Sampling Methods

Sampling methods or sampling techniques.

1. Probability Sampling Techniques
   Simple Random Sampling
   Systematic Sampling
   Stratified Sampling
   Cluster Sampling
2. Non-probability Sampling Techniques
   Convenience Sampling
   Judgmental Sampling
   Snowball Sampling
   Quota Sampling

Probability Sampling Techniques

Simple Random Sampling

In this sampling technique, each member of the population has an equal chance of being selected. Samples are randomly drawn from the population.

There are two approaches: Simple random sampling without replacement (SRSWOR) and Simple random sampling with replacement (SRSWR).

When to use simple random sampling? When the population is homogeneous or when the population is smaller. 
However, when the population is heterogeneous, dispersed or large, other sampling techniques are more appropriate.

Systematic Sampling

Let us understand this with an example. If the population size (N)=100 and sample size (n) = 20, then sampling interval (k) = N/n, k=100/20=5.

A random starting point between 1 to k (5 here) is selected - let us say 3. From 3, every kth element (5th in this case) is selected. That means our sampling units are 3, 8, 13, 18,.......93, 98.

So systematic sampling involves selecting every kth element from the population after establishing a random starting point.
Systematic sampling is particularly useful when the population is ordered, spatially distributed, or time-based.
E.g. Sampling every 10th house on a street, sampling every 20th customer, etc.

Systematic sampling is not useful when there are hidden periodic patterns in the population. For example, if every 10th house is a corner house, then our samples would be either corner houses or all non-corner houses if the sampling interval (10 in this case) coincides with that periodic pattern.

Stratified Sampling
The population is subdivided into homogeneous subgroups or strata (plural of stratum). Then samples are drawn from those strata. 

For example, if in a class of 100 students, there are 60 boys and 40 girls, then the population is subdivided into strata of boys and girls. If the sample size needed is 10, then 6 samples are drawn from boys and 4 from girls, depending on the proportion (proportional stratified sampling). Sometimes, disproportional stratified sampling is also used. 

Cluster Sampling

The population is subdivided into clusters. Then, simple random samples of these clusters are taken. And from the selected clusters, samples are drawn.

Difference between clusters and strata:
Clusters are mutually homogeneous but internally heterogeneous.
Strata are mutually heterogeneous but internally homogeneous.
Examples for clusters could be a geographical area. E.g. states or districts could be clusters.

Non-probability Sampling Techniques

Convenience Sampling

In this most common non-probability sampling technique, samples are drawn based on what is convenient to the researcher. May be sample respondent is nearby, easily accessible or in the same office. Selecting friends or coworkers is an example. 

Judgmental Sampling

Samples are drawn based on the researcher's judgement on what could be the representative sample for the study

Snowball Sampling

As the snowball rolls downhill, its size increases. Similarly, initial participants are asked to refer other participants in this sampling technique.

Quota Sampling

Initially population is segmented into subgroups. Based on specified proportion, samples are drawn for each subgroups until the quota is reached.