İki ortalama farkına ilişkin hipotezlerin z ve t testi ile kontrolünde, verilen kararların doğruluk dereceleri
Files
Date
1993-03-29
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Uludağ Üniversitesi
Abstract
Biyoistatistik, kurulan hipotezlerin kontrolü toplum yapısı, örneklerdeki birim sayılan ve toplumlara ilişkin varyansların bilinip, bilinmemelerine göre değişik yollarla yapılmaktadır. Bu araştırmada, yapay olarak oluşturulan ve parametreleri u. = 50, u= 10, N= 2996 olan normal ve u. = 48.734, u = 11.36, çarpıklık katsayısı Çk = 0.1533, basıklık katsayısı Km = - 0.136 ve N = 3000 olan anomial, iki toplumdan nx ny (nx ve ny 5'ten 50'ye kadar 5 ve 5'in katlarını alacak şekilde) birimli 100'er örnek rastgele seçilerek toplam 10000 örnek incelendi. Her iki toplumda, H0 hipotezin a = 0.05 düzeyinde z ve t ye göre red oranlan, nx ve ny'nin değişik kombinasyon gruplarında hesaplanıp birbirleriyle karşılaştırdılar.
In biostatistics, testing of hypotheses are made in various ways. Choosing the way depends on structure of population, sample sizes are known or not. In this study, two populations were formed artificially. The first population with u. = 50, u = 10 N = 2996 was nonnal and the other was witlı u. = 48.734, u = 11.36, coefficient of skewness (Çk = 0.1533, coefficient of kurtosis (Km) = - 0.136, N = 3000 was not normal. 100 sanıples of ng ny (nx and ny aries from 5 unit) were drawn randomly from each x and y population. The total sanıpfes investigated were 10000. For both population, at the significant level of a = 0.05 the rejection ratios of zero hypotheses according to z and t tests were calculated for each combination group of ng ny and comparisons were made among them.
In biostatistics, testing of hypotheses are made in various ways. Choosing the way depends on structure of population, sample sizes are known or not. In this study, two populations were formed artificially. The first population with u. = 50, u = 10 N = 2996 was nonnal and the other was witlı u. = 48.734, u = 11.36, coefficient of skewness (Çk = 0.1533, coefficient of kurtosis (Km) = - 0.136, N = 3000 was not normal. 100 sanıples of ng ny (nx and ny aries from 5 unit) were drawn randomly from each x and y population. The total sanıpfes investigated were 10000. For both population, at the significant level of a = 0.05 the rejection ratios of zero hypotheses according to z and t tests were calculated for each combination group of ng ny and comparisons were made among them.
Description
U.Ü. Sağlık Bil. Enst. Yüksek Lisans Tezi olarak verilmiştir.
Keywords
Ortalama, Fark, Hipotez, Z testi, T testi, Mean, DiiTerence, Hypothes, Z test, T test
Citation
Ediz, B. ve Kan, İ. (1993). ''İki ortalama farkına ilişkin hipotezlerin z ve t testi ile kontrolünde, verilen kararların doğruluk dereceleri''. Uludağ Üniversitesi Tıp Fakültesi Dergisi, 20(2), 127-130.