Abstract:
In this paper, multistage median ranked set sampling (MMRSS) is used for esti-
mating the median of a distribution. Five estimators of the population median
based on MMRSS are proposed. The performance of the new proposed estima-
tors are compared with their conventional counterpart in simple random sampling.
The comparison is carried out by a simulation study based on two measures of
performance; bias and e ciency. Moreover, the impact of the set size, number of
cycles, and the number of stages on the e ciency of the proposed estimators is
studied.
