Science and Spirit

In the banter of words between religions and science through the ages, the only thing that could be said was "Science will prove God." The added piece to that needs to be. "When God knows man can accept the knowledge and when to have it used in the right context for all, human, animal, mineral, the planet.The universes.

As a working medium I look to all areas of proof. Why should I be any different? I often say to mediums, we need to look to improving our comunication and information given to those in need. Platitudes are not enough.
I went to a church service recently, from the platform the medium said. " I have a man in uniform, was in the services." So, I said "Which one?" All three were represented in our family."The Army, and he had medals which were not talked about." "Yes." I said I do know of such a man in the family. "He was in the SAS." "No he wasn't,I said "It was not formed while he was in the forces."
The asumption, without proof of names etc, and the relationship means nothing. Spiritualist Churches are always complaining of such standards we need to improve, I am no different.

Paul Rowlands, a friend I have worked with in varied paranormal situations is an inventor of paranormal related equipment. Visual, audio, and electricty (aura and current natured). Having 5 of these machines, one connected to a plant and which can trigger a photograph if anything moves passed it. Two of which are crystal orientated, others which are electiric coil motivated, all of which can be linked to visual and audio devices, for human or spiritual detection.

A group of people were asked to sit with the equipment, and the results were that people could actualy get the light indicators to move right to left. the plant would register movement in anything up to four foot away.
Closer testing revealed varied readings with different people. the same with the crystals, but not in such a wider distance range. The electric coils not so conclusive. Not forgetting that the human body is contolled by the elecricty which eminates from the brain. the aura, our energy fields. That by which we can "read" each other.

In the tests we did not touch the objects and those who could move the indicator lights, well make up you own minds if it's a form of kentics, or as I think we are looking at in a basic manner. "The inter- connectedness of all, human, mineral plant, the planet." That which connects all things, makes it our brother and sister, no matter what.

I was kindly recommended to the people doing the Trisha's Great Psychic Challenege, on Channel 5. having attended, took some of the supposed tests. I was dismayed by the lack of use of the right words. Psychic/ Mediumship. Our auric field is connectedto the chakra system. The bottom three in the main system , the psychic, the earth, our anchor while in the physical. the strongest being the base, the Kundalini. That which the ancients knew had to rest in the heart centre in balance to achieve, the "enlightened" state. The chakra's from the heart centre up operate the spiritual aspect of man. The chakra's above move in their gentle way to bring this about when the love and warmth of the heart is in motion. None of this was exlained when the "contestants" were using psychometry, the art of scrying with objects, from the psychic energies they hold, to the spiritual information which can be given.

I was offered a part in the second series, should there be one. I have declined. No explanations, just a type of X Factor thing, to make it look as if those taking part had come through some sort of testing? Doh!!! as Homer would say.

I have now been asked to partake in experiments at the Hope University, Liverpool. This I'm sure will be on the same scientific lines as work done with Paul. Under the auspices of Cairan O Keefe in his capacity of the Crime, Psychology, Paranaormal Dept of that University. Results good or not so good will appear here.

"Here I am at Hope University.Liverpool 25.01.06 I do look as if I'm asleep, but I promise you this was a different afternoon. From 150pm until 530pm I am assured there were 5 different people I could not see or hear in another part of the University, who could hear the information I was giving to them on a mediumship level. This was controlled by Steve a paranormal investigator by walky talky when we needed to stop or start the readings and his conterpart with the sitters in the other building. I await with interest the results from Cairan O'Keefe and his co scientific investigators. And, as stated will produce them in honesty here, Good, Bad, or Indifferent"

ASSESSING MEDIUMSHIP
Brief Report

(Authored by Ciarán O’Keeffe ©2006)

Background

During a traditional mediumistic reading the medium usually produces a large number of statements and the sitter has to decide whether these statements accurately describe themselves or departed loved ones or relatives. In order to assess the extent to which such statements may provide evidence of the possibility of post-mortem survival, it is important to employ protocols that are effective in ruling out alternative explanations (such as chance or ‘cold reading’). The proposed research intends to build upon recent studies that have taken steps to rule out these kinds of alternative explanations yet still yielded results that support mediumistic claims.

Early Work

Within the archives of the SPR, there are literally hundreds of journal articles and proceedings explicitly devoted to the examination of the verbal material produced by mediums (as well as physical phenomena). Many of these studies are purely descriptive, primarily the experimenter’s subjective assessment of readings given in séance scenarios during which the experimenter would be present. In these situations the experimenter would often be the target person, and little attempt would be made to prevent the medium divining information via alternative means (e.g. Hodgson, 1892; Hyslop, 1901; Salter, 1922).

Quantitative examination of such readings effectively began with Hyslop’s (1919) work with Leonora Piper. Hyslop’s protocol did not differ, to a large extent, from the earlier descriptive work, but Hyslop chose further quantitative analysis following public criticism (Schouten, 1994). The most vocal critic, Frank Podmore (1901-2) reduced Piper’s impressive results to issues of guessing and chance coincidence. Hyslop attempted to counter this theory using a control group to assess the probability value for an entire session. The medium’s statements were rephrased, into 105 questions, and sent to 1,500 people. The 420 replies provided a probability of correctness for each statement (e.g. if 42 people out of 420 rated a question as applicable, the probability of a statement being true would be 1 in 10). The final p - value for the entire session (calculated by multiplying all probabilities) was astonishingly low, apparently ruling out chance as a likely explanation of his results. One problem with Hyslop’s method, however, is that it wrongly assumes all statements are independent. For example, a statement concerned with ‘marriage’ and another statement concerned with ‘children’ would not be independent of each other as the presence of one would increase the possibility of the other, especially around this period in history.

Saltmarsh (1929) reported on a year long study with Mrs Warren Elliott. Two series of studies were carried out. In the first series (53 sessions) the target person was present but not identified, in the second (89 sessions) the target person was absent. In addition to testing several theories on ESP and the survival hypothesis, Saltmarsh, reminiscent of Hyslop, aimed to test the criticism that results were due to chance. He introduced the concept of control studies where statements were judged on correctness by subjects other than the target person. These control studies relied on the pre-selection of statements from particular sessions and a weighting system calculated by examining comments given by target and control persons. A problem with this procedure lies in the experimenter’s active role in the judging process. Saltmarsh was aware that since he was the selector and ultimate judge of statement correctness, a potential bias could result. The lack of any statistical assessment was another weakness.

Saltmarsh and Soal (1930) rectified earlier statistical problems by seeking advice from the statistician, R. A. Fisher. In attempting to answer the question ‘what is the probability of the statement being correct?’ the authors devised a formula that would yield a standard-normal score (z – score) with an associated probability of significance. Taking a session from Saltmarsh’s (1929) study they applied the formula and produced a highly significant score. Committees of investigators were also employed to evaluate for what proportion of the population the statement holds. They were provided with the medium’s statements and the comments on these statements made by the target persons. A disadvantage of this method is that the committee members could have had knowledge of the accuracy of the statement as indicated by positive or negative remarks (Schouten, 1994). The probability of the statements could also be affected by interdependency as in the aforementioned Hyslop (1919) study, therefore influencing the correctness or falseness of previous statements.

At this stage in the development of research on mediumship, a number of weaknesses were still apparent. The most important ones were (a) that the people involved in rating the statements (researchers, target persons, controls) knew for whom the statements were intended, which could bias their judgements, and (b) the unreliability of estimating probabilities for the individual statements.

To eliminate these two main weaknesses, Pratt (1969; Schouten, 1994) introduced a different method during his quantitative evaluation of sessions with the medium Eileen Garrett. After a first series of sessions in which the subject, or target person, was in the same room as the medium, a second series had the target person in an adjacent room. In addition, Pratt remained blind to the relationship between sessions and target persons. The verbal material was divided into separate statements and then individually assessed by the target persons as to its personal applicability. He then used the responses of 14 control subjects, from which a mean standard deviation for each session’s scores was derived. The outcome of the study was striking (z = 5.10, p < 0.01).

However, despite Pratt’s attempts to improve judging, there is still a potential bias since the control subjects were aware, as was the experimenter, that they were control subjects (i.e., they knew the statements were not intended to apply to them) and may have rated statements accordingly. Pratt and Birge (1948; Pratt, 1969) countered this problem by developing a methodology and analysis that has remained as standard. The essence of this method is that a medium makes statements about a number of absent target persons, and each target person judges all statements on personal applicability without knowing for whom each statement was intended and without the experimenter being aware of the target person’s identity. A potential problem that still remains with this method, however, is the presence of ‘temporal’ cues. For example, in the Pratt and Birge (1948) study five subjects (target persons) were given a reading by the same medium and the outcome assessed on their responses. Subject ‘A’ present in the morning, however, may receive all statements from the medium in which a few refer to the morning. These particular statements would be rated highly by subject ‘A’ and lower by those subjects in afternoon sessions. The same also applies for ‘environmental’ cues that may show up in comments made by the medium. These may include any session-specific incidents such as rain outside, sounds etc.. These ‘temporal’ and ‘environmental’ cues rely on independent awareness by the medium and subject, but sensory leakage could mean an increase in the number of possible cues (e.g. subject hearing medium cough or voice loud exclamation). Several studies have been conducted since the late 1940s but none have tackled the latter problems (e.g., Schmeidler, 1958; West, 1949; see Schouten, 1994 for a review).

Recent Work

Within the last few years, there has been renewed interest in the systematic study of mediumistic communications. Most notable is the research programme led by Gary Schwartz (Schwartz, Russek, Nelson & Barentsen, 2001; Schwartz & Russek, 2001; Schwartz, Russek & Barentsen, 2002) and the studies by Tricia Robertson and Archie Roy (Robertson & Roy, 2001; Roy & Robertson, 2001; Robertson & Roy, 2004). We now review this recent work in more detail.

Schwartz, Russek, Nelson & Barentsen (2001) reported two studies in which sitters and ‘highly skilled’ mediums were separated by a large opaque screen. The mediums conducted readings with questions restricted so sitters could only give ‘yes’ or ‘no’ answers. Five mediums and two sitters were involved. All five mediums gave a reading under these conditions to the first sitter, whilst the second sitter received readings from only two mediums due to time constraints.

A few months after the sittings, both sitters were asked to rate each of the statements made by the mediums on a 7-point scale from -3 (definitely an error) to +3 (definitely correct) to each of the statements made by the mediums. The two sitters gave +3 ratings to 83% and 77% of the statements, respectively. To assess the extent to which this level of accuracy could be explained in terms of educated guess-work by the mediums, the authors adopted a similar procedure to that used by Hyslop (1919). Seventy statements were selected from the readings given to the first sitter and turned into questions. For example, if the medium had said ‘your father loved dancing’, the question became ‘who loved to dance?’. These 70 questions, along with a photograph of the sitter, were presented to 68 undergraduates who were asked to guess the answers. Schwartz et al. reported that the average proportion of items guessed correctly was just 36%, and argued that the high level of accuracy obtained by the mediums could not be due to chance guessing. In addition, the authors report a number of compelling qualitative examples of apparently highly accurate and specific items of information provided by the mediums.

In another study, three of the same mediums each gave readings to each of five sitters. As before, mediums were blind to the identity of the sitters and mediums and sitters were separated by a floor-to-ceiling opaque screen. However, an added control was that mediums were not permitted to ask questions, and sitters did not speak. Once the readings had been transcribed, sitters rated a total of 528 statements from all 15 readings using the same -3 to +3 scale as before. On average, sitters gave +3 ratings to 40% of the statements from their own readings but only 25% of others’ (control) readings (Schwartz et al., 2001). In a later study, with just one medium, Schwartz and Russek (2001) focused on increasing the number of ‘blinds’ by including a reading that was given prior to the scheduled time of the reading, and ensuring there was a complete absence of visual, auditory, or other potential sensory cues by using a muted telephone. The medium was in Arizona whilst the sitter (who was also a medium) was in California. The experimental design involved three phases. Phase 1 was a pre-reading contemplation period where the medium attempted to receive information about the sitter's deceased loved ones before the phone link was made. Phase 2 was a sitter-silent period where the telephone was on mute so that the sitter could not hear the information given by the medium. Phase 3 was an actual telephone reading involving dialogue between the medium and the sitter. The authors report that specific information regarding names and relationships provided during phases 1 and 2 (in which there was no opportunity for the medium to pick up cues from the sitter) was over 90% accurate (Schwartz & Russek, 2001). Information obtained during phase 3 included four pieces of specific information that were unknown to the sitter but were later confirmed. The possibility of sitter rater bias was reduced since, although the report focuses on the results with one sitter, the actual experiment involved several sitters who were required to ‘blind’ rate a selection of readings without being informed which were their own. There is little detail, however, given to the presentation of these transcripts and potential problems could arise with the inclusion of Phase 3 (i.e. not blind) readings (Schwartz, 2004).

A separate programme of research conducted by Archie Roy and Tricia Robertson has tested the extent to which the statements given by mediums may be applicable to people for whom the information was not intended (Roberston & Roy, 2001; 2004). The authors used different platform mediums (i.e., individuals who give stage demonstrations of mediumship), and audiences of up to thirty, in order to examine this question.

In their first paper, Robertson and Roy (2001) examined the performance of 10 mediums. The mediums faced the audience and gave readings to the target person who was selected by the medium. Each set of statements were distributed to the audience and asked to tick statements they could accept as personally relevant. They report a highly significant difference between fractions of statements accepted by 44 ‘recipients’ and 407 ‘non-recipients’ (Robertson & Roy, 2001). Recognising alternative explanations for these results, the authors devised a protocol that attempted to (a) eliminate possible cues from the target person, (b) control for rater bias, and (c) take into account the specificity of each statement. Cues from the target person were eliminated by keeping the medium blind to the identity of the target person who was in a separate room. Rater bias was controlled for by keeping the subject blind as to whether or not he or she was the intended recipient of the medium’s statements. This was achieved by randomly selecting the target person from a group of participants moments before this group were seated together in one of the testing rooms. All participants indicated whether or not they could ‘accept’ each statement as being potentially relevant to themselves. Finally, statements were weighted in order to differentiate the vague, ambiguous statements acceptable to many participants from the highly specific statements acceptable to perhaps only a single person. The full details of this protocol, referred to as the Robertson-Roy Protocol, and variations upon it, are reported in Roy and Robertson (2001). The authors tested 10 mediums who had performed well in their earlier research using this more stringent protocol. They report that, even under these conditions, the mediums were still able to provide information that was more likely to be accepted by the intended recipient than the ‘control’ participants (Robertson & Roy, 2004).

Thus, recent attempts to revive quantitative assessment of mediumistic communications under controlled conditions have yielded evidence that seems to support the mediums’ claims. That is, some mediums are able to provide accurate and specific information for unseen target persons about whom they have no knowledge. One interpretation of these findings is that the mediums are indeed receiving this information directly from deceased individuals known to the target person.

However, before this interpretation of the findings is likely to be accepted by critics, there is a need for this research to be independently replicated. Moreover, we feel that the methodologies employed by both Schwartz et al. and Robertson & Roy may be improved to further reduce the likelihood of rater bias (i.e., target persons giving inflated ratings for their own readings as a consequence of them knowing the reading was intended for them). Rater bias is a potential problem with the Schwartz et al. research, especially with the earlier studies as the sitter could hear the reading and so he or she would know which statements were intended to apply to them when they were giving their ratings. Their later research, with the medium Laurie Campbell, went some way to removing this weakness by having the sitter rate statements given by the medium during a sitter-silent period which was not heard by the sitter (Schwartz & Russek, 2002). The authors contend that rater bias could not account for the large number of specific pieces of verifiable information, especially names, rated as accurate by the sitter. Even so, such findings require replication under conditions that would rule out rater bias as a potential ‘out’ for critics.

The work of Robertson and Roy goes much further in reducing the possible influence of rater bias, by selecting a target person at random from a group of participants and keeping all participants blind to the identity of the target person (the intended recipient of the medium’s statements). However, even this procedure may introduce potential bias at the rating stage due to the fact that participants have knowledge of the other participants. Whilst this may not seem to be problematic, it means that participants’ ratings of statements are not independent of each other. For example, hypothetically, if all participants present, except one, are long-haired, any information provided by the medium relating to the recipient’s short hair would result in a potential ‘stacking effect’ (i.e. that particular statement would not be chosen by the majority). This would be the case, whether the short-haired subject was the recipient or not. This could, therefore, skew results in the wrong direction even for accurate information.

In short, previous research examining mediumistic communications has provided evidence that lends support to the notion that some mediums are able to provide accurate and specific information for sitters under conditions that appear to rule out conventional explanations. However, there is a need for these findings to be replicated. Furthermore, we propose employing a protocol that builds upon previous protocols which we argue is even more effective at controlling for conventional explanations such as experimenter bias and rater bias. As has been noted by others, it is important that the tightest possible protocols are used to assess the abilities of mediums so as to prevent critics from resorting to ‘outs’ to explain significant results and to provide assurance to the mediums themselves that it would not be possible to explain away positive results in terms of chance or other conventional explanations (e.g., Hansen, 1990; Wiseman & Morris, 1995). Only then will research with mediums be able to provide a basis from which to establish a useful theory of survival after death.

Searching for a replicable mediumship study (years 1-2)

We propose to replicate and extend previous research by employing a double-blind protocol that involves multiple mediums and multiple sitters. The protocol aims to eliminate the potential for cold reading, rater bias, experimenter bias, and coincidence as likely explanations of positive results. Although the protocol may operate, in principle, with any number of mediums and any number of sitters (minimum two of each), we anticipate employing up to a maximum of five mediums and five sitters at any one time. The protocol, and subsequent analysis, allows for the possibility of highlighting ‘good’ mediums and/or ‘good’ sitters and then reusing a combination thereof in further studies.

< Back