UTILIZING ADAPTIVE CLUSTER SAMPLING DESIGN FOR PRODUCING COTTON IN IRAN

In most of statistical studies, we are dealing with populations that don’t have any frame,or the characteristic of participants in population that we intend to study may be rare or in some populations like ecological populations, participants have not scattered randomly in population. In such cases if we use classical sampling designs, like simple random, stratified, or cluster sampling, we may estimate the parameters of populationwith large error. One of the designs that have been suggested to solve this problem is adaptive cluster sampling design. In the present paper, this design has been studied and unbiased estimators of mean and total will be presented. Finally, with regard to the fact that Iranian cities which produce cotton are clustered and dispersed and with the supposition that any list of these cities are not available, the amount of cotton producedin 2002-2003 agriculture year will be estimated by using this method and in order to compare this design with others, a simulation study on cotton-producing cities in Iran will be carried out.

Keywords:

adaptive cluster sampling , sequential sampling , clustered population