Statistical Sciences 2141A/B Lecture Notes - Lecture 27: Binomial Distribution, Null Hypothesis, Test Statistic

• School: Western University
• Department: Statistical Sciences
• Course: Statistical Sciences 2141A/B
• Professor: Joe Raaymakers

Document Summary

However, each pair is independent of the others and the differences di = xi - yi are independent. T e di "s are assumed to e normally distri uted (usually as a result of t e xi "s and yi "s eing independent) with mean = a - b. As a result, the one-sample techniques (for the standard normal and t-distributions) discussed in the last c apter can e applied to t e mean, , of t e di "s. In particular, we can use these techniques to test the following: H0 : = 0 s. ha : 0 i. e. , a b (two-tailed) H0 : = 0 s. ha : > 0 i. e. , a > b (upper-tailed) H0 : = 0 s. ha : < 0 i. e. , a < b (lower-tailed) We can also use the one-sample tec niques to construct confidence inter als for . The test statistic and p-value for the t-distribution with 39 degrees of freedom are.