Environmental Sampling Design - Fundamentals of Environmental Sampling and Analysis - page 53

needs to be quantified, judgment sampling is not applicable. Keith (1990) further
indicated that samples collected for research studies allow use of prior knowledge in
order to obtain as much useful data as possible, but sampling for legal purposes often
requires absolutely random samples.

3.3.2 Simple Random Sampling

Simple random sampling, also referred to as random sampling, is the arbitrary
collection of samples by a process that gives each sample unit in the population (e.g.,
a lake) the same probability of being chosen (Fig. 3.5a). The word ''arbitrary'' is
probably misleading because a random process (e.g., random number table) is used
and each sampling point is selected independently from all other points in random
sampling. This differs from the nonstatistical haphazard sampling, in which the
sampling person may consciously or subconsciously favor the selection of certain
units of the population. The haphazard sampling claims ''any sampling location will
do'' and hence encourages taking samples at convenient locations or times. Simple
random sampling is the simplest and most fundamental probability-based sampling
design. It is also the benchmark against which the efficiency and cost of other
sampling designs are often compared.
Simple random sampling assumes that variability of a sampled medium is
insignificant, and hence it is appropriate for relatively uniform or homogeneous
populations such as holding vessel, lagoon, and so forth. This random sampling
approach also applies for sites with little background information or for site where
obvious contaminated areas do not exist or are not evident. One advantage of using
simple random sampling is that statistical analysis of the data is simple and
straightforward. Explicit formulas as well as tables and charts are readily available
to estimate the minimum sample size needed to sup po rt many statistical analyses.
The formula for the calculation of (arithmetic) mean (x) is the same as Eq. 2.12 and
the standard deviation (s) can be obtained from the sample variance (s 2 ) in Eq. 2.17.
Both were discussed in Section 2.2:

t

X

t

X

X ! 2

n n

n
i¼1 ðx i xÞ

2

x i x i = n

i¼1 i¼1

s ¼ ¼ ð3:1Þ

n1 n1

(a)

(b) (c) (d)

L

k = 2

k = 1

k = 3

Figure 3.5 Three common probability-based sampling designs for sampling in two-dimensional
space: (a) simple random sampling; (b) stratified sampling (three strata); (c) systematic grid sampling;
and (d) systematic random sampling (L¼spacing)