Total Survey Error Model for Estimating Population Total in Two Stage Cluster Sampling

Abstract: This study is based on the total survey error paradigm in which the purpose is to examine a variety of sources of errors in a survey. The perspective taken follows an error model based on a finite population model, in which the main objective of this study is to propose a total survey error model in two stage cluster sampling. Our input is the demonstration that survey estimates have been presented with only one source of error measured, error due to sampling, resulting from the fact that survey estimates would have different values had another sample been drawn using the same sampling design. Other variable errors like the response error are ignored, and biases are rarely mentioned. The presence of a total survey error model offers a rare opportunity to measure and quantify a large set of variable errors and biases that are normally assumed to be negligible in survey data analysis. The estimators used for the population parameter are seen to be subject to both variable errors and biases