is able to represent or stand for the entire population. Then the sample is studied and the results obtained for it are generalised to the entire population. The statistical properties of a scientifically selected sample ensure that the characteristics of the sample will closely resemble the characteristics of the population it is drawn from. There may be small differences, but the chance of such deviations occuring can be specified.
This is known as the margin of error, or sampling error. It arises not due to any mistakes made by researchers but because we are using a small sample to stand for a large population. When reporting the results of sample surveys, researchers must specify the size and design of their sample and the margin of error. The main strength of the survey method is that it is able to provide a broad overview representative of a large population with relatively small outlays of time and money.
The bigger the sample the more chance it has of being truly representative; the extreme case here is that of the census, which includes the entire population. In practice, sample sizes may vary from - to many thousands. (See the box on the National Sample Survey). It is not only the size of the sample that matters; its mode of selection is even more important.
Of course, decisions on sample selection can often be based on practical considerations. In situations where a census is not feasible the survey becomes the only available means of studying the population as a whole. The unique advantage of the survey is that it provides an aggregated pictur e , that is, Activity Discuss among yourselves some of the surveys you have come across. These may be election surveys, or other small surveys by newspapers or television channels.
When the results of the survey were reported, was the margin of error also mentioned? Were you told about the size of the sample and how it was selected? You must always be suspicious of surveys where these aspects of the research method are not clearly specified, because without them,