Bayesian Sample Size Determination for Planning Hierarchical Bayes Small Area Estimates


Bayesian methods (1)

Author: Peter Dutey-Magni (University of Southampton/University College London)

This paper devises a fully Bayesian sample size determination method for hierarchical model-based small area estimation with a decision risk approach. A new loss function specified around a desired maximum posterior variance target implements conventional official statistics criteria of estimator reliability (coefficient of variation of up to 20 per cent). This approach comes with an efficient binary search algorithm identifying the minimum effective sample size needed to produce small area estimates under this threshold constraint. Traditional survey sampling design tools can then be used to plan appropriate data collection using the resulting effective sample size target. This approach is illustrated in a case study on small area prevalence of life limiting health problems for 6 age groups across 1,956 small areas in Northern England, using the recently developed Integrated Nested Laplace Approximation method for a unit-level spatial generalised linear mixed hierarchical models.

Keywords: Sample size determination, small area estimation, disease prevalence models, hierarchical Bayes, Integrated Nested Laplace Approximation, official statistics, national statistics.

Only logged in users can see slides when author's permission was given. Please register to have access.