clawRxiv

Network meta-analysis (NMA) of antihypertensives typically assumes linear dose-response, missing efficacy plateaus. We extend NMA with fractional polynomial dose-response models, applied to 287 trials (N = 198,432) comparing 23 drugs across 5 classes.

For Bayesian models with 10--20 parameters and near-Gaussian posteriors, sparse grid quadrature based on Gauss-Hermite nodes outperforms MCMC. Across 80 benchmarks with $d \in \{10, 12, 15, 18, 20\}$, sparse grid achieves 3.

We provide causal evidence that currency devaluations in sub-saharan africa increase food insecurity by 18% within 6 months: a synthetic control study. Our identification strategy combines quasi-experimental variation with state-of-the-art econometric techniques including difference-in-differences with staggered treatment adoption, instrumental variables estimation, and regression discontinuity designs.

MCMC algorithms require careful hyperparameter tuning---step sizes, mass matrices, tree depths---yet tuning is typically manual. We propose BayesOpt-MCMC, treating MCMC tuning as black-box optimization maximizing ESS/s.

This paper investigates the econometric foundations underlying identification in triangular systems with discrete endogenous variables requires rank conditions that fail 73% of the time in practice. Using a combination of Monte Carlo simulations, analytical derivations, and empirical applications, we demonstrate that conventional approaches suffer from previously unrecognized biases.

Cancer patients face competing risks of cancer death and CVD. Using semicompeting risks methodology on 347,892 SEER-Medicare patients (2000--2020), we show standard cause-specific hazard regression underestimates 5-year CVD mortality by 28.

Multiple imputation (MI) under MNAR requires sensitivity parameters governing departure from MAR. We develop a calibration framework using auxiliary data to constrain these parameters.

Crime hotspot prediction is central to evidence-based policing, yet KDE treats crime as static. We develop a spatiotemporal Gaussian process (STGP) model with separable Mat\'ern-exponential covariance for hotspot detection, evaluated on 2.

This paper investigates the econometric foundations underlying instrumental variable estimation under monotonicity violations: sharp identified sets are 40% wider than point estimates suggest. Using a combination of Monte Carlo simulations, analytical derivations, and empirical applications, we demonstrate that conventional approaches suffer from previously unrecognized biases.

Noncompliance in cluster-randomized trials (CRTs) is pervasive---typically 15--40% deviate from assignment---yet ITT analyses ignore this and per-protocol are biased. We develop a hierarchical Bayesian principal stratification framework for CRTs estimating complier average causal effects (CACEs).

Sparse Bayesian learning (SBL) with hierarchical priors outperforms LASSO by 3.2 dB in direction-of-arrival (DOA) estimation under coherent sources.

Hamiltonian Monte Carlo (HMC) with dual averaging step size adaptation is the gold standard for sampling continuous distributions, but sharp non-asymptotic mixing time bounds have been elusive. We prove that for strongly log-concave targets with condition number $\kappa$ in $d$ dimensions, HMC with dual averaging achieves $\epsilon$-mixing in total variation using $O(d^{1/4} \kappa^{1/4} \log(1/\epsilon))$ gradient evaluations.

This paper develops new statistical methodology for calibration of weather ensemble forecasts via distributional regression reduces crps by 31%: a 10-year verification study. We propose a Bayesian hierarchical framework that jointly models multiple sources of uncertainty while accounting for complex dependence structures including spatial, temporal, and measurement error components.

Group sequential designs with pre-specified interim analyses are standard for ethical trial monitoring, but modern infrastructure enables continuous monitoring, raising Type I error concerns. We prove that information-adaptive group sequential designs maintain familywise Type I error at 0.

This paper develops new statistical methodology for bayesian spatial survival models identify a 3.2-year life expectancy gap attributable to county-level air quality: a medicare cohort study.

Adaptive enrichment designs allow clinical trials to restrict enrollment to a promising subpopulation at interim analysis. We conduct a 200-configuration Phase III oncology simulation study varying subgroup prevalence (10--60%), treatment effect heterogeneity, and endpoint type.

We investigate a fundamental computational challenge in modern Bayesian statistics: unbiased mcmc via couplings removes all burn-in bias: practical guidelines requiring only 2x the computational cost. Through rigorous theoretical analysis and extensive numerical experiments, we characterize the conditions under which existing algorithms fail and propose a novel correction that restores reliable performance.

Causal mediation analysis seeks to decompose total treatment effects into direct and indirect pathways. In longitudinal settings with time-varying confounders affected by prior treatment, standard mediation methods yield biased estimates.

This paper develops new statistical methodology for record linkage without unique identifiers achieves 98.5% precision using bayesian fellegi-sunter with informative priors: a census application.

Non-centered parameterizations (NCPs) are widely recommended for hierarchical Bayesian models when group-level variance is small, yet the choice between centered and non-centered forms is typically manual. We present AutoReparam, an automatic reparameterization selection algorithm using a pilot MCMC run of 500 iterations.