M-quantile models with application to poverty mapping
RIS ID
26201
Abstract
Over the last decade there has been growing demand for estimates ofpopulation characteristics at small area level. Unfortunately, cost constraints in thedesign of sample surveys lead to small sample sizes within these areas and as a resultdirect estimation, using only the survey data, is inappropriate since it yields estimateswith unacceptable levels of precision. Small area models are designed to tackle thesmall sample size problem. The most popular class ofmodels for small area estimationis random effects models that include random area effects to account for between areavariations. However, such models also depend on strong distributional assumptions,require a formal specification of the random part of the model and do not easily allowfor outlier robust inference. An alternative approach to small area estimation thatis based on the use of M-quantile models was recently proposed by Chambers andTzavidis (Biometrika 93(2):255-268, 2006) and Tzavidis and Chambers (Robust predictionof small area means and distributions.Working paper, 2007).Unlike traditionalrandom effects models, M-quantile models do not depend on strong distributional
Publication Details
Tzavidis, N., Salvati, N., Pratesi, M. & Chambers, R. L. (2008). M-quantile models with application to poverty mapping. Statistical Methods and Applications, 17 (3), 393-411.