Abstract
Abstract Tukey's [4] T-(maximum modulus) method of multiple comparisons is applicable to homoscedastic and equally correlated normally distributed random variables. For multiple comparisons on interaction effects (in factorial experiments), the allied random variables are not equally correlated. Nevertheless, it is shown that the T-method holds in this situation. The theory is illustrated by an example on a partially balanced incomplete block (PBIB) design.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have