การเปรียบเทียบการทดสอบเอฟและการทดสอบมอนติคาร์โลด้วยอัตราส่วนภาวะน่าจะเป็น สำหรับแผนการทดลองแบบสุ่มตลอดที่ปัจจัยทดลองคงที่ / อรไท สงวนสินธ์ = A comparison on F-test and monte carlo likelihood ratio test for fixed-effect completely randomized design /Orathai Sanguansin
To compare the methods of hypothesis testing on the difference of treatment effects by 2 methods; F-test and Monte Carlo likelihood ratio test. To generate the data for this study, the Monte Carlo simulation technique is done using S-plus 2000 package. The number of treatments is specified at 2,3,4 and 5 treatments. The sample size on each treatment is at 2,4,6 and 8. The coefficient of variation is specified at 10%, 20% and 30%. The significance levels for this study are at 0.01 and 0.05 level. The proportion of null hypothesis rejection and the power of the test are a measure for comparison for both methods. The results of this study can be summarized as follow 1. Proportion of null hypothesis rejection. Almost all of cases, Monte Carlo likelihood ratio test gives proportion of null hypothesis rejection less than F-test. In the case that significance level is 0.05, F-test gives proportion of null hypothesis rejection less than Monte Carlo likelihood ratio test when the number of treatments, the sample sizes on each treatment and the coefficient of variation increase. 2. Power of the test. When the difference of treatment effects is less, Monte Carlo likelihood ratio test gives the highest power of the test. When the difference of treatment effects is moderate, Monte Carlo likelihood ratio test gives the highest power of the test except in the case that the number of treatments, the sample sizes on each treatment and the coefficient of variation increase, F-test gives the highest power of the test. When the difference of treatment effects is high, both statistics gives approximately the same power of the test level.