Balanced Random Sample and Its Selection Method

Objective of sampling is to make conclusions about the whole population based on the information of a part of it. Generally, sample desirability can be evaluated from two points of view. The first view considers representative of the sample, i.e., the extent it consists main structure of the population. The second view pays attention to randomness of the sample, i.e., whether both selected and non-selected units have positive and certain selection probably. In this paper primarily balanced sampling has been described as a probability sampling method that makes it possible to select a representative sample from the target population. Then the cube method for selection of a balanced sample is described. At last a simulation is performed using a real data set and efficiency of this method is compared with PPS sampling method. This simulation is implemented for three different groups of variables as basis of balancing. The results show that balanced sampling improves estimations efficiency and by adding proper balance variables the rate of efficiency increase becomes more meaningful.