25th Anniversary Speakers and Abstracts

Session: Alumni Panel
CSSS Track: Political Science, 2008

John Ahlquist is a professor of political economy in UC San Diego's School of Global Policy and Strategy. He is also a founding faculty member of UCSD's Halıcıoğlu Data Science Institute.  In addition to substantive courses on labor and social welfare policy, Ahlquist teaches a Masters'-level Introduction to Data Science for Public Policy as well as a PhD-level courses on the generalized linear model and survey methods. His past research has appeared in numerous scholarly outlets, including the American Economic Journal (policy), American Political Science Review, American Journal of Political Science, Daedalus, and Political Analysis. He is the co-author of two books, In the Interest of Others (2013) and Maximum Likelihood Strategies for Social Science (2018). Prior to joining UC San Diego, Ahlquist held faculty positions at the University of Wisconsin—Madison and Florida State University. Ahlquist completed the political science CSSS track during his PhD at the University of Washington, graduating in 2008. John Ahlquist's website.

Session: Morning Scientific Session - Advances in Social Network Analysis
Title: Social Networks and Health: Measurement, Contagion, and Interventions
Abstract: In this talk, I will provide a high-level overview of my research on social networks and health. First, I show how to combine peer-reports and self-reports to provide more accurate measurement of sensitive behaviors. I propose a framework for examining the accuracy of peer-reports and show that weighting peer-reports by peers’ network centrality can lead to more accurate peer-reports that can be used to more effectively correct potentially biased self-reports. Second, I study social contagion in social networks. I identify a list of methodological challenges in doing so and implement a design-based instrumental variable approach that helps to estimate causal peer effects with “validated” instrumental variables. Lastly, I introduce multilevel meta network analysis and apply it to study network dynamics in a social network-based smoking prevention intervention. The findings highlight the importance of examining network outcomes in evaluating health interventions and the potential of using social networks to design more powerful interventions.

Dr. Weihua An is Associate Professor of Sociology & Quantitative Theory and Methods and associated faculty of the East Asian Studies Program, the Goizueta Business School, and the Rollins School of Public Health at Emory University. He received a Ph.D. in Sociology and an A.M. in Statistics from Harvard University and was a doctoral fellow and a postdoc fellow at Harvard Kennedy School. His research advances theories and methods for network analysis and causal inference with applications to studying inequality and social policy, health, and organizations. Weihua An's website. 

Session: Afternoon Scientific Session - Bias, Fairness, and Inequality in an Algorithmic Age
Title: Algorithmic Reinterpretations: College Rankings and Socioeconomic Self-Sorting
Abstract: Public metrics like college rankings are algorithms that incorporate multiple inputs into single quantities, and a recurring criticism is that they impose excessive uniformity in evaluations. A less explored possibility is that single quantities enable users to draw diverse algorithmic reinterpretations that reflect their own background. I investigate the inequality implications of this possibility in the context of educational rankings. In conjoint experiment conducted among a diverse sample of U.S. adults (n=1,968), I show that college rankings are stronger signals exclusivity, academic rigor, safety, and stress for those from higher socioeconomic status (SES) backgrounds, while serving as stronger signals of exclusion (unwelcoming) for those from lower SES backgrounds. In a second experiment among a diverse sample of U.S. adolescents (n=800), I find that first-generation students perceive rankings as a stronger signal for college cost than their more advantaged peers, with no corresponding change in perceived financial aid. These cost differences partially explain why lower SES students prefer not to attend higher prestige colleges. In the case of college rankings, algorithms produce quantities that are differentially reinterpreted along SES cleavages, likely contributing to self-sorting where lower SES students prefer colleges of lower prestige, and vice versa.

James Chu studies economic and organizational sociology, social stratification, and political polarization. His primary line of research investigates how status is defined and allocated among social actors, and how varying ways of organizing status competitions translate to different patterns of inequality and conflict. James Chu's website.

Session: Short Course 2
Title: Sensitivity Analysis for Causal Inference in the Social Sciences
Abstract: The past few decades have witnessed rapid and unprecedented theoretical progress on the science of causal inference. Most of this theoretical progress, however, relies on strong, exact assumptions, such as the absence of unobserved common causes (ignorability assumptions), or the absence of certain direct effects (exclusion restrictions). Unfortunately, more often than not these assumptions are very hard to defend in practice, especially in the social sciences. This leads to two undesirable consequences for applied quantitative work: (i) important research questions may be neglected, simply because they do not exactly match the requirements of current methods; or, (ii) researchers may succumb to making the required “identification assumptions” simply to justify the use of available methods, but not because these assumptions are truly believed (or understood).  In this course, we will cover new theory, methods, and software for permitting causal inferences under more flexible and realistic settings. We will focus on a flexible suite of sensitivity analysis tools for common identification strategies, such as confounding adjustment, instrumental variables, and difference-in-differences. These tools can be immediately put to use to improve the robustness and transparency of current applied research, and students are encouraged to bring their own examples. 

Carlos is an assistant professor at the Department of Statistics at the University of Washington. He is also a data science fellow in the eScience Institute, and the Consulting Director of the Center for Statistics and the Social Sciences. He obtained my Ph.D. in Statistics at the University of California, Los Angeles, advised by Chad Hazlett and Judea Pearl. His research focuses on developing new causal and statistical methods for transparent and robust causal claims in the empirical sciences. He is particularly interested in the inferential challenges faced by social and health scientists, as well as the intersections of causality with machine learning and artificial intelligence. Carlos Cinelli's website.

Session: Alumni Panel
UW PhD: Statistics, 2006

Adam Glynn is Professor in the Department of Political Science and the Department of Quantitative Theory and Methods with a secondary appointment in the Department of Biostatistics and Bioinformatics at Emory University. Glynn's research focuses on causal inference and measurement for social science applications. He is a co-founder and co-director of the Politics of Policing Lab at Emory. Prior to joining Emory, Glynn held faculty positions at Harvard University. He received a PhD in Statistics from the University of Washington in 2006. Adam Glynn's website. 

Session: Afternoon Scientific Session - Bias, Fairness, and Inequality in an Algorithmic Age 
Title: Fair inference in multilevel data analysis
Abstract: Mixed effects models are used routinely in the social sciences to share information across multiple groups, such as schools or counties. The statistical properties of such models are often quite good on average across groups, but may be poor for any specific group. For example, commonly-used confidence interval procedures may maintain a target coverage rate on average across groups, but have a near zero coverage rate for a group that differs substantially from the others. As such, this unfairness in statistical performance can most adversely affect groups about which there is most concern. In this talk, we review some basic mixed effects modeling tools, discuss their group-specific properties, and present some new tools for multiple testing and inference problems that permit information sharing, while maintaining equal coverage rates for each group.

Peter Hoff does research in multivariate statistics, Bayesian methods and multilevel modeling. Before joining the Department of Statistical Science at Duke University in 2016, he was a Professor of Statistics at UW and core faculty member of CSSS since 2000. Peter Hoff's website.

Session: Discussant, Morning Scientific Session - Advances in Social Network Analysis

Yuan Hsiao is an Assistant Professor in the Department of Communication at the University of Washington. His major research explores the intersection of political communication, social media, and social networks. He is particularly interested in bringing a social network perspective to understanding a variety of communication and social processes, such as how networks on social media contribute to protest mobilization, how social interactions shape the production of misinformation and public opinion, how spatial and social relationships affect the spread of religion, or how community networks affect health behavior. He then combines multiple sources of data, such as “big” digital data, survey experiments, or historical archives, to glean insight into general theoretical processes. His work spans the disciplines of communication, sociology, political science, and public health, and he is deeply interested in inter-disciplinary dialogues. Yuan Hsiao's website. 

Session: Discussant, Afternoon Scientific Session - Bias, Fairness, and Inequality in an Algorithmic Age

Kosuke Imai is Professor in the Department of Government and the Department of Statistics at Harvard University. He is also an affiliate of the Institute for Quantitative Social Science. Before moving to Harvard in 2018, Imai taught at Princeton University for 15 years where he was the founding director of the Program in Statistics and Machine Learning. Imai specializes in the development of statistical methods and machine learning algorithms and their applications to social science research. His areas of expertise include causal inference, computational social science, and survey methodology. Imai leads the Algorithm-Assisted Redistricting Methodology Project (ALARM) and served as an expert witness for several high-profile legislative redistricting cases. In addition, he is the author of Quantitative Social Science: An Introduction (Princeton University Press, 2017). Outside of Harvard, Imai was the President of the Society for Political Methodology from 2017 to 2019. Kosuke Imai's website. 

Session: Alumni Panel
CSSS Track: Political Science, 2013

Carolina is a data scientist committed to developing ethical public sector data capacity. She has been working with King County for over six years working to support cross-system data integration, equity-centered data governance, and creative problem-focused data uses, as well as supporting technical development of a large and growing team of evaluators and data scientists. Before joining King County she completed a PhD in Political Science on the CSSS track at UW, with a focus on understanding the civic effects of participatory budgeting in local communities. Carolina Johnson's website.

Session: Short Course 3
Title: Text as Data
Abstract: This short course will offer a introduction and overview into the use of text as data in quantitative social scientific analyses. The course will focus on outlining a number of the ways that social scientists are combining machine learning tools with social scientific research designs to conduct a range of statistical analyses.

Benjamin Mako Hill is an Associate Professor in the Department of Communication at the University of Washington and a founding member of the Community Data Science Collective. He is also an Adjunct Associate Professor in UW’s Department of Human-Centered Design & Engineering, Paul G. Allen School of Computer Science & Engineering, and Information School. He is also a Faculty Associate at the Berkman Klein Center for Internet and Society and an affiliate at the Institute for Quantitative Social Science—both at Harvard University. Benjamin Mako Hill's website.

Session: Alumni Panel & Morning Scientific Session - Advances in Social Network Analysis
CSSS Track: Anthropology, 2010
Title: Mosuo social networks do not support universalist theories of gendered social relationships 
Abstract: Gender is often thought to constrain activities and social relationships. Evolutionary theorists link such constraints to gendered differences in payoffs from childcare versus activities focused on politicking and securing additional reproductive partners. According to this “universalist” view, women’s networks should often be smaller, more stable, more focused on dyadic interactions, and less hierarchical than men’s. To investigate these propositions, we collected social network data from Mosuo women and men residing in the Hengduan Mountains of Southwest China. Mosuo people share much in common but reside within two strikingly different kinship systems – a matrilineal one emphasizing the centrality of women and a patrilineal one emphasizing the centrality of men. We show that women’s friendship networks are larger than men’s in matrilineal communities, but smaller than men’s in patrilineal ones. Our results are therefore inconsistent with universal gender differences in social networks and speak instead to flexibility in gendered social interactions. 

This talk summarizes collaborative work. Siobhán Mattison, the speaker, is an associate professor of evolutionary anthropology at the University of New Mexico and Director of the Human Family and Evolutionary Demography Laboratory. She is interested in understanding apparent paradoxes in human family structures and how kinship impacts health. She works in Vanuatu and in Southwest China. Siobhán Mattison's website. 

Session: Afternoon Scientific Session - Bias, Fairness, and Inequality in an Algorithmic Age
Title: Advancing Algorithmic Fairness: A Statistical Learning Approach with Fairness Constraints 
Abstract: Statistical machine learning algorithms, crucial in sectors like hiring, finance, and healthcare, risk reinforcing societal biases based on gender, race, religion, among others. To combat this, it's vital to design models adhering to fairness norms. This involves embedding fairness constraints such as 'equal opportunity' [Hardt et al., 2016], ensuring uniform true positive rates across groups, and 'path-specific counterfactual fairness' [Nabi and Shpitser, 2018, Nabi et al., 2019], which restricts the effect of the sensitive feature on the outcome along certain user-specified mediating pathways. Without favoring a specific fairness criterion, we propose a general framework for deriving optimal prediction functions under various constraints. It conceptualizes the learning problem as estimating a constrained functional parameter within a comprehensive statistical model, using a Lagrange-type penalty. Key contributions of our work include a flexible framework for solving constrained optimization problems, closed-form solutions for specific fairness constraints, and an algorithm-neutral approach to fair learning.  

Dr. Razieh Nabi is an endowed Rollins Assistant Professor in the Department of Biostatistics and Bioinformatics at Emory Rollins School of Public Health, with the secondary appointment in the department of Computer Science at Emory. She earned her PhD from Johns Hopkins University in 2021. Dr. Nabi's methodological research encompasses a range of topics including addressing both measured and unmeasured confounders in causal inference from observational studies, mediation analysis, ensuring algorithmic fairness, and strategies for dealing with missing and censored data. Her work primarily utilizes graphical models and employs both non-parametric and semi-parametric statistical methods. Razieh Nabi's website.

Session: Alumni Panel
CSSS Track: Statistics, 2010

Gail Potter is a mathematical statistician in NIAID’s Biostatistics Research Branch and the Deputy Section Chief of the Clinical Trials Research Section (CTRS).  She provides statistical leadership in the design and analysis of clinical trials and observational studies and oversees the development and implementation of statistical standard operating procedures for CTRS.  Her methodological research relates to statistical issues in clinical trial design and interpretation as well as statistical modeling of face-to-face social networks to understand epidemic transmission.  She served in the Peace Corps in Guinea and Nepal and performed field work in Malawi and Senegal. Gail Potter's website. 

Session: Morning Scientific Session - Advances in Social Network Analysis
Title: Population-Level Balance in Signed Networks
Abstract: In many real-world networks,  relationships often go beyond simple presence or absence; they can be positive (e.g., friendship, alliance, and mutualism) or negative (e.g., enmity, disputes, and competition). These negative relationships display substantially different properties from positive ones, and more importantly, their presence interacts in unique ways. The balance theory originating from social psychology, illustrated by proverbs like "a friend of my friend is my friend'' and "an enemy of my enemy is my friend'', provides insight into the formation mechanism of positive and negative connections. In this talk, we characterize the balance theory with a novel and natural notion of population-level balance. We propose a nonparametric inference method to evaluate the real-world evidence of population-level balance in signed networks. Inspired by the empirical findings, we further develop a general latent space framework for modeling signed networks while accommodating the balance theory. 

Weijing Tang is an Assistant Professor in Statistics and Data Science at Carnegie Mellon University. She has been working on developing statistical methodology and theory for network analysis, machine learning, and survival analysis with applications to health and social sciences. Weijing Tang's website. 

Session: Short Course 1
Title: Introduction to Missing Data Methods for Observational Data
Abstract: Missing data are common in many disciplines including the social sciences, and therefore understanding the impact of missing data on estimation and inference as well as the strengths and weaknesses of approaches to handling missing data is crucial. This course aims to provide participants with the knowledge and tools to assess and mitigate the impact of missing data in their research. We will begin with an overview of common missing data mechanisms and traditional methods to handling missing data before discussing more contemporary approaches including multiple imputation. Methods will be illustrated using R. This short course is targeted towards individuals with little or no prior experience with modern missing data methods. Prior introductory applied statistical coursework (SOC 504-505-506 or equivalent) and experience using regression methods to analyze data (e.g., linear regression, logistic regression) is important background for this course.

Katie Wilson is an Assistant Teaching Professor in the Department of Biostatistics at the University of Washington. Katie Wilson's website.

Session: Statistical Demography Seminar
Title: Measuring and understanding the dynamics of populations of scholars
Abstract: For a quarter of a century, the Center for Statistics and the Social Sciences (CSSS) has fostered interdisciplinary research. This talk focuses on a project that started thanks to CSSS, and that builds on the spirit of the Center, in terms of blending statistics, demography (and social sciences more broadly), and modern data science practices. We measure migration of scholars based on information on changes in their institutional affiliations over time, using metadata on over 36 million journal articles and reviews indexed by Scopus. Specifically, we produce a database of annual international migration flows of scholars, for all countries, from 1998-2018 (the “Scholarly Migration Database”). We then use the newly generated database to provide evidence on the relationship between economic development and the emigration propensity of scholars, to assess patterns and trends of gender inequalities in international mobility, and the impact of policies on migration. This talk highlights initial key results as well as further ongoing developments to fully model and understand the dynamics of populations of scholars, including entry and exit from the academic system, in addition to mobility across space at different levels of spatial granularity.  

Emilio Zagheni is Director of the Max Planck Institute for Demographic Research (MPIDR) and Affiliate Associate Professor of Sociology at the University of Washington. He received his Ph.D. in Demography (2010) and M.A. in Statistics (2008) from U.C. Berkeley. Zagheni is best known for his work on combining digital trace data and traditional sources to track and understand migrations and to advance population science. In 2016 he received the Trailblazer Award for Demographic Analysis from the European Association for Population Studies for his role in developing the field of Digital and Computational Demography. Emilio Zagheni's website.