Abstract:
Many differnt methods of variance estimation in complex surveys have been proposed by many researchers.Quenoville(1996)introduced the jackknife procedure as a means of reducing bias of ratio astimates, hamesen, hrwitz and madow (1953) describe the random groups procdure which is closest to the original concept of replicating sampling.the bootsrap method, introducted originally by Rfron(1979) for the independent and identically distributed case, relies on recomputing the estimate O a large number B of times by resampling the original sample. simple replication is both easy to compute and by replicating at both phase,truly independent samples are formed.chapter 1 of this project reviews the meaning of samplling survey and sample survey design.it also contains the definition of integraed survey with its advantages and disadvantages.i have includes brife notes on master sampl and its usefulness.chapter2 includes the description of the 1980-1986 nigh design. i have applied the 1987/88 survey data in order to compare their relative performance. chapter 3 sample for obtainig variance in large scale surveys. they includes the jackkinife and bootstrap methods. in chapter4 the methodes of using the means of the repliance ratio estimate has been considered in order to compare conveniently with the bootstrap method. finally chapter 5 concludes and bringe cut suggestion based on my observation.