Over randomness
In the inferential statistics, a relationship is described as over- randomness whose proven significance is high enough to only occur with a low, predetermined frequency if it occurs randomly . A direct correlation does not necessarily follow from over-randomness .
The term comes from the category of time-series analytical procedures in the field of descriptive (i.e. descriptive) statistics, but is sometimes also used outside of scientific literature for a process whose occurrence has not yet been shown to be regular, but which can nevertheless be observed more frequently. than would be expected from the statistical normal distribution . Over randomnesses are determined with the help of statistical test procedures, e.g. B. with the chi-square test or with test procedures for cross-spectral analysis .
swell
- Over randomness in the test procedure for cross-spectral analysis in the article by the University of Frankfurt ( Memento from December 11, 2012 in the web archive archive.today )
literature
- Markus Pospeschill: Practical Statistics . Psychologie Verlags Union, Weinheim 1996, ISBN 3-621-27345-X .
- Thomas Köhler: Binomial Distribution; Checking of randomness and "over-randomness" by means of probability calculation . Kohlhammer-Verlag, 2004, p. 116 ff., ISBN 3-17-018183-1 .
Web links
- Hartmut Wewetzer: "It could also be coincidence" - the head of the Institute for Biometry at the University of Mainz on the plausibility of connections, over-randomness and the consequences of the last study on nuclear power plants. Tagesspiegel from December 10, 2007.