Selected work

Representative Papers

Spatial Dependence in the Self-Response: Spatial Dependence, Modeling, and Operational Consequences

Jun 29, 2026

This study addresses the persistent spatial autocorrelation in residuals from the U.S. Census Bureau’s Low Response Score (LRS) model after ordinary least squares regression, which undermines predictive accuracy. Leveraging 2010 data from 71,076 census tracts—including mail nonresponse rates and 25 predictors—the authors systematically compare a suite of spatial autoregressive models. They find that spatial dependence arises primarily from the error term rather than global endogenous lags. Employing queen-contiguity spatial weights, they evaluate the spatial error model (SEM/SDEM), spatial Durbin model (SDM), and spatial lag model, complemented by spatial block cross-validation for robust generalization assessment. Results indicate that the SDEM achieves optimal performance while preserving interpretability, revealing that local neighborhood demographic characteristics influence response behavior through spatial spillover effects. Findings prove robust to alternative weight specifications and heteroskedasticity.

0 citationsRead paper

Using Linked Micromaps to Explore Complex Structures in Official Statistics

Apr 30, 2026

Official statistics often exhibit complex structures across geographic and subpopulation dimensions that traditional tabular formats struggle to convey effectively, thereby hindering policymakers’ comprehension and application. This study introduces linked micromaps as a visualization framework that systematically integrates descriptive statistics, multivariate relationships, ranking structures, and spatiotemporal heterogeneity to enable intuitive exploration of high-dimensional official data. The approach substantially enhances the interpretability and readability of statistical information, uncovering latent patterns while also offering new avenues for subsequent modeling and uncertainty quantification. By doing so, it expands the potential of linked micromaps in public policy analysis and social science research.

0 citationsRead paper

A Shiny micromapST App

Apr 14, 2026

This study addresses the limitations of traditional linked micromaps—namely, cumbersome data preparation and a lack of intuitive interactivity—that hinder efficient exploration of geostatistical data. To overcome these challenges, the authors introduce, for the first time, a graphical user interface (GUI) for the micromapST package built within the R Shiny framework. This interactive application streamlines data input and visualization workflows, enabling users to rapidly generate statistical graphics such as scatterplots, boxplots, and time series through an accessible visual interface. Crucially, these charts are dynamically linked to arbitrary geographic regions, facilitating coordinated exploration. Validation through real-world case studies demonstrates that the tool substantially lowers the technical barrier to entry and enhances both the efficiency and intuitiveness of geostatistical analysis.

0 citationsRead paper

Self-Tuned Rejection Sampling within Gibbs and a Case Study in Small Area Estimation

Sep 21, 2025

In Gibbs sampling, drawing from nonstandard univariate conditional distributions—lacking closed-form samplers and resisting efficient proposal construction for rejection sampling—remains challenging. Method: This paper proposes the Adaptive Vertical Weighted Strip (VWS) method, which constructs a finite mixture proposal distribution and dynamically optimizes component weights online to achieve low rejection rates with substantially reduced computational overhead. Contribution/Results: The key innovation lies in embedding VWS within the Gibbs framework, enabling iterative pruning of ineffective components and self-tuning of the proposal structure. Applied to small-area estimation, VWS enables efficient and accurate Bayesian inference for the posterior distribution of school-age children living in poverty at the county level. It thus enhances both the feasibility and practicality of Gibbs sampling for large-scale, complex hierarchical models.

0 citationsRead paper

SafeTab-P: Disclosure Avoidance for the 2020 Census Detailed Demographic and Housing Characteristics File A (Detailed DHC-A)

May 02, 2025

Fine-grained race/ethnicity statistics in the 2020 U.S. Census Detailed Demographic and Housing Characteristics File A (DHC-A) pose significant privacy risks due to re-identification vulnerabilities. Method: We propose the first differentially private framework for releasing multi-level geographic and cross-tabulated group statistics, featuring an adaptive granularity control mechanism that dynamically adjusts the number of statistics and the resolution of geographic and categorical dimensions based on group size, coupled with discrete Gaussian noise injection under zero-concentrated differential privacy (zCDP). Contribution/Results: Implemented and deployed via Tumult Analytics within the official census release pipeline, our approach achieves strong formal privacy guarantees (ε = 0.48 zCDP) while substantially improving statistical utility over prior methods. Empirically tuned and budget-validated, it balances practical deployability with regulatory compliance, establishing a scalable paradigm for privacy-preserving large-scale official statistics.

0 citationsRead paper

Recent publications

Latest Papers

Spatial Dependence in the Self-Response: Spatial Dependence, Modeling, and Operational Consequences

Jun 29, 2026

This study addresses the persistent spatial autocorrelation in residuals from the U.S. Census Bureau’s Low Response Score (LRS) model after ordinary least squares regression, which undermines predictive accuracy. Leveraging 2010 data from 71,076 census tracts—including mail nonresponse rates and 25 predictors—the authors systematically compare a suite of spatial autoregressive models. They find that spatial dependence arises primarily from the error term rather than global endogenous lags. Employing queen-contiguity spatial weights, they evaluate the spatial error model (SEM/SDEM), spatial Durbin model (SDM), and spatial lag model, complemented by spatial block cross-validation for robust generalization assessment. Results indicate that the SDEM achieves optimal performance while preserving interpretability, revealing that local neighborhood demographic characteristics influence response behavior through spatial spillover effects. Findings prove robust to alternative weight specifications and heteroskedasticity.

0 citationsRead paper

Using Linked Micromaps to Explore Complex Structures in Official Statistics

Apr 30, 2026

Official statistics often exhibit complex structures across geographic and subpopulation dimensions that traditional tabular formats struggle to convey effectively, thereby hindering policymakers’ comprehension and application. This study introduces linked micromaps as a visualization framework that systematically integrates descriptive statistics, multivariate relationships, ranking structures, and spatiotemporal heterogeneity to enable intuitive exploration of high-dimensional official data. The approach substantially enhances the interpretability and readability of statistical information, uncovering latent patterns while also offering new avenues for subsequent modeling and uncertainty quantification. By doing so, it expands the potential of linked micromaps in public policy analysis and social science research.

0 citationsRead paper

A Shiny micromapST App

Apr 14, 2026

This study addresses the limitations of traditional linked micromaps—namely, cumbersome data preparation and a lack of intuitive interactivity—that hinder efficient exploration of geostatistical data. To overcome these challenges, the authors introduce, for the first time, a graphical user interface (GUI) for the micromapST package built within the R Shiny framework. This interactive application streamlines data input and visualization workflows, enabling users to rapidly generate statistical graphics such as scatterplots, boxplots, and time series through an accessible visual interface. Crucially, these charts are dynamically linked to arbitrary geographic regions, facilitating coordinated exploration. Validation through real-world case studies demonstrates that the tool substantially lowers the technical barrier to entry and enhances both the efficiency and intuitiveness of geostatistical analysis.

0 citationsRead paper

Self-Tuned Rejection Sampling within Gibbs and a Case Study in Small Area Estimation

Sep 21, 2025

In Gibbs sampling, drawing from nonstandard univariate conditional distributions—lacking closed-form samplers and resisting efficient proposal construction for rejection sampling—remains challenging. Method: This paper proposes the Adaptive Vertical Weighted Strip (VWS) method, which constructs a finite mixture proposal distribution and dynamically optimizes component weights online to achieve low rejection rates with substantially reduced computational overhead. Contribution/Results: The key innovation lies in embedding VWS within the Gibbs framework, enabling iterative pruning of ineffective components and self-tuning of the proposal structure. Applied to small-area estimation, VWS enables efficient and accurate Bayesian inference for the posterior distribution of school-age children living in poverty at the county level. It thus enhances both the feasibility and practicality of Gibbs sampling for large-scale, complex hierarchical models.

0 citationsRead paper

SafeTab-P: Disclosure Avoidance for the 2020 Census Detailed Demographic and Housing Characteristics File A (Detailed DHC-A)

May 02, 2025

Fine-grained race/ethnicity statistics in the 2020 U.S. Census Detailed Demographic and Housing Characteristics File A (DHC-A) pose significant privacy risks due to re-identification vulnerabilities. Method: We propose the first differentially private framework for releasing multi-level geographic and cross-tabulated group statistics, featuring an adaptive granularity control mechanism that dynamically adjusts the number of statistics and the resolution of geographic and categorical dimensions based on group size, coupled with discrete Gaussian noise injection under zero-concentrated differential privacy (zCDP). Contribution/Results: Implemented and deployed via Tumult Analytics within the official census release pipeline, our approach achieves strong formal privacy guarantees (ε = 0.48 zCDP) while substantially improving statistical utility over prior methods. Empirically tuned and budget-validated, it balances practical deployability with regulatory compliance, establishing a scalable paradigm for privacy-preserving large-scale official statistics.

0 citationsRead paper