Larger sample sizes and cluster sampling are often used in household surveys to produce reliable estimates of such indicators while planning and such large sample survey requires a sound system in the organization especially when resources are limited. As the staffing requirement would be much less for a smaller sample size, it would be possible to use well trained and motivated enumerators, specialized equipment and exercise, intensive supervision, all of which are expected to improve the quality of the data collected and the accuracy of results. BDHS used large sample to produce very précised estimates for most of the socio economic indicators. Four reports of BDHS state basic national indicators of social progress and the S.E. these indicators remain unchanged or have little variation. Increasing of cluster in each time of their survey, the benefit of these increases in terms of variance was not analyzed. This book attempted to examine the sample size in terms of number of clusters or the PSUs/precision measured by standard error (S.E.) or coefficient of variation (CV) of some major indicators usually attempted in a household, socio-economic and demographic survey.