![IconGems-Print](images/IconGems-Print.png)
292 – Small Area Estimation with Small Samples
Choice of Small Area Models Based on Sample Designs and Availability of Auxiliary Data in PIAAC Study
Jianzhu Li
Westat
Tom Krenzke
Westat
Leyla Mohadjer
Westat
Wendy VanDeKerckhove
Westat
Lin Li
BioStat Solutions
Weijia Ren
Westat
Henok Adbaru
Westat
The Programme for the International Assessment of Adult Competencies (PIAAC) is an international survey conducted in about 40 countries to measure adult proficiency in key information-processing skills. Although the survey was designed to produce nationally representative estimates with adequate precision, most countries are also interested in estimates for local areas where sample size is very small. The application of small area estimation (SAE) approaches provides an affordable option. SAE methods are a set of model-dependent approaches that employ a statistical model using auxiliary information and the survey data to produce indirect estimates when survey data alone are inadequate for direct estimation. The PIAAC participating countries adopted various sample designs from single-stage stratified sample from a registry to multiple-stage area sample involving stratification and clustering. This paper discusses the impact of different design features and sources of auxiliary data on the choice of appropriate small area models to estimate the proportion of adults lacking basic literacy skills in local areas. We used a few countries to demonstrate how the decisions are made.