Örnek büyüklüğüne bağlı olarak z ve t değerleri arasındaki ilişki
Loading...
Date
1988
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Uludağ Üniversitesi
Abstract
Yapay olarak oluşturulan 1997 birimli bir normal dağılım ile 2200 birimli bir normal dağılımın her birinden n = 3 den başlayarak n= 50 ye kadar her n değeri için 100 örnek rastgele seçilip bu örnekleri için z ve t testleri kullanılarak hipotez kontrolleri yapıldı. Normal dağılım da, 3 ,n 50 için, z testine göre sıfır hipotezi kabul edildiğinde, bunun t testi ile kabul edilmesi % 95, kabul edilmemesi ise % 5 olarak bulundu. n ,30 için bu oran% 5.5 olmaktadır. Normal dağılımda yerine s konularak yapılan hipotez testlerinde ise yapılan hata % 6.5 olarak bulunmuş olup, bu hata n, 30 için % 7.3 olmaktadır.
Two populations were made artificially. One of them is distributed normally and has 1997 variables, the other is not distributed normally and has 2200 variables. When n varies from 3 to 50 for each value of n 100 random samples were drawn from two populations. For these samples zero hypotheses were controlled. For normal population when 3 ,n ,50 95 % of the hypotheses were accepted both z and t test. 5 % of the hypotheses were accepted according to z test and rejected according to t test. This ratio for n 30 was found as 5 %. For a normal population, when the test was made by putting s instead of a in z formula the error made is 6.5%. This error is 7.3 % for n ,30.
Two populations were made artificially. One of them is distributed normally and has 1997 variables, the other is not distributed normally and has 2200 variables. When n varies from 3 to 50 for each value of n 100 random samples were drawn from two populations. For these samples zero hypotheses were controlled. For normal population when 3 ,n ,50 95 % of the hypotheses were accepted both z and t test. 5 % of the hypotheses were accepted according to z test and rejected according to t test. This ratio for n 30 was found as 5 %. For a normal population, when the test was made by putting s instead of a in z formula the error made is 6.5%. This error is 7.3 % for n ,30.
Description
Keywords
Örnek büyüklük, Z ve t değerleri arasındaki ilişki, Relationship between z and t values, Sample size
Citation
Kan, İ. (1988). ''Örnek büyüklüğüne bağlı olarak z ve t değerleri arasındaki ilişki''. Uludağ Üniversitesi Tıp Fakültesi Dergisi, 15(3), Supplementum 3, 1-17.