Stratified Random Sampling

It is another restricted type of random sampling in which the different numbers of samples are drawn at random from different strata, or divisions of the universe. For this, the entire universe is first divided into certain numbers of strata on the basis of certain known as stratifying factors such as age, sex, income, education, status, geographical area, economic condition and sociological character, While dividing, and sub-dividing a universe into certain strata care must be taken to see :

(i)  That there is a remarkable heterogeneity between the various strata;

(ii)  That there is a remarkable homogeneity between the different units of each stratum ;

(iii) That there is no overlapping between any strata which means that no unit of the universe finds place in more than one stratum or subdivision, and

(iv)  That the number of the strata, or sub-divisions is not too large, and it remains preferably within 9 or else, it may give rise to various complicacies.

Size of the Sample

After the universe is divided into a large number of strata the next thing to do is to fix the size of the sample for the entire universe in a reasonable manner. After this, the next important thing to be done is to allocate the sample size to the various strata from which the data are to be collected. Such allocation can be done in any of the following manners:

(i) Proportionate manner

(ii) Disproportionate manner

(iii) Disproportionate manner