The probability that the over-all deviations in the direction of intention from chance expectation that have been found are due to chance are smaller than 5%. Therefore we may conclude that paranormal effects have been obtained in this experiment.
The enhancement in effect size, if the apparently most sensitive hit count of individual samples is replaced by the the hit count on the majority vote of hits over run pairs is larger than can be expected on basis of mathematical considerations. Similar effects have been observed in a number of other experiments (Brier & Tyminski, 1970; Houtkooper, 1967) and may be interpreted in terms of the goal oriented character of psi.
The scoring on the prerecorded samples is important for the experimental verification of the Observational hypothesis of Psi Phenomena. The present experiment is to be considered as an exploratory investigation rather than a crucial test because the direction of intentional efforts of the subjects (hitting or missing) cannot yet be predicted with any reliability. The result of the prerecorded samples in part A were only observed by the two experimenters. These samples do not show any deviation form chance expectation. On the other hand the prerecorded samples of part B were observed by the subjects at a later time in which the goal specified by the computer was opposite to the goals specified during production and recording of the samples. According to a traditional 'force-like' model the samples are biased during production while according to the observational hypothesis they will be biased 'retro-actively' during observation. The results seem to support the latter model because the deviations were into the direction specified during observation rather than during production.
We have compared the efficiency, expressed as the normal score squared divided by the duration of the experiment in this study, with that of efficiencies obtained in Schmidt's experiments. From the table it is obvious that the efficiencies in the present experiment are rather low.
Schmidt, however used selected subjects in his experiments. We prefer to work with unselected subjects in order to keep generalizability as high as possible. It is unknown how much time Schmidt invested to trace these subjects. Furthermore his procedure was such that experimenter and apparatus had to be available any time the subjects considered themselves to be 'hot'. Our experiment took 7 hours in total including the introduction of the subjects to the experiment. As we took for each subject just half an hour, the atmosphere was rather impersonal and hurried. It is our feeling that a further study with Ss 2,3 and 14 may enhance the efficiency to a level comparable to that found by Schmidt. Further research with systematic manipulation of sample frequency, run duration and 'observational history' is needed to further the theoretical understanding of the apparent psychokinetic effects on Random Number Generators as found in the present experiment.