A short manual on sampling

As its title implies, this manual is not designed for the sampling expert; nor is it a substitute for the excellent textbooks which are listed in the bibliography and which must be consulted for a proper appreciation and understanding of the technical procedures indicated here. The manual owes its origin to a close study of the methods actually in use, mainly by government statistical offices. These methods confirm to but a few patterns and it is assumed that these patterns, largely imposed by circumstances, will, in the main, persist. Though cost and local administrative arrangements usually act as constraints in choice of method, there have been many inquiries studied in which it seemed likely that increased efficiency could have been attained with closer attention to the tenets of sampling theory and practice, with little or no increase in expense. This manual is then designed for statistical authorities, regarded as non-mathematicians, who are aware of statistical shortcomings in their countries which could be made good by the use of the sample survey method and whose responsibility it is to make decisions to this end. While formulae for estimation and for sampling error variances associated with common processes are displayed, mathematics are generally avoided. Certain mathematical aspects are, however, dealt with in appendices.