Analyzing Longitudinal and Panel Data

Abstract

6 Data often comes from observations made at multiple points in time. Obtaining repeated observations on the same units allows the researcher to access a richer information set about observed units than would be possible with single observations and to map the evolution of the phenomenon over multiple periods for both individual units and overall as a trend. (For example, relationships between two variables may strengthen, weaken, or even disappear over time.) Longitudinal data can be gathered via survey instruments or archival databases that offer repeated measures on the same variables at different times. Longitudinal data differ from cross-sectional data in two important ways: There are multiple observations for the same units, and there is a time dimension that can be exploited because the phenomenon is observed at different points in time. Therefore, data is nested across two dimensions: units and time. As a result, not all observations are independent of each other, thus violating one of the assumptions of multivariate data analysis. So, how do we deal with longitudinal data? What kind of tests and tools are available to exploit such richness? This chapter examines the key features of longitudinal and panel data analysis by comparing it with the classic (OLS) regression models discussed in Chap. 3. The chapter provides guidance on data structures with multiple observations for the same units (N) at different points in time (T), on discerning between instances in which panel-data estimations are preferred over pooled-OLS estimations, on running fixed-effects or random-effects models, and on interpreting and reporting estimations from these models. Because these topics are complex, we discuss the choice of the appropriate method toward the end of the chapter.