PS 2740: Time-Series-Cross-Section Data Analysis

Course Description

The use of time-series cross-section (TSCS) data in political science has grown steadily over the last forty years. TSCS data contain information on cross-sectional units (indexed by n=1,2...N) over time (indexed by t=1,2...T). The typical TSCS dataset in political science has a large T and aggregate spatially-organized units (e.g., counties, states, countries). These features distinguish TSCS analysis from micro-behavioral panel analysis.

Time series and spatial econometric methods are critical for drawing valid statistical inferences from samples of TSCS data and are useful for understanding how outcomes respond dynamically to stimuli and diffuse geographically. This course covers methods for the analysis of temporal and spatial relations in TSCS data. The primary focus is on models for continuous dependent variables observed at discrete points in time and space.

Parts I and II of the course begin with the fundamentals of time series and spatial analysis, using pure time series and cross-sectional data respectively to introduce the workhorse autoregressive models from time-series and spatial econometrics. Students will learn how to estimate these models and calculate their implied dynamic and diffusive effects. Following this introduction, in Part III, participants are taught to integrate the different analytical frameworks using spatio-temporal models. In Part IV, a set of important topical extensions—parameter heterogeneity, binary outcomes, networks, and missing data—are covered.

Requirements and Grading: Students will complete homework assignments and write the analysis section of either a methods or substantive paper using time series and spatial econometrics. Homework is due one week after assigned; the analysis write-up will be due at the end of the semester. The homework (cumulatively) and analysis write-up will each comprise 50% of your final grade. I expect students to write their course assignments in R Markdown.

Required Texts:

• Enders, Walter. 2010. Applied Econometric Time Series. Third Edition. New York: Wiley.
• LeSage, James and R. Kelley Pace. 2009. Introduction to Spatial Econometrics. New York: Chapman and Hall/CRC.

Supplemental Texts:

• Box-Steffensmeier, Janet M., John R. Freeman, Matthew P. Hitt, Jon C. W. Pevehouse. 2014. Time Series Analysis for the Social Sciences. Third Edition. New York: Wiley.
• Pickup, Mark. 2014. Introduction to Time Series Analysis. Beverly Hills, CA: Sage Publications.
• Ward, Michael D., and Kristian Skrede Gleditsch. Spatial Regression Models. Second Edition. Sage, 2018.
• Brandt, Patrick and John Williams. 2006. Multiple Time Series Models. Beverly Hills, CA: Sage Publications.
• Elhorst, J. Paul. 2014. Spatial Econometrics: From Cross-sectional Data to Spatial Panels. Berlin: Springer.
• Shumway, Robert H. and David S. Stoffer. 2011. Time Series Analysis and Its Applications: With R Examples. Third Edition. New York: Springer.
• Pesaran, Hashem. 2015. Time Series and Panel Data Econometrics. Oxford: Oxford University Press.

Online Resources:

• Forecasting: Principles and Practice
• Time-Series Cross-Validation
• R Package, "ARDL"
• R Package, "Forecast"
• R Package, "Orcutt"
• R Package, "Sandwich"
• R Package, "Spatialreg"