2nd Edition
Analyzing Spatial Models of Choice and Judgment
With recent advances in computing power and the widespread availability of preference, perception and choice data, such as public opinion surveys and legislative voting, the empirical estimation of spatial models using scaling and ideal point estimation methods has never been more accessible.The second edition of Analyzing Spatial Models of Choice and Judgment demonstrates how to estimate and interpret spatial models with a variety of methods using the open-source programming language R.
Requiring only basic knowledge of R, the book enables social science researchers to apply the methods to their own data. Also suitable for experienced methodologists, it presents the latest methods for modeling the distances between points.
The authors explain the basic theory behind empirical spatial models, then illustrate the estimation technique behind implementing each method, exploring the advantages and limitations while providing visualizations to understand the results.
This second edition updates and expands the methods and software discussed in the first edition, including new coverage of methods for ordinal data and anchoring vignettes in surveys, as well as an entire chapter dedicated to Bayesian methods. The second edition is made easier to use by the inclusion of an R package, which provides all data and functions used in the book.
David A. Armstrong II is Canada Research Chair in Political Methodology and Associate Professor of Political Science at Western University. His research interests include measurement, Democracy and state repressive action.
Ryan Bakker is Reader in Comparative Politics at the University of Essex. His research interests include applied Bayesian modeling, measurement, Western European politics, and EU politics.
Royce Carroll is Professor in Comparative Politics at the University of Essex. His research focuses on measurement of ideology and the comparative politics of legislatures and political parties.
Christopher Hare is Assistant Professor in Political Science at the University of California, Davis. His research focuses on ideology and voting behavior in US politics, political polarization, and measurement.
Keith T. Poole is Philip H. Alston Jr. Distinguished Professor of Political Science at the University of Georgia. His research interests include methodology, US political-economic history, economic growth and entrepreneurship.
Howard Rosenthal is Professor of Politics at NYU and Roger Williams Straus Professor of Social Sciences, Emeritus, at Princeton. Rosenthal’s research focuses on political economy, American politics and methodology.
1. Introduction
The Spatial Theory of Voting
Theoretical Development and Applications of the Spatial Voting Model
The Development of Empirical Estimation Methods for Spatial Models of Voting
The Basic Space Theory
Summary of Data Types Analyzed by Spatial Voting Models
Conclusion
2. Analyzing Issue Scales
Aldrich-McKelvey Scaling
The basicspace Package in R
Example : European Election Study (French Module)
Example : American National Election Study Urban Unrest and Vietnam War Scales
Estimating Bootstrapped Standard Errors for Aldrich- McKelvey Scaling
Basic Space Scaling: The blackbox Function
Example : Convention Delegate Study
Example : Swedish Parliamentary Candidate Survey
Estimating Bootstrapped Standard Errors for Black Box Scaling
Basic Space Scaling: The blackbox transpose Function
Example : and Comparative Study of Electoral Systems (Mexican Modules)
Estimating Bootstrapped Standard Errors for Black Box Transpose Scaling
Using the blackbox transpose Function on Datasets
Ordered Optimal Classi_cation
Using Anchoring Vignettes
Conclusion
Exercises
3. Analyzing Similarities and Dissimilarities Data
Classical Metric Multidimensional Scaling
Example : Nations Similarities Data
Metric MDS Using Numerical Optimization
Metric MDS Using Majorization (SMACOF)
The smacof Package in R
Non-Metric Multidimensional Scaling
Example : Nations Similarities Data
Example : th US Senate Agreement Scores
Individual Di_erences Multidimensional Scaling
Example : European Election Study (French Module)
Conclusion
Exercises
4. Unfolding Analysis of Rating Scale Data
Solving the Thermometers Problem
Metric Unfolding Using the MLSMU Procedure
Example : Interest Group Ratings of US Senators Data
Metric Unfolding Using Majorization (SMACOF)
Example : European Election Study (Danish Module)
Comparing the MLSMU and SMACOF Metric Unfolding Procedures
Conclusion
Exercises
5. Unfolding Analysis of Binary Choice Data
The Geometry of Legislative Voting
Reading Legislative Roll Call Data into R with the pscl Package
Parametric Methods - NOMINATE
Obtaining Uncertainty Estimates with the Parametric Bootstrap
Types of NOMINATE Scores
Accessing DW-NOMINATE Scores
The wnominate Package in R
Example : The th US House
Example : The First European Parliament (Using the Parametric Bootstrap)
Nonparametric Methods - Optimal Classi_cation
The oc Package in R
Example : The French National Assembly during the Fourth Republic
Example : American National Election Study Feeling Thermometers Data
Conclusion: Comparing Methods for the Analysis of Legislative Roll Call Data
Identi_cation of the Model Parameters
Comparing Ideal Point Estimates for the th US Senate
Exercises
6. Bayesian Scaling Models
Bayesian Aldrich-McKelvey Scaling
Comparing Aldrich-McKelvey Standard Errors
Bayesian Multidimensional Scaling
Example : Nations Similarities Data
Bayesian Multidimensional Unfolding
Example : American National Election Study Feeling Thermometers Data
Parametric Methods - Bayesian Item Response Theory
The MCMCpack and pscl Packages in R
Example : The Term of the US Supreme Court (Unidimensional IRT)
Running Multiple Markov Chains in MCMCpack and pscl
Example : The Con_rmation Vote of Robert Bork to the US Supreme Court (Unidimensional IRT)
Example : The th US Senate (Multidimensional IRT)
Identi_cation of the Model Parameters
MCMC or a-NOMINATE
The anominate Package in R
Ordinal and Dynamic IRT Models
IRT with Ordinal Choice Data
Dynamic IRT
EM IRT
Conclusion
Exercises
Biography
Dave Armstrong (http://quantoid.net) is Canada Research Chair in Political Methodology and Associate Professor of Political Science at Western University in Ontario, Canada. He received a Ph.D. in Government and Politics from the University of Maryland in 2009 and was a post-doctoral fellow in the Department of Politics and Nuffield College at the University of Oxford. His research interests revolve around measurement and the relationship between Democracy and state repressive action. His research has been published in the American Political Science Review, the American Journal of Political Science, the American Sociological Review and the R Journal among others. Dave is an active R user and maintainer of a number of packages. DAMisc has a number of functions that ease interpretation and presentation of GLMs.
Ryan Bakker is Reader in Comparative Politics at the University of Essex. He received his Ph.D. in Political Science from the University of North Carolina at Chapel Hill in 2007. His research and teaching interests include applied Bayesian modeling, measurement, Western European politics, and EU elections and political parties. He is a principal investigator for the Chapel Hill Expert Survey (CHES), which measures political party positions on a variety of policy-specific issues in the European Union. His work has appeared in Political Analysis, Electoral Studies, European Union Politics, and Party Politics.
Royce Carroll is Professor in Comparative Politics at the University of Essex, where he teaches graduate and undergraduate courses on comparative politics and American politics. He received his Ph.D. in Political Science at the University of California, San Diego in 2007. In addition to political methodology, his research focuses on comparative politics of legislatures, coalitions and political parties, as well as measurement of ideology. Carroll is also Director of the Essex Summer School in Social Science Data Analysis.
Keith T. Poole is Philip H. Alston Jr. Distinguished Professor, Department of Political Science, University of Georgia. He received his Ph.D. in Political Science from the University of Rochester in 1978. His research interests include methodology, political-economic history of American institutions, economic growth and entrepreneurship, and the political-economic history of railroads. He is the author or coauthor of over 50 articles as well as the author of multiple books. He was a Fellow of the Center for Advanced Study in Behavioral Sciences 2003-2004 and was elected to the American Academy of Arts and Sciences in 2006.
Howard Rosenthal is Professor of Politics at NYU and Roger Williams Straus Professor of Social Sciences, Emeritus, at Princeton. Rosenthal's coauthored books include Political Bubbles: Financial Crises and the Failure of American Democracy, Polarized America: The Dance of Ideology and Unequal Riches, Ideology and Congress, and Prediction Analysis of Cross Classifications. He has coedited "What Do We Owe Each Other?" and "Credit Markets for the Poor." Rosenthal is a member of the American Academy of Arts and Sciences. He has been a Fellow of the Center for Advanced Study in Behavioral Sciences and a Visiting Scholar at the Russell Sage Foundation.
"This book will have broad appeal across the social sciences, but especially in political science and psychology. An obvious audience is scholars doing work in attitudinal scaling, or psychometrics. However, the application of spatial models of the sort addressed in this text is certainly not limited to survey data or other types of data for which people are the units of analysis. These methods can be used to assess and describe the structure of relationships between variables or units wherever such relationships can be conceptualized as distances in some abstract space. I expect that this book will be used mostly as a reference guide, but only because courses in spatial models of this sort are (unfortunately) fairly limited. However, more advanced courses in multivariate analysis, latent variable modeling, dimensional analysis, and measurement across the social sciences would likely find this text extremely useful. (Adam Enders, University of Louisville)
"This book provides excellent coverage of spatial models of choice and judgment…Overall the manuscript is technically correct and clearly written. The biggest strength of the book is the deliberately informal and applied nature of the approach of the book, where both code and output are shown. This makes it very easy for researchers to quickly get these models running on their own data quickly." (James Lo, USC)
"I find the manuscript technically sound, clearly written, and at an appropriate level of difficulty for quantitative social scientists. It has several strengths. First, it is a comprehensive and up-to-date survey of spatial models for scaling preferential and perceptual data (including dyadic data measuring similarities/distances). Second, it is replete with interesting examples from political science, which greatly increases the readability of the material. Third, by including many chunks of R code for data analysis and visualization, it greatly reduces barriers to implementing these methods for practitioners." (Xiang Zhou, Harvard University)