Time Series Analysis of Political Change

Abstract

This chapter considers the use of time-series analysis in the study of politics. Time-series analysis, as traditionally understood, requires a moderate number of observations on a single unit at approximately equal intervals. Data of this kind are rarely available at the individual level, so this chapter will focus on the macro level. Quantitative analysis of individual change is usually based on panels rather than time series (see Hsiao 1986 for a comprehensive discussion). Examples of time series relevant to the study of politics include election results, the number of protests, ratings of governments in terms of various qualities, and averages of individual opinions as measured by surveys. The chapter will focus on dependent variables that can be treated as at least approximately continuous, although it will give some attention to issues involving small counts. The analysis of single events calls for different techniques, which are discussed in Chap. 32.