APPROXIMATE SRIVASTAVA ESTIMATION IN FORESTRY

Srivastava estimation formalizes the introduction of additional prior and posterior information in sampling to increase efficiency in estima tion. An unbiased new estimator is given for Poisson sampling which is approximately as efficient as the standard slightly biased estimator without using any additional information. The increased efficiency pos sible with this method or approximations to it are illustrated for Poi sson sampling and point-Poisson sampling using simulations. Srivastava estimation may be applicable in many forest inventory situations. Bes ides efficiency, the seriousness of estimation bias and reliability of variance estimators need to be considered in all applications.