Sunday, Aug 3: 4:00 PM - 5:50 PM
4023
Contributed Papers
Music City Center
Room: CC-105B
This session explores cutting-edge statistical methods for analyzing disparities, crime, and social inequities in the justice system and beyond. Talks will cover diverse topics, including racial bias in prosecutorial decisions and policing, forensic error rates, measurement error in mortgage discrimination, and the statistical modeling of cartel violence and food insecurity. By leveraging nonparametric methods, causal inference, and advanced modeling techniques, these studies provide new insights into systemic inequities and the role of data in shaping policy and reform.
Main Sponsor
Social Statistics Section
Co Sponsors
Caucus for Women in Statistics
History of Statistics Interest Group
Presentations
Analyzing crime events is crucial to understand crime dynamics and it is largely helpful for constructing prevention policies. Point processes specified on linear networks can provide a more accurate description of crime incidents by considering the geometry of the city. We propose a spatio-temporal Dirichlet process mixture model on a linear network to analyze crime events in Valencia, Spain. We propose a Bayesian hierarchical model with a Dirichlet process prior to automatically detect space-time clusters of the events and adopt a convolution kernel estimator to account for the network structure in the city. From the fitted model, we provide crime hotspot visualizations that can inform social interventions to prevent crime incidents. Furthermore, we study the relationships between the detected cluster centers and the city's amenities, which provides an intuitive explanation of criminal contagion.
Keywords
crime data
Dirichlet process
linear network
Markov chain Monte Carlo
spatio-temporal point processes
Racial disparities in cognitive health reflect entrenched structural inequalities. The Mortgage Density Index Ratio (MDIR) quantifies census-tract level housing and lending discrimination, but it may be unstable in hypersegregated areas. To address this, we developed a joint modeling approach that simultaneously estimates cognitive outcomes and latent mortgage rates for Black and White households. In simulations, joint modeling showed notably lower bias and greater robustness in small- to moderate- sized census tracts compared to traditional regression approaches. Applying joint modeling to six cognitive domains in Michigan Cognitive Aging Project (MCAP) data (N = 644), we identified a significant association between MDIR and processing speed only among Black participants, with a one-unit MDIR increase (i.e., greater racial parity in mortgage lending) corresponding to a 0.48 SD improvement in processing speed (95% CI: 0.05-0.93). Traditional regression failed to detect this effect. These findings underscore the importance of advanced statistical methods in quantifying structural racism and highlight the disproportionate effects of mortgage discrimination in Black adults.
Keywords
Measurement error
Joint modeling
Hypersegregation
Health disparities
Studying racial bias in policing is a critically important problem, but one that involves inherent difficulties due to the nature of available data. In this manuscript, we tackle multiple key issues in the causal analysis of racial bias in policing. First, we formalize race and place policing, the idea that individuals of one race are policed differently in neighborhoods primarily made up of individuals of other races. We develop an estimand to study this rigorously, outline assumptions necessary for causal identification, and conduct sensitivity analyses to assess robustness to key assumption violations. Additionally, we investigate difficulties with existing estimands targeting racial bias in policing. We show that estimation for these estimands, and those developed in this manuscript, benefits from incorporating mobility data into analyses. We apply these ideas to a study in New York City, finding substantial racial bias, race and place policing, and robustness to large violations of untestable assumptions. We additionally show that mobility data can make substantial impacts on the resulting estimates, suggesting it should be used whenever possible in subsequent studies.
Keywords
Causal inference
Mobility data
Racial discrimination
Race and place
Sensitivity analysis
Peremptory strikes allow attorneys to remove jurors without providing a reason, raising concerns about racial bias despite rulings like Batson v. Kentucky (1986) that prohibit race-based exclusions. Using data from the Fifth Circuit Court of Mississippi, we flexibly estimate racial disparities in peremptory strikes by the state and find that Black jurors are struck at a 37% higher rate than white jurors, even when accounting for characteristics like gender, attitudes toward punishment, and trial-level factors such as crime type and defendant race. To explore variability in these conditional differences in strike rates, we employ flexible nonparametric methods, including DR-learners, and assess the importance of specific covariates using variable importance measures. We assess the robustness of our conclusion to unmeasured or partially measured covariates through various sensitivity models. This work introduces a nonparametric, flexible, and robust framework for quantifying racial disparities in jury selection, contributing to the broader goal of ensuring equity in legal proceedings.
Keywords
nonparametric statistics
racial disparity
peremptory challenge
jury selection
sensitivity analysis
Cartel violence in Mexico remains a critical issue, affecting millions. Understanding its escalation is key for public policy and law enforcement. This study applies the Univariate Hawkes Process to analyze the self-excitation of cartel conflicts, revealing how violent events increase the likelihood of future incidents. Using UCDP data (1989 - 2021), we identify key dyads, with Jalisco-Sinaloa accounting for 20.06% of all conflicts. The half-life of cartel violence varies, from 1.89 days (Government vs. Civilians) to 59.75 days (Los Ardillos vs. Los Rojos).
To enhance accuracy, we implement a Hierarchical Hawkes Process, estimating global parameters: μ ≈ 0 (low background violence), α= 0.98 (high self-excitation), and β = 0.03 (moderate decay). This model improves predictive accuracy by capturing global patterns in violence escalation. We plan to compare the results from the Hawkes Process against the Poisson Point Process, which assumes that violent events occur independently over time, without any triggering effect. Future work will refine this comparison, integrate a Multivariate Hawkes Model, and assess external factors such as law enforcement actions and economic conditions.
Keywords
Cartel Conflicts
Univariate Hawkes Process
Hierarchical Hawkes Process
Jalisco-Sinaloa
Self-excitation
Poisson Point Process