ESTIMATION FOR DOMAINS PARAMETERS IN DOUBLE SAMPLING FOR STRATIFICATION WITH NON-LINEAR COST FUNCTION IN THE PRESENCE OF NON RESPONSE

Abstract

In sample surveys separate estimates of a parameter maybe required for sub-populations into which a population is subdivided without separately sampling from these sub- populations. Such sub-populations are called domains of study. Most studies have been carried out on domain estimation using linear cost function. This study there- fore estimates domains parameters in double sampling for strati cation in the pres- ence of non-response with non-linear cost function which will minimize cost for a xed value of variance or minimize variance for xed cost in the presence of non- response. The optimum stratum sizes of a given set of non-linear unit costs are derived using double sampling for strati cation. The relative precision of the es- timators are empirically compared with corresponding existing estimators. Data simulation will be done and then analysis and evaluation done using R-package.