Journal of the Royal Statistical Society: Series A

High recidivism risk found for 20% of ever-jailed. 10% cycle back to jail up to twice per year. Little association to age, gender, crime type and race.

Most recidivism studies focus on prisons and occurrence in a discrete framework. Little is known about jail recurrence & time-to-recidivism. Barone and Farcomeni use novel latent class multi-state quantile regression with cure fraction methods on >550,000 US (2020–2023) jail records

Latent class multi-state quantile regression with a cure fraction: application to jail recidivism in the U.S. Abstract. We propose a multi-state quantile regression model that admits a cure-fraction for each possible transition, so that individuals may not experien

🆕Latent class multi-state quantile regression with a cure fraction: application to jail recidivism in the USA

💡@barross993.bsky.social and
@afarcome.bsky.social
analyze time spent after release with a quantile regression approach
doi.org/10.1093/jrss...

They also illustrate the application of the proposed tree-based MI method using data from a cellphone survey on COVID-19 vaccination in Uganda, which represents a subcohort sample drawn from the 2020 Uganda Population-based HIV Impact Assessment Survey.

In this paper, authors propose a Bayesian tree-based multiple imputation (MI) approach for estimating population means using the Phase II sample, where the parent survey was conducted using a complex survey design, and with simulations they test the advantages of the approach

🆕Improving survey inference in two-phase designs using Bayesian machine learning

🪶Wang, Chen, et al with a very clear purpose in the title of their paper

doi.org/10.1093/jrss...

GitHub - yuzhouguangc/MTGCL Contribute to yuzhouguangc/MTGCL development by creating an account on GitHub.

Empirical evaluations demonstrate that MTGCL outperforms existing graph contrastive learning models in classification accuracy across multiple time periods while maintaining competitive computational efficiency.

Data and Code are here : github.com/yuzhouguangc...

Fraud detection in blockchain networks presents unique challenges due to decentralized and pseudonymous nature of transactions. This study introduces a novel Multilayer Topology-Aware Graph Contrastive Learning (MTGCL) framework to detect fraudulent activity within the Ethereum transaction network

Multilayer topology-aware graph contrastive learning for fraud detection in the Ethereum transaction network Abstract. Fraud detection in blockchain networks presents unique challenges due to the decentralized and pseudonymous nature of transactions. This study in

🆕Multilayer topology-aware graph contrastive learning for fraud detection in the Ethereum transaction network

💱Chen et al delve in Blockchain networks

doi.org/10.1093/jrss...

Their results suggest that overlooking latent homophily can lead to either underestimation or overestimation of causal peer influence, accompanied by considerable estimation uncertainty.

In this paper, authors address this challenge by leveraging latent locations inferred from the network to disentangle homophily from causal peer influence, and extend this approach to multiple networks by adopting a Bayesian hierarchical modelling framework

Researchers have focused on understanding how an individual’s behaviour is influenced by their peers behaviours. Identifying causal peer influence, is challenging due to confounding by homophily (people tend to connect with those who share similar characteristics)

🆕A Bayesian approach to estimate causal peer influence accounting for latent network homophily

💡Um , Sweet and Adhikari show a framework to disentangle homophily from causal peer influence

doi.org/10.1093/jrss...

Bernadette: Bayesian Inference and Model Selection for Stochastic Epidemics Bayesian analysis for stochastic extensions of non-linear dynamic systems using advanced computational algorithms. Described in Bouranis, L., Demiris, N., Kalogeropoulos, K., and Ntzoufras, I. (2022) ...

Bonus: Model estimation can be carried out with the R package Bernadette, available on CRAN

(cran.r-project.org/web/packages...)

The temporal evolution of transmission rates in populations containing multiple types of individual is reconstructed via an appropriate dimension-reduction formulation driven by independent diffusion processes.

🆕Bayesian analysis of diffusion-driven multi-type epidemic models with application to COVID-19
💡Bouranis et al. develop a flexible Bayesian evidence synthesis approach to model age-specific transmission dynamics of COVID-19 based on daily death counts

Authors propose practical guidelines, and present the performance of the proposed estimators in numerical studies in two sets of real data: exit polls from the 19th South Korean election and public data collected from the Korean Survey of Household Finances and Living Conditions

When survey non-response isn't random but depends on the unobserved answer itself, standard methods give biased results. Previous solutions required hard-to-find "instrumental variables" that researchers can't easily identify beforehand. 📊

Identifying enhanced generalized linear model estimation with nonignorable missing outcomes Abstract. Missing data often result in undesirable bias and loss of efficiency. These issues become substantial when the response mechanism is nonignorable

🆕Identifying enhanced generalized linear model estimation with nonignorable missing outcomes

💡Beppu, Choi, Morikawa & Im try to respond what happens when people don't respond to surveys because of their actual (unobserved) answer?

doi.org/10.1093/jrss...

They use a unique form of resampling for valid estimates of our test statistic's null distribution even under violations of standard assumptions. This GeoRDD procedure gives substantially different results in the analysis of NYC arrest rates than those that rely on standard assumptions.

They study variation in policing outcomes attributable to differential policing practices in NYC using geographic regression discontinuity designs (GeoRDDs).

Robust inference for geographic regression discontinuity designs: assessing the impact of police precincts Abstract. We study variation in policing outcomes attributable to differential policing practices in New York City (NYC) using geographic regression discon

🆕Robust inference for geographic regression discontinuity designs: assessing the impact of police precincts
💡Kendall et al analyze how smaller, sub-municipal boundaries like police districts, precincts, and service areas also influence police outcomes
doi.org/10.1093/jrss...

Overall, men, previous SNAP participants, under-50-year-olds, and those with previous non-urgent, primary care treatable ED visits drive the average effect of Medicaid on ED use

Who increases ED use after Medicaid? New causal ML methods reveal effect a small share of Oregon Medicaid experiment recipients drive overall ED use increase, masking wide variation.

Who increases emergency department use? New insights from the Oregon health insurance experiment Abstract. We provide new insights regarding the headline result that Medicaid increased emergency department (ED) use from the Oregon experiment. Using cau

Who increases emergency department (ED) use? New insights from the Oregon health insurance experiment
💡Austin Denteh and Helge Liebert
estimate the heterogeneous impacts of Medicaid on ED use and characterize them
doi.org/10.1093/jrss...

Bonus: The processed county-level weekly maximum ozone data and the R code for changepoint detection and copula-GEV model fitting can be found at github.com/mintaek0764/...

When accounting for changepoints, 27.56% of counties had their trend estimates change by >0.03 ppm. While overall ozone levels dropped (thanks to air quality policies), extreme ozone trends actually increased in 45.82% of counties after adjusting for data disruptions.

The authors created a sophisticated statistical model that handles "long memory" patterns in ozone data (where past values influence future ones for extended periods) and changepoints in the data. They used genetic algorithms to detect when these disruptions occurred.

How to accurately measure long-term ozone trends when data keeps getting disrupted? Changes in air quality policies, monitor locations, instruments, and sampling methods create "changepoints" that can make trend analysis misleading if you don't account for them properly.

Long-term trends of US county-level extreme ozone concentrations with long memory and changepoint considerations Abstract. We quantify long-term linear trends in weekly maximum ground-level ozone concentrations of 395 US counties by considering two critical issues: ch

🆕Long-term trends of US county-level extreme ozone concentrations with long memory and changepoint considerations
🌬️Lee and Lee develop a long-memory extreme series model using a copula transformation to quantify long-term trends of extreme ozone
doi.org/10.1093/jrss...