ABSTRACT
This chapter discusses the ways to work with panel data, that is, how to add a temporal dimension into linear regressions. It shows how to estimate models with fixed effects and random effects, how to diagnose unit roots, how to create variables with lags or leads, and how to calculate panel-corrected standard errors. One way of approaching the abstract concept of “political proximity” in international relations is through the convergence of votes in the United Nations General Assembly. There is an outstanding tool called unvotes in R for all of those who study the voting history of countries in the United Nations Assembly. In political science, it is common to encounter the distinction between panel and Time Series Cross-Sectional data. In political science, both specifications for modeling variation between groups in panel data are models of fixed or random effects. Many political-data time series exhibit trending behavior or non-stationarity behavior in the mean.
