A0565
Title: Advancing precision in survey estimates: A statistical framework for small area estimation using hierarchical models
Authors: Ananya Sen - Birla Institute of Technology and Science (India) [presenting]
Abstract: Large-scale household and health surveys are critical for informing policy decisions, yet their utility often diminishes at finer geographic or demographic resolutions due to limited sample sizes and increased variability. A statistical framework is presented for small area estimation (SAE) using hierarchical generalized linear models designed to produce more reliable estimates for sub-populations and small domains. The methodology integrates classical survey sampling principles with modern model-based approaches, leveraging Bayesian and empirical Bayes techniques to "borrow strength" across related areas. Covariate information from auxiliary sources is incorporated to improve estimation efficiency, while uncertainty is quantified using both posterior variance and design-based diagnostics. The framework is implemented using R and Stata, and its flexibility is demonstrated through simulations and real-world public health and demographic data examples. The contribution is to the growing intersection of survey methodology and computational statistics, offering scalable solutions for improving local-level estimation. The proposed approach holds relevance for statisticians, demographers, and policymakers seeking to enhance evidence-based decision-making under resource and data constraints.
