APST 207 Lecture Notes - Lecture 12: Null Hypothesis, Type I And Type Ii Errors, Dependent And Independent Variables
56 views5 pages
Document Summary
Know how to compare sample to population. Know how to compare two samples from different groups. The world is complex so we often have more than one group to compare. Need to find a way to compare the means of all groups at once. Live in a null universe (where there is no difference) Pull samples from a universe in which the null is true. The risk of this is that we"ll say there is a difference when there really isn"t. If =. 05 then our risk type error is 5% 5% of the time we"re going to say there"s a difference when there isn"t. If 1 in 20 t-tests give type i error, that means we should try to conduct a few t-tests as possible. In fact, that"s not true of t-tests, it true of all statistic. It can tests all the groups at once to determine if there is a difference. Born of the need to test experimental differences.