are Less Than Some of the Estimates?

An organization was sampling a small
area. When they calculated the "K+Z" value to control sample size, it was less
than many of the estimated volumes. How could they solve this problem without getting into
complicated statistics, biased answers or smaller sample sizes? The answer is to use several random number lists. For example, create 2 lists, each designed to select half of the sample size you need. For each tree, compare the estimated volume to both |
of the lists, and choose as a sample
tree if the estimate is larger than (or equal to) the number on either list. In the rare
instance where the tree is chosen by both lists, use that sample tree ratio twice in the
calculations. This technique allows you to create a list having a K+Z value larger than any single estimate, but still have the desired sample size. The logic could of course be extended to 3, 4, or more lists if necessary. This is a simple field solution requiring no statistical, programming or compilation changes. |