Applied Ba Yes
Applied Ba Yes
Applied Ba Yes
Teaching assistants
Herrissa Lamothe Neil Williams
Princeton University University of Georgia
hlamothe@princeton.edu snpwill@uga.edu
Office: Helen Newberry Building Office: Helen Newberry Building
Office hours: 2pm–3pm Office hours: 12pm–1pm
1
2
Course description
This workshop introduces the basic theoretical and applied principles of Bayesian statistical analysis in a manner
geared toward students and researchers in the social sciences. The Bayesian paradigm is particularly useful for
the type of data that social scientists encounter given its recognition of the mobility of population parameters,
its ability to incorporate information from prior research, and its ability to update estimates as new data are
observed. The workshop begins with a discussion of the strengths of the Bayesian approach for social science
data and the philosophical differences between Bayesian and frequentist analyses. Next, the workshop covers
the theoretical underpinnings of Bayesian modeling and provides a brief introduction to the primary estima-
tion algorithms. The bulk of the workshop focuses on estimating and interpreting Bayesian models from an
applied perspective. Participants are introduced to the Bayesian forms of the standard statistical models taught
in regression and MLE courses (i.e., linear, logit/probit, poisson, etc.). Additional topics include measurement
models, model comparison, and an in-depth treatment of multilevel modeling.
Participants should have a solid understanding of the linear model and matrix algebra and some exposure to
models with limited dependent variables. The Mathematics for Social Scientists III lecture with Don Eckford
is an ideal companion lecture for this workshop.
The workshop relies mostly on R and WinBUGS/JAGS for estimation, with a short segment on Stan, a newer,
but rapidly growing tool for Bayesian inference. Prior experience with R is preferred but not assumed or neces-
sary. We offer several lab sessions to familiarize participants with R, WinBUGS, JAGS, and Stan.
Goals. Upon conclusion of this workshop, we aim for participants to be able to:
· appreciate the fundamental differences and similarities between frequentist and Bayesian approaches to
inference
· apply Bayes’ rule to statistical inference
· formulate linear and generalized linear models in the Bayesian framework
· estimate linear and generalized linear models in the Bayesian framework using customizeable code
· exploit the advantages of Bayesian estimation with regard to
– incorporating prior information
– estimating uncertainty in parameter estimates
– dealing with missing data
– measuring latent concepts
– incorporating variance at multiple levels of observation
· present and communicate results from Bayesian (and frequentist) estimation in an effective manner
· have fun learning new methods and better understanding familiar ones!
A note on computing. This workshop mostly uses JAGS and WinBUGS (and its open-source twin Open-
BUGS) as the primary software options to fit Bayesian models, with one unit toward the end dedicated to Stan.
We access JAGS and Win/OpenBUGS through R. Most lectures build on JAGS and Win/OpenBUGS. The lan-
guages of these two programs are nearly identical. Win/OpenBUGS run on Macs only with Windows emulation
software, but can be a bit buggy. JAGS runs on all platforms, including Macs. We offer special Mac-friendly lab
sessions and support both JAGS and Win/OpenBUGS. JAGS code for all models encountered in this workshop
and other JAGS-specific code and examples are provided.
Workshop resources
M-Box: All slides, code used in workshop sessions, and problem sets will be posted on M-Box. Details will be
provided on the first day of the workshop. The TAs are also always available to point you to these materials.
3
Google Group: We will ask you to sign you up for a Google Group for this workshop. We will use this Google
Group to share information related to the workshop during the ICPSR Summer Program.
Website with additional materials: Additional code, tutorials, and other materials for weeks 3–4 are posted
on Johannes’ website: http://www.jkarreth.net/bayes-icpsr.html.
Reading materials
Books
The main required texts for this workshop are:
· Gelman, Andrew and Hill, Jennifer. 2007. Data Analysis Using Regression and Multilevel/Hierarchical
Models. New York, NY: Cambridge University Press.
· Gill, Jeff. 2014. Bayesian Methods: A Social and Behavioral Sciences Approach, Third Edition. Boca Raton,
FL: Chapman / Hall/CRC.
You may also find the following titles useful for many of the topics discussed in this workshop. They are available
in the ICPSR Summer Program Library for borrowing:
· Congdon, Peter D. 2003. Applied Bayesian Modelling. Chichester: Wiley.
· Congdon, Peter D. 2010. Applied Bayesian Hierarchical Methods. Boca Raton, FL: Chapman / Hall/CRC.
· Gelman, Andrew et al. 2013. Bayesian Data Analysis, Third Edition. Boca Raton, FL: Chapman & Hall/CRC.
· Jackman, Simon. 2009. Bayesian Analysis for the Social Sciences. Chichester: Wiley.
· Kruschke, John. 2014. Doing Bayesian Data Analysis, Second Edition: A Tutorial with R, JAGS, and STAN.
Oxford: Academic Press / Elsevier.
· Lunn, David et al. 2012. The BUGS Book: A Practical Introduction to Bayesian Analysis. Boca Raton, FL:
Chapman / Hall/CRC.
· Ntzoufras, Ioannis. 2009. Bayesian Modeling Using WinBUGS. Hoboken, NJ: Wiley.
Articles
All articles listed in the syllabus are made available to you on M-Box and through the University of Michigan
library website from the campus network. Please contact the TAs if you have trouble locating any of the articles
listed in the syllabus.
4
Software
This workshop relies mostly on R, JAGS/WinBUGS/OpenBUGS, and Stan. We may also briefly discuss Stata
as an alternative for some applications. We provide assistance installing R and JAGS/WinBUGS/OpenBUGS
on your computers during labs in the first week of the workshop. There is a learning curve for these programs,
but you need not have any computer programming background to learn them rather easily—just patience and
desire. Our goal is to make you as comfortable as possible with these programs by the end of this workshop so
that you will be able to use them with ease at your home institutions and in your own work.
Mac and JAGS users: See Johannes’ website for more information on installing JAGS.
Problem sets
We assign problem sets in class. Our goal is to make sure participants receive sufficient feedback to complete
all assignments successfully. We distribute between 2 and 4 assignments per week. They are mostly computer-
based with the exception of the first assignment. We ask that you include all code you used to complete your
assignments; the TAs will show you how to best do this. The TAs will aim to return graded assignments to you
within 5 days with comments via email. We (the instructors and TA) are more than happy to provide help with
assignments during office hours: don’t be afraid to come by and ask.
Labs
We offer several labs with guided hands-on exercises. Lab sessions will be held at the times in the room(s) listed
in the schedule below (watch for updates). Planned topics:
Monday, June 25
No course meeting
Recommended: Introduction to the LATEX Text Processing System, 5:30pm–7:30pm.
Wednesday, July 4
No lecture
Instructor & TAs are available for office hours by appointment.
Sample application:
· Blais, André, Guntermann, Eric, and Bodet, Marc A. 2017. “Linking Party Preferences and the Composi-
tion of Government: A New Standard for Evaluating the Performance of Electoral Democracy.” Political
Science Research and Methods 5 (2): 315–331.
HW 3 assigned: Linear model.
HW 4 assigned: Debugging BUGS/JAGS code.
· Gelman, Andrew et al. 2008. “A weakly informative default prior distribution for logistic and other re-
gression models.” Annals of Applied Statistics 2 (4): 1360–1383.
Sample application:
· Karreth, Johannes. 2018. “The Economic Leverage of International Organizations in Interstate Disputes.”
International Interactions 44 (3): 463–490.
HW 5 assigned: Logistic regression model.
· Neelon, Brian H, O’Malley, A James, and Normand, Sharon-Lise T. 2010. “A Bayesian model for repeated
measures zero-inflated count data with application to outpatient psychiatric service use.” Statistical Mod-
elling 10 (4): 421–439.
HW 7 assigned: Poisson model.
– Lopes, Hedibert Freitas. 2014. “Modern Bayesian Factor Analysis.” In Bayesian Inference in the Social
Sciences, ed. by Jeliazkov, Ivan and Yang, Xin-She, 115–153. John Wiley & Sons, Inc., sections 5.1 and
5.2.
· On IRT models:
– Jackman, Simon. 2001. “Multidimensional Analysis of Roll Call Data via Bayesian Simulation: Iden-
tification, Estimation, Inference, and Model Checking.” Political Analysis 9 (3): 227.
– Clinton, Joshua D. and Jackman, Simon. 2009. “To Simulate or NOMINATE?.” Legislative Studies
Quarterly 34 (4): 593–621.
Sample applications (please read one):
· Bakker, Ryan. 2009. “Re-measuring Left–Right: A Comparison of SEM and Bayesian Measurement Mod-
els for Extracting Left–Right Party Placements.” Electoral Studies 28 (3): 413–421.
· Bakker, Ryan and Poole, Keith T. 2013. “Bayesian Metric Multidimensional Scaling.” Political Analysis 21
(1): 125–140.
· Benson, Brett V. and Clinton, Joshua D. 2016. “Assessing the Variation of Formal Military Alliances.”
Journal of Conflict Resolution 60 (5): 866–898.
· Campbell, Susanna P., Findley, Michael G., and Kikuta, Kyosuke. Forthcoming. “An Ontology of Peace:
Landscapes of Conflict and Cooperation with Application to Colombia.” International Studies Review.
· Caughey, Devin and Warshaw, Christopher. 2015. “Dynamic Estimation of Latent Opinion Using a Hi-
erarchical Group-Level IRT Model.” Political Analysis 23 (2): 197–211.
· Caughey, Devin and Warshaw, Christopher. 2016. “The Dynamics of State Policy Liberalism, 1936-2014.”
American Journal of Political Science 60 (4): 899–913.
· Clinton, Joshua D. and Jackman, Simon. 2009. “To Simulate or NOMINATE?.” Legislative Studies Quar-
terly 34 (4): 593–621.
· Copelovitch, Mark S., Gandrud, Christopher, and Hallerberg, Mark. “Financial Regulatory Transparency
and Sovereign Borrowing Costs.”
· Fariss, Christopher J. 2014. “Respect for Human Rights has Improved Over Time: Modeling the Changing
Standard of Accountability.” American Political Science Review 108 (2): 297–318.
· Fox, Jean-Paul and Glas, Cees. 2001. “Bayesian Estimation of a Multilevel IRT Model Using Gibbs Sam-
pling.” Psychometrika 66 (2): 271–288.
· Fox, Jean-Paul and Glas, Cees A.W. 2003. “Bayesian modeling of measurement error in predictor variables
using item response theory.” Psychometrika 68 (2): 169–191.
· Garrett, Elizabeth S. and Zeger, Scott L. 2000. “Latent Class Model Diagnosis.” Biometrics 56 (4): 1055–
1067.
10
· Gray, Julia and Slapin, Jonathan B. 2012. “How Effective are Preferential Trade Agreements? Ask the
Experts.” Review of International Organizations 7 (3): 309–333.
· Hare, Christopher et al. 2015. “Using Bayesian Aldrich-McKelvey Scaling to Study Citizens’ Ideological
Preferences and Perceptions.” American Journal of Political Science 59 (3): 759–774.
· Hollyer, James R., Rosendorff, B. Peter, and Vreeland, James Raymond. 2014. “Measuring Transparency.”
Political Analysis 22 (4): 413–434.
· Linzer, Drew A. and Staton, Jeffrey K. 2015. “A Global Measure of Judicial Independence, 1948-2012.”
Journal of Law and Courts 3 (2): 223–256.
· Manatschal, Anita and Bernauer, Julian. 2016. “Consenting to Exclude? Empirical Patterns of Democracy
and Immigrant Integration Policy.” West European Politics 39 (2): 183–204.
· Rosas, Guillermo, Shomer, Yael, and Haptonstahl, Stephen R. 2015. “No News Is News: Nonignorable
Nonresponse in Roll-Call Data Analysis.” American Journal of Political Science 59 (2): 511–528.
· Selin, Jennifer L. 2015. “What Makes an Agency Independent?” American Journal of Political Science 59
(4): 971–987.
· Slapin, Jonathan B. and Proksch, Sven-Oliver. 2008. “A Scaling Model for Estimating Time-Series Party
Positions from Texts.” American Journal of Political Science 52 (3): 705–722.
· Treier, Shawn and Jackman, Simon. 2008. “Democracy as a Latent Variable.” American Journal of Political
Science 52 (1): 201–217.
HW 8 assigned: Factor or IRT model.
as well as any of these empirical articles using MLMs that is/are in your area of interest:
12
· Pang, Xun. 2010. “Modeling Heterogeneity and Serial Correlation in Binary Time-Series Cross-sectional
Data: A Bayesian Multilevel Model with AR(p) Errors.” Political Analysis 18:470–498.
· Pang, Xun. 2014. “Varying Responses to Common Shocks and Complex Cross-Sectional Dependence:
Dynamic Multilevel Modeling with Multifactor Error Structures for Time-Series Cross-Sectional Data.”
Political Analysis 22 (4): 464–496.
· Ward, Michael D., Siverson, Randolph M., and Cao, Xun. 2007. “Disputes, Democracies, and Dependen-
cies: A Reexamination of the Kantian Peace.” American Journal of Political Science 51 (3): 583–601.
· Blaydes, Lisa and Linzer, Drew A. 2012. “Elite Competition, Religiosity and Anti-Americanism in the
Islamic World.” American Political Science Review 106 (2): 225–243.
· Lock, Kari and Gelman, Andrew. 2010. “Bayesian Combination of State Polls and Election Forecasts.”
Political Analysis 18 (3): 337–348.
· Masters, Ryan K., Hummer, Robert A., and Powers, Daniel A. 2012. “Educational Differences in U.S. Adult
Mortality.” American Sociological Review 77 (4): 548–572.
· Stegmueller, Daniel. 2013a. “How Many Countries for Multilevel Modeling? A Comparison of Frequentist
and Bayesian Approaches.” American Journal of Political Science 57 (3): 748–761.
· Chaudoin, Stephen, Milner, Helen V., and Pang, Xun. 2015. “International Systems and Domestic Politics:
Linking Complex Theories with Empirical Models in International Relations.” International Organization
69 (2): 275–309.
· Beazer, Quintin H. and Woo, Byungwon. 2016. “IMF Conditionality, Government Partisanship, and the
Progress of Economic Reforms.” American Journal of Political Science 60 (2): 304–321.
· Danneman, Nathan and Ritter, Emily Hencken. 2014. “Contagious Rebellion and Preemptive Repression.”
Journal of Conflict Resolution 58 (2): 254–279.
· Quaranta, Mario and Martini, Sergio. 2016. “Does the economy really matter for satisfaction with democ-
racy? Longitudinal and cross-country evidence from the European Union.” Electoral Studies 42:164–174.
· Eagle, David. 2016. “The Negative Relationship between Size and the Probability of Weekly Attendance
in Churches in the United States.” Socius 2.
· Cao, Xun and Ward, Hugh. 2017. “Transnational Climate Governance Networks and Domestic Regula-
tory Action.” International Interactions 43 (1): 76–102.
· Helgason, Agnar Freyr and Mérola, Vittorio. 2017. “Employment Insecurity, Incumbent Partisanship,
and Voting Behavior in Comparative Perspective.” Comparative Political Studies 50 (11): 1489–1523.
· Mummolo, Jonathan and Peterson, Erik. 2018. “Improving the Interpretation of Fixed Effects Regression
Results.” Political Science Research and Methods.
Overview and applications of multilevel regression with poststratification (MRP):
· Park, David K., Gelman, Andrew, and Bafumi, Joseph. 2004. “Bayesian Multilevel Estimation with Post-
stratification: State-Level Estimates from National Polls.” Political Analysis 12 (4): 375–385.
· Lax, Jeffrey R. and Phillips, Justin H. 2009. “How Should We Estimate Public Opinion in The States?”
American Journal of Political Science 53 (1): 107–121.
· Kastellec, J. P., Lax, J. R., and Phillips, J. H. (2016). Estimating State Public Opinion With Multi-Level Re-
gression and Poststratification using R. Working paper. Available at http://www.princeton.edu/~jkastell/
MRP_primer/mrp_primer.pdf.
· Selb, Peter and Munzert, Simon. 2011. “Estimating Constituency Preferences from Sparse Survey Data
Using Auxiliary Geographic Information.” Political Analysis 19 (4): 455–470.
· Warshaw, Christopher and Rodden, Jonathan. 2012. “How Should We Measure District-Level Public
Opinion on Individual Issues?” Journal of Politics 74 (1): 203–219.
· Buttice, Matthew K. and Highton, Benjamin. 2013. “How Does Multilevel Regression and Poststratifica-
tion Perform with Conventional National Surveys?” Political Analysis 21 (4): 449–467.
13
· Toshkov, Dimiter. 2015. “Exploring the Performance of Multilevel Modeling and Poststratification with
Eurobarometer Data.” Political Analysis 23 (3): 455–460.
· Flores, Andrew R, Herman, Jody L, and Mallory, Christy. 2015. “Transgender inclusion in state non-
discrimination policies: The democratic deficit and political powerlessness.” Research & Politics 2 (4).
· Leemann, Lucas and Wasserfallen, Fabio. Forthcoming. “Extending the Use and Prediction Precision of
Subnational Public Opinion Estimation.” American Journal of Political Science.
· Broockman, David E. and Skovron, Christopher. Forthcoming. “Bias in Perceptions of Public Opinion
among Political Elites.” American Political Science Review.
· Kiewiet De Jonge, Chad P, Langer, Gary, and Sinozich, Sofi. Forthcoming. “Predicting State Presidential
Election Results Using National Tracking Polls and Multilevel Regression With Poststratification (MRP).”
Public Opinion Quarterly.
Day 18: Friday, July 20: Using Bayesian Modeling in Your Applied Work
Please read:
· Gelman, Andrew. 2008. “Objections to Bayesian Statistics.” Bayesian Analysis 3 (3): 445–450.
· Humphreys, Macartan and Jacobs, Alan M. 2015. “Mixing Methods: A Bayesian Approach.” American
Political Science Review 109, no. 4 (04): 653–673.
Also have a look at:
· Stan example models “ARM Models Sorted by Chapter” at https://github.com/stan-dev/example-models/
wiki/ARM-Models-Sorted-by-Chapter.