Category:Introduction to sampling

If every element in the population is listed or enumerated we call that a census or complete enumeration. Censuses are rare in forest inventory. They are feasible, for example, when interviews are to be made with forest owners and all owners are being visited; or when in a tropical forest only the large commercial trees are to be observed. The normal case, however, is that statistical sampling needs to be applied because the populations are simply too big. Sampling is essentially an “information collection tool”. It is done to produce estimates, wherever a census is not possible or reasonable. Those estimates should be such that they go along with the project’s requirements and expectations. Sampling is being applied in all empirical disciplines so that it is a real asset for a researcher and manager working in forestry, agriculture, biology, sociology and others to know the principles of sampling well. In forest inventories, samples are taken in the field, but samples are also selected in remote sensing imagery, or when interview partners are to be selected, or pieces of wood for drying and wood density measurements.

We may state that sampling is a central element in forest inventories. Sampling (planning /implementation/ presentation) determines largely how credible and/or reliable the results are. It is important to know how to do “statistically valid” sampling (this is actually not a good term; we get more specific later on …). Only with a sound knowledge of the principles of statistical sampling you will be able to design and plan a forest inventory properly, make a proper interpretation of the data and results and present and defend the results properly to an audience who has good knowledge or interest in sampling or statistics.

Sampling studies should achieve a defined goal with minimum cost or produce the best (i.e. most precise) product for given cost. However, most important is probably that the forest inventory and its results make it to generate overall credibility; only then will the results be used for improved decision making. This overall credibility is reached to a large extent through strictly adhering to statistical principles of sampling and through clear definitions and full transparency of the planning, implementation and analysis process. In many occasions there are various options of how to design the inventory. A recommendable guideline for decision is then to adhere to the well known principle keep-it-simple (KIS) or keep-it-short-and-simple (KISS); but, of course, it must not be oversimplified.