The accuracy of your findings largely depends upon the way you select your sample. The basic objective of any sampling design is to minimise, within the limitation of cost, the gap between the values obtained from your sample and those prevalent in the study population.
The underlying premise in sampling is that a relatively small number of units, if selected so that they genuinely represent the study population, can provide – with a sufficiently high degree of probability – a fairly true reflection of the sampling population that is being studied.
When selecting a sample you should attempt to achieve two key aims of sampling: (i) the avoidance of bias in the selection of a sample; and (ii) the attainment of maximum precision for a given outlay of resources.
There are three categories of sampling design (Chapter 12): random/probability sampling designs, non-random/non-probability sampling designs, and the ‘mixed’ sampling design.There are several sampling strategies within the first two categories. You need to be acquainted with these sampling designs – the strengths and weaknesses of each and the situations in which they can or cannot be applied – in order to select the one most appropriate for your study. The type of sampling strategy you use will influence your ability to make generalisations from the sample findings about the study population, and the type of statistical tests you can apply to the data.