Bayesian Learning about Ideal Points of U.S. Supreme Court Justices, 1953-1999 (PDF; or gzipped postscript) by Andrew Martin and Kevin Quinn. In this manuscript we employ Markov chain Monte Carlo (MCMC) methods to fit Bayesian measurement models of judicial preferences for all justices serving on the U.S. Supreme Court from 1953 to 1999. We are particularly interested in considering to what extent ideal points of justices change throughout their tenure on the Court, and how the proposals over which they are voting also change across time. To do so, we fit four longitudinal item response models that include dynamic specifications for the ideal points and the case-specific parameters. Our results suggest that justices do not have constant ideal points, even after controlling for the types of cases that come before the Court. A version of this paper presented at the 2001 Political Methodology Summer Meeting was awarded the 2001 Gosnell Prize for best paper in political methodology presented at a conference in 2000-2001.
The Dimensions of Supreme Court Decision Making: Again Revisiting The Judicial Mind (PDF; or gzipped postscript by Andrew Martin and Kevin Quinn. In this manuscript we employ Markov chain Monte Carlo (MCMC) methods to fit Bayesian measurement models of judicial preferences for the justices of the seventh Burger Court (1981-1985). In so doing we simultaneously estimate an ideal point for each justice in a multi-dimensional space and the cutting hyperplane for each non-unanimous case. In general, estimating these models for the Supreme Court is quite difficult due to the small number of justices. The models we propose are thus developed to address aspects of this 'micro-committee problem'. The Bayesian approach allows us to include additional information about the cases being studied in a straightforward manner, and also allows us to gauge the uncertainty of each estimate. We present results for several models, compare the estimated bliss points across models, and utilize the findings to study two cases of substantive importance: Garcie v. San Antonia Metropolitan Transit Authority (469 US 528) and Dixson v. United States (465 US 482).
Visualizing Multivariate Outliers and Leverage Points (PDF; or gzipped postscript by Kevin Quinn. Data encountered by social scientists are notoriously ``dirty'' in the sense that they often contain numerous outliers and bad leverage points that can distort results obtained from a naive statistical analysis. When working with a simple bivariate dataset, one of the most effective means of identifying aberrant observations is to examine the data with a simple scatterplot. However, when (as is usually the case) one is interested in a multivariate dataset direct visualization becomes much more difficult. Low-dimensional summaries of the multivariate dataset might seem to hold some promise except that most of these methods are not robust to outliers themselves. This is an example of the well-known masking problem which prevents such summaries from successfully identifying outlying observations. In this paper I review recent advances in the field of robust statistics and data visualization and demonstrate how such methods can be used to identify and visualize multivariate outliers and leverage points. Multivariate data visualization offers advantages over purely statistical methods of identifying outliers. Visualization allows one to see the relationship between aberrant observations and the remainder of the data, and allows one to determine if outliers tend to cluster in regions of the data space that might be explained by unmeasured variables or background knowledge. These methods are applied to the Alvarez, Garrett, and Lange (1991) dataset on economic performance and labor organization in 16 OECD countries. I show that the AGL data contain numerous outliers and bad leverage points. These outliers and leverage points exhibit some temporal dependence and cluster in the years 1974, 1975, and 1982. One possible explanation of these outlying values is has to do with the large shocks to the world economy brought on by the OPEC oil embargo of 1973 and the steep rise in oil prices in the early 1980s resulting from the Iran-Iraq war. Reanalysis of the cleaned data shows no relationship between leftist cabinets, labor organization, and economic performance.
An Integrated Computational Model of Multiparty Electoral Competition (PDF; or gzipped postscript by Andrew Martin and Kevin Quinn. Most theoretic models of multiparty electoral competition make the assumption that party leaders are motivated to maximize their voteshare or seatshare. In plurality rule systems this is a sensible assumption. However, in proportional representation systems, this assumption is questionable since the ability to make public policy is not strictly increasing in voteshares or seatshares. We present a theoretic model in which party leaders choose electoral declarations with an eye toward the expected policy outcome of the coalition bargaining game induced by the party declarations and the parties' beliefs about citizens' voting behavior. To test this model, we turn to data from the 1989 Dutch Parliamentary Election. We use Markov chain Monte Carlo methods to estimate the parties' beliefs about mass voting behavior, and to average over measurement uncertainty and missing data. Due to the complexity of the parties' objective functions and the uncertainty in objective function estimates, equilibria are found numerically. Unlike previous models of multiparty electoral competition, the equilibrium results are consistent with the empirical declarations of the four major Dutch parties.