Seminar Series - Autumn 2013
Listing of past seminars
Autumn 2013 Seminar Schedule
October 4, 2013
While there is a small literature espousing a more nuanced view, conventional wisdom---reflected in literature reviews as well as newspaper and magazine stories---holds that poor children are at risk for a lifetime of obesity. This wisdom is reflected in programs and interventions targeting the nutrition and exercise habits of poor children. In fact, there appears to be far more literature discussing mechanisms for these relationships, from relative food prices to access to parks to epigenetic programming, than literature supporting a causal effect of childhood economic resources on later body mass. This project set out to investigate what it is about poverty in childhood that is related to later excess body fat. I investigated different poverty definitions, demonstrating a methodology for empirically deriving optimal poverty thresholds. I analyzed the total effect of childhood poverty via models for moving beyond crude measures whereby a single indicator of economic circumstances at some unspecified point in childhood stands in for all childhood experiences. The aim was to isolate what aspect of childhood poverty at what point in childhood accounts for the apparent effect of relative economic deprivation on body mass. Fixed effects models would show that the total effect of child poverty was robust to unobserved heterogeneity. Instead, where I find effects consistent with the conventional wisdom, they are quite small. Fixed effects models suggest only protective effects---lower adult BMI---of more of childhood spent in poverty, the opposite of the conventional wisdom.
October 11, 2013
Research on immigration, educational achievement, and ethnoraciality has followed the lead of racialization and assimilation theories by focusing empirical attention on the immigrant- origin population (immigrants and their children), while overlooking the effect of an immigrant presence on the third-plus generation (U.S.-born individuals of U.S.-born parents), especially its white members. We depart from this approach by placing third-plus-generation individuals at center stage to examine how they adjust to norms defined by the immigrant- origin population. We draw on fieldwork in Cupertino, California, a high-skilled immigrant gateway, where an Asian immigrant-origin population has established and enforces an amplified version of high-achievement norms. The resulting ethnoracial encoding of academic achievement constructs whiteness as having lesser-than status. Whiteness represents low-achievement, laziness, and academic mediocrity; Asianness, in contrast, stands for high- achievement, hard work, and success. We argue that immigrants can serve as a foil against which the meaning and status of an ethnoracial category is recast, upending how the category is deployed in daily life. Our findings call into question the largely taken-for-granted analytic position that treats the third-plus generation, especially whites, as the benchmark population that sets achievement norms and to which all other populations must adjust.
October 18, 2013
Brian Dillon is an economist whose work focuses on risk and uncertainty in agriculture, health, and education markets in east Africa. Other research interests include the link between food markets and energy markets, traditional land tenure regimes in east Africa, and partial identification problems in econometrics. He works primarily in Tanzania and Kenya, but has participated in numerous multi-country projects involving other countries in Africa. Brian received his Ph.D. in economics from Cornell University in 2011. He earned a BS in mathematics and a BA in philosophy from Loyola University Chicago, and an MPhil in economics as a Gates Scholar at the University of Cambridge. Prior to joining the Evans School faculty he spent two years as a post-doctoral researcher at the Harvard Kennedy School and the Cornell Dyson School of Applied Economics and Management.
October 25, 2013
Seminar Cancelled this Week
November 1, 2013
Quantifying Crime Displacement After A Hot-Spot Intervention
Research shows that urban crime concentrates in specific areas, leading scholars and practitioners to combine criminological theory and spatial methods to help inform policies and strategies on how to address this phenomenon. Outside of measuring the success of an intervention in a focal area, there is concern that the operation will simply move crime to nearby areas. Alternatively, some scholars argue that the positive benefits of the operation spread into nearby areas, reducing crime and improving the neighborhood. This study attempts to address previous methodological concerns by measuring the geographic extent to which crime might displace as well as the lasting impact of an operation. Results suggest that crime may move further from a focal area than previously expected, the positive effects of an operation may not last, and there may be unintended effects leading to neighborhood change.
November 8, 2013
Research suggests that Latino immigrant men face many difficulties in adapting to life in the United States, and previous studies suggest that unhealthy alcohol use is one coping mechanism in response to these stressors. The aim of this study was to understand the use and context for use of alcohol among Latino immigrant men as a basis for development of an intervention to reduce heavy alcohol use in this population. Our study used a mixed methods approach. We conducted in-depth interviews with 27 Latino immigrant men who were current drinkers, along with 11 social service providers serving this population. Results from the qualitative study helped inform a quantitative survey which was administered to 104 Latino immigrant men at a day labor worker center. Results from both studies indicate that unhealthy alcohol use is common in this population and is tied to migration-related stressors. Men were more likely to drink when feeling lonely, social isolated, and guilty about being away from their families. Drinking also helped men feel more strong, confident, friendly and conversational. Negative consequences of drinking included strained family relationships, poor work performance, and job loss. Community-based interventions that address migration challenges and unhealthy alcohol use may promote the health of this vulnerable population.
November 15, 2013
Climate change has been and will likely be accompanied by an additional increase in the frequency and severity of rapid- and slower-onset events such as floods and droughts. When in situ adaptive capacity is strained, migration can serve as an adaptation strategy to climate-induced vulnerability in several ways described in the literature on adaptation and vulnerability as well as by some migration theories. Scholars generally agree that, with few exceptions, displacement of individuals from a given country led by increasing climatic variability will largely take place within national borders. However, as prior work shows that the choice of an internal or international destinations is to some extent contingent on pre-existing connections to either of these locales, it is likely that some "climate-induced" migration from rural Mexico may be directed towards the United States. I present selected results from three different studies carried out in collaboration with scholars at the University of Colorado showing that the association between relative rainfall deficits possibly signaling drought and U.S.-bound migration are indeed highly contingent on some contextual characteristics. I discuss some of the implications of these results for migration theory and for social and immigration policy on both sides of the border.
November 22, 2013
Laina Mercer & Tyler McCormick
"Estimation of Child Mortality in the Absence of Vital Registration Systems"
- Laina Mercer, Graduate Student in Statistics, UW
Complex surveys are commonly used to collect health and population data in the absence of vital registration systems in low- to middle-income countries. Generally these surveys are designed to provide estimates on a national or regional level. However, reliable small area-level estimates would provide the information needed to inform resource allocation and target interventions. When estimation is carried out, one must acknowledge the sampling scheme in order to address bias. In this talk I will discuss methods to accurately estimate the uncertainty in estimates and to reduce bias and variance through hierarchical modeling. These methods will be used to estimate child mortality in the districts of Tanzania from 1990-2010.
"InSilicoVA: A Statistical Model for Automated Coding of Verbal Autopsies"
- Tyler McCormick, Assistant Professor in Statistics and Sociology, UW
Verbal autopsies (VA) are widely used to provide cause-specific mortality estimates in developing world settings where vital registration does not function well. VAs assign cause(s) to a death by using information describing the events leading up to the death, provided by care givers. Typically physicians read VA interviews and assign causes using their expert knowledge. Physician coding is often slow, and individual physicians bring bias to the coding process that results in non-comparable cause assignments. These problems significantly limit the utility of physician-coded VAs. A solution to both is to use an algorithmic approach that formalizes the cause-assignment process. This ensures that assigned causes are comparable and requires many fewer person-hours so that cause assignment can be conducted quickly without disrupting the normal work of physicians. We develop a statistical model and computational implementation for automated cause of death assignment from VA data – provisionally named InSilicoVA. We model the individual-level cause assignments and the population-level CSMFs to be consistent with one another, and results from our model are presented as distributions that communicate uncertainty. InSilicoVA uses InterVA’s widely-used input data format, which corresponds to the WHO 2012 VA standard, produces estimates for the same WHO-defined cause categories and utilizes the same expert knowledge; this ensures that current users of InterVA can begin using InSilicoVA with no important changes to their workflows. Like InterVA, InSilicoVA can be used with general knowledge of major disease prevalence (e.g. HIV, malaria) and a standard distribution of deaths by cause as a starting point to identify likely causes of death for an individual. We validate InSilicoVA’s statistical model through simulation and use the ALPHA Network’s data that includes HIV status to fine-tune InSilicoVA’s ability to assign HIV-related causes, and future work will seek to validate and possibly re-calibrate InSilicoVA using additional VA reference data when those become available. We have built and will provide open source implementations of both InSilicoVA and InterVA in the statistical software R and Stata. InSilicoVA produces consistent estimates of population-level CSMFs and individual cause of death assignments with uncertainty bounds for both, and current users of InterVA will be able to use InSilicoVA easily.
November 29, 2013
No Seminar - Thanksgiving
December 6, 2013
No Seminar This Week