CLA260H1S – Lecture 23
Getting the measure of the land 3
Survey sampling strategies
It is rarely possible to survey a complete region under study – and even
where it might be feasible, it may well not be the best use of available
resources. It is usually necessary instead to look at only a sample of the
terrain. How large a sample is required is the first decision, followed by
the shape of the study areas.
Sample size is based on a number of factors according to the situation
(including logistical constraints and the variability of the terrain). It is
important to remember that sample size is not necessarily directly
correlated with sample quality. Grid squares are the simplest units to
choose, but in large surveys it is common to use transects (or a
combination of transects and randomly-distributed squares).
A number of different sampling strategies can be adopted, including:
Simple random sample is exactly what it sounds like: the areas
sampled are chosen using random number tables.
Stratified random sample is where the region to be surveyed is first
divided into ‘strata’ or natural zones (e.g.