Yığın varyansları heterojen iken ortalamaların eşitliği hipotezi için bazı testler ve karşılaştırmaları

---

Thesis Type: Postgraduate

Institution Of The Thesis: Gazi University, Fen Bilimleri Enstitüsü, İSTATİSTİK ANABİLİM DALI, Turkey

Approval Date: 2019

Thesis Language: Turkish

Student: AHMET METİN YAVUZ

Supervisor: HAMZA GAMGAM

Open Archive Collection: AVESIS Open Access Collection

Abstract:

In the problems of the test of the hypothesis indicating the equality of population means, it is commonly encountered that the assumption of the homogeneity of population variances is not assured. In the literature, there are many tests proposed for the test of hypothesis indicating the equality of means for the populations with normal distributions when the population variances are not homogeneous. In this study, Cochran Test, James Test, Welch Test, Brown-Forsythe Test, Alexander and Govern Test, Brown-Forsythe Test, Modified Welch Test, One-Stage Test, Fiducial Approach Test and Parametric Bootstrap Test have been briefly introduced and compared to Monte Carlo Simulation Method. In these comparisons, various simulation scenarios have been used with design parameters such as population number, sample size, variance values and mean values. In this simulation study, these tests have been compared in terms of experimental type I error rate and the power of the test. According to the simulation results, the James Test has given better results than the other tests in terms of type I error rate. According to the power results, it has been seen that the Cochran Test is more powerful than the other tests in general.