Answer Happy

Select the situations when it's acceptable to test Ho: M1 = Hy against one-sided alternative H1:41 > H2 The populations are such that Hy <Hy is impossible X1 > Ãin the samples We don't care about fly < Hz: for example, if X1 is the effect of a new drug, and X, is the effect of placebo, it's against out interests to use two-sided alternative, because we don't want to show that the drug is worse than placebo. Select all correct statements about Behrens-Fisher problem The problem is when comparing means of two independent samples with unknown variances, we cannot even approximate the number of degrees of freedom for the null distribution of t-test statistic Approximate solution to the problem (given by t test) is quite accurate when [n1 > n2] = [01 <02] Approximate solution to the problem (given by t test) is quite accurate when ni = 12

3. Which of the following statements are true about Wald's, score, and likelihood ratio tests for hypothesis about a single Number of (1.1) pairs Number of (0,0) pairs Number of (0.1) pairs Number of (1.0) pairs 5. What would be the consequence of using test for the equality of two proportions in independent samples on paired samples problem? Lower power Going to hell for statisticians