Discussion
The present experiment was designed to explore the idea that hidden theory based manipulations could be used as a control against normal explanations of apparent anomalous outcomes of psi experiments. In that respect the experiment succeeded. Furthermore it was hoped that a special subject working in his own home environment with no experimenter present would succeed to produce effects of a magnitude comparable to those obtained under completely uncontrolled conditions, i.e. effects that are much stronger than those that are generally observed in the rather artificial laboratory environment.
In that respect the experiment failed because the effect size of ~ 0.15 obtained for the delayed correct feedback conditions, though somewhat larger than the mean effect size of ~ 0.04 reported for forced choice lab experiments (Honorton & Ferrari, 1989), is much smaller than the effect size of at least this subject working in a non-controlled setting.
Although in the controlled condition psi apparently ocurred while the pattern in the data strongly suggest that there was no cheating, some might interpret this discrepancy between controlled and uncontrolled conditions as the result of cheating in the uncontrolled condition on the part of A.K. However an alternative explanation that suggests itself might be that the lower effect size is due to the delay in feedback, for while feedback for the a, b, c, d-sets was immediate, feedback for A, B, C, D, took three weeks to reach A.K.. Such an explanation would be in accordance with the smaller effect sizes obtained in meta-analysis of precognition experiments as compared to clairvoyance experiments, for here too the main difference between the precognition and clairvoyance conditions lies in the delayed feedback in the latter case. This explanation is also in accordance with some outcomes of recent research. In a modified Ganzfeld procedure the subject gets feedback of the target picture immediately after mentation, while staying in the state of consciousness of his mentation. The results were better (43,75%) than the expected effect size (33%) (Wezelman et al, 1996b).
If delay in feedback is an important variable then one would expect that within the sets of 10 packs with direct feedback at the end of completing the set, performance on the last few packs would be better than on the very first packs since the latter were farther away from the feedback moment. A.K. always did one series at a time in one effort. Indeed, as can be seen from table 1, there is an incline in scoring across a set with the first 5 packs of sets a, b, c, and d, resulting in a deviation of +16 while the last 5 packs of these sets give a total deviation of +41 (two sample t = 2.14, df=38, p < 0.05). The incline in itself seems not to fit in any normal explanation of the results in the uncontrolled condition. A similar incline is not visible in the over-all scoring on the delayed feedback condition sets A, B, C, and D. However if one considers set A and D, for which actually the correct feedback was given, there is also a suggestive incline in deviations from the first half to the second half of the sets (+5 and +11 respectively).
Further research is needed on the effects of delay between calls and feedback in psi studies.
Acknowledgement
We are greatly indebted for the co-operation from our special subject A.K..