Stratified Sampling vs Cluster Sampling

Stratified sampling is a better representative of the population. Stratified sampling is a better choice of sampling method when you anticipate that the subgroup you are studying will have different mean values for variables.  

Let’s look through a stratified sampling example to see when it’s used.

Example 1: In the case of a political survey or say, a socio-economic class survey, you can use stratified sampling.

As a researcher, you would need to include participants of various groups like race, religion, or economic class in the case of the overall political survey. In the socio-economic survey, you can make a stratified sample for each category, lower class, middle class, and upper class. 

However, you need to ensure that the stratified samples are based on their proportionality to the total population. This will guarantee a total representation of the entire population while also reducing any sampling errors. 

Example 2: You want to understand what effect an MBA degree can have on the income gap between different gender identities in your company. Relying on the employee list you figure only a small proportion of company employees are MBA graduates. So, you use stratified sampling to compare the difference between men, women, and other gender identities with an MBA degree against those without one.