3B Dynamic Models for Social Scientists
Harold D. Clarke, University of Texas at Dallas & Guy D. Whitten, University of Texas A&M
5 - 16 August (two week course / 35 hrs)
This is an applied course which focuses on statistical methods for conducting dynamic analyses of economic, political and social data. A variety of important models are considered including ARFIMA, Fractional Error Correction, GARCH and Dynamic Conditional Correlations, Duration Models, Markov Switching Models, Dynamic Panel Models for Time Series Cross-Sectional (TSCS) Data, VAR and Vector Error Correction. Special attention is given to specifying and analyzing State Space models of the latent dynamics of processes of interest. Both frequentist and Bayesian approaches to model specification, analysis, and interpretation are employed. The course will provide working knowledge of how to use Stata, R and Winbugs to analyze various dynamic models. Students are invited to bring their own data sets for analyses in daily lab sessions.
The course will benefit anyone who is interested in conducting multivariate dynamic analyses of economic, political, and social processes from frequentist or Bayesian perspectives. The aim is to teach course participants how to undertake and evaluate sophisticated dynamic analyses of economic, political, and social data. The methods considered will be helpful to graduate students and faculty in the social sciences, as well as researchers working in the public and private sectors.
Participants should be familiar with applied multiple regression analysis and with the standard Windows computing environment. Basic knowledge of the Stata, R and Winbugs programs is helpful but not required.
Commandeur, Jacques and Siem Jan Koopman. 2007. An Introduction to State Space Time Series Analysis. Oxford: Oxford University Press.
Ntzoufras, Ioannis. 2009. Bayesian Modeling Using Winbugs. New York: Wiley.
Plaff, Bernhard. 2006. Analysis of Integrated and Cointegrated Time Series With R. New York: Springer.