It's common for marketers insist on a stratified sample design to ensure adequate representation of the relevant segments for analysis, even when it would lead to a rift with random methods. Some insist on resorting to procedures as closely as possible attached to a probabilistic method and, in any case, carry out post-stratification processes if it is requiee. What is your regular survey roll, if any, and what your view about it?
I agree with Barbara and George! It depends on... The main thing is to solve business problems involved professionals.
For getting representative and random sample, stratification is better prior to the survey.
It is a scientific part of research and survey plan.
George leads the thread. I don't think any of the approaches is a priori superior to the other. It depends on the problem and data available.
I agree with George Stoica; Subhash Kundu; Napoleon Ono Imaah and Barbara Sawicka who have answered the question most adequately.
I agree with George and all other contributors .The choice of a particular sampling process may depend on the population of interest and the type of data to be derived. An expert advice maybe needed in such a situation
The "rift" with random sampling methods should not be a problem if you sample randomly from each strata, and assign the proper sample sizes to those strata.
In drawing a stratified sample, the size of your subsample from each strata should be proportionate to the size of that strata in the population as a whole. This will maintain your representativeness to the population as a whole.
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Sampling Design