Figure 2. Systematic random sampling

ERDC TN-DOER-C15

July 2000

essentially the prescriptive approach described in Olin-Estes and Palermo (2000a). A systematic

approach involves imposing a uniform grid over the area of interest and sampling from the same

location in each grid. A random approach involves selecting sampling points within a gridded area

using a random number generator to choose from among the alternative sample locations. The

random approach is optimum from a statistical standpoint but in environmental sampling may not

be the best choice. If the number of samples being taken over a large area is small, purely random

sampling could well miss an area of known contamination. The judgmental and systematic

approaches help to compensate for this, but may violate the assumption of randomness required in

statistical analysis. In practice, the following three principal sampling approaches that are used in

environmental sampling incorporate a combination of these elements:

Systematic random.

Judgmental random.

Systematic judgmental.

As previously mentioned, a uniform grid is imposed over the area to be sampled in the systematic

methods. In the systematic random method, a random number generator is then used to pick the

locations within the grid that are to be sampled

(Figure 2). Alternatively, one may select only

every nth member from the sampling grid with the

starting element randomly selected (Lubin, Wil-

liams, and Lin 1995). Systematic judgmental

methods focus attention on the area most likely to

be contaminated, a grid is imposed, and a sample

is taken from the center of each grid (Figure 3).

The judgmental random method involves separat-

ing the area of interest into blocks that are ex-

pected to contain similar samples (such as similar

levels of contaminants). A grid is imposed over

these areas and sample sites are selected randomly

Figure 2. Systematic random sampling

within each block (Figure 4). This is also referred

Figure 3. Systematic judgmental sampling

Figure 4. Judgemental random sampling