Biostats Notes Section 2: Error I and II: State of nature: 1% of population has a disease Type I error rate =0.05 (false positive rate) Type II error rate = 0.10 (detection rate = 0.90) If someone has a positive test, what is the probability of the person having the disease? What are consequences of Type I and II errors? Testing a drug for a potential benefit? Testing a drug for harmful side effects? Which error is more serious? 10/24 T-Test: One Sample T-Test: T-statistic: Null Hypothesis: ? = C The sample mean does not differ from the known population mean. Degrees of freedom: df=n-1 Independent samples / Two-sample T-Test: Null Hypothesis: ?1 = ?2 The means do not differ. Degrees of Freedom (df=n1+n2-2):