My research topic is to analyse the perceptions of tax payers on the new indirect tax introduced by the government.
i want to apply cluster sampling. i have the data of tax payers in the city.
how should i frame clusters and how to draw sample.
thanks in advance
I assume you are using probability-of-selection-based (i.e., design-based) sampling and Inference, selecting clusters with either equal or unequal probabilities, depending upon your design.
In stratified random sampling, you consider each stratum as different, and you need to sample from each. When you use the less efficient cluster sampling method, you consider each cluster to be a member of a population made up of these clusters, and you can sample them without worrying about randomly leaving out entire clusters, as those clusters are the sampling units. (You can then either census each probabilistically selected cluster, or randomly sample from within each selected cluster.)
To form such clusters, you generally consider geography, because travel cost considerations are usually the reason for cluster sampling. At any rate, you form clusters based on what is logistically advantageous, as long as you give every cluster a positive probability of being selected. (For example, putting aside some clusters you expect will not give you a favorable answer, and making sure you do not select them would be a disqualifying procedure.)
thank you prof. James R Knaub for your detailed explanation to my question.
Dear prof. James R Knaub
From your reply, I understood that each cluster is similar to one another and so, we can select a few clusters randomly.
but, as i understood in cluster analysis, a multivariate technique, each cluster represents a unique profile of the sample units on a set of variables.
if so, how can we select a few clusters randomly leaving other clusters.
i am very much confused with the meaning assigned to clusters in cluster sample and in cluster analysis.
please clarifty my doubt.
- Badges
- Science topic
- Similar topics
- Computer Science
- Research Topics