Random selection in ranked set sampling and its applications (Q1298897)

The concept of \textit{rank set sampling} is to take \(n\) samples each of size \(n\) and to base statistical inference on \((X_{11},X_{22},\ldots,X_{nn})\), where \(X_{jk}\) is the \(j\) th order statistic in the \(k\) th sample. The new concept random selection in ranked set sampling is to base the inference on \((X_{1k_1},X_{2k_2},\ldots,X_{rk_r})\), where \((k_1,k_2,\ldots,k_r)\) is a random selection of indices \((1,2,\ldots,n)\), \(r\leq n\) fixed. Estimators of the population mean and population variance are proposed and their efficiency is discussed. Applications to normal, exponential, and logistic distributions are given.