ABSTRACT
Social researchers usually begin with a plethora of hypotheses and half-formed ideas. Statistical tests of hypotheses, however, seem to be of quite limited aid in building theoretical social science. There is no intention of suggesting that statistical inference is generally irrelevant in social research. It is certainly relevant for descriptive studies, to attach some measure of precision to statements made about a population on the basis of measuring a sample. Statistical tests are used for a number of different purposes. One use is to indicate the precision of a descriptive statement about a population. When a random sample of a population is measured in terms of some attribute, then the sample distribution is used to make a statement about the population distribution. The statement may be made in terms of confidence intervals or in terms of a mean value with certain limits.
