Also found in: Dictionary, Thesaurus, Medical, Financial, Wikipedia.
nonparametric statistics[¦nän ¦par·ə′me·trik stə′tis·tiks]
A class of statistical methods applicable to a large set of probability distributions used to test for correlation, location, independence, and so on.
McGraw-Hill Dictionary of Scientific & Technical Terms, 6E, Copyright © 2003 by The McGraw-Hill Companies, Inc.
nonparametric statisticsstatistical methods used for the analysis of ordinal and categorical level sample data which do not require assumptions about the shape of the population distribution from which the samples have been drawn. Such statistics are often referred to as ‘distribution-free statistics’. In contrast to PARAMETRIC STATISTICS, assumptions underlying the use of the methods are lenient and the formulae involved are simple and easy to use. Examples are the runs tests, the signs test and Cramer's V. Although such measures are popular in sociology, they have the disadvantages that they waste information if interval data is ‘degraded’ into categorical data, and that the tests are not as powerful as parametric tests (see also SIGNIFICANCE TEST). Against this, they are often more robust, i.e. they give the same results in spite of assumptions being violated. Hence, if the assumptions of a parametric test are not met, the use of an equivalent nonparametric test will still be valid.
Collins Dictionary of Sociology, 3rd ed. © HarperCollins Publishers 2000