2015-2016 P2C Pilot Projects
Studying Immigration through Network Sampling and Memory
Immigration will be the driving force in the future growth of the US population, political debates surrounding the subject are intensifying, and the immigration system needs reform — all of which underscore the need for better understanding of the foreign-born population. The study of immigration to the United States is complicated by the difficulty and costs of obtaining analytical samples for inference of small groups and hidden segments (undocumented immigrants) of the population for which no sampling frame exists. This leads to sampling frames with incomplete coverage and to partial inclusion of immigrants in official data (i.e., CPS or ACS).
Using conventional survey methods to collect data on specific immigrant groups is difficult because each group represents a small portion of the overall population. a large number of screening interviews are necessary to recruit a sufficient sample (McKenzie and Mistaien, 2009). In addition, undocumented migrants may constitute a "hidden" or stigmatized population (Kalton 2009) that is reluctant to respond to conventional surveys. The goals of this pilot are: 1) to field-test a novel sampling approach, network sampling with memory (NSM); to recruit an analytical sample of Chinese immigrants in a circumscribed area of North Carolina, the Raleigh-Durham area; and to evaluate how closely it represents the target population through comparison with ACS fata; 2) to reduce the costs of implementing NSM on a large geographic scale and in transnational networks by assessing a dual data collection mode: a conventional version of NSM implemented through personal referrals and in person interviews and a Web version that recruits respondents through online ties; and 3) to use network data collected through NSM to analyze immigrants' network-related social processes.
With respect to Aim 1, while a principal advantage of network sampling is to draw in hidden populations and since ACS may undercount these, we pick a population for testing where this is less likely to be an issue, because comparatively few Chinese immigrants are believed to be undocumented (Hoefer et al., 2011). With respect to Aim 2, we plan to develop a mobile-enabled survey Web site that would allow for the survey to be accesses on a browser or smartphone platform for which the digital divide is smaller (Rainie 2013). With respect to Aim 3, data collected through the pilot will begin to shed light on important issues related to immigrant social networks, for example the role of co-ethnic social networks in job search; the effect of social networks on the social incorporation of recent immigrants; the role fo peer effects in health behaviors and health-related knowledge transfers; and the differential impacts of immigration on gender relations.
This project supports DPRC's theme of social connectedness by using network sampling methods to assemble representative samples and experimenting to extend these cost-effective Web surveys and network analyses, with a focus on using network relationships to understand immigrant settlement and assimilation.
Chemicals in Latin America have largely been unregulated for decades, resulting in high environmental exposures early and throughout one’s life, particularly in areas where resource extractive industries exist. In the past 10 years, the rapidly expanding field of environmental epigenetics has discovered that chemical exposures during early life development can increase risk to adverse health outcomes in adulthood via deregulation of epigenetic marks that govern gene expression (Dolinoy, Huang, & Jirtle 2007; Heijmans, Tobi, Stein et.al 2008). Pollutants such as mercury and select pesticides may be absorbed more readily in malnourished individuals, or, for early-life exposures, may increase risk for obesity and metabolic dysfunction. Unfortunately, obesity and chronic diseases are growing problems in Latin America that co-exist with severe malnutrition. The high exposure risk and differences in nutritional vulnerability suggest significant health impacts from early life chemical exposures in the Amazon. In this pilot project, we propose to study epigenetic outcomes of environmental exposures in communities near small scale and artisanal gold mining in Madre de Dios, Peru. This Health and Human Development research highlights two DUPRI NIH center grant funding priorities: life course studies and developmental processes of population health and well-being.
Our long-term goal is to study the effects of large-scale development projects on human health. The proposed pilot project would provide preliminary data to enable our group to apply for a larger collaborative proposal to study epigenetic effects and metabolic health outcomes related to environmental toxicant exposures resulting from multiple large-scale development projects (LSDP), including plantation agriculture in Rondonia, Brazil, oil extraction in the Northern Ecuadorian Amazon, and gold mining in Madre de Dios, Peru. In our future proposal, we will establish birth cohorts in our three study sites to prospectively assess differential environmental LSDP exposures in early life, which will enable validation of the exposure index proposed in this pilot project .
In recent years, an explosion in the field of microbiome research has highlighted how the intimate relationship between the human body and its microbial inhabitants shapes our health. Microbes train our immune systems, help resist pathogens, and contribute substantially to energy acquisition. Already, targeted experimental manipulations of the microbiome have been shown to produce dramatic effects on host phenotypes, including obesity, autoimmune disease susceptibility, and, in some animal models, behavior itself. These findings strongly motivate discovery of the factors that predict microbiome structure across the life course—especially given emerging evidence that it may constitute a novel route through which social relationships influence health. For example, I recently led a study that showed that social networks predict gut microbiome composition in wild baboons, and more so for specific types of bacterial taxa (Tung et al 2015, eLife).
Understanding the relationship between social interactions, microbiome composition, and its potential health-related outcomes requires repeated, longitudinally collected samples over the lifetime of known individuals. Such samples will take decades to collect in humans, but are already available for the baboon population I study in Kenya, an important model for human social behavior, demography, and evolution. Specifically, as a result of banked samples originally collected for other purposes, we have over 20,000 samples available for 613 individuals—a sample size that vastly outstrips all published work on the microbiome to date (Figure 1). Detailed data on diet, space use, social interactions, social status, early life adversity, and changes with age are available for the same animals, making this system ideal for understanding the causes and consequences of natural variation in gut microbiome composition. Specifically, we are interested in whether (i) variation in the gut microbiome predicts fertility and survival; and (ii) the extent to which early life conditions and social interactions throughout life contribute to long-term variation in microbiome composition.
Our unique sample set has already attracted substantial interest, including from the Earth Microbiome Project—one of the largest consortia focused on microbiome studies assembled to date. With the EMP, we are currently in the process of profiling all 20,000 samples, in a large-scale effort that involves collaborations between multiple PIs and multiple institutions (myself, Dr. Elizabeth Archie at University of Notre Dame, Dr. Ran Blekhman at University of Minnesota, Dr. Luis Barreiro at University of Montreal, and the EMP, led by Dr. Jack Gilbert at University of Chicago and Dr. Rob Knight at UCSD). The total cost of data generation itself, excluding the original costs of sample preparation, is expected to be >$150,000, of which I will contribute approximately $25,000. Given the unique value of the resulting data to understanding how early life conditions influence this key component of human health, and to how the microbiome itself changes with age as a result of social and demographic conditions, investing pilot/seed grant funding from DUPRI into this project would be ideal, allowing me to reallocate the funds earmarked for this project (from my start-up) for a grad student or post-doc dedicated to analyzing the resulting data.
Sustainable Development and Population: Making the Most of Spatial Data for Demographic Studies
Economic development and change in exposures to environmental toxins go hand in hand. High-frequency data from satellite pictures provide unique opportunities to pinpoint location and timing of exposure events. The combination of these data with information on production techniques and the availability of administrative records on births, death, and scholastic achievement allow social scientists to study the impact of economic development for the population directly exposed to its costs and benefits.
This project proposes to examine the health effects of traditional sugarcane harvesting techniques in Brazil. Ethanol from sugarcane is increasingly seen as a viable, renewable energy source that can help meet rising global energy demand, but traditional farming practices involve pre-harvest field burning, which elevates air pollution levels around producing regions. This study will quantify the consequences for infants exposed to smoke before and after birth. In addition to contributing to the burgeoning economics literature on the negative effects of air pollution on infant health (Currie, Graff, Mullins, and Neidell, 2014), the study will respond to debates about pollution released in the production of many energy sources, not just ethanol. In the United States, such debates have recently focused on the natural gas extraction process of hydraulic fracturing; longer-standing controversy surrounds nuclear power, coal, and oil. In most cases, these production externalities pose distributional questions, with nonproducing areas benefiting from cheaper fuels or reduced emissions even as producing areas suffer. Distributional questions are especially pertinent in Brazil, where automobile fuel must be at least one-fourth ethanol, but sugarcane fields cover less than one percent of the country's land mass.
This project aims to combine data from satellite imagery with administrative and survey data to measure economic development and its environmental challenges to population and demographic processes, including fertility, mortality, birth outcomes, human capital accumulation, and migration flows. Data are employed from the Fire Information for Resource Management System (FIRMS) of NASA's Earth Observing System Data and Information System to track the use of fires in agricultural activities in Brazil. These are combined with weather data from geographically disperse weather stations and administrative data on birth outcomes, death records, hospitalizations, and school performance to examine the impact of modernization of agricultural production (as well its expansion) on various dimensions of welfare.
The innovations of this project center on the extensive usage of satellite imagery information combined with geocoded administrative records, which allow us to infer causal relationships and to detect heterogeneity in effects. Satellite imagery provides an important tool for population research and social policy and offers a wealth of scientific possibilities when the granularity of such data (both in terms of time and location) is fully explored but its use remains modest among population scientists. The use of "big data" that capture variations in exposure and outcomes over time and space can provide opportunities for detailed analysis of policies.
The results will also speak to policy debates around the benefits of changing technology in order to reduce small but repeated health insults that stem from a production process many see as crucial for enabling increased global reliance on sustainable energy sources. Both in Brazil and elsewhere, the sugarcane industry is gradually adopting mechanized harvesting methods that do not require fires. We will quantify the benefits of limiting emissions of particulate matter, the primary byproduct of field fires.
This project supports two DPRC research themes: 1) biodemography, with its focus on the developmental consequences of exposure to various environmental hazards; and 2) estimation of causal impacts, with its assembling of unique data that combines satellite-imaging data on the environment with Brazilian administrative records that will measure health outcomes and indicators of populations, especially children, by geography.