Hypothesis testing: A procedure based on sample evidence and probability theory to determine whether the hypothesis is a reasonable statement. State null H0 and alternative hypotheses H1. Select a level of significance, α. Identify the test statistic. Formulate a decision rule. Critical-value Approach P-value Approach Confidence interval Approach Take a sample, arrive at decision. Do Not Reject, H0 Reject H0 and accept H1 Type I error: Rejecting the null hypothesis, H0, when it is true. (False positive, 假阳性) Type II error: Not to reject the null hypothesis when it is false. (False negative, 假阴性) Two-tailed test: Θ0 = {µ0} H0 : µ = µ0, Ha : µ = µ0. Upper-tailed test: Θ0 = (−∞, µ0] H0 : µ ≤ µ0, Ha : µ > µ0. Lower-tailed test: Θ0 = [µ0, ∞) H0 : µ ≥ µ0, Ha : µ < µ0. Variance is known Z-test Under H0, the distribution of z is known, i.e., z ∼ N(0, 1) Population proportion Bernoulli(p): reject H0 : p ≤ p0 z ≥ zα Significance level = P(Reject H0|H0) = P(|z| ≥ c|H0) p-value: the probability of observing a sample value as extreme as, or more extreme than, the value observed, given that H0 is true. reject H0 : µ = µ0, if p-value ≤ α T test Under H0, the distribution of t is known, i.e., t ∼ tn−1, where n − 1 is the degree of freedom. Other same as Z test When variance is unknown The power of a hypothesis test is 1 minus the probability of a Type II error. Two sample test: H0 : µ1 − µ2 = D0, Ha : µ1 − µ2 = D0. variance 1,2 are known, Two population proportions: H0 : p1 = p2 Ha : p1 = p2 Variance unknown: Variance unknown: use use Matched samples: new list: µd=µ1 − µ2 Non-parametric: sign test: H0 : p = 0.5, Ha : p = 0.5 Normal approximation: t~tdf