BioEpi :

Contact : Other Links

Overview

The main thrust of this research is to improve understanding of estimation of realized random effects. Such estimates are used for domain means, subject effects, and cluster effects. This research is pursued by using superpopulation sampling prediction theory based on a superpopulation generated from a random permutation probabilities. The background and overview of the ideas are given.

The intention of this Website is to share the basic results of this work at their current stage of development. Much of the development requires abstract representation using matrix algebra, which at times is complex. For this reason, very simple, but perhaps trivial, problems are developed first. We wish is to share the ideas in an accessible form, so as to provide insight on more complicated problems. Much of this work has not been published in the peer reviewed literature, although manuscripts are being submitted on a regular basis. A listing of working documents by calendar year is given under "Results". Presentations on the results have been given at national and international meetings. We are happy to share this work at this stage of development, and encourage the sharing of it with others who may be interested. We also welcome collaboration on future work.

We ask that these results not be plagiarized. This work has been developed with support from an NIH grant on Cluster Randomized Studies.

---

Comments and suggestions for the authors are welcome (send to Ed Stanek at stanek@schoolph.umass.edu).

Link to the Cluster WEB Site