การเปรียบเทียบอำนาจการทดสอบของตัวสถิติทดสอบค่าเฉลี่ยของการแจกแจงแบบเบ้ขวา / อัญชนา ลีลาจรัสกุล = A comparison on power of the tests for testing the mean of positive skewed distributions / Anchana Lelajaratkul
The purpose of this research is to investigate the power of tests of Student’s t test, Johnson’s t test, Ling Chen’s t test and Sutton’s composite test for testing the mean of a population which is a positively skewed distribution. The distributions under study are Gamma distribution, Weibull distribution and Log-normal distribution with six levels of skew (0.25, 0.50, 1.00, 1.50, 2.00 and 2.50). The sample sizes are 10, 15, 20, 30, 50 and 70 respectively andlevels of significance are 0.01, 0.05 and 0.10 . For this research, the Monte Carlo technique is used by repeating 5,000 times for each case. Results of the study are as follows : At the level of significance is 0.01, the Sutton’s composite test has the highest power for all levels of skew [0.25,2.50] and all sample sizes. At the level of significance is 0.05 or 0.10, sample sizes are small and medium (10≤n≤30), the Sutton’s composite test has the highest power when level of skew is in [0.25,0.50] and Ling Chen’s t test has the highestpower when level of skew is in (0.50,2.50], For the large sample size (30≤n≤70), the Sutton’s composite test has the highest power when level of skew is in [0.25,2.50], The power of the test varies according to the level of significance and sample size.