In the earlier post we said that if a large population is to be considered some individuals from it should be selected that represent the population. How does one achieve that? How do we make sure that the sample we are choosing is representative of the population? This post is going to deal with exactly that – The Representativeness of a sample.
While choosing a sample, it is important that we define the characteristics of the population which should appear in the sample, i.e. we should know what it is that should be represented in the sample.
For example if a person wants to study the opinion of students on the current education system, in this the only characteristic of the population that is observed is that the population is going to be of students, but then selecting the sample becomes very difficult as there are so many students, nothing is specified with respect to their location, level etc, so technically the whole possible population of students existing comes into picture. To avoid such a situation one should try to present relevant characteristics of a population. On the other hand it should also be kept in mind that considering too many characteristics will make choosing a sample difficult as each item in the sample will need to exhibit those characteristics. In this case it could begin by stating characteristics that are going to be dealt with by the researcher, for e.g. the academic level of students, their intelligence level, their location, their gender etc.
As we go on listing the characteristics of the population we should also decide up to what extent the sample must represent the population. It means we need to decide whether all or some of the listed characteristics should appear in the sample. Ideally if all the desired characteristics appear in the sample it would render a highly representative sample. But this may not be possible sometimes due to limited costs or convenience.
Representativeness of a sample has implications on the generalisability of results, because if a sample is not representative of a population, then the results for that sample might not generalise to that population.
Or it could also be said that the results of that sample will be applicable only to that sample and other samples similar to them.