how to understand the difference between cluster samplying and stratified sampling? can anybody explain it with a simple illustration.
thanks in advance
Both are a form of systematic sampling, and considered probabilistic (random sampling).
Stratified sampling, from the name, is when you enroll a sample according to a specific criteria. For example, if there are 100 individuals in a room (30 boys and 70 girls) and you want to randomly select 20 sample. Stratified sampling based on sex would require you to select 14 girls and 6 boys—this way you’ve maintained the original gender distribution in your sampling. This is called proportionate stratified sampling. If you pay no mind to the original gender distribution and decide to take 10 boys and 10 girls, that’s is non-proportionate stratified sampling.
Cluster sampling on the other hand is when you have a large population say 50 districts and you want to study say behavioral trait and you can afford only 5 districts in your sampling so you randomly select 5 districts from the 50. If you go further to select sub-districts from the 5 districts in your sampling that is called multi-stage cluster sampling.
If you don’t understand, let me know so I can send you a self-explanatory diagram.
You do cluster sampling as an approach to efficiently increase sample size while sticking to random sampling procedures. Cluster sampling is the more problematic the more homogeneous your clusters are and the more differences there are between clusters.
One motiviation for stratification as a sampling strategy is to oversample small population groups in order to have sufficient power for statistical estimates of these subgroups, and sufficiently large numbers for building complex models. Stratification is all the more useful/necessary the more homogeneous "strata" are, and they more they differ from each other. So in a way, it`s just the other way round ...
Both are a form of systematic sampling, and considered probabilistic (random sampling).
Stratified sampling, from the name, is when you enroll a sample according to a specific criteria. For example, if there are 100 individuals in a room (30 boys and 70 girls) and you want to randomly select 20 sample. Stratified sampling based on sex would require you to select 14 girls and 6 boys—this way you’ve maintained the original gender distribution in your sampling. This is called proportionate stratified sampling. If you pay no mind to the original gender distribution and decide to take 10 boys and 10 girls, that’s is non-proportionate stratified sampling.
Cluster sampling on the other hand is when you have a large population say 50 districts and you want to study say behavioral trait and you can afford only 5 districts in your sampling so you randomly select 5 districts from the 50. If you go further to select sub-districts from the 5 districts in your sampling that is called multi-stage cluster sampling.
If you don’t understand, let me know so I can send you a self-explanatory diagram.
thank you Mr. Mohammed M. Alhaji
you have explained the difference between the two terms in simple and very clear manner.
can you please the self-explanatory diagram also?.
once again thank you for your reply
Simply the difference is that stratified sampling is to choose samples from a level or strata, such as from different age groups (20-25, 26-30, 31-35, 36-40), gender (male and female), education (elementary and upper), whereas cluster sampling is to choose samples from units that could be based on, such as cities and districts.
A helpful explanation has been provided by Sekaran & Bougie(2016).
Ref:
Sekaran, U., & Bougie, R. (2016). Research Methods for Business A Skill-Building Approach (7th ed). John Wiley & Sons.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis