ABSTRACT
T o CONCLUDE the discussion of sampling, a random sample design used in a national market research survey in 1952 will be described in detaii.I No single design can illustrate all the
problems and refinements of practical sampling, since so much depends on the population to be covered, the resources available and the subject matter of the survey. However, the sample will serve as well as any to illustrate the process which links the initial decisions on design with the final allocation of names and addresses to the interviewers.•
The subject of the survey-the readership of newspapers and periodicals-is fairly typical of the general run of market research surveys; but in one respect this survey did differ from many others. Its prime purpose was to check the accuracy of the results achieved by a quota sample survey on the same subject, and its own reliability was therefore vital. Broadly speaking, the limits to the size and precision of this sample were time (only four weeks were allowed for the fieldwork) and the number of interviewers available, rather than financial resources. We may add that the results from this random sample showed those of the quota survey to have been substantially correct.
