16 Panel Data
A panel follows the same units, such as workers, firms, regions or countries, over several periods. Its great advantage over a single cross-section is that it can control for characteristics of the units that are not observed but do not change over time: ability, management quality, geography, culture. Such characteristics are the commonest source of omitted-variable bias in cross-section work, and the panel removes them by comparing each unit with itself. This section develops the two main models, fixed and random effects, the tests that choose between them, standard errors that allow for correlation within units, and the special problems of dynamic panels.
16.1 Unobserved heterogeneity
16.2 Fixed effects
16.3 Random effects
16.4 Choosing between fixed and random effects
16.5 Cluster-robust standard errors
16.6 Dynamic panels
16.7 Panel data in software
16.8 Practice Problems
16.2 Fixed effects
16.3 Random effects
16.4 Choosing between fixed and random effects
16.5 Cluster-robust standard errors
16.6 Dynamic panels
16.7 Panel data in software
16.8 Practice Problems
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.