We present ensemble Bayesian model averaging (EBMA) and illustrate its ability to aid scholars in the social sciences to make more accurate forecasts of future events. In essence, EBMA improves prediction by pooling information from multiple forecast models to generate ensemble predictions similar to a weighted average of component forecasts. The weight assigned to each forecast is calibrated via its performance in some validation period. The aim is not to choose some “best” model, but rather to incorporate the insights and knowledge implicit in various forecasting efforts via statistical postprocessing. After presenting the method, we show that EBMA increases the accuracy of out-of-sample forecasts relative to component models in three applied examples: predicting the occurrence of insurgencies around the Pacific Rim, forecasting vote shares in U.S. presidential elections, and predicting the votes of U.S. Supreme Court Justices.

Multivariate count models are rare in political science despite the presence of many count time series. This article develops a new Bayesian Poisson vector autoregression model that can characterize endogenous dynamic counts with no restrictions on the contemporaneous correlations. Impulse responses, decomposition of the forecast errors, and dynamic multiplier methods for the effects of exogenous covariate shocks are illustrated for the model. Two full illustrations of the model, its interpretations, and results are presented. The first example is a dynamic model that reanalyzes the patterns and predictors of superpower rivalry events. The second example applies the model to analyze the dynamics of transnational terrorist targeting decisions between 1968 and 2008. The latter example's results have direct implications for contemporary policy about terrorists' targeting that are both novel and innovative in the study of terrorism.

Identifying causal effects attributable to network membership is a key challenge in empirical studies of social networks. In this article, we examine the consequences of endogeneity for inferences about the effects of networks on network members' behavior. Using the House office lottery (in which newly elected members select their office spaces in a randomly chosen order) as an instrumental variable to estimate the causal impact of legislative networks on roll call behavior and cosponsorship decisions in the 105th–112th Houses, we find no evidence that office proximity affects patterns of legislative behavior. These results contrast with decades of congressional scholarship and recent empirical studies. Our analysis demonstrates the importance of accounting for selection processes and omitted variables in estimating the causal impact of networks.

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.

We examine the trade-offs associated with using Amazon.com's Mechanical Turk (MTurk) interface for subject recruitment. We first describe MTurk and its promise as a vehicle for performing low-cost and easy-to-field experiments. We then assess the internal and external validity of experiments performed using MTurk, employing a framework that can be used to evaluate other subject pools. We first investigate the characteristics of samples drawn from the MTurk population. We show that respondents recruited in this manner are often more representative of the U.S. population than in-person convenience samples—the modal sample in published experimental political science—but less representative than subjects in Internet-based panels or national probability samples. Finally, we replicate important published experimental work using MTurk samples.

Recent research finds that, for social desirability reasons, uninformed individuals disproportionately give “neither agree nor disagree” type responses to survey attitude questions, even when a “do not know” option is available. Such “face-saving” responses inflate the indifference (i.e., middle) categories of ordered attitude variables with nonordered responses. When such inflation occurs within the middle category of one's ordered dependent variable, estimates from ordered probit (and ordered logit) models are likely to be unreliable and inefficient. This article develops a set of mixture models that estimate and account for the presence of “face-saving” responses in middle categories of ordered survey response variables, and applies these models to (1) simulated data and (2) a commonly studied survey question measuring attitudes toward European Union (EU) membership among individuals in EU-candidate countries. Results from the survey data set and the Monte Carlo experiments suggest that, when middle category inflation is present in one's ordered dependent variable, the estimates obtained from middle category mixture models are less biased than—and in some cases substantively distinct from—the estimates obtained from “naive” ordered probit models.

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 loglinear 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 United States federal government to the states, and against simulated data from several data-generating processes.

The relationship between a party's popular vote share and legislative seat share—its seats—votes swing ratio—is a key characteristic of democratic representation. This article introduces a general approach to estimating party-specific swing ratios in multiparty legislative elections, given results from only a single election. I estimate the joint density of party vote shares across districts using a finite mixture model for compositional data and then computationally evaluate this distribution to produce parties' expected change in legislative seats for plausible changes in their vote share. The method easily extends to systems with any number of parties, employing both majoritarian and proportional electoral rules. Applications to legislative elections in the United States, United Kingdom, Canada, and Botswana demonstrate how parties' swing ratios vary both within countries and over time, indicating that parties under majoritarian electoral rules are subject to unique and possibly divergent geographic—political constraints.