ABSTRACT
The previous chapters in this book have been predominantly con cerned with the estimation and testing of mean values. The only exceptions to this have been the estimation and testing of the vari ances of random effects in Chapters 3 and 4. This preoccupation with means is caused by most questions in applications being concerned with differences in location of different data sets. However, questions about variability do arise either as the primary issue, as in random effects variance component problems and in deciding which of sev eral measurements is more reproducible, or as a secondary issue, as in deciding whether to pool sample variances.
