Sampling is the process of selecting a subset of individuals from a larger group of individuals, the selection being done with a view of drawing inferences about the larger group on the basis of information obtained from the subset. The larger group of individuals is known as population while the subset is known as sample. The idea of sampling is in fact very frequently used in our everyday life, although usually in crude form.
For
example: a physician tests only a few drops of blood of his
patient and then the test results are generalized for the entire volume of
blood in his body. The generalization is possible since every drop of blood is
the sample as every other drop in terms of all possible characteristics, that
is, all drops of blood are homogeneous to the extent that any drop perfectly
represents any other drop. Here the physician is selecting a perfectly
representative sample of a few drops of blood from the population of all drops
of blood conceivable in the patient’s body, and then based on the results of
the test of these sample drops, is drawing inference about the entire
population of blood drops.
Likewise,
a rice dealer testing a handful of rice from a sack to determine the quality of
the whole lot of rice in it, or a cook pressing a few boiled grains between
fingers to declare whether they are ready to be served, are all selecting
samples, like the physician, from extremely homogeneous populations. Because of
this homogeneity, any sample is enough to draw a valid conclusion about the
total population. When the population is not homogeneous, inference based on
just any portion from this population cannot be so easily taken as valid.
For
example, if a customer encounters high prices of commodities
in a few shops in a large shopping complex comprising a hundreds of shops and
then based on his experiences, infers that the prices of commodities in the
whole shopping complex are higher than usual, it would be difficult to take his
inference as granted. This is because, the shops in the complex are likely to
vary in prices of commodities unlike for example, the drops of blood in a body
which are all alike. A drop of blood in a body is the sample as any other drop,
but a shop may or may not be the sample as any other shop in terms of prices of
commodities. In other words, population of blood drops is homogeneous while the
shop population in terms of prices of commodities is not. For a homogeneous
population, whatever method is used to draw the sample, it represents the
population and hence inference based on it can be taken as valid. For a
heterogeneous, population, on the other hand, in order to choose a
representative sample that should essentially contain the same variation as in
the population so that the inference based on it can be taken as useful to
certain degree of confidence, the method of choice becomes important. A number
of methods based on varying circumstances have been developed to obtain
representative samples to enable the researcher to draw sound inferences about
the population on the basis of the sample results.
Source: Methods and Techniques of Social Research by Abu Jafar Mohammad Sufian