Abstract:
In the present article, we suggest nonlinear regression models for variance function in cluster sampling. The estimates of parameters of suggested and existing models have been obtained. The suggested and existing models have been fitted to data sets given in Sukhatme et al (1984) and Zakkula Govindarajulu (1999). An improvement has been shown over existing models in the sense of best fitting, that is having lesser residual mean Square S2 for the relationship between variance within cluster and size of the cluster M.
