TOWARDS A COMPLEX SYSTEMS MODEL OF PSI PERFORMANCE

Abstract

If psi is like any other human abiliry, and widely disrribured in rhe popularion, rhan psi performance may be modulated by many of the same facrors that affect other forms of human performance. To test rhis notion, 105 people participated in a series of four psi tests. Participants represented a broad range of age, belief. experience and personality type; data of 96 participants was adequate for further analysis.

A rotal of 149 variables were tracked per participant, including psi performance, psychological, sociological, meteorological, and solar-geophysical facrors. The first test relied upon background ionizing radiation as the physical target, the second and third tests used electronic noise, and the fourth test involved a riming task. A single score was used to summarize each individual's psi performance in the four tests.

An artificial neural network was trained on the data of 72 participants to see if psi performance could be predicted based on eight variables representing an aggregation of 24 environmental, sociological and psychological factors. The network successfully learned to predict the performance of the remaining 24 people (correlation between prediction and actual performance was r 0.47, p ; 0.01), suggesting that psi performance in the laboratory, and probably in life, is influenced by some of the same environmental factors that influence other forms of human behavior.