Browsing by Author "Naimani, Godwin M"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    A study on the national masters’ sample for Tanzania: issues concerning sample design and quality of statistics
    (University of Dar es Salaam, 1994) Naimani, Godwin M
    This study is concerned with the design of sample and quality of statistics issues in the National Master Sample of Tanzania (NMS).The sampling design aspects such as the sample allocation, selection procedures and estimation procedures have been considered. It has been found out that the present design of NMS did not make use of optimal its criteria when allocating sampling units to the different strata. For characters which are poorly correlated with the selection probability of units the use of Amahia, et al (1989)estimator for estimating population total has been recommended instead of the usual pps unbiased estimator. The theoretical consideration on the possible use of rotation design for primary stage units (PSUs) and secondary stage units (SSUs) have been given. It has been observed that the rotation scheme which retains a fraction of PSUs in first occasion to be used in the second occasion and retains a fraction of SSUs to be used in the second occasion give smaller variance when estimating change in mean from previous occasion and mean on current occasion. But for estimating the overall mean, scheme which requires retaining only a fraction of the PSUs, with their samples of SSUs, selected in first occasion and selecting a fresh fraction of PSUs in the second occasion is more precise. The study has revealed that estimates of most of the agricultural and livestock characteristics, for AGSASU, have large sampling errors. This indicates inadequacy of the NMS design for these characteristics. A review of small area estimation techniques has been done and appropriate techniques to be used for some agricultural variables in agricultural sample surveys have been considered.