ABSTRACT

Summarizes information scattered in the technical literature on a subject too new to be included in most textbooks, but which is of interest to statisticians, and those who use statistics in science and education, at an advanced undergraduate or higher level. Overviews recent research on constructin

chapter 1|4 pages

Introduction

chapter |4 pages

= Tj

= Yii - a}, and Expand

chapter 2|10 pages

General Concepts

chapter |9 pages

= + + +

chapter 3|20 pages

General Results for Balanced Designs

+ crla + + (KIJ- Q=4, c

chapter |3 pages

= 2, Q =

= 4,

chapter |4 pages

= 0, or

chapter |3 pages

= .90 si-

= 3.592, F.=

chapter 4|7 pages

The One-Fold Nested Design

chapter |5 pages

= 2a

chapter |5 pages

= 0. If it is known that ~ is

= I;Y-

chapter |3 pages

> .25, it can be liberal in other cases.

2a confidence interval on is = c against Ha: > c. The

chapter |9 pages

2a confidence interval on

chapter |1 pages

= = (6

> As an example, the test of

chapter |6 pages

+ + +

ee is represented as [L*;

chapter |1 pages

= 1 and Kij =

= K, ww = 1K, = 0, if = 1 = K.

chapter |13 pages

= where dij are known

chapter 6|3 pages

Crossed Random Designs

chapter |4 pages

= 4. These distributional assumptions allow

2a confidence

chapter |3 pages

> Fa:n.,n > Fa:n

> 0 and hence an interval on crllcri is not of + + crk +

chapter |3 pages

= Fos:s, =

chapter |5 pages

= 0 was

chapter |3 pages

crklcri provides an exact test of H

I';if

chapter |1 pages

2a conservative

chapter |7 pages

+ = + =

+ = = +

chapter 7|14 pages

Mixed Models

chapter |5 pages

= -Y* **) -

chapter |6 pages

= kkiY

chapter |7 pages

APPENDIX A F-Tables

chapter |5 pages

s-F.

chapter |6 pages

andj N -

haver> 0 andj> 0. = I, anN x 1 vector of the Y X is anN x 1 vector of ls, = x 1 vector of the A; random variables, and §.is anN x 1