inferential statistics that assume that the population from which the SAMPLE has been drawn has a particular form, i.e. they involve hypotheses about population parameters. These assumptions are generally that the populations involved have a NORMAL DISTRIBUTION, that they have equal variances (see MEASURES OF DISPERSION) and that the data are at interval level (see CRITERIA AND LEVELS OF MEASUREMENT). Examples are the PEARSON PRODUCT MOMENT CORRELATION COEFFICIENT, multiple regression, and analysis of variance. Such procedures use all available information and tests are more powerful than nonparametric tests. In sociology, the problem of data that are not normally distributed in the population frequently arises. A transformation of scale, a reliance on the robustness of the technique, or a move to a nonparametric equivalent are the available solutions. Compare NONPARAMETRIC STATISTICS.