Presents computing methods for variance component estimation. These methods are designed to handle large sample survey or animal breeding experiments in which the random effects may have thousands of levels. The underlying theoretical technique, MIVQUEO (with V=1), has been shown to yield unbiased, locally best, admissible, and asymptotically consistent estimators.
13 pages.
To order SAS Institute Publications, call SAS Institute book sales at 1-800-727-3228. (Note: This toll-free number is for ordering books in the U.S. only.) You may also contact book sales by E-mail at sasbook@unx.sas.com.
To order SAS Institute Publications, contact your local SAS office. International orders sent to the U.S. office will be faxed to the appropriate SAS Subsidiary or Distributor office. Pricing, delivery, and terms for all international documentation orders are determined by each local office.