Examining the behaviors of population mean estimators and their variances for a finite population in design-based adaptive sampling

In the cases where population is of rare and clustered character with respect to the variable of interest, using adaptive sampling design rather than classical sampling methods allows for obtaining more effective estimators. It is known that the effectiveness of the estimators obtained through adaptive sampling varies according to a large number of factors. Of these factors, the most important are the number and the size of the networks. This study aims to examine the relative effectiveness of the variables of the number of networks and networks for the finite populations of geographic character on the estimators. To this end, the software was designed. The graphics illustrating the changes concerning the relative effectiveness were drawn for four different finite populations. The effect of the numbers of samples and trials was also observed. Special importance was attached to the examination of the relative effectiveness of estimators developed and Rao-Blackwell estimators. vii It was investigated, especially with respect to overcoming the difficulty of calculation, how reliable it would be to use the developed estimators in the place of Rao-Blackwell estimators.