Title: Combining national and constituency polling for forecasting

Authors: Chris Hanretty, Benjamin E Lauderdale, Nick Vivyan

Status: Published online, Electoral Studies

Abstract: We describe a method for forecasting British general elections by combining national and constituency polling. We reconcile national and constituency estimates through a new swing model.

Download Paper

Journal Website

Title: Dyadic representation in a Westminster system

Authors: Chris Hanretty, Benjamin E Lauderdale and Nick Vivyan

Status: Forthcoming, Legislative Studies Quarterly

Abstract: Is policy representation in contemporary Westminster systems solely a function of programmatic national parties, or does the election of legislators via single member districts result in MPs whose policy positions are individually responsive to public opinion in their constituencies? We generate new measures of constituency opinion in Britain and show that, in three different policy domains and controlling for MP party, the observed legislative behaviour of MPs is indeed responsive to constituency opinion. The level of responsiveness is moderate, but our results do suggest a constituency-MP policy bond which operates in addition to the well-known bond between voters and parties.

Download Paper

Title: Simulating Counterfactual Representation

Authors: Andrew C Eggers and Benjamin E Lauderdale

Status: Forthcoming, Political Analysis

Abstract: We show how to use multilevel modeling and post-stratification to estimate legislative outcomes under counterfactual representation schemes that e.g. boost the representation of women or translate votes into seats differently. We apply this technique to two research questions: (1) Would the U.S. Congress be less polarized if state delegations were formed according to the principle of party proportional representation? (2) Would there have been stronger support for legalizing same-sex marriage in the U.K. House of Commons if Parliament more closely reflected the population in gender and age?

Download Paper

Title: Measuring Political Positions from Legislative Speech

Authors: Benjamin Lauderdale and Alexander Herzog

Status: Published online, Political Analysis

Abstract: Existing approaches to measuring political disagreement from text data perform poorly except when applied to narrowly selected texts discussing the same issue and written in the same style. We demonstrate the first viable approach to scaling the entire speech corpus of a legislature, producing valid legislator-specific scores as well as extensive information about the evolution of speech polarization and politically loaded language. In the Irish Dail, we show that the dominant dimension of speech variation is government-opposition, with ministers more extreme on this dimension than backbenchers. In the US Senate, we estimate a dimension that has moderate within-party correlations with scales based on roll-call votes and campaign donation patterns, however we observe greater overlap across parties in speech positions than roll-call positions and partisan polarization in speeches varies more clearly in response to major political events.

Download Paper

Download Appendix

Title: Crowd-Sourced Text Analysis: Reproducible and Agile Production of Political Data

Authors: Kenneth Benoit, Drew Conway, Benjamin E. Lauderdale, Michael Laver and Slava Mikhaylov

Status: Forthcoming, American Political Science Review

Abstract: Empirical social science often relies on data that are not observed in the field, but are transformed into quantitative variables by expert researchers who analyze and interpret qualitative raw sources. While generally considered the most valid way to produce data, this expert-driven process is inherently difficult to replicate or to assess on grounds of reliability. Using crowd-sourcing to distribute text for reading and interpretation by massive numbers of non-experts, we generate results comparable to those using experts to read and interpret the same texts, but do so far more quickly and flexibly. Crucially, the data we collect can be reproduced and extended transparently, making crowd-sourced datasets intrinsically reproducible. This focuses researchers’ attention on the fundamental scientific objective of specifying reliable and replicable methods for collecting the data needed, rather than on the content of any particular dataset. We also show that our approach works straightforwardly with different types of political text, written in different languages. While findings reported here concern text analysis, they have far-reaching implications for expert-generated data in the social sciences.

Download Paper

Title: Comparing Strategies for Estimating Constituency Opinion from National Survey Samples

Authors: Chris Hanretty, Benjamin E. Lauderdale, and Nick Vivyan

Status: Conditionally accepted, Political Science Research and Methods

Abstract: Political scientists interested in estimating how public opinion varies by constituency have developed several strategies for supplementing limited constituency survey data with additional sources of information. We present two evaluation studies in the previously unexamined context of British constituency-level opinion: an external validation study of party vote share in the 2010 general election and a cross-validation of opinion toward the European Union. We find that most of the gains over direct estimation come from the inclusion of constituency-level predictors, which are also the easiest source of additional information to incorporate. Individual-level predictors combined with post-stratification particularly improve estimates from unrepresentative samples, and geographic local smoothing can compensate for weak constituency-level predictors. We argue that these findings are likely to be representative of applications of these methods where the number of constituencies is large.

Journal Website

Title: Partisan Disagreements Arising from Rationalization of Common Information

Authors: Benjamin E. Lauderdale

Status: Published online, Political Science Research and Methods

Abstract: Why do opposing partisans sometimes disagree about the facts and processes that are relevant to understanding political issues? One explanation is that citizens may have a psychological tendency toward adopting beliefs about the political world that rationalize their partisan preferences. Previous quantitative evidence for rationalization playing a role in explaining partisan factual disagreement has come from cross-sectional covariation and from correction experiments. In this paper, I argue that these rationalizations can occur as side effects when citizens change their attitudes in response to partisan cues and substantively relevant facts about a political issue. Following this logic, I motivate and report the results of a survey experiment that provides US Republicans and Democrats with information that they will be inclined to rationalize in different ways, because they have different beliefs about which political actors they should agree with. The results are a novel experimental demonstration that partisan disagreements about the political world can arise from rationalization.

Journal Website

Download Pre-Publication Version

Download Appendix

Title: Who Controls Opinion Content? Testing Theories of Authorship using Case-Specific Preference Estimates for the US Supreme Court
Authors: Benjamin E. Lauderdale and Tom S. Clark
Status: Forthcoming, Journal of Politics

Abstract: Recent research has demonstrated that the preferences of US Supreme Court justices are not simply unidimensional. We demonstrate a new approach to Bayesian preference estimation that estimates case-specific preferences for justices, using a conditional autoregressive model with legal similarity determining the correlation between justices’ preferences across cases. We employ the overlap in legal topics addressed in each case to identify the most relevant precedent cases to describe variation in revealed preferences across areas of the law. In applications that test theories of bargaining on the Court, these estimates enable stronger identification of variation in preferences while holding the composition of the Court constant. We show that Chief Justices from 1946 to 2005 strategically assign authorship to their colleagues in cases where those colleagues are more closely in agreement with the Chief Justice, and patterns of other justices joining those opinions are consistent with the idea that authors shape the content of opinions.

Download Paper

Title: Under-performing, Over-performing, or Just Performing? The Limitations of Fundamentals-based Presidential Election Forecasting

Authors: Benjamin E. Lauderdale and Drew A. Linzer

Status: International Journal of Forecasting (2015)

Abstract: U.S. presidential election forecasts are of widespread interest to political commentators, campaign strategists, and the public. We argue that most fundamentals-based political science forecasts overstate what historical political and economic factors can tell us about the likely outcome of a forthcoming presidential election. Existing approaches generally overlook uncertainty in coefficient estimates, decisions about model specification, and the translation from popular vote shares to Electoral College outcomes. We introduce a novel Bayesian forecasting model for state-level presidential elections that accounts for each of these sources of error, and allows for the inclusion of structural predictors at both the national and state levels. Applying the model to presidential election data from 1952 to 2012, we demonstrate that, for covariates with typical amounts of predictive power, the 95% prediction intervals for presidential vote shares should span approximately +/- 10% at the state level and +/- 7% at the national level.

Download Pre-Publication Version

Journal Website

Title: A Partisan Gap in the Supply of Female Potential Candidates

Authors: Melody Crowder-Meyer and Benjamin E. Lauderdale

Status: Research & Politics, 1:1 (2014)

Abstract: A partisan disparity in women representatives in the US House emerged in the 1980s and has continued to grow in magnitude. We show that this pattern closely mirrors the emergence of a partisan disparity in the fraction of women in the US public with the typical characteristics of high-level officeholders. Our analysis indicates that the fraction female in the Democratic pool of potential candidates is now two to three times larger than in the Republican pool of potential candidates. Given the current association of party identification with gender and other characteristics, this gap is more likely to increase than decrease over the coming decade, with potential consequences for the descriptive and substantive representation of women in American politics.

Download Paper

Title: Scaling Meaningful Political Dimensions Using Texts and Votes

Authors: Benjamin E. Lauderdale and Tom S. Clark

Status: American Journal of Political Science, 58:3, 754-771 (2014)

Abstract: Item response theory models for roll-call voting data provide political scientists with parsimonious descriptions of political actors’ relative preferences. However, models using only voting data tend to obscure variation in preferences across different issues due to identification and labeling problems that arise in multidimensional scaling models. We propose a new approach to using sources of metadata about votes to estimate the degree to which those votes are about common issues. We demonstrate our approach using votes and opinions from the U.S. Supreme Court, using Latent Dirichlet Allocation to discover the extent to which different issues were at stake in different cases and estimating justice preferences within each of those issues. This approach can be applied using a variety of unsupervised and supervised topic models for text, community detection models for networks, or any other tool capable of generating discrete or mixture categorization of subject matter from relevant vote-specific data.

Download Paper

Title: Does Inattention to Political Debate Explain the Polarization Gap Between the U.S. Congress and Public?

Authors: Benjamin E. Lauderdale

Status: Public Opinion Quarterly, 77:S, 2-23 (2013)

Abstract: Recent studies of the U.S. Congress have demonstrated a substantial difference in partisan polarization between legislators’ votes and citizens’ survey responses about those votes. But perhaps public polarization would increase if citizens were more attentive to political debates in Congress? Using matching techniques on natural variation in citizens’ political information levels, I show that citizens who are informed about the partisan alignment of issues have a similar preference distribution to Congress, even after the former are re-weighted to resemble the entire public along salient political, social, and demographic dimensions. In contrast, using a survey experiment, I show that cue and argument treatments only partially reduce the discrepancy between the views expressed by the public and the voting behavior of Congress on the same issues. Both experimental and observational studies have significant limitations for measuring counterfactuals involving public opinion, and so our understanding of the polarization gap remains unfortunately limited.

Download Paper

Title: The Supreme Court’s Many Median Justices

Authors: Benjamin E. Lauderdale and Tom S. Clark

Status: American Political Science Review, 106:4, 847-866 (2012).

Abstract: Theoretical and empirical research on the US Supreme Court has become increasingly channeled through a unidimensional spatial model of judicial decision-making. We argue that limitations in quantitative measurement of justices’ preferences have obscured the ways in which justices’ preferences vary across areas of the law and constrained the set of questions the research can pursue. We introduce a new approach, using multiple indices of substantive similarity among cases and a kernel-weighted optimal classification estimator to recover estimates of judicial preferences that are localized to particular legal issues as well as periods of time. Allowing preference variation across legal areas significantly improves the predictive power of estimated preference orderings versus a model that only allows for variation in preferences over time. We find judicial preferences are not reducible to simple left-right ideology and, as a consequence, there is substantial variation in the identity of the median justice across areas of the law during all periods of the modern court. These results suggest a need to reconsider empirical and theoretical research that hinges on the existence of a unitary and well-identified median justice.

Download Paper

Title: The Genealogy of Law

Authors: Tom S. Clark and Benjamin E. Lauderdale

Journal: Political Analysis, 20:3, 329-350 (2012).

Abstract: Many theories of judicial politics have at their core the concepts of legal significance, doctrinal development and evolution, and the dynamics of precedent. Despite rigorous theoretical conceptualization, these concepts remain empirically elusive. We propose the use of a genealogical model (or “family tree”) to describe the Court’s construction of precedent over time. We describe statistical assumptions that allow us to estimate this kind of structure using an original data set of citation counts between Supreme Court majority opinions. The genealogical model of doctrinal development provides a parsimonious description of the dependencies between opinions, while generating measures of legal significance and other related quantities. We employ these measures to evaluate the robustness of a recent finding concerning the relationship between ideological homogeneity within majority coalitions and the legal impact of Court decisions.

Download Paper

Title: Compound Poisson-Gamma Regression Models for Dollar Outcomes That Are Sometimes Zero

Authors: Benjamin E. Lauderdale

Journal: Political Analysis, 20:3, 387-399 (2012)

Abstract: Political scientists often study dollar-denominated outcomes that are zero for some observations. These zeros can arise because the data-generating process is granular: the observed outcome results from aggregation of a small number of discrete projects or grants, each of varying dollar size. This article describes the use of a compound distribution in which each observed outcome is the sum of a poisson distributed number of gamma distributed quantities, a special case of the Tweedie distribution. Regression models based on this distribution estimate log-linear marginal effects without either the ad hoc treatment of zeros necessary to use a log-dependent variable regression or the change in quantity of interest necessary to use a tobit or selection model. The compound poisson-gamma regression is compared with commonly applied approaches in an application to data on high-speed rail grants from the U.S. federal government to the states, and against simulated data from several data-generating processes.

Download Paper

Note: The MLE code to fit the model as described in the paper is available on request. However, I recommend estimating a Bayesian version of the model by MCMC using R and JAGS (or WinBUGS with slight modifications), as it is substantially faster.

Download R and JAGS Script Files and Example Data

Title: Distinguishing Between Influences on Brazilian Legislative Behavior

Authors: Cesar Zucco Jr. and Benjamin E. Lauderdale

Journal: Legislative Studies Quarterly, 36:3, 363-396 (2011).

Abstract: Ideal point estimators hold the promise of identifying multiple dimensions of political disagreement as they are manifested in legislative voting. However, standard ideal point estimates do not distinguish between ideological motivations and voting inducements from parties, coalitions, or the executive. In this article we describe a general approach for hierarchically identifying an ideological dimension using an auxiliary source of data. In the case we consider, we use an anonymous survey of Brazilian legislators to identify party positions on a left-right ideology dimension. We then use this data to distinguish ideological motivations from other determinants of roll-call behavior for eight presidential-legislative periods covering more than 20 years of Brazilian politics. We find that there exists an important nonideological government-opposition dimension, with the entrance and exit of political parties from the governing coalition appearing as distinct shifts in ideal point on this second dimension. We conjecture that the Brazilian president’s control over politically important resources is the source of this dimension of conflict, which has recently become far more important in explaining roll-call voting than the ideological dimension.

Download Paper

Title: Locating Supreme Court Opinions in Doctrine Space

Authors: Tom S. Clark and Benjamin E. Lauderdale

Journal: American Journal of Political Science, 54:4, 871-890 (2010).

Abstract: We develop a scaling model to estimate U.S. Supreme Court opinion locations and justice ideal points along a common spatial dimension using data derived from the citations between opinions. Citations from new opinions to precedent opinions usually apply and endorse the doctrine of the precedent opinion; however, sometimes they implicitly or explicitly dispute the precedent opinion. We collect original datasets classifying citations from search and seizure and freedom of religion opinions written between 1953 and 2006 into these different types and develop a model relating the similarity of the doctrine embodied in the citing and cited opinions to the relative probability of these different types of citations. The resulting spatial estimates of opinion location are used to evaluate theories of Supreme Court bargaining and opinion writing. We find empirical support for theoretical models that predict the majority opinion will fall at the ideal point of the median member of the majority coalition. Given the centrality of theories of judicial policymaking to various substantive problems in political science, the method of scaling opinions developed in this article can facilitate a range of future research.

Download Paper

Title: Unpredictable Voters in Ideal Point Estimation

Authors: Benjamin E. Lauderdale

Journal: Political Analysis 18:2, 151-171 (2010).

Abstract: Ideal point estimators are typically based on an assumption that all legislators are equally responsive to modeled dimensions of legislative disagreement; however, particularistic constituency interests and idiosyncrasies of individual legislators introduce variation in the degree to which legislators cast votes predictably. I introduce a Bayesian heteroskedastic ideal point estimator and demonstrate by Monte Carlo simulation that it outperforms standard homoskedastic estimators at recovering the relative positions of legislators. In addition to providing a refinement of ideal point estimates, the heteroskedastic estimator recovers legislator-specific error variance parameters that describe the extent to which each legislator’s voting behavior is not conditioned on the primary axes of disagreement in the legislature. Through applications to the roll call histories of the U.S. Congress, the E.U. Parliament, and the U.N. General Assembly, I demonstrate how to use the heteroskedastic estimator to study substantive questions related to legislative incentives for low-dimensional voting behavior as well as diagnose unmodeled dimensions and nonconstant ideal points.

Download Paper

“If you like political theory, statistical analysis, complicated equations that look like the set dressing on “Good Will Hunting”, and the facile utilization of words like “homoskedasticity,” then this paper is basically your “Eat Pray Love”.” – Jason Linkins, The Huffington Post, May 6, 2010.

Research Note: John McCain is No Longer a Maverick

Note: the MCMCpack code discussed in the paper for fitting the model no longer works. Fortunately, improvements in computer performance have largely obviated the need for custom MCMC code for this class of models. For an example script using JAGS, download this combined R + JAGS script.

Title: Pass the Pork: Measuring Legislator Shares in Congress

Authors: Benjamin E. Lauderdale

Journal: Political Analysis 16:3, 235-249 (2008).

Abstract: Linear regression models are frequently used to analyze distributive politics in the U.S. Congress; however, authors have used a variety of specifications with different implicit assumptions about how bicameralism shapes legislative bargaining. I derive a model that describes district or state spending authorizations as the aggregation of spending secured by multiple legislators working on behalf of overlapping constituencies. This bicameral shares model allows the disaggregation of House and Senate influence through simultaneous estimation of the relative bargaining power of the two chambers and the advantages that accrue to legislators holding partisan, committee, and other relevant affiliations. In the 2005 transportation bill, the model better predicts the functional form of small state advantage than recently employed specifications in the literature.

Download Paper