![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Targeted Sampling
Description: This category of sampling is something of a catch-all. Targeted sampling may also be referred to as judgment sampling and selects sites according to a particular condition or to test a scientific hypothesis. Most non-random sampling plans would probably be included in this category. Targeted sampling is the best approach when a specific question is being evaluated, e.g., the effect of restoration or best management practices on stream condition.
Simple example: You manage a state forest and want to test whether erosion barriers and curtains reduce sediment deposition in streams during logging. You must identify streams with active timber harvest and then divide stream sites into treatment groups (with barriers) and control groups (without barriers). Random selection is inappropriate for initial site selection because only sites with active harvest schedules are relevant; however, once a set of appropriate sites is identified, sites should be randomly allocated to treatment and control.
MAIA example: For this project, hand-picked reference and test sites were included with the randomly selected sites to ensure that the entire range of conditions was sampled for the region. Click on this link to see the comparison of the targeted reference and test sites used to compare macroinvertebrate metric values.
How the method works: Sites are selected for a specific feature, treatment, or condition. Examples include selecting reference and test sites for metric testing, choosing sites upstream and downstream of a pollution point source, or comparing the same sites before and after some type of treatment. Any number of statistical methods would apply depending on the type of questions being asked. The key point to remember is that results observed for targeted sites cannot be extended to sites not sampled, as the results can be for random sampling.