Optimum sample size in stratified two-stage sampling design for estimating planted area and production of corn
1990
Montri Muangprom
Objective of this study was to find the optiumum number of sample villages and households for estimating cron planted area and production in Saraburi province. The sample design used in this study was a stratified two-stage sampling. The villages, the primary sampling units, were stratified into 4 strata according to the reported planted area of corn in the village. In all there were 311 corn growing villages. From these villages, 93 villages were randomly selected from each stratum. The sample villages were allocated among these strata proportional to the number of the corn growing villages in each stratum. In this study, two methods were employed: the simulation and direct calculation methods. For the first method, a subsample of 10, 20, 30 and 40 percent of sample villages or 3, 6, 9 and 12 percent of total corn growing villages and a subsample of 5, 10, 20 and 30 households per village were simulated 100 times for each combination to get the estimated planted area and production of corn. These estimated values were then compared with the estimated value obtained from 93 villages in order to measure the size of the error and the reliability of the estimates. For the second method, the cost function was developed and the optimum sample size was determined by using two criteria: minimize cost given error and minimize orror given cost. The optimum sample size which is also feasible in practice, using one enumerator, was 74 villages of 24 percent of the total corn growing villages and 10 households per village. With this sample size, the estimated field cost would be 28,113 Baht and the coefficient of variation of the planted area was 9.60 percent at 95 percent confident level.
显示更多 [+] 显示较少 [-]