For binary data, if sample size is large( n>30) , you can choose normal approximation, or its variation, also called McNemar test, Motherwise you can use exact binomial distribution for constructing confidence interval for inference or for computing exact p-value. Refering to the 4-fold table, you can consult any books about categorical data analysis.
davidhaitaopan 发表于 2010-11-21 09:36
For binary data, if sample size is large( n>30) , you can choose normal approximation, or its variation, also called McNemar test, Motherwise you can use exact binomial distribution for constructing confidence interval for inference or for computing exact p-value. Refering to the 4-fold table, you can consult any books about categorical data analysis.