Statistical Sciences 2244A/B Lecture Notes - Lecture 13: Type I And Type Ii Errors, Null Hypothesis, Confidence Interval

• School: Western University
• Department: Statistical Sciences
• Course: Statistical Sciences 2244A/B
• Professor: Jennifer Waugh

Document Summary

Lecture 013 hypothesis testing iii - one-sample test for mu. Suppose we set a significance level of alpha = 0. 05 to use as our criterion of what"s too small to happen due to chance alone". P (true proportion) = p 0 (predicted proportion) If the probability (under a given assumption) of a particular observed result is small, then the assumption is probably not correct. If the p-value is very small, reject the null hypothesis. However, a non-zero p-value indicates that it is still possible to get an. Ex. if =0. 05, reject null when the p-value <= to 0. 05 ( ) extreme value. So, even when the null is rejected at <=0. 05, there is still a 5% chance that the actual value is not included inside the confidence interval. Can be controlled by selecting of various sizes. If is extremely small, an extremely large interval will be generated. Null hypothesis essentially would never be rejected.