ABSTRACT
Sampling is the selection of part of a population to represent the whole population. The law of statistical regularity states that a set of subjects taken at random from a large group tends to reproduce the characteristics of that large group. The validity of the sample depends whether the errors introduced by the sampling process are sufficiently small not to invalidate the results for the purposes for which they are required. The essence of a good sample is no ‘bias’ and minimum ‘random errors’. Bias occurs whenever the characteristic to be measured is influenced or lie in the same direction. Selecting a sample size involves a trade-off between precision and cost. Larger samples provide greater precision and cost. Stratified sampling is useful when the population can readily be divided into many groups or strata, so that each member of the population falls in one and only one stratum. Systematic sampling or quasi-random sampling consists of taking every element from subsets.
