< Back to Volume 17, Number 3
Volume 17: Number 3: Article 4
Pattern Count Statistics for the Analysis of Time Series in Mind-Matter Studies
Werner Ehm, Institut für Grenzgebiete der Psychologie und Psychohygiene,
Wilhelmstr. 3a, D-79098 Freiburg, Germany
A class of statistical tests is proposed for testing the null hypothesis
that a given time series is random in the sense of pattern-free. The
basic idea is to look whether certain patterns occur more frequently
in the time series than expected under the null hypothesis. The construction
and theoretical basis of the tests are an application of martingale
theory. The method is applied to data from an experiment designed to
study possible correlations between the intention of human operators
and a physical random number generator. In order to account for possible
heterogeneity of the deviations from the null hypothesis, a new combination
technique is introduced for the aggregation of individual results across
agents and/or intentions. Significant results are found with respect
to oscillatory patterns and intention "baseline," but are due to one
single operator, essentially. Some methodological issues related to
multiple testing are addressed.
Keywords: (testing for) mind-matter interrelations, random number generator,
martingale, pattern analysis, order pattern, local variation pattern,
combination of p-values
FULL TEXT:
Pattern Count Statistics for the Analysis of Time Series in Mind-Matter Studies
To purchase back issues contact Allen Marketing & Management: 1-800-627-0629
