MULTISTAGE & CLUSTER SAMPLING

Multi-stage Sampling:  
This  sampling procedure is carried out in several stages. The population is first divided into large groups known as first stage units.  Every first stage unit is divided into smaller groups, known as second stage units.

  These second stage units are further divided into third stage units and so on until the final stage units are reached.  A sample from the first stage units will be selected by any suitable method such as random sampling. Then a second stage unit is selected from each of the first stage units again by applying a suitable technique (the technique applied may be the same as in the case of selecting first stage units or an altogether new technique can be employed). This process is repeated at every successive stage until the final sampling units are drawn.

• This technique is considered highly scientific as selection procedures follow the scientific norms at every stage.   
• There is little or no chance to ignore any character of the universe. 
• Stratification of the universe at different stages helps identify all the significant homogeneous characters of the universe.   
• This is the most accepted procedure, especially when enquiry is to be made, based on a sample drawn from a comparatively large universe.  For example, to introduce a new agricultural scheme in 1000 villages in a state with 30,000 villages we need identify sample villages apply three-stage sampling.
• In the first stage, the state is divided into a number of districts and first stage selection of districts will be made on a random basis applying some acceptable probability sampling techniques. As a result, a few sample districts can be drawn.
• In the second stage, these districts are divided into mandals so as to make second stage selection of mandals. A few mandals are selected following the scientific statistical sampling techniques. In the third stage, these selected mandals will be subdivided again into villages so as to make finally the third or the final stage sampling units, viz., and villages.
• At every stage, sampling procedures are clearly followed to ensure that there is no element of bias in selection. Further, this sort of sampling introduces flexibility in sampling method. This will help cover a large area of the universe, stratification at each stage will lead to accurate sub-division and creation of reasonably small sampling units.

CLUSTER SAMPLING

1. Certain categories of units are concentrated in certain areas or geographically located in certain areas.  
2. The cluster sampling technique is a relatively easy device to select sample units on the premise that units belonging to one category exist at a specified geographical location.
3. It is presumed that units available at a locality or area belong to one homogeneous group. 
4. Thus, by applying simple principles of random sampling the sample units can be select easily so that they belong to the desired or specified category. 
5. This technique can be applied in two stages for selection of a sample. In the first stage, various homogeneous groups are selected from the population depending on the nature of inquiry. 
6. These are termed as primary sampling units. In the second stage, elementary sampling limits are selected from each of these clusters (groups). 
7. The technique of selecting sample observations from each cluster is known as sub-sampling. On geographical consideration, if specified groups exist at specified locations and random sampling technique is applied when selecting sample observations accordingly, it is then termed as area sampling