Alternative Estimators for Point-Poisson Sampling
A new jackknife-based variance estimate is derived for point-Poisson (point-3P) sampling. This variance estimate is shown to be superior to most estimates derived assuming point-pps with or without replacement sampling, for a population with a very strong linear relationship between the variable of interest and the covariate used for sample selection and estimation. Only a variance estimate suggested by L. Grosenbaugh is equally good. This variance estimate is recommended since it is simplest to compute. Both the standard point-Poisson and a proposed unbiased jackknife estimate of total volume are about equally efficient. Forest Sci. 30:803-812.
No Supplementary Data
No Article Media