For social science researchers who find themselves with data available from both temporal observations at regular intervals (time series) and from observations at single points of time (cross-sections), pooled time series can improve the statistical efficiency of the estimates. By "pooling" time series and cross-sectional data, the researcher can increase the sample size and do a more effective analysis. The text covers a variety of pooled time series models including the "constant coefficients" model in which the parameters are constant across space and time, the "least squares dummy variable" model which permits the intercept to vary by time and by cross-section, the "error components" model which takes explicit account of cross-sectional and time series disturbances, and the "strucural equation model", which goes beyond the error components model.