SAMPLE DESIGN

The sample design influences the precision of the subset of measurements obtained from the sample.   Sampling strategies can be categorized as random probability, non-random probability and mixed sampling.   A random probability sample design provides for an equal and independent chance of selection, i.e. each individual has a known probability of selection set by the sampling procedure.   Non-random probability sample designs are used when the population size is either unknown or cannot be discretely identified.   A mixed sampling design has characteristics of both random and non-random probability sampling.   Systematic sampling is the most common strategy used, based on a known population with the sample size determined a priori.   The advantages of systematic sampling over simple random sampling is that it is more efficient, in terms of information per respondent cost, and easier to perform thus reducing sampling error.   Systematic sampling will generally be an adequate form of random sampling to the extent that the placing of any sampling subject or unit is independent of the placing of other sampling individuals or units, i.e. there is no systematic bias introduced into the listing of sampling units.   Should such a risk of sampling bias be known, it may be avoided by choosing a suitable sampling interval; or, after a predetermined number of units have been drawn, a fresh random start can be made.

Thanks