张俊妮《概率统计》北大光华-lecture20.pptVIP

The 90% Confidence Interval Suppose n=10 and Y=6. The equivalent confidence interval is * * Lecture 20 The Paired t test Equivalence of Tests and Confidence Intervals A Discrete Example Matched-sample design: e.g. We want to compare 2 production methods. Paired: e.g. In a production experiment, each alloy is divided into 2 parts. Production method 1 is used on one part, and production method 2 is used on another part. Variation between alloys is eliminated as a source of sampling error. Before/after: e.g. In a production experiment, each worker first uses one production method and then uses production method 2. Variation between workers is eliminated as a source of sampling error. The Paired t Test Consider Assuming then Assume you work in the finance department. Is the new financial package faster (0.05 level)? You collect the following data entry times: Paired t Test: Example User Current Leader (1) New Software (2) Difference Di C.B. 9.98 Seconds 9.88 Seconds .10 T.F. 9.88 9.86 .02 M.H. 9.84 9.75 .09 R.K. 9.99 9.80 .19 M.O. 9.94 9.87 .07 D.S. 9.84 9.84 .00 S.S. 9.86 9.87 - .01 C.T. 10.12 9.86 .26 K.T. 9.90 9.83 .07 S.Z. 9.91 9.86 .05 Is the new financial package faster (0.05 level)? A level a0 =.05 t test reject H0 if The test statistic is calculated as Reject a 0 =.05 1.8331 Decision: Reject H0 Conclusion: The new software package is faster. Suppose that a random sample X=(X1,…,Xn) is to be taken from a distribution that depends on an unknown parameter . If a random set w(X) satisfies for every , we call w(x) a coefficient g confidence interval for q. Equivalence of Tests and Confidence Intervals Consider testing the hypotheses: Suppose that for every point and every value of g (0g1), we can construct a level 1-g test such that