Unformatted text preview: his blocking approach (moving
data around as needed so that within each block the elements are closely matched) should improve the
accuracy of the test of the brands (the main treatment test). In this case, the blocking or matching is the
best it can be: each element within a block is the same expert, so each expert constitutes her own block.
(The data have already been matched so that blocking applies.)
Our testing will be at alpha = 0.05. The tests are:
We should verify whether the brand populations are normally distributed, using CHI-SQ, but we'll
assume they ARE normally distributed. Thus, just state that we assume the brand populations are
View Full Document
This note was uploaded on 01/27/2013 for the course STAT 385 taught by Professor Szatrowski during the Spring '08 term at Rutgers.
- Spring '08