Accuracy & Efficiency Comparisons Between Two-stage Stratified Random Sampling And Grid Random Sampling

Abstract

Stratified random sampling is the most accurate random sampling for behavioral science and social science researches while grid random sampling method is widely used for conducting physical science researches on population living in large area. Therefore, this paper is to compare accuracy and effectiveness between two-stage stratified random sampling and grid random sampling techniques by using O-NET (Ordinary National Educational Test) score database in 2005 from 2,584 schools in Thailand. The average score is 161.67 and the standard deviation is 21.74. To compare the two methods mentioned above, 335 schools are sampled by each method and then find average and standard deviation. With two-stage stratified random sampling method, the average score is 156.56 and the standard deviation is 15.81. With grid random sampling method, the average score is 158.49 and the standard deviation is 16.51. The results indicate that the two random sampling methods statistically have the same accuracy; however, the grid method is more efficient due to the fact that samples cluster around the grid points. Close distances among the samples facilitate data collection process. In contrast, samples obtained by the stratified sampling method scatter and need more resource to collect.