Does cluster sampling depend on the order of clusters in the sampling frame? Won't there be a selection bias? For example, if the clusters are ordered/numbered alphabetically, the selected clusters would be different from if the clusters are ordered/numbered ascending or descending.
Dear Avinav Prasad Maddury,
Sampling will begin at a randomly selected starting point. So, we choose a random number as our starting between 1 and the sampling interval should be total population divided by the number of clusters you need.
Hence, cluster sampling does not depend on the order of clusters in the sampling frame.
I hope it helps. All the best !
Dear Avinav Prasad Maddury,
Sampling will begin at a randomly selected starting point. So, we choose a random number as our starting between 1 and the sampling interval should be total population divided by the number of clusters you need.
Hence, cluster sampling does not depend on the order of clusters in the sampling frame.
I hope it helps. All the best !
Your question is incomplete in my opinion. For instance, how are you selecting clusters? Simple random sampling of clusters? systematic sampling of clusters with a random start or sampling with probability proportional to cluster size? The only sampling method that I believe may be impacted by order is systematic sampling with a random start. Even then, I am not sure that can result in a bias. You can verify that with a simple simulation study by repeatedly sampling clusters of known response in order to find out if there are any differences between orders. As you have been told already, there shouldn't be any difference as long as numbering order does not modify the inclusion probability of a particular cluster or clusters.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis