I’d never heard of that technique until now. This link on software for Ranked Set Sampling just uses a sample size formula for a pure randomly selected sample:
https://vsp.pnnl.gov/help/vsample/Design_Ranked_Set_Sampling.htm
My intuition is that a better sample size and power analysis formula would need to know the order statistics or the extreme value distribution of the data.
https://en.wikipedia.org/wiki/Order_statistic
https://en.wikipedia.org/wiki/Extreme_value_theory
Let me know if you would like to derive the formulas/ develop an algorithm together and publish a paper together.
I did not noticed any work on optimum selection of set sizes on unbalance ranked set sampling. If there is no work it can be an idea to carry out research for selecting optimum sizes under imbalance ranking. I have already worked on a proposal suggesting the same idea not completed yet.
Dear Rehman,
In the balanced RSS, the efficiency of RSS estimators typically increases with set size while total sample size is kept fixed.
For unbalanced RSS, the following references can help:
https://www.researchgate.net/publication/242997745_Optimal_rank_set_sampling_estimates_for_a_population_proportion
Article On ranked-set sample quantiles and their application
- Badges
- Science topic