Abstract/Results: | RESULTS:
Despite the high-quality research on mediumship a century ago, this line of research has been almost entirely neglected since then. The purpose of this study was to identify contemporary mediums who can produce evidential material under controlled conditions.
To rule out the two common normal explanations – that the medium fishes for information (“cold reading”), whether deliberately or inadvertently, or that the medium’s statements are so general or vague that they can apply to many people or be interpreted in a variety of ways – the present study was conducted using double-blind conditions. All sittings were done, not with the “real” sitters, but with a proxy sitter who knew little or nothing about the deceased persons, and, because the real sitters were not present for the readings, their judging of transcripts was blind.
In Stage 1, 4 mediums and 12 real sitters participated; each medium did 3 readings. The results were not significant. In Stage 2, therefore, I made 3 major changes: (1) In the hope of focusing the reading on the intended deceased person, I sent photographs of the deceased to the mediums, but provided no other information; (2) to see whether the proxy might play some role in the success or failure of a reading, another experimenter and I shared the role of proxy, each doing half the readings; and (3) I simplified the evaluation method and asked the real sitters to rate and rank transcripts globally rather than item by item.
In Stage 2, 9 different mediums and 40 different sitters participated; each medium did 4 or 6 readings. Each sitter evaluated 6 transcripts – the intended one, as well as 5 controls. 37 ratings were returned. 14 of the 37 were ranked #1; 27 were ranked in the top half. Analysis of the results with a sum-of-ranks method gives a z score of 4.09 (p < .00003). The person serving as the proxy sitter had no effect on the results. In the 19 readings for one proxy, the sum of ranks was 46.5 (.005 < p < .004); in the 18 readings for the other proxy, the sum of ranks was 40 (p < .001). The difference is not significant.
|