STAT368 Lecture : Inferences; multiple comparisons.pdf
5 Apr 2014
School
Department
Course
Professor
Document Summary
Multiple comparisons - comparing several means with each other. 2 is esti- is estimated mated by , whose variance by. Similarly, a con dence interval on one di erence is. Typically we are interested in contrasts of the treat- ment means: =p. , where p = 0: assess the di erence between treatments 1 and 2: = 1. = p = p . We might end up constructing many cis. Before sampling the probability that any one of them will be wrong is. , so that the probability that at least one will be wrong (the experimentwise error rate ) is much more than. We might not know which comparisons we would like to make until after looking at the data. We need a method which accommodates this kind of. One such method is sche e s method for comparing all contrasts. In sche e s method we seek a number so that max .
