The Sample Size Requirements Of Equivalence Test Of The Two Population And Simulation Study

The hypothesis of equivalence test is defferent from the tranditional hypothesis test since the null hypothesis on the parameter is not a point but a interval. it becomes much more applicable than the traditional hypothesis test, and have been widely used in many industries and sectors, such as whether domestic products can replace imported products, the new teaching method is superior to the old teaching method, etc. At present there are many discussions on the equivalence of the mean of normal continuous data and rate in the literature, These problems involved such as interval estimation, hypotheses test, sample size estimation, the Type I errors and efficacy calculation.In this paper the sample size formulas of equivalence test of continuous normal data and rate data and simulation testification are discussed under various conditions, the relationship between traditional hypothesis test and equivalence test are also made a detailed discussion. normal approximation algorithm and iterative algorithm are presented in this paper at continuous situation, under discrete conditions the maximum likelihood estimation under constraint conditions are used to estimated the parameters and eventually the calculation of the sample size are solved by normal distribution. The influence of sample size calculation between the dependence and the independence are also discussed in detail.