Sampling Approaches

Links:

Successful regional monitoring programs typically include status, trend, and targeted sampling of sites within a region. The three approaches answer different questions and differ in the way sampling sites are selected from the population of all possible locations.

Probability-based sampling vs. other sampling approaches

Probability sampling means that every element, or sampling unit, of a population, has a known probability of being included in the survey sample. Once the population is defined, sampling sites are randomly selected. More sampling sites may be selected from different subsets, or strata, of the population, but within the strata, selection is random. In contrast, judgment sampling refers to selection of sites based on experience or scientific criteria. Convenience sampling selects sites for their availability or ease of access.

When sites are randomly selected from a larger population, results for the sample represent the larger population, and are not biased. In some situations, judgment sampling may be more appropriate, but no matter how "representative" he sampling sites may be, the results for those sites cannot be generalized to the larger population. Similarly, any change observed for sites in a judgment sample cannot be applied to sites not sampled, no matter how similar those sites may be.

Probability sampling for status assessment

Status refers to the current condition of a resource and is best assessed by randomly selecting sites each year from the region of interest. The only better approach for status assessment is census sampling which would visit every possible sampling site; however, the cost associated with census sampling at the regional level is prohibitive. Random sampling in a survey design means that not all sites will be sampled; consequently, estimates of resource condition will have an estimate of variance associated with them.

Probability sampling for trend assessment

Trend refers to the change in resource condition over time and is best assessed by randomly selecting sites during the first year and then revisiting those same sites during each time period. Fixed station sampling is similar in that the same locations are sampled each year, but differs from probability-based sampling in that the sites are not randomly selected at the outset. This means that results from fixed stations cannot be generalized to assess regional conditions.

Targeted Sampling

Targeted sampling goes by many names, and refers to the selection of sampling locations based on specific criteria. Targeted, or judgment, sampling, is appropriate for many situations, but a poor choice for the assessment of regional status and trend. Specific sites may be targeted for assessment because of restoration programs or known point-sources that must be monitored.