Many variables that are of interest to foresters are difficult and/or expensive to measure but are related to other variables that are less difficult or expensive to evaluate. If the relationship between such variables can be established, it becomes possible to use data from the latter to make estimates of the former. This process is known as double sampling. Usually the term
is reserved for situations where estimates are made of population means or totals. However, whenever regressions are developed to obtain estimates of difficult-to-measure quantities as, for example, the volume of standing trees using dbh and the height, the procedure could be considered a form of double sampling.