Geoscience Reference
In-Depth Information
FIGURE 10.8
A BAS sample of size n = 100 in the Canadian province of Alberta. The BAS sample was drawn
using the SDraw package in R.
and second-order inclusion probabilities can be approximated by simulation
using code similar to that used for GRS samples (Robertson et al . , 2013).
10.5 Summary
In this chapter, concepts important when designing a sampling scheme
were discussed. Concepts included the definition of scientific and nonsci-
entific designs. Scientific designs are those that draw probability samples;
nonscientific designs are those that do not. It is suggested that environmen-
tal studies, and monitoring studies in particular, should always use scien-
tific designs.
Another concept discussed was the distinction between research and
monitoring studies. This distinction was made because the sampling design
needs of both studies are generally different. In particular, the sampling
design of research studies must usually provide the highest possible statisti-
cal precision for a particular parameter or hypothesis. The sampling design
Search WWH ::




Custom Search