Monte Carlo Simulation and Resampling Methods for Social Science

Taking the topics of a quantitative methodology course and illustrating them through Monte Carlo simulation, this book examines abstract principles, such as bias, efficiency, and measures of uncertainty in an intuitive, visual way. Instead of thinking in the abstract about what would happen to a particular estimator "in repeated samples," the book uses simulation to actually create those repeated samples and summarize the results. The book includes basic examples appropriate for readers learning the material for the first time, as well as more advanced examples that a researcher might use to evaluate an estimator he or she was using in an actual research project. The book also covers a wide range of topics related to Monte Carlo simulation, such as resampling methods, simulations of substantive theory, simulation of quantities of interest (QI) from model results, and cross-validation. Complete R code from all examples is provided so readers can replicate every analysis presented using R.

Table of Contents:
1. Introduction 2. Probability 3. Introduction to R 4. Random Number Generation 5 .Statistical Simulation of the Linear Model 6. Simulating Generalized Linear Models 7. Testing Theory Using Simulation 8. Resampling Methods 9. Other Simulation-Based Methods 10. Final Thoughts

About the Author :
Thomas M. Carsey was the Thomas J. Pearsall Distinguished Professor of Political Science and Director of the Odum Institute for Research in Social Science at the University of North Carolina at Chapel Hill. His research interests revolved around representation in American politics and quantitative methods. Within American politics, Carsey’s work focused on state politics, campaigns and elections, public opinion and mass behavior, partisanship and party polarization, and legislative politics. His methodological interests included all aspects of computational social science with specific interests in Monte Carlo simulation, resampling methods, clustered and pooled data, and methods for contextual analysis. Carsey’s research was funded by several grants from the National Science Foundation, and he published articles in journals such as American Political Science Review, American Journal of Political Science, Journal of Politics, State Politics & Policy Quarterly, and many others. Jeffrey J. Harden is an assistant professor in the Department of Political Science at the University of Colorado, Boulder specializing in political methodology and American politics. He received his PhD in political science from the University of North Carolina at Chapel Hill. His methodology interests include model selection, robust regression methods, multilevel data, and the use of Monte Carlo simulation to better understand issues that arise in applied analysis. His research agenda in American politics focuses on political representation, mass/elite linkages, and state politics. Harden has published articles in Political Analysis, Sociological Methods & Research, Legislative Studies Quarterly, State Politics & Policy Quarterly, and Public Choice.

Review :
There is no text like this that is geared toward a social science market. [The] writing is direct and to the point... I can’t underemphasize that part. Too many methods books try to soften the technical edge by throwing in lots of commentary. Bradley Efron discussed the newly-invented bootstrap and other computationally intensive statistical techniques in a 1979 article entitled "Computers and the Theory of Statistics: Thinking the Unthinkable." But as computer power grew exponentially and software for simulation greatly improved, what was once unthinkable has become routine. Carsey and Harden have performed a service by making modern tools for random simulation and resampling methods (like the bootstrap) accessible to a broad readership in the social sciences, developing these methods from first principles, and showing how they can be applied both to understand statistical ideas and in practical data analysis. Statistical simulation has become an essential tool of modern statistics and data analysis—useful for evaluating estimators, calculating features of probability distributions, transforming difficult-to-interpret statistical results into meaningful quantities of interest, and even helping with alternative theories of inference. Simulation perspectives also offer a terrific way to learn many aspects of statistical modeling. Join Tom Carsey and Jeff Harden for a clearly written and deeply practical book on this crucial topic. Your scholarly work will be better for it. Carsey and Harden have written an intuitive and practical primer to a radical—but increasingly widely used—approach to statistical inference: Monte Carlo and resampling. They focus on using these techniques to evaluate more standard statistical approaches, but in the process, they convey their broader use and importance. They also teach the reader about statistical inference at a much more basic level than do most social science treatments of empirical methods. Their book is destined to be used widely in graduate social science statistics classes around the world. Monte Carlo simulation and resampling are the workhorse of modern methods. Carsey and Harden provide the perfect, accessible guide to learn this  fundamental, must-have skill for social scientists.