The objectives of some experiments are to compare the variances of two or more treatments, products, or techniques. If the investigator is more concerned about within-unit variances rather than between-unit variances, then a repeated measurement design is needed. We invoke a random effects model with heterogeneous within-unit variances for certain repeated measurements designs. We do not impose any distributional assumptions for the random effects, whereas we assume either a normal or multivariate t distribution for the random errors. We propose a partial likelihood analysis for population-based inference and individual-based inference. We illustrate the methodology with an example from a trial comparing serum cholesterol measurements from a routine laboratory analyzer to those of a standardized method.
All Science Journal Classification (ASJC) codes
- Statistics and Probability
- Biochemistry, Genetics and Molecular Biology(all)
- Immunology and Microbiology(all)
- Agricultural and Biological Sciences(all)
- Applied Mathematics