| บทคัดย่อ(English) |
The objective of this study is to illustrate the basic concept ofestimating population proportion and probability of success under Bernoullidistribution. Clear understanding on this matter leads to determination forappropriate or correct choice of statistical method. The study is dividedinto two parts, i.e. the analysis of basic concept on theory of statisticsfrom examples in statistical textbooks, and a simulation study forinvestigate on properties of point and interval estimator. It was founded that many statistical textbooks lack of the clearexplanation on the basic concept of the estimating populationcharacteristic under Theory of Sample Survey and the estimating ofparameter under Theory of Statistical Inference. In addition, thedistribution of statistics "a" (number of interest in the sample)approaches normal distribution when "n" (sample size) and "N-n" (differencebetween of population size and sample size) are large. For intervalestimation, the estimator given by M.E.Thompson (1997) has the closetcoverage probability to the confidence level. When stratified randomsampling is used, the weighted estimator has lower mean absolute errorcomparing to unweighted estimator. |
