Recently, empirical research in economics has been enriched by the availability of a wealth of new sources of data: cross sections of individuals observed over time. These allow us to construct and test more realistic behavioral models that could not be identified using only a cross section or a single time series data set. Nevertheless, the availability of new data sources raises new issues. New methods are constantly being introduced, and points of view are changing. An author preparing an introductory monograph has to select the topics to be included. My selection involves controlling for unobserved individual and/or time characteristics to avoid specification bias and to improve the efficiency of the estimates. The more basic and more commonly used methods are treated here, although to some extent the coverage is a matter of taste. Some examples of applications of the methods are also given, and the uses, computational approaches, and interpretations are discussed.
I am very much indebted to C. Manski and to a reader for Cambridge University Press, as well as to G. Chamberlain and J. Ham, for helpful comments and suggestions. I am also grateful to Mario Tello Pacheco, who read through the manuscript and made numerous suggestions concerning matters of exposition and corrections of errors of every magnitude. My appreciation also goes to V. Bencivenga, A.C. Cameron, T. Crawley, A. Deaton, E. Kuh, B. Ma, D. McFadden, D. Mountain, G. Solon, G. Taylor, and K.Y. Tsui, for helpful comments, and Sophia Knapik and Jennifer Johnson, who patiently typed and retyped innumerable drafts and revisions.