From the viewpoint of efficiency, sampling error contributes most to the effort in making a stereological estimate. The more work that goes into a stereological estimate, the lower the sampling error. Consider a study to quantify the total number of cells in a defined reference space. Instead of an estimate of the total cell number obtained by sampling the reference space with a theoretically unbiased stereological probe (disector), suppose we count all the cells; that is, we make a determination. In this case there is no sampling error because there is no sampling.

Such a high level of precision and effort is unnecessary, however, because the biological variation from one individual to the next is typically on the order of 20 to 30%, and frequently 50% or higher in some cases.

Thus, the more rational strategy, also known as "Do More Less Well", is to use relatively light sampling within each individual, while analyzing more individuals from the population.

This approach will efficiently capture the majority of variation within each individual, and rapidly converge the mean value of the estimate on the true value for the population.