About the Committee
Our new concentration in Quantitative Methods and Social Analysis (QMSA) draws from the interdisciplinary faculty of the University-wide Committee on Quantitative Methods in Social, Behavioral, and Health Sciences. Dramatic advances in statistical modeling, experimental design, and statistical analysis have created unprecedented opportunities for advancing knowledge across a wide range of disciplines. There is an ever-greater demand for scholars who can innovate methodologically, who understand how to use the theory of statistical inference to tackle really hard problems in social, behavioral, and health science. We look to theory-based models for populations and societies, examining biological, behavioral, and environmental factors and the way they interact. We are primarily interested in (a) theoretical controversies, questions, and hypotheses that arise in scientific discourse and the formal models that can make these precise, (b) systematic measurement of key theoretical constructs with known and consistent psychometric properties, (c) the design of research to test these models, (d) the specification of assumptions that are required for linking analytic results to theoretical claims, and (e) the validity of statistical inferences.
The QMSA Concentration
This concentration is for students who seek rigorous training and critical exposure to the latest techniques of quantitative social science. Admitted candidates may participate on research teams or conduct independent projects applying quantitative thinking and analysis to important research questions. Our goal is to prepare students for PhD study in quantitative social science, and for professional positions at research institutions and government or nongovernment agencies.
Students who declare an interest in QMSA admission must have a minimum quantitative GRE of 75%. They must also furnish a statement of purpose outlining their intended research and the two QMSA faculty members they most hope to work with.
If accepted, students will select a minimum of 5 courses in theoretical modeling, research design, causal inference, and statistical analysis, and write their MA thesis with a member of the QMSA faculty.
One of those 5 courses will be Design and Analysis in Social, Behavioral, and Health Sciences, taught in the Fall quarter.
In addition, students will take Perspectives in Social Science Analysis and up to 3 electives in their social science field.
Finally, QMSA students must attend the biweekly Workshop on Quantitative Methods in Education, Health, and Social Sciences (QMeHSS). That Workshop invites leading methodologists to present their work, and offers an ideal venue for students to get up to speed with the latest developments in quantitative research.
STAT 24400. Statistical Theory. (Rina Foygel Barber, Stephen Stigler) This is a systematic introduction to the principles and techniques of statistics, as well as to practical considerations in the analysis of data, with emphasis on the analysis of experimental data. (1/2)
STAT 24500. Statistical Theory II. (Weibiao Wu) This sequence is a systematic introduction to the principles and techniques of statistics, as well as to practical considerations in the analysis of data, with emphasis on the analysis of experimental data. (2/2)
PPHA 31200. Mathematical Statistics for Public Policy I. (Jeffrey Grogger) This course focuses on concepts used in statistical inference. This course will introduce students to basic principles of probability and statistics: random variables, standard distributions, and hypothesis testing. Lectures will explore uses of these principles in policy analyses. This course seeks to prepare students for PPHA 31300. This course will assume a greater mathematical sophistication on the part of students than is assumed in PPHA 31000.
PPHA 31300. Mathematical Statistics for Public Policy II. (Jeffrey Grogger) A continuation of PPHA 31200, this course focuses on the statistical concepts and tools used to study the association between variables and causal inference. This course will introduce students to regression analysis and explore its uses in policy analyses. This course will assume a greater mathematical sophistication on the part of students than is assumed in PPHA 31100.
CHDV 30102, MACS 51000, PBHS 43201, PLSC 30102, STAT 31900, SOCI 30315. Introduction to Causal Inference. (Guanglei Hong and Kazuo Yamaguchi) This course is designed for graduate students and advanced undergraduate students from the social sciences, education, public health science, public policy, social service administration, and statistics who are involved in quantitative research and are interested in studying causality. The goal of this course is to equip students with basic knowledge of and analytic skills in causal inference. Topics for the course will include the potential outcomes framework for causal inference; experimental and observational studies; identification assumptions for causal parameters; potential pitfalls of using ANCOVA to estimate a causal effect; propensity score based methods including matching, stratification, inverse-probability-of-treatment-weighting (IPTW), marginal mean weighting through stratification (MMWS), and doubly robust estimation; the instrumental variable (IV) method; regression discontinuity design (RDD) including sharp RDD and fuzzy RDD; difference in difference (DID) and generalized DID methods for cross-section and panel data, and fixed effects models. Intermediate Statistics or equivalent such as STAT 224 is a prerequisite. This course is a prerequisite for “Advanced Topics in Causal Inference” and “Mediation, moderation, and spillover effects.”
CHDV 30102. Advanced Topics in Causal Inference. (Guanglei Hong) This course provides an in-depth discussion of selected topics in causal inference that are beyond what are covered in the introduction to causal inference course. The course is intended for graduate students and advanced undergraduate students who have taken the intro course and want to extend their knowledge in causal inference. Topics include (1) alternative matching methods, randomization inference for testing hypothesis and sensitivity analysis; (2) marginal structural models and structural nested models for time-varying treatment; (3) Rubin Causal Model (RCM) and Heckman's scientific model of causality; (4) latent class treatment variable; (5) measurement error in the covariates; (6) the M-estimation for the standard error of the treatment effect for the use of IPW; (7) the local average treatment effect (LATE) and its problems, sensitivity analysis to examine the impact of plausible departure from the IV assumptions, and identification issues of multiple IVs for multiple/one treatments; (8) Multi-level data for treatment evaluation for multilevel experimental designs and observational designs, and spilt-over effect; (9) Nonignorable missingness and informative censoring issues.
CHDV 32411, CCTS 32411, PBPL 29411, PSYC 32411, STAT 33211. Mediation, Moderation, and Spillover. (Guanglei Hong) This course is designed for graduate students and advanced undergraduate students from social sciences, statistics, public health science, public policy, and social services administration who will be or are currently involved in quantitative research. Questions about why a treatment works, for whom, under what conditions, and whether one individual’s treatment could affect other individuals’ outcomes are often key to the advancement of scientific knowledge. We will clarify the theoretical concepts of mediated effects, moderated effects, and spillover effects under the potential outcomes framework. The course introduces cutting-edge methodological approaches and contrasts them with conventional strategies including multiple regression, path analysis, and structural equation modeling. The course content is organized around application examples. The textbook “Causality in a Social World: Moderation, Mediation, and Spill-Over” (Hong, 2015) will be supplemented with other readings reflecting latest developments and controversies. Weekly labs will provide tutorials and hands-on experiences. All students are expected to contribute to the knowledge building in class through participation in presentations and discussions. Students are encouraged to form study groups, while the written assignments are to be finished and graded on an individual basis. Intermediate Statistics, Introduction to Causal Inference, and their equivalent are prerequisites.
PLSC 30901. Game Theory I. (Monika Nalepa) This is a course for graduate students in Political Science. It introduces students to games of complete information through solving problem sets. We will cover the concepts of equilibrium in dominant strategies, weak dominance, iterated elimination of weakly dominated strategies, Nash equilibrium, subgame perfection, backward induction, and imperfect information. The course will be centered around several applications of game theory to politics: electoral competition, agenda control, lobbying, voting in legislatures, and coalition games. This course serves as a prerequisite for Game Theory II.
PLSC 31000. Game Theory II. (John Patty) This is a course for graduate students in Political Science. It introduces students to games of incomplete information through solving problem sets. We will cover the concepts of Bayes Nash equilibrium, perfect Bayesian equilibrium, and quantal response equilibrium. In terms of applications, the course will extend the topics examined in the prerequisite, PLSC 30901. Game Theory I to allow for incomplete information, with a focus on the competing challenges of moral hazard and adverse selection in those settings.
PLSC 40801. Social Choice Theory. (Elizabeth Penn) This course will provide you with an introduction to the field of social choice theory, the study of aggregating the preferences of individuals into a "collective preference." It will focus primarily on classic theorems and proof techniques, with the aim of examining the properties of different collective choice procedures and characterizing procedures that yield desirable outcomes. The classic social choice results speak not only to the difficulties in aggregating the preferences of individuals, but also to the difficulties in aggregating any set of diverse criteria that we deem important to making a choice or generating a ranking. Specific topics we will cover include preference aggregation, rationalizable choice, tournaments, sophisticated voting, domain restrictions, and the implicit trade-offs made by game theoretic versus social choice theoretic approaches to modeling.
PLSC 40815. New Directions in Formal Theory. (Elizabeth Penn) In this graduate seminar we will survey recent journal articles that develop formal (mathematical) theories of politics. The range of topics and tools we touch on will be broad. Topics include models of institutions, groups, and behavior, and will span American politics, comparative politics, and international relations. Tools include game theory, network analysis, simulation, axiomatic choice theory, and optimization theory. Our focus will be on what these models are theoretically doing: What they do and do not capture, what makes one mathematical approach more compelling than another, and what we can ultimately learn from a highly stylized (and necessarily incomplete) mathematical representation of politics. The goal of the course is for each participant, including the professor, to emerge with a new research project. Some background in formal modeling, such as a prior course in game theory, is required.
PLSC 43100. Maximum Likelihood. (John Brehm) The purpose of this course is to familiarize students with the estimation and interpretation of maximum likelihood, a statistical method which permits a close linkage of deductive theory and empirical estimation. Among the problems considered in this course include: models of dichotomous choice, such as turnout and vote choice; models of limited categorical data, such as those for multi-party elections and survey responses; models for counts of uncorrelated events, such as executive orders and bookburnings; models for duration, such as the length of parliamentary coalitions or the tenure of bureaucracies; models for compositional data, such as allocation of time by bureaucrats to task and district vote shares; and models for latent variables, such as for predispositions. The emphasis in this course will be on the extraction of information about political and social phenomena, not upon properties of estimators.
PLSC 43200. Maximum Likelihood II. (John Brehm) This course furthers and expands upon the methods covered in the first Maximum Likelihood course (PLSC 43100). The format of the course will be a special topics course, focused around detailed discussion about the implementation and interpretation of applications of ML methods in the social sciences. In particular, we are likely to cover multiple equation models, event history, treatment of censored/ unmeasured observations, and item response theory. The course will incorporate alternative methods of computation of results beyond strict optimization of likelihoods. PLSC 43100 is a mandatory prerequisite, no exceptions.
PLSC 35801. Formal Theory and Comparative Politics. (Monika Nalepa)In this course we will discuss 10 newly published or still in press papers in Comparative Politics that employ formal modeling. We will study models of the state and its security agencies (by Dragu and Tyson), models of state-building (by Robinson and Lessing), models of authoritarianism and regime change (Svolik, Little, and Miller) and models of corruption and clientelism (by Stokes, Nichter, and Rueda). Because of its topical breadth, this course may therefore be also taken as a field survey in comparative politics.
STAT 34700. Generalized Linear Models. (Peter McCullagh) This applied course covers factors, variates, contrasts, and interactions; exponential-family models (i.e., variance function); definition of a generalized linear model (i.e., link functions); specific examples of GLMs; logistic and probit regression; cumulative logistic models; log-linear models and contingency tables; inverse linear models; Quasi-likelihood and least squares; estimating functions; and partially linear models.
SOCI 30253. Introduction to Spatial Data Science. (Luc Anselin) Spatial data science is an evolving field that can be thought of as a collection of concepts and methods drawn from both statistics and computer science. These techniques deal with accessing, transforming, manipulating, visualizing, exploring and reasoning about data where the locational component is important (spatial data). The course introduces the types of spatial data relevant in social science inquiry and reviews a range of methods to explore these data. The types of data considered include observations at the point level (e.g., locations of crimes, commercial establishments, traffic accidents), data gathered for aggregate units, such as census tracts or counties (e.g., unemployment rates, disease rates by area, crime rates), and data measured at spatially located sampling points (such as air quality monitoring stations and urban sensors). Specific topics covered include the implementation of formal spatial data structures, geovisualization and visual analytics, spatial autocorrelation analysis, variogram analysis, cluster detection, regionalization, point pattern analysis and spatial data mining. An important aspect of the course is to learn and apply open source geospatial software tools, such as R and GeoDa.
SOCI 40217. Spatial Regression Analysis. (Luc Anselin) This course covers statistical and econometric methods specifically geared to deal with the problems of spatial dependence and spatial heterogeneity in cross- sectional and panel (space-time) data. The main objective of the course is to gain insight into the scope of spatial regression methods, to be able to apply them in an empirical setting, and to properly interpret the results of spatial regression analysis. While the focus is on spatial aspects, the types of methods covered have general validity in statistical practice. The course covers the specification of spatial regression models in order to incorporate spatial dependence and spatial heterogeneity, as well as different estimation methods and specification tests to detect the presence of spatial autocorrelation and spatial heterogeneity. Special attention is paid to the application to spatial models of generic statistical paradigms, such as Maximum Likelihood and Generalized Method of Moments. An important aspect of the course is an emphasis on computation, with the application of open source software tools such as R, GeoDa/GeoDaSpace and PySAL to solve empirical problems, and the use of R or Python to design simulation exercises.
SOCI 40204. Categorical Data Analysis. (Xi Song) It is expected that the students have a good understanding of basic descriptive statistics such as means, variances and expectation, of the inferential notions of estimate, confidence intervals and significance or hypothesis testing. Familiarity with one statistical package, e.g. R, Splus, SAS, SPSS, Stata or Minitab, and ability to access Web sites and to download files from the Web are required. The free statistical package R will be used in this course. This course is an introduction to the theory and applications of statistical methods for investigating the relationships among discrete variables. The course will present methods for analyzing categorical data, including standard methods for contingency tables such as odds ratios, tests of independence and various measures of association, generalized linear models for binary data and count data, logistic regression for binomial data, loglinear models for Poisson data, and models for paired samples with categorical responses. The statistical techniques discussed will be presented in many real examples involving physical, biological and social science data.
SOCI 40212. Demographic Technique. (Xi Song) Introduction to methods of demographic analysis. Topics include demographic rates, standardization, decomposition of differences, life tables, survival analysis, cohort analysis, birth interval analysis, models of population growth, stable populations, population projection, and demographic data sources.
SOCI 40103. Event History Analysis. (Kazuo Yamaguchi) An introduction to the methods of event history analysis will be given. The methods allow for the analysis of duration data. Non-parametric methods and parametric regression models are available to investigate the influence of covariates on the duration until a certain even occurs. Applications of these methods will be discussed i.e., duration until marriage, social mobility processes organizational mortality, firm tenure, etc.
PBHS 33300, STAT 36900. Applied Longitudinal Data Analysis. (Donald Hedeker) Longitudinal data consist of multiple measures over time on a sample of individuals. These types of data occur extensively in both observational and experimental biomedical and public health studies, as well as in studies in sociology and applied economics. This course will provide an introduction to the principles and methods for the analysis of longitudinal data. Whereas some supporting statistical theory will be given, emphasis will be on data analysis and interpretation of models for longitudinal data. Problems will be motivated by applications primarily in mental health, public health, prevention research, and health services research.
SOCI 30112. Applications of Hierarchical Linear Models. (Stephen Raudenbush). A number of diverse methodological problems such as correlates of change, analysis of multi-level data, and certain aspects of meta-analysis share a common feature--a hierarchical structure. The hierarchical linear model offers a promising approach to analyzing data in these situations. This course will survey the methodological literature in this area, and demonstrate how the hierarchical linear model can be applied to a range of problems.
PBHS 33400. Multilevel Modeling. (Donald Hedeker) This course will focus on the analysis of multilevel data in which subjects are nested within clusters (e.g., health care providers, hospitals). The focus will be on clustered data, and several extensions to the basic two-level multilevel model will be considered including three-level, cross-classified, multiple membership, and multivariate models. In addition to models for continuous outcomes, methods for non-normal outcomes will be covered, including multilevel models for dichotomous, ordinal, nominal, time-to-event, and count outcomes. Some statistical theory will be given, but the focus will be on application and interpretation of the statistical analyses.
PBHS 33501. Statistical Applications. (Robert Gibbons) This course provides a transition between statistical theory and practice. The course will cover statistical applications in medicine, mental health, environmental science, analytical chemistry, and public policy. Lectures are oriented around specific examples from a variety of content areas. Opportunities for the class to work on interesting applied problems presented by U of C faculty will be provided. Although an overview of relevant statistical theory will be presented, emphasis is on the development of statistical solutions to interesting applied problems.
PSYC 37300. Experimental Design. (Steven Shevell) This course covers topics in research design and analysis. They include multifactor, completely randomized procedures and techniques for analyzing data sets with unequal cell frequencies. Emphasis is on principles, not algorithms, for experimental design and analysis.
PPHA 34600. Program Evaluation. (Juan Pantano) The goal of this course is to introduce students to program evaluation and provide an overview of current issues and methods in impact evaluation. We will focus on estimating the causal impacts of programs and policy using social experiments, panel data methods, instrumental variables, regression discontinuity designs, and matching techniques. We will discuss applications and examples from the fields of education, demography, health, crime, job training, and others. Prerequisites: PP31001 and PP31101 or equivalent statistics coursework.
PPHA 42600. Survey Research Methods and Analysis. (Colm O’Muircheartaigh) The goal of this course is to learn about the methods used to collect publicly available survey data that can be used for policy research so that students can appropriately use these data to answer policy relevant questions. Students will learn about the methods used to collect survey data, how to develop researchable policy questions that can be answered with the survey data, and about the limitations of the survey data for answering policy research questions. In order to analyze policy questions using available survey data, students will also learn about actual survey instruments, survey sample designs, survey data processing, and survey data systems that the major public policy relevant surveys use. The course will also examine specific measurement and analysis issues that are of interest to policy research (e.g., measuring public program enrollment and public program eligibility simulation). By the end of the course each student will understand the methods used to collect survey data, have developed a researchable policy question, carried out the appropriate analysis to answer the question, produced high quality analytical tables, and written up descriptions of the methods used to produce the numbers in the tables in a style that is consistent with professional policy research.