About

Wendy K Tam Cho is a professor in the Departments of Political Science, Statistics, Mathematics, Computer Science, Asian American Studies, and the College of Law at the University of Illinois at Urbana-Champaign. She is also a Senior Research Scientist at the National Center for Supercomputing Applications, a faculty member in the Illinois Informatics Institute, and an affiliate of several research centers including the Cline Center for Advanced Social Research, the CyberGIS Center for Advanced Digital and Spatial Studies, the Computational Science and Engineering Program, and the Program on Law, Behavior, and Social Science. Her professional affiliations include being a Fellow of the John Simon Guggenheim Memorial Foundation, the Society for Political Methodology, and the Center for Advanced Study in the Behavior Sciences at Stanford University, as well as a Visiting Fellow at the Hoover Institution at Stanford University.

Research topics

• Computer Science
• Machine Learning
• Sociology
• Econometrics
• Demography
• Data science
• Medicine
• Gerontology
• Mathematics

Selected publications

• Sociodemographic Disparities in Preoperative Visual Acuity and Cataract Surgery Utilization in the San Francisco Bay Area

  Journal of Racial and Ethnic Health Disparities · 2024-02-08 · 4 citations

  article1st authorCorresponding
  PublisherDOI

• A GPU-Accelerated Population Generation, Sorting, and Mutation Kernel for an Optimization-Based Causal Inference Model

  2023-08-07

  articleOpen access1st authorCorresponding

  We develop a GPU-accelerated machine learning generative adversarial network model that can be used with observational data for the purpose of constructing causal inferences. The theoretical basis of our machine learning model is novel and is conceptualized to be operable and scalable for high performance computing platforms. Our GPU-accelerated code enables large-scale parallelization of the computation within a common and accessible computing environment. This will expand the reach of our model and empower research in new substantive domains while maintaining the underlying theoretical properties.

  PublisherOA PDFDOI

• Racial/Ethnic, Biomedical, and Sociodemographic Risk Factors for COVID-19 Positivity and Hospitalization in the San Francisco Bay Area

  medRxiv · 2022-04-06

  preprintOpen access1st authorCorresponding

  Abstract BACKGROUND The COVID-19 pandemic has uncovered clinically meaningful racial/ethnic disparities in COVID-19-related health outcomes. Current understanding of the basis for such an observation remains incomplete, with both biomedical and social/contextual variables proposed as potential factors. PURPOSE Using a logistic regression model, we examined the relative contributions of race/ethnicity, biomedical, and socioeconomic factors to COVID-19 test positivity and hospitalization rates in a large academic health care system in the San Francisco Bay Area prior to the advent of vaccination and other pharmaceutical interventions for COVID-19. RESULTS Whereas socioeconomic factors, particularly those contributing to increased social vulnerability, were associated with test positivity for COVID-19, biomedical factors and disease co-morbidities were the major factors associated with increased risk of COVID-19 hospitalization. Hispanic individuals had a higher rate of COVID-19 positivity, while Asian persons had higher rates of COVID-19 hospitalization. Diabetes was an important risk factor for COVID-19 hospitalization, particularly among Asian patients, for whom diabetes tended to be more frequently undiagnosed and higher in severity. CONCLUSIONS We observed that biomedical, racial/ethnic, and socioeconomic factors all contributed in varying but distinct ways to COVID-19 test positivity and hospitalization rates in a large, multiracial, socioeconomically diverse metropolitan area of the United States. The impact of a number of these factors differed according to race/ethnicity. Improving over-all COVID-19 health outcomes and addressing racial and ethnic disparities in COVID-19 out-comes will likely require a comprehensive approach that incorporates strategies that target both individual-specific and group contextual factors.

  PublisherOA PDFDOI

• Differential Effects of Race/Ethnicity and Social Vulnerability on COVID-19 Positivity, Hospitalization, and Death in the San Francisco Bay Area

  Journal of Racial and Ethnic Health Disparities · 2022 · 9 citations

  1st authorCorresponding
  • Sociology
  • Demography
  • Gerontology
  PublisherOA PDFDOI

• Testing Causal Theories with Learned Proxies

  Annual Review of Political Science · 2022 · 35 citations

  Senior authorCorresponding
  • Computer Science
  • Machine Learning
  • Econometrics

  Social scientists commonly use computational models to estimate proxies of unobserved concepts, then incorporate these proxies into subsequent tests of their theories. The consequences of this practice, which occurs in over two-thirds of recent computational work in political science, are underappreciated. Imperfect proxies can reflect noise and contamination from other concepts, producing biased point estimates and standard errors. We demonstrate how analysts can use causal diagrams to articulate theoretical concepts and their relationships to estimated proxies, then apply straightforward rules to assess which conclusions are rigorously supportable. We formalize and extend common heuristics for “signing the bias”—a technique for reasoning about unobserved confounding—to scenarios with imperfect proxies. Using these tools, we demonstrate how, in often-encountered research settings, proxy-based analyses allow for valid tests for the existence and direction of theorized effects. We conclude with best-practice recommendations for the rapidly growing literature using learned proxies to test causal theories.

  PublisherDOI

• Racial/Ethnic, Biomedical, and Sociodemographic Risk Factors for COVID-19 Positivity and Hospitalization in the San Francisco Bay Area

  Journal of Racial and Ethnic Health Disparities · 2022-07-19 · 2 citations

  article1st authorCorresponding
  PublisherDOI

• Differential Effects of Race/Ethnicity and Social Vulnerability on COVID-19 Positivity, Hospitalization, and Death in the San Francisco Bay Area

  medRxiv · 2022-01-05 · 5 citations

  preprintOpen access1st authorCorresponding

  Abstract BACKGROUND Higher COVID-19 incidence and morbidity have been documented for US Black and Hispanic populations but not as clearly for other racial and ethnic groups. Efforts to elucidate the mechanisms underlying racial health disparities can be confounded by the relationship between race/ethnicity and socioeconomic status. OBJECTIVE Examine race/ethnicity and social vulnerability effects on COVID-19 out-comes in the San Francisco Bay Area, an ethnically and socioeconomically diverse region, using geocoded patient records from 2020 in the University of California, San Francisco Health system. KEY RESULTS Higher social vulnerability, but not race/ethnicity, was associated with less frequent testing yet a higher likelihood of testing positive. Asian hospitalization rates (11.5%) were double that of White patients (5.4%) and exceeded the rates for Black (9.3%) and Hispanic patients (6.9%). A modest relationship between higher hospitalization rates and increasing social vulnerability was evident only for White patients. Hispanic patients had the highest years of expected life lost due to COVID-19. CONCLUSIONS COVID-19 outcomes were not consistently explained by greater social vulnerability. Asian individuals showed disproportionately high rates of hospitalization regardless of social vulnerability status. Study of the San Francisco Bay Area population not only provides valuable insights into the differential contributions of race/ethnicity and social determinants of health to COVID-19 outcomes but also emphasizes that all racial groups have experienced the toll of the pandemic, albeit in different ways and to varying degrees.

  PublisherOA PDFDOI

• AI and Redistricting: Useful Tool for the Courts or Another Source of Obfuscation?

  The Forum · 2022-11-08 · 1 citations

  articleOpen access1st authorCorresponding

  Abstract Redistricting is a politically fraught exercise. Recently, new technology has emerged that has the potential to improve the redistricting process by providing information to aid in judicial oversight. These scientific advances create the potential to improve societal governance but are also potentially manipulable by partisan interests. To avoid these negative externalities, we must thoughtfully design the processes, implement safeguards, and have clear policies that regulate and steer the emerging AI toward democratically favorable goals. We propose institutional changes toward these aims.

  PublisherOA PDFDOI

• Recovering Vote Choice from Partial Incomplete Data

  Journal of Data Science · 2021-07-11 · 5 citations

  articleOpen access1st authorCorresponding

  In voting rights cases, judges often infer unobservable individ ual vote choices from election data aggregated at the precinct level. That is, one must solve an ill-posed inverse problem to obtain the critical information used in these cases. The ill-posed nature of the problem means that tradi tional frequentist and Bayesian approaches cannot be employed without first imposing a range of assumptions. In order to mitigate the problems result ing from incorporating potentially inaccurate information in these cases, we propose the use of information theoretic methods as a basis for recovering an estimate of the unobservable individual vote choices. We illustrate the empirical non-parametric likelihood methods with some election data.

  PublisherOA PDFDOI

• Recovering Vote Choice from Partial Incomplete Data

  Journal of Data Science · 2021-07-11 · 2 citations

  articleOpen access1st authorCorresponding

  In voting rights cases, judges often infer unobservable individual vote choices from election data aggregated at the precinct level.That is, one must solve an ill-posed inverse problem to obtain the critical information used in these cases.The ill-posed nature of the problem means that traditional frequentist and Bayesian approaches cannot be employed without first imposing a range of assumptions.In order to mitigate the problems resulting from incorporating potentially inaccurate information in these cases, we propose the use of information theoretic methods as a basis for recovering an estimate of the unobservable individual vote choices.We illustrate the empirical non-parametric likelihood methods with some election data.

  PublisherOA PDFDOI

Frequent coauthors

• Yan Liu

  22 shared

• George G. Judge

  University of California, Berkeley

  14 shared

• James G. Gimpel

  12 shared

• Bruce E. Cain

  Stanford University

  11 shared

• Brian J. Gaines

  10 shared

• David G. Hwang

  9 shared

• Joanne Lee

  NIHR Leicester Biomedical Research Centre

  7 shared

• Shaowen Wang

  University of Illinois Urbana-Champaign

  5 shared

Education

• Ph.D

  U.C. Berkeley

  1997

Awards & honors

• Fellow of the John Simon Guggenheim Memorial Foundation
• Fellow of the Society for Political Methodology
• Fellow of the Center for Advanced Study in the Behavior Scie…
• Visiting Fellow at the Hoover Institution at Stanford Univer…