Sampling for population estimation
Studies of natural populations of animals and plants almost inevitably involve sampling programmes. Sometimes these programmes are relatively straightforward. For example, to determine the proportion of trees that are infested in a large orchard of numbered trees, a random sample of tree numbers can be drawn and the chosen trees inspected. The sample proportion of infested trees is then an unbiased estimator of the orchard proportion and standard statistical methods allow limits to be put on the likely error of estimation. However, the situation is much more complicated if the problem is to estimate the number of insects in different life stages in a population that covers a large heterogeneous area. In this case, the different stages are likely to occupy different parts of the habitat and sampling all stages with the same intensity may be difficult.
KeywordsSampling Unit Population Estimation Ratio Estimation Unbiased Estimator Simple Random Sample
Unable to display preview. Download preview PDF.