In most marketing research, it is seldom necessary to conduct a complete census; i.e. to talk to 100 per cent of the target segment. To do so is time-consuming and expensive. For this reason most marketing surveys make use of samples. A sample is a group of elements (persons, stores, financial reports) chosen from among a "total population" or "universe". The value of a research project is directly affected by how well the sample has been conceived and constructed.5
The selection of the sample to be investigated requires a master list, or a framework, from which they may be selected. The sampling frame is the "population" or statistical "universe" from which the sample units will be selected. The frame for a survey of attitudes of credit customers of a department store would be the company's list of customers using charge accounts.
Although there are many kinds of sample designs, all of them can be classified as either probability samples or nonprobability samples. In a probability sample, each unit has a known chance of being selected for inclusion in the sample. Its simplest version is the simple random sample, in which each unit in the sample frame has exactly the same chance of selection. Examples of this include flipping a fair coin, whose sides have a 50 per cent chance of turning up and throwing an unloaded die, whose sides have a 16 ⅔ per cent chance of turning up. This same principle can be applied to the previous department store example. A sample of names could be selected from the company's list of charge customers according to a random process, such as that of using a table of random digits.
While in a probability sample the sampling units have a known chance of being selected, in a nonprobability sample the sampling units are selected arbitrarily. To return to our department store example, instead of using a table of random numbers to select a sample of charge customers, an arbitrary and more convenient method would be to take the first 50 or 60 names on the list.