Author/Authors :
UZUNOĞLU, Arda Gazi Üniversitesi - Fen Fakültesi - İstatistik Bölümü, Turkey , GÖKPINAR, Fikri Gazi Üniversitesi - Fen Fakültesi - İstatistik Bölümü, Turkey , GÖKPINAR, Esra Gazi Üniversitesi - Faculty of Science - Department of Statistics, Turkey
Title Of Article :
Comparison of Some Tests for the Equality of the Coefficient of Variations in Terms of Type One Error and Power of Test
Abstract :
Coefficient of variation measures the uniformity or the consistency of the observations of a random variable. Since coefficient of variation is standard deviance per mean, it is a measurement of relative variability. For this reason coefficient of variation is used to determine the variability of the groups. In literature, some tests are given to compare the coefficients of variations k groups. In this study, the modified Bennett’s test, Score test, Non-central t test, Likelihood ratio test, Generalized p value are investigated. Also, these test statistics are compared according to their type I-error rates and powers under different of sample sizes and different number of groups. From the results obtained from simulation studies, the power of the Modified Bennett test is greater than the other tests when sample size are equal and group size is small. The power of the Score test is greater than the other tests when sample size are equal and group size is getting larger.
NaturalLanguageKeyword :
Coefficients of variation Modified Bennett’s test. Likelihood ratio test. Non central t test. Score test. Generalized p approach.
JournalTitle :
Journal Of Natural and Applied Sciences