Survey Research

Scientific Bases of Survey Research

Survey research is based on sampling techniques founded on probability theory. People attribute different meanings to the concept of probability. Most often it is interpreted as the relative frequency of events, such as the number of red or black cards in a deck. Its precise meaning is centered on the concept of a “normal distribution.” The most important characteristic of the normal distribution is that even when the actual values of a variable do not have a normal distribution, the mean of that variable estimated from a large sample can be regarded as having come from a normal distribution of sample means. This is what the central limit theorem states. If researchers draw a large number of independent random samples of any variable, the mean value of the means obtained in each of the random samples will be close to the actual mean of the population. Moreover, they will be able to calculate the standard error of the mean. Hence, researchers can draw samples of populations and infer that the mean of the variables observed will be close to the mean of the population with a known margin of error.

Good, reliable surveys follow basic scientifically founded rules to select the sample of the population [Page 2567]to be studied. First, the researcher has to precisely define the population he or she wants to study. The sample of individuals to be interviewed is selected in such a way that every member of the total population has a not necessarily equal, but a known, chance of selection. If all members of the total population have the same chance to be selected, every individual in the sample represents the same number of persons in the population. These are self-weighted samples. If the selected individuals have different chances of being selected, weights are used to estimate the number of persons in the population they represent.

...