Applied Regularization Methods for the Social Sciences

Social scientists often face data sets with a modest number of cases but a large number of variables, a setting known as high-dimensional data. This book addresses that problem through regularization, presenting a range of models and showing how each one works on real analytic problems.

Each chapter is devoted to one application of regularization. The method is first explained in accessible terms, and then extended worked examples demonstrate it in use. The approaches cover univariate and multivariate models common in social science research, and the examples are carried out in R, SAS, and SPSS. Both frequentist and Bayesian versions of the techniques are included.

All datasets and software scripts used in the examples are available on an accompanying website, so readers can rerun the analyses themselves. Each chapter also ends with application exercises, which instructors can assign in class and independent researchers can use to practice the material.