AGAPÈ: GROUP TELEPATHY. A LONG-TERM EXPERIMENTAL SERIES

We hypothesized that a majority vote process would favor the ESP success between two groups by enhancement of the signal to noise ratio. We supposed that the repetition of the trials and the use of an individual and group feed-back could improve the telepathy between these two groups of individuals, and obtain results higher than chance expectation. The participants consisted of 418 unselected volunteers, who participated either in the ‘agent’ group or the ‘percipient’ group. The agent group and percipient groups were each placed in acoustically isolated rooms. The size of the agent group varied from 0 to 15 persons and the size of the percipient group from 1 to 16. We also varied several target parameters: target-type (pictures, words), and size of target packs (judging packs consisting of 2, 3 or 5 targets). For each trial, the computer randomly selects a target from a database of pictures or words, and displays it to the agents. In an isolated room, percipients had to guess the target among randomly selected decoys, using an individual keypad connected to the main computer. Each individual guess, the target, the date and hour of the trial, were recorded on the main computer. Over the course of seven years, participants contributed a total of 240 sessions, involving 27,845 collective trials, which represent over 250,000 individual trials. The Overall hit rate was at chance, both for individual trials and for collective (majority vote) trials. Some interesting variance results are noted, though their interpretation remains ambiguous. The seemingly randomness of our data could belong to the Type II error. This is why we undertook a post hoc covariance analysis on the collective trials which deviated significantly from expectation (p < .05). This study made it possible to highlight some interesting parameters, but the variable linked to better results are also linked to worse results. Because that could be linked to a stacking effect, we used a resampling method to control it. The creation of faked targets allowed us to select new significant collective trials. The regression conducted on these resampled data gathered results very similar to those excerpted from the genuine data. We observed the same thing about the sociometric data. Indeed, the replacements of the genuine targets by fake targets does not fundamentally change the results of the covariance analysis: Thus, we are most probably facing a stacking effect.