第五章 假设检验的功效与样本量（The fifth chapter is the function and sample size of hypothesis testing）.docVIP

第五章 假设检验的功效与样本量（The fifth chapter is the function and sample size of hypothesis testing） The fifth chapter is the function and sample size of hypothesis testing When testing the hypothesis that the test does not reject H0, the probability of inferring the correct probability is called the test function. In clinical research, it is necessary to evaluate the efficacy of the tests if the hypothesis tests are not statistically significant. 5.1, two types of errors and effects 1., the probability of two types of errors H0: (= 0, H1: () (0 (5.1)) Z = EMBED, Equation.3 (5.2) Any hypothesis test may have two types of errors and be measured by two probabilities The first kind of error probability (P = H0, H0 refused to be true) = ((5.3) Type II error probability (P = H0, H1 is really not rejected) = ((5.4A) Can also be understood as Type II error probability (P = H0, H0 = false reject = ((5.4b)) Whether the diagnosis is a disease is considered a hypothesis testing problem: H0: no disease, H1: ill The first kind of error: false positive / misdiagnosis probability P (positive, no disease (()) Type II error: false negative / misdiagnosis probability P (negative, sick) (() Background of two types of errors: A class I error may occur when you deny H0 Failure to reject H0 may cause a category II error Consequences of two types of errors: The type I error may be mistaken for invalid real effective / misdiagnosis Type II error may be mistaken for real effective invalid / missed (general (numerical) in scientific research, design, current affairs first 2. effect (power) (hypothesis testing shows that the effect of real difference is no less than 1- beta, i.e. The power of the test (H0, H1 = P to true) than 1- beta (5.5) The power of the test (H0 = P to H0 is false) or 1- beta (5.5) (the probability / disease effect is the real effective drug was found to be diagnosed 5.2, the four factors affecting efficacy (suppose the efficacy of a test is affected by at least four factors, see (5.