Balanced unequal probability sampling with maximum entropy
(English)Manuscript (Other academic)
This paper investigates how to perform balanced unequal probability sampling with maximum entropy. Focus is on balancing conditions having the form of known marginal sums in a cross-stratification table. Since only marginal sums are fixed, the sample sizes for one or more cells in the table are random. The probability distribution for those sample sizes can be expressed explicitly but there are computational difficulties except for very small cases. Markov Chain Monte Carlo methods are proposed for obtaining good distribution approximations, as well as sample selection. Some large-sample Gaussian approximations are also considered. Iterative procedures for obtaining sampling probabilities yielding specified inclusion probabilities are discussed.
Probability Theory and Statistics
Research subject Mathematical Statistics
IdentifiersURN: urn:nbn:se:umu:diva-22457OAI: oai:DiVA.org:umu-22457DiVA: diva2:216704