' records across censuses. IPUMS data can be combined with data from the U.S. Census Bureau’s data linkage infrastructure to link historical records to numerous recent censuses, surveys and administrative data that measure social, economic and health outcomes.  The Canadian Peoples (TCP) is a comprehensive public research database of 40 million coded and georeferenced records found in Canadian censuses from the middle of the nineteenth century until after the First World War. The data include personal, family and household characteristics of every individual enumerated in each census from 1852 to 1921. TCP investigators are exploring record linkage technology to solve challenges presented with these historical records. The workshop included one presentation from each organization, with these objectives: introduce the data collections, explain the challenges and opportunities of census linkage, describe the linking strategies, and provide an overview on how to access the data.

Multiple births strain mothers’ and families’ resources in ways that should highlight preferences for family size, birth spacing, and parity-dependent stopping behavior. Couples with surviving twins reach their target family size sooner than other couples and should be more likely to practice family limitation. Twins are also a greater burden on the mother's time and health, which could lead to postponing the next birth, even among couples who want additional children. We examine these hypotheses by analyzing families with twins in the 1900 and 1910 U.S. Censuses. Using reconstructed birth histories for more than 7 million women in the IPUMS full-count 1900 and 1910 datasets and event-history methods (Kaplan–Meier curves, cure models), we find clear evidence of family limitation following a multiple birth. Couples who had twins or triplets were more likely to stop childbearing, and those who continued having children delayed their next birth. Responses to multiple births were larger in groups previously identified as leaders in the transition to smaller families, and roughly one third of couples stopped after one or two children. We find no evidence that some groups relied primarily on birth spacing to reduce family size while others relied primarily on stopping.

In contrast to earlier United States policies of open war, forcible removal, and relocation to address the “Indian Problem,” the Dawes Act of 1887 focused on assimilation and land severalty — making American Indians citizens of the United States with individually-titled plots of land rather than members of collective tribes with communal land. Considerable scholarship shows that the consequences of the policy differed substantially from its stated goals, and by the time of its repeal in 1934, American Indians had lost two-thirds of all native land held in 1887 (86 million acres)— and nearly two-thirds of American Indians had become landless or unable to meet subsistence needs. Complementing rich qualitative history, this paper provides new quantitative evidence on the impact of the Dawes Act on mortality among American Indian children and adults. Using 1900 and 1910 U.S. population census data to study both household and tribe-level variation in allotment timing, we find that assimilation and allotment policy increased various measures of American Indian child and adult mortality from nearly 20% to as much as one third (implying a decline in life expectancy at birth of about 20%) —confirming contemporary critics’ adamant concerns about the Dawes Act.

Quantitative analysis of Black business districts and evidence on the magnitude of social change leading up to the passage of the Civil Rights Act of 1964, in particular as it relates to the accessibility of public accommodations, is limited. We combine newly digitized data on the precise geocoded location of nearly 6000 Green Book establishments – public accommodations that were friendly towards African American clientele – across major urban areas with existing and new sources of data on social change to understand the dynamics of Black-friendly businesses within cities during the middle of the twentieth century. In doing so, we document a new set of facts. First, we show that the location and growth of Green Book establishments responded to economic forces. Second, we show that there was a large increase in the number of Green Book establishments in cities between 1939 and 1955. Third, for Green Book establishments located in cities for which the Home Owner’s Loan Corporation (HOLC) drew residential security maps, the vast majority (nearly 70 percent) were located in the lowest-grade, redlined neighborhoods. And finally, we show that 1950s urban renewal projects were related to the contraction of non-discriminatory businesses. Collectively, these facts suggest that more research on Black-owned and Black-friendly businesses is needed to fully understand the economics of urban change in the twentieth century.

We present results from a nationally representative survey of American adults, guided by a simple theoretical model expressing health care-seeking behavior as a function of economic and behavioral fundamentals and highlighting the role of trust. We report several findings. First, we document a strong association between higher levels of trust in the health care system and reported care-seeking behavior, both retrospective and anticipated. This relationship holds across several care scenarios, from routine checkups to vaccinations. Second, the impact of trust on health care utilization is similar in magnitude to that of factors such as income and education, long recognized as crucial in the existing literature. Third, the relationship between trust and care-seeking behavior appears to be mediated by key mechanisms from our theoretical framework, notably individuals' beliefs about the system's effectiveness in managing their health and their personal disutility tied to medical visits. Fourth, we ask respondents about trust in specific health care system sectors, and we find important heterogeneity in the associations between trust and care-seeking behavior, notably between trust in the Centers for Disease Control and Prevention and the likelihood to receive flu and COVID-19 vaccinations. Finally, we find no differential relationship between trust and care-seeking for Black respondents, but we find important differences by age and political affiliation. Our findings hold significant implications for policy, particularly given that trust in medical and, more broadly, scientific expertise is increasingly difficult to establish.

Lasting changes in women’s employment followed the 1918 influenza pandemic in the United States. In the decades before the pandemic, consistently fewer women reported an occupation in cities that would go on to have longer interventions targeted at curbing influenza. This gap narrowed after the pandemic, and by 1930 cities with longer interventions experienced a 3.9 percentage point improvement in women’s employment rates on average, relative to cities with shorter interventions. These gains were concentrated in cities in which women had the right to vote prior to 1920.

In this dissertation I analyze patterns of maternal-infant health in developing contexts. My first chapter uses child hemoglobin as a bio-marker for maternal malnutrition in Senegal during the 2008 food price crisis. In early 2008, world rice prices skyrocketed, causing people around the world to plunge into poverty. Senegal, in particular, depends heavily on imported foodstuffs. I find that the crisis had a large, negative impact on child anemia in urban Senegal, most likely reflecting a deterioration of maternal nutrition caused by rising food prices. In the second and third chapters, we introduce a novel indicator of maternal-infant health: childhood sex ratios. Because infant females have lower rates of mortality than infant males, the sex ratio of the surviving population reflects the level of infant mortality. Childhood sex ratios are widely available from census data, meaning we can use them to shed new light on historical populations who lack traditional sources of data on infant mortality, like birth and death records. We apply this new method to the 19th-century US, where the lack of vital statistics has left uncertainty over even approximate levels of infant mortality. We find that the level of infant mortality in the pre-industrial US was much lower than previously thought, but that racial health disparities were much greater.

A burgeoning social scientific literature on the place-based legacy of slavery has mostly overlooked the effect of slavery on incarceration, despite the fact that the intensity and racial disparity of US incarceration is often attributed to its history of slavery. I analyze data on incarceration from 1840 to 2020 and show that the historic prevalence of slavery tends to be negatively associated with Black incarceration, especially under Reconstruction and Jim Crow (1870-1940). In line with recent work by Christopher Muller, I argue this is at least partly explained by white planters paying the fines of Black convicts, who would then have to work off the debt or suffer imprisonment. I conclude that the existing literature is not wrong to assume that Southern incarceration was shaped by slavery. But it shaped it in surprising ways that previous work has often failed to identify.

During the Second Industrial Revolution, in the late nineteenth century, the proliferation of automation technologies coincided with substantial job creation but also a “hollowing out” of middle-skilled job opportunities, which historically offered reliable paths to prosperity. We use recently linked U.S. census data to document three main facts: (i) declining demand for middle-skilled labor in manufacturing corresponded to greater reallocation of workers into comparatively less-skilled occupations; (ii) older workers were more likely to switch to unskilled physical labor; (iii) younger workers led switching into growing occupations affected by automation technologies.

The idea that labor scarcity can induce economic development has been long hypothesized (Hicks, 1932; Habakkuk, 1962), but the evidence is scarce, especially on non-agricultural development. In this paper, I assess the role of the Second Great Migration (1940-1970) on the subsequent structural change in the American South between 1970 and 2010. Empirical results using shift-share instruments show that out-migration incentivized physical capital investment and capital-augmenting technical change, increasing capital per worker and output in both agriculture and manufacturing at least until 2010. Labor reallocated from agriculture to non-agriculture. I then develop and calibrate a dynamic spatial equilibrium model that allows substitution between factors of production, factor-biased technical change, and the Heckscher-Ohlin force in trade. The contribution analyses using the model suggest that labor-capital substitution played a key role in economic adjustments to the South-to-North migration.

This article describes four new IPUMS datasets constructed from the 1850, 1860, 1870, and 1880 Censuses of Mortality of the United States. We discuss the creation of the datasets, the variables included in each census year, and their potential for social science research. We highlight several limitations in the data and caution users about potential biases. Finally, we illustrate the usefulness of the new data by analyzing the relationship between household wealth and child mortality in 1870. All four datasets and associated documentation are distributed for public use via the IPUMS website.