Monte Carlo

This paper develops and provides software for a class of spatial hierarchical (multilevel) models for binary outcomes, motivated by situations where outcome dependence arises both from diffusion among lower-level units and from spatially structured unobserved group effects. The authors propose two model variants—one that treats unobserved group effects as spatially independent and another that treats them as spatially dependent—and use a Bayesian estimation approach. They evaluate performance with Monte Carlo simulations and illustrate the approach by analyzing causes of civil rights protests in the United States in the 1960s.

Across the social sciences, pooling effects over long periods can produce faulty inferences when the underlying data generating process is dynamic. This work introduces a permutation-based, two-stage changepoint procedure designed for time series cross-sectional data to give researchers a principled way to detect when effect sizes truly change versus when they reflect routine stochastic fluctuation.

A new theoretical approach models voters on a lattice where districts are formed by partitioning that lattice. This framework makes the spatial distribution of voters explicit and supports construction of districts subject to population and connectivity constraints.

Information about the ideological positions of political actors is central to questions of representation, polarization, and voting behavior. Surveys that ask respondents to place actors on a common ideological scale are common, but respondents often introduce systematic biases into those placements.

The spatial weights matrix W sits at the heart of spatial regression models, yet there are few techniques for validating a chosen specification of W. Reframing misspecification of W as a measurement error reveals a distinctive form of endogeneity that matters for spatial inference.

Political and economic distributions vary depending on the political subdivision used for analysis—nations, subnational regions, urban versus rural areas, or electoral districts. The article identifies problems that arise when measuring and comparing geographic distributions across these different units, emphasizing how the choice of aggregation can alter observed values.

Unobserved heterogeneity—systematic differences across actors that are not measured—poses a persistent threat to studies of social processes. This threat is especially acute for statistical models of networks, where complex dependencies in tie formation combine with restrictive modeling assumptions to magnify bias in inference. The paper introduces a frailty extension to the exponential random graph model (FERGM) to explicitly account for such unobserved heterogeneity.

Monte Carlo experiments vary list-experiment design choices and introduce small amounts of non-strategic respondent error to evaluate two estimators: the item count technique maximum likelihood estimator (ICT-MLE) and the simple difference-in-means. The simulations focus on conditions created by common best practices that reduce strategic misrepresentation and thus produce very few respondents at the extrema (choosing no items or all items).

This article sorts remaining disagreements from a symposium on time-series political modeling and highlights practical issues that deserve clearer explanation for applied researchers.

Most multidimensional spatial models specify the systematic component of voter utility as additively separable. That assumption implies voters do not care how positions on multiple policy dimensions combine—an implication that is too restrictive in the context of mass elections and can mischaracterize vote choice.

A method to estimate spatial-lag duration models when the outcome is subject to right censoring, the most common form of censoring in duration data. This fills a gap because existing spatial duration models do not account for censoring in the likelihood function.

Latent class analysis (LCA) is widely used in political science for substantive applications and to estimate measurement error. A common "three-step" practice relates estimated class assignments from an LCA to external variables while effectively ignoring classification error. Vermunt (2010, Latent class modeling with covariates: Two improved three-step approaches, Political Analysis 18:450–69) demonstrated that this omission produces inconsistent parameter estimates and proposed a bias correction that is now implemented in standard software. Inconsistency, however, is not the only problem.