Mediumship — triple blind study

[FatPhil]

granny needs to wait for judgment day to get into heaven

Nope, Lazarus, who died just before the rich man (who went to hell, and wanted to warn his family), went to heaven right after death. This is common knowledge! (That's my justification for not including a link, I'm busking here, not thumbing through sources.)

 

[LilWabbit]

granny needs to wait for judgment day to get into heaven

That's a Lutheran and Calvinist doctrine and doesn't account for all of Christianity. I understand you're from majority-Lutheran Germany so your sin of generalization is hereby graciously forgiven.

Bold added:

Article:

Christian mortalism is the Christian belief that the human soul is not naturally immortal[1][2][3][4][5] and may include the belief that the soul is "sleeping" after death until the Resurrection of the Dead and the Last Judgment,[6][7][8][9][10] a time known as the intermediate state. "Soul sleep" is often used as a pejorative term,[11][a][14] so the more neutral term "mortalism" was also used in the nineteenth century,[15] and "Christian mortalism" since the 1970s.[16][17][18][19][20][21][22] . . .

Christian mortalism stands in contrast with the traditional Christian belief that the souls of the dead immediately go to heaven, or hell, or (in Catholicism) purgatory. Christian mortalism has been taught by several theologians and church organizations throughout history while also facing opposition from aspects of Christian organized religion. The Catholic Church condemned such thinking in the Fifth Council of the Lateran as "erroneous assertions". Supporters include the sixteenth-century religious figure Martin Luther and the eighteenth-century religious figure Henry Layton, among many others.

Source: https://en.wikipedia.org/wiki/Christian_mortalism

 

[Mendel]

Nope, Lazarus, who died just before the rich man (who went to hell, and wanted to warn his family), went to heaven right after death. This is common knowledge! (That's my justification for not including a link, I'm busking here, not thumbing through sources.)

it's a parable

27 "He answered, 'Then I beg you, father, send Lazarus to my family, 28 for I have five brothers. Let him warn them, so that they will not also come to this place of torment.'

29 "Abraham replied, 'They have Moses and the Prophets; let them listen to them.'

30 "'No, father Abraham,' he said, 'but if someone from the dead goes to them, they will repent.'

31 "He said to him, 'If they do not listen to Moses and the Prophets, they will not be convinced even if someone rises from the dead.'"

this makes a strong case that the dead are not going to talk to the living—>no mediums

 

[Ann K]

it's a parable

27 "He answered, 'Then I beg you, father, send Lazarus to my family, 28 for I have five brothers. Let him warn them, so that they will not also come to this place of torment.'

29 "Abraham replied, 'They have Moses and the Prophets; let them listen to them.'

30 "'No, father Abraham,' he said, 'but if someone from the dead goes to them, they will repent.'

31 "He said to him, 'If they do not listen to Moses and the Prophets, they will not be convinced even if someone rises from the dead.'"

this makes a strong case that the dead are not going to talk to the living—>no mediums

You can find support for any position from the Big Book-o-Multiple Choice, and (considering how few people actually know what all the points of doctrine are in their nominal religion) even knowing what church the subject of the original post attends cannot tell you what HER understanding is. But it is well documented that mediums, clairvoyance, conversing with the dead etc are neither a science not a religion; those involved in it are either the scammers or the scammed. There's nothing harmless about either of those positions, and any attempts to dissuade her are laudable.

 

[Landru]

I do not want this thread to drift (no careen) to a discussion of Christian doctrine. Can Mediums' ability be supported by evidence.

 

[deirdre]

Training to be a medium, eh? Ask her to ask your closest dead relative to reveal the thing that you both agreed, whilst they were still alive, they would communicate from beyond the grave. If you didn't have such a message agreed, and being a medium was a real thing, then she'd be able to inform you that you were making stuff up.

Yes, this is the Houdinis idea, from almost a century back. I didn't claim to be original.

that doesnt actually work. the psychic could just be reading the sitter. the dead person would have to write down a phrase and lock it in a bank vault with directions that the lawyer cant open it until a written phrase from a medium is presented.

of course you can only do this once as the lawyer would then know. i guess you could included like 50 envelopes numbered and the medium would have ask the dead person "what is in envelope number 3" etc. Provided the ghost can remember that many envelopes.

 

[FatPhil]

it's a parable

How does that make it not canon?

this makes a strong case that the dead are not going to talk to the living—>no mediums

Firstly, that wasn't the argument you were making. You were arguing that granny doesn't go to heaven.
Secondly, that passage implies that Lazarus won't corporeally be *sent* back - explicitly "rises from the dead", that's even in the bit you quoted. There's nothing about spirit communication in it.

 

[Ann K]

that doesnt actually work. the psychic could just be reading the sitter.

Oh, please. First we are going to need evidence that ANYONE is really "psychic", and as yet that has not happened.

 

[FatPhil]

Oh, please. First we are going to need evidence that ANYONE is really "psychic", and as yet that has not happened.

You could just read @deirdre's comment as saying "so-called 'psychic'" instead, and it would make the same point.

But I don't fully subscribe to deirdre's point of view - it depends on what level of certainty is desired, and to whose satisfaction. It's only a weak disagreement, not a fundamental one. I've worked on verification protocols in the past; I even tried to get James Randi to expand the remit of his million dollar prize once to expand it into a field of pseudoscience that I and some other like minded debunkers was familiar with. Alas, as he had no experience in that particular field, he wasn't interested. That was a shame, as that particular wing of pseudoscience was particularly lively at the time, and in order to "win" you wouldn't just need to break the laws of physics, you'd have had to break the laws of mathematics, which is much harder. To have a single clearly-defined test protocol with the heft of having Randi's name on it might have reduced the noise level in the field. (Who am I kidding?)

However, back to the matter in hand: if the sitter is going in skeptically - which is a premise I have explicitly adopted - it shouldn't be too hard for them to come up with both real and fake questions (that have no correct answer) to trap and fool the medium into bullshitting something. Cold reading really isn't as powerful as people think it is, there's a hell of a lot of dancing going on, and a skeptical sitter should be aware of that. Naming the sweet - and the particular flavour - that gramps used to share with you without letting grandma know, for example, involves no dancing - it's either named or not. Any dancing can be considered disproof, to the satisfaction of the sitter.

 

[FatPhil]

Who is doing the training?

This is a very interesting question. A line of friendly discussion that gets her to volunteer that detail could possibly be found.

 

[deirdre]

This is a very interesting question. A line of friendly discussion that gets her to volunteer that detail could possibly be found.

i googled it. apparently there are online courses. or card packs you can buy on amazon.

Spoiler

Article:

Questions to ask an Animal Communicator or Pet Medium:


An animal communicator is someone who specializes in animal communication and will typically be able to communicate with both living animals and pets that have passed. Many Evidential Mediums will also communicate with animals - it is common for me to connect with all types of animals in my sessions with clients. If you're interested in connecting to pets you've lost, be sure to ask the Medium you're thinking about working with if connecting with pets is a specialty of theirs.



1. What does my pet want me to know about their passing?

Source: https://www.sherylwagnermedium.com/blog/questions-to-ask-a-medium

 

[Ann K]

i googled it. apparently there are online courses. or card packs you can buy on amazon.

Spoiler

Article:

Questions to ask an Animal Communicator or Pet Medium:


An animal communicator is someone who specializes in animal communication and will typically be able to communicate with both living animals and pets that have passed. Many Evidential Mediums will also communicate with animals - it is common for me to connect with all types of animals in my sessions with clients. If you're interested in connecting to pets you've lost, be sure to ask the Medium you're thinking about working with if connecting with pets is a specialty of theirs.

1. What does my pet want me to know about their passing?

Source: https://www.sherylwagnermedium.com/blog/questions-to-ask-a-medium

 

[John J.]

From the 2nd Beischel (et al.) study, usefully brought to our attention by Amathia
https://www.windbridge.org/papers/BeischelEXPLORE2015vol11.pdf

Scoring. For the Experiment 2 readings, sitters provided individual scores for each item in each of two readings: a target and a decoy. Each item received one of the following scores:

5: Obvious fit (used if the item is a direct or concrete hit that does not require interpretation to fit)

4: Fit requiring minimal interpretation (used if the item indirectly applies and needs minimal interpretation or symbolism to fit)

3: Fit requiring more than minimal interpretation (used if the item indirectly applies and needs a greater degree of interpretation or symbolism to fit)

2: Other fit (used if the item does not fit the named discarnate or the rater, but does fit someone else that the rater is/was close to and that is likely to be the subject of the statement)

1: No fit (used if the information is a concrete miss—is clearly wrong—or if it is information for which there is no reasonable interpretation)

0: Donʼt know (used if the rater does not understand the item or does not have enough information to judge its accuracy)

Percentage accuracy was calculated in Experiment 2 by tallying the number of items that received scores of 4 or 5 and dividing that total by the total number of items minus the items scored as 0ʼs ([4ʼs/5ʼs]/[total 0ʼs]) and calculated separately for the Five-Questions and the Free-Form sections. A more conservative tally was also examined in which only the items scored as 5ʼs (hits) or 1ʼs (misses) were totaled.

Not totally happy about all this.

And I'm not happy at all that first names, correct answers, and readings are unavailable.

Agreed- and would add, the questions that the mediums put are unavailable (AFAIK).

From the paper,

During Phase 2(unblinded but regulated), the medium was introduced to the sitter by first name. In this 20-min phase, the medium was permitted, if she chose, to ask the sitter yes-or-no questions to which the sitter could respond "yes," "no," "maybe," "sort of, "or "I donʼt know."

An interesting secondary experimental question might be,
"is there a relationship between the score rated by the sitter and the number of questions asked by the medium?"
No control group of sitters who are not asked for feedback is used in Experiment 2.
The researchers write,

The Phase 2 sections in which the mediums were permitted to request limited feedback to yes-or-no questions if they chose received a mean score of 4.20 +/- 0.25 and also ranged from two to six. The scores from the two phases are not significantly different (P = .08); i.e., a significant scoring increase did not occur once the WCRMs were able to request feedback during the reading. This suggests that the accurate information reported during Phase 1 may have been acquired from a non-local source

Alternatively, it suggests the Windbridge-approved mediums (remember, Julie Beischel is one of Windbridge's two managing directors) didn't perform significantly better when they could ask Yes or No questions of their sitters.
One wonders if this finding has been communicated to the Windbridge mediums so that they don't have to waste time/ bother their clients with pointless questions in the future. Somehow I doubt it (but I'd be happy to be proved wrong).

The scoring system is problematic. It is obviously subjective, and conducted by people who wish to communicate with a deceased person, have accessed the Windbridge website, completed an online form (whose content- and questions, if any- we don't know; it's generally considered good practice to publish recruitment materials in an appendix if these might be relevant to a study) and who then have been selected "at random" ...

...from the general sitter pool using www.random.org to select participant ID numbers.

(My emphasis).

How it 's determined who should be in the "general sitter pool" is not stated. Presumably "general sitter pool" does not mean all applicants who completed the Windbridge questionnaire, so what criteria were applied to determine who should be in that category?
I get the impression that there has been an element of pre-selection before www.random.org was used.

I don't know if this is relevant: I've seen a couple of stage hypnotists; they initially (and openly) engaged the whole audience in a series of "tasks" to narrow down the field to the most "suggestible".

I'm not sure about the scoring system.
Without knowing what statements were made by the mediums, or what questions were asked of the sitters, we are free to speculate.

"This person [or more likely, the name supplied to the medium] was special to you..."
"[Name of deceased] has known that you wanted to be in touch...",
"They were sad to leave you, but want you to be happy"
...and similar phrases, to subjects who wish to contact the deceased and who have pro-actively contacted an organisation advocating mediumship, are easy "5's".

For "4's",

...if the item indirectly applies and needs minimal interpretation or symbolism to fit

...who determines what minimal interpretation or symbolism is? Are any guidelines provided to the sitter?
Much the same apples to "3's".

"2's",

...if the item does not fit the named discarnate or the rater, but does fit someone else that the rater is/was close to

...so any remark that might fit anyone that the rater is close to "and is likely to be the subject of the statement" is valid.

"1's" include

...information for which there is no reasonable interpretation

-which again is highly subjective. The sitters/ raters are not asked to not give "5" for "hits" which might have a "reasonable [non-paranormal] interpretation" (e.g., statements that might be applicable to most potential sitters).

Edited to add, Saturday 09/09: Not sure I put my concern about "1's" (above) very clearly:
I was trying to say, scoring a "1" requires a medium to say something "for which there is no reasonable interpretation".
But symbolic meanings, "interpretation", and information which might be about anyone "close" to the sitter are all accepted in the scoring system as not 1's.
The criteria for a "1" seem to set a high bar for the sitter, a possible believer- they might well rate an objectively incorrect statement as being partially correct following "interpretation", or after considering possible possible symbolic meanings.

0: Donʼt know (used if the rater does not understand the item or does not have enough information to judge its accuracy)

If the rater is prepared to give the medium "the benefit of the doubt"- and the selection process might favour potential subjects so inclined- then it's easy to see how wrong answers ("1's") might end up in this category.

Despite my, ahem, considerable scepticism about mediumship, I don't know how we can account for the significant differences in ratings for "target" and "decoy" readings.
Of course (and remembering posting guidelines) we must assume that the above experiments were conducted in good faith.
Anyone got any ideas how those results were obtained?

 

[Mendel]

From the 2nd Beischel (et al.) study, usefully brought to our attention by Amathia
https://www.windbridge.org/papers/BeischelEXPLORE2015vol11.pdf

I didn't see anything on how the decoy readings were chosen. Pick a maximum non-fit, and the real reading will seem that much more accurate.

Of course (and remembering posting guidelines) we must assume that the above experiments were conducted in good faith.

I don't agree.

 

[Mendel]

The scoring in that second study is also suspect.

Note that "correct" ranges from 3 to 6, while "incorrect" ranges from 1 to 2. (1 was never chosen in the exploratory experiment. Since 1 means "much worse than random chance", that's not surprising.) If you have 1 rater choosing "6" and 3 raters choosing "2", the mean is still 3!

Note that "good reading with some incorrect information" and "good reading with little incorrect information" are essentially the same, and exhibit a ChatGPT level of trustworthiness, which means you can't really rely on anything the medium tells you => waste of money.

Yet only "1" is later counted as a "miss".

They seem to have used a 0-5 scale in experiment 2, with "fit" ranging from 2 to 5, and "no fit" being 1. A "5" and a "1" results in a "mean" of "3".

"2" implies that if the reading fits the rater, but not the deceased, it is scored at "3" or higher, but not considered a miss; and in my opinion, "2" itself should be considered a miss as well. This illustrates considerable bias and a "we'll take anything we can get" mindset.

Note also the inclusion of "sybolism". I have no idea how that even applies to the questions:

You're going to symbolize someone's favorite food?

Notwithstanding the 0-5 scale, all "global scores" were taken on the 0-6 scale.

The averages for experiments 1 and 2 are below 3, which suggests that the majority of scores indicated no communication.

Note also that the 3 experiments comprised 28, 28 and 40 readings respectively; this means that 1 reading was dropped from experiment 1, and 9 readings (22.5%) were dropped from experiment 2, for unclear reasons.

There is no sharing of data, not even tables etc.

 

[John J.]

There is no sharing of data, not even tables etc.

Good point. It shouldn't have been difficult to tabulate the sitters and the scores they gave.
Beischel must be familiar with "Extraordinary claims require extraordinary evidence"; publishing raw data would have been helpful.

... 9 readings (22.5%) were dropped from experiment 2, for unclear reasons.

Yes... somehow I doubt that those readings, whatever problems arose, were heading in a "positive" (sitter confirms accuracy) direction.
Maybe it was established that the people being contacted weren't actually dead yet...

 

[Mendel]

Maybe it was established that the people being contacted weren't actually dead yet...

They didn't care about that. Remember, their scoring has a category for "does not fit the rater", which implies they have categories that do. And if they're using dead raters, well, that'd just be like the mediums scoring themselves.

 

[Skalman]

Didn't they email the readings to the sitters? My first assumption would be that the dropouts were people who simply didn't reply for some reason.

 

[Mendel]

Didn't they email the readings to the sitters? My first assumption would be that the dropouts were people who simply didn't reply for some reason.

You shouldn't have to assume that. It should have been clearly stated, and if the fringe journal this was published in had competent reviewers, they would have pointed this out.

Also, "not replying" constitutes a self-selected sample: e.g. it could be that people mostly didn't reply because they felt the readings they had received were trash, and that would bias the remainder of the responses in favor of readings that looked valid—especially with a 22.5% loss of responses, that bias would be considerable.

But ultimately, we do not know why these responses were dropped.

 

[Skalman]

You shouldn't have to assume that. It should have been clearly stated, and if the fringe journal this was published in had competent reviewers, they would have pointed this out.

Also, "not replying" constitutes a self-selected sample: e.g. it could be that people mostly didn't reply because they felt the readings they had received were trash, and that would bias the remainder of the responses in favor of readings that looked valid—especially with a 22.5% loss of responses, that bias would be considerable.

But ultimately, we do not know why these responses were dropped.

Yeah, I completely agree with you. It's such a small sample to begin with, and selected from people who volunteered and were picked by the researchers, not a survey sent to random households through the mail (where most people usually won't bother to reply).

 

[FatPhil]

fringe journal

I just had a sniff around, and it does look like it carries a lot of articles in woo-attracting fields, and that such field are explicitly in the journal's remit:

About the journal

The Journal of Science & Healing

EXPLORE: The Journal of Science & Healing addresses the scientific principles behind, and applications of, evidence-based healing practices from a wide variety of sources, including conventional, alternative, and cross-cultural medicine. It is an interdisciplinary journal that explores the healing …

https://www.sciencedirect.com/journal/explore

It also seems as if the editorial board do a lot of self-publishing in it. For example the current first 3 "Recent Articles" listed on the above page are:
- Mini review "Asking about guns in the home should be pre-requisite for play dates" https://www.sciencedirect.com/science/article/pii/S1550830723001349 is by Angie Lillehei who is "Associate Editor" and "Editorial Director";
- Short communication "Covid-19: A case study on the psychology of exploitation" https://www.sciencedirect.com/science/article/pii/S1550830723001337 is by Stephan A. Schwartz who is "Columnist";
- Research article "Worldviews and environmental ethics: Contributions of brain processing networks" is by Marjorie Woollacott, ... Natasha Tassell-Matamua, where that final author is "Associate Editor".
roles listed here: https://www.sciencedirect.com/journal/explore/about/editorial-board

The impact factor isn't terrible, 2.4, however, that's not particularly good either. I do notice that many of the articles I could see reference lists for were citing other papers from the same journal, so it seems like there might be a self-boosting clique that's making the impact factor seem higher than it really ought to be.

 

[John J.]

My first assumption would be that the dropouts were people who simply didn't reply for some reason.

That was my first thought too (if we're talking about subjects dropping out), subjects do drop out of studies post-recruitment.
But I don't think that's the case:

Of the 40 Experiment 2 readings performed, 31 were returned and contained usable data Content from External Source https://www.windbridge.org/papers/BeischelEXPLORE2015vol11.pdf

...so there were 40 subjects at the start of experiment 2, not the "n = 31" stated several times on page 140.

40 readings were performed- so no subjects dropped out- so why is the data from 9 (22.5%) missing or not usable?

Each of the twenty Windbridge-approved medium performed two readings for experiment 2,
so between five (4 mediums x 2 readings, + 1 medium x 1 reading) and nine (9 mediums x 1 reading) mediums out of twenty failed to supply usable data from at least one of their readings.

40 readings were performed for 40 subjects (sitters) in Experiment 2 of Beischel et al.s' 2015 paper (link in quote above).

9 out of 40 reading results either weren't returned or didn't contain usable data. The authors do not provide further information.
It's hard to think of how a sitting could be conducted with a (hopefully) competent subject and not generate usable data.

Maybe some of the transcripts of sittings were lost (which in itself would be problematic). Maybe mediums deviated from the trial protocols for some sittings, or sitters inadvertently gave away information, invalidating their sitting.

I think a "failure" rate of 22.5% should have been explained by Beischel et al., as less charitable interpretations are possible.

Edited to add: I've just realised I didn't consider the possibility of sitters dropping out after their sitting, i.e. not rating the medium's report. *Doh!* I overlooked that Mendel has already considered this,

Also, "not replying" constitutes a self-selected sample: e.g. it could be that people mostly didn't reply because they felt the readings they had received were trash, and that would bias the remainder of the responses in favor of readings that looked valid—especially with a 22.5% loss of responses, that bias would be considerable.

 

[Ann K]

9 out of 40 reading results either weren't returned or didn't contain usable data. The authors do not provide further information.

I think there is such an enormous difference between "9 were not returned" and "9 didn't contain usable data" that having such an ambiguous description with them all lumped into one sentence is sufficient to make the experimental results highly suspect. Nine out of forty is a very high proportion to ignore, and clarification should certainly have been made.

 

[NorCal Dave]

From Mendel's post #54:

If I'm understanding this correctly, if I was a sitter, I would have provided the name of a discarnate, like my late mother. First name, maiden name, full name, married and maiden name I don't know, but married initials would be JT.

Likewise, a number of other sitters do the same thing. Say FatPhil is a sitter, and he gives the name of his great grandfather to be read.

Then the mediums give readings for the various names they are given. As one of those readings was for my mom JT, that is one of the readings I will be asked to evaluate. The other reading is a "decoy" reading that was for another sitter. So, it's still a reading about someone. How is my decoy assigned to me and how is the reading about my mom assigned as a decoy to the other sitters?

They know that sitter Dave is related to discarnate JT because they have to provide me with the reading about JT as one of my 2 blind readings to rate. What else do they know about me? If they have my approximate age (almost 60 now) and that I'm from the US is my decoy reading going to be about FatPhils great grandfather? I pretty sure Phil is younger than me and I think he was born in Europe, the UK or maybe Finland.

Or would my decoy reading be from someone that is closer to my age and lives in the US? Maybe even someone that listed a female discarnate like I did. If that's the case, it's a bit like the old Astrology experiment where a group is asked to rate their personal horoscope as applying to them. After everyone gives their personal horoscope a good to great rating, it is revealed that the group all got the exact same horoscope and there was nothing personal about it. The more in common I have with the sitter who's reading is my decoy, the more likely we'll have some common traits in our discarnates.

While I used FatPhil as an example, it appears that most or all of the sitters were US based, in addition having a belief in mediumship, so there are already multiple layers of commonality between the sitters.

In fact, shouldn't a control reading, that ideally is completely fictitious, be the decoy reading that everybody gets along with their personal one. I'd want to know how those 40 people all rated a fictional reading.

 

[John J.]

First name, maiden name, full name, married and maiden name I don't know

At the start of each reading, the first name of a discarnate was given to the medium.

… In cases in which the names provide overt evidence about the discarnatesʼ ethnicities and in turn their probable physical descriptions… or provide other identifying information (e.g., religion)… a pair is chosen to include two discarnates of the same ethnicity, religion, etc.

Beischel et al.s' 2015 paper

In fact, shouldn't a control reading, that ideally is completely fictitious, be the decoy reading

To play devil's advocate, the researchers probably feel that a "genuine" reading is a better control than a fictitious one.

How on Earth (if that's the right phrase here) the mediums establish "contact" with the specific deceased person that the sitter wants on the basis of a first name is anyone's guess, but then the researchers don't need to explain plausible mechanisms.

...the old Astrology experiment where a group is asked to rate their personal horoscope as applying to them.

I think that's a good comparison.
It'd be brilliant if we had access to the transcripts (or better yet, recordings!) of the sittings; it would be interesting if the "5" scores might be more frequent for readings returning more "generic" /widely applicable statements than readings with lower scores.

The audio-recordings of the readings were then transcribed, formatted into lists of definitive statements, and blinded to remove any references to the discarnatesʼ names

(Beischel 2015).

It would also be interesting to compare the original transcripts of the recordings to the lists of "definitive statements".
I don't expect that we (or anyone else not associated with the trials) will get that opportunity.
But it seems the sitters are not actually seeing a transcript of the medium's statements "in the cold light of day" as it were.

During Phase 2... ... the medium was permitted, if she chose, to ask the sitter yes-or-no questions to which the sitter could respond "yes," "no," "maybe," "sort of, "or "I donʼt know."

(Beischel et al. 2015, link as above).

This phase lasted 20 minutes. Cold reading techniques can elicit information of surprising specificity; although the transcripts of the mediums notes were edited to remove names, if a medium successfully "hit" upon something quite specific- which non-mediums demonstrating cold reading have done successfully- and that ended up in the transcript/ "formatted definitive statements", it could make the sitter's "target" sitting/ transcript much more likely to be identifiable.

Here, illusionist Derren Brown demonstrates cold reading. Just an example, Beischel's mediums didn't have all the feedback that Brown has with "live" sitters. The program was filmed in the USA because Brown isn't widely known there. (The rest of the show, "Messiah", is well worth watching IMHO [Edited to add] despite the stupid choice of title).


Source: https://www.youtube.com/watch?v=6UE-1Thzj_0&t=36m39s

 

[NorCal Dave]

To play devil's advocate, the researchers probably feel that a "genuine" reading is a better control than a fictitious one.

Maybe. I think of the decoy reading as the equivalent to a placebo in clinical trials. It's not the real medicine but nobody knows that, so we see how people react when they think they might be getting the real thing.

If they are testing for the ability of mediums to contact dead people and the criteria is the scores they receive from various sitters, I would want to know how those sitters ranked a fictional reading. If the sitters are just as likely to rank a fictional reading as a "real" one, then it shows the sitters have a high level of conformation biases and their scores for the "real" readings is questionable.

Even if they used a "genuine" reading as the decoy it should be for a person not affiliated with any of the sitters and all the sitters should receive the same decoy reading. Using other sitter's readings as decoys can have the effect of making some of the readings being very similar or in the opposite order the decoy reading is so vastly different that it obviously does not apply to the sitter, thus making the one that does score that much higher.

Using the same decoy reading for all the sitters wherever it comes from lets us know how all of the sitters responded to their own reading and the same decoy reading.

As you point out, none of the relevant questions and readings are available so it just creates a "trust us" scenario, and I don't.

 

[Merle]

Arriving a little late to the party, but I'm glad to see this discussion on the Internet. I struggled with this issue several months before this thread. I ended up in the Psience Quest forum, where folks are fanatical about woo. I decided to take them on—listening to their arguments and responding as best I could.

I think I had a good case against all their arguments, but the one argument I always felt I needed more information on was the claim made by certain medium studies. In particular, the Beischel study referenced above (http://www.patriziotressoldi.it/cmssimpl…ter_22.pdf) is the most recent and appears to have good methodology.

The claim was that subjects were given two readings, one of which was supposedly for the deceased person they referenced. They guessed the correct reading 65 times out of 100. That is statistically significant. If I flip a coin 100 times and get 65 heads, I would conclude that the coin is probably not fair. So how did they get these results?

My basic conclusion is that there must have been either some sort of fraud involved or the readings were real. Since the claim that the readings are real is so extraordinary, I concluded that fraud is more likely than a legitimate finding. That was not an easy argument to make when I was the only skeptic there and the others in the forum were fairly well-informed and dogmatic in favor of woo.

By the way, there is a dataset available at: https://figshare.com/articles/dataset/Mediumship/13311710 — including names, a "list of common questions," and actual readings (in Italian).

An AI translation of the first reading listed says:

She shows me a photo that her family is looking at, on which she had to make some choices

She tells me: "I wasn't able to express my feelings"

She shows me white flowers or small white objects

Her house wasn't exactly in the city because she shows me a lot of greenery from her home

She makes me hear the sound of railways or a train in the distance

She often shows me her ring finger, maybe she wants to show me a ring

Which sure sounds like cold reading—although there was supposedly no way for the mediums to get feedback from the subjects.

Since the readings were done over the Internet, one possible source of fraud is that the mediums had assistants search obituaries for the name to find tidbits to add to the cold readings. However, there were probably so many obituaries for names like "Teresa," for instance, that this would be difficult—but it remains a possibility.

 

[FatPhil]

I think I had a good case against all their arguments, but the one argument I always felt I needed more information on was the claim made by certain medium studies. In particular, the Beischel study referenced above (http://www.patriziotressoldi.it/cmssimpl…ter_22.pdf) is the most recent and appears to have good methodology.

Not Found

The requested URL was not found on this server.

Not necessarily your fault, and please don't interpret it that way, the modern internet has a way of simply deciding it does not want to propagate information. Berners-Lee would turn in his grave, were he dead.

 

[Merle]

Not Found

The requested URL was not found on this server.

Not necessarily your fault, and please don't interpret it that way, the modern internet has a way of simply deciding it does not want to propagate information. Berners-Lee would turn in his grave, were he dead.

Try this link: http://www.patriziotressoldi.it/cmssimpled/uploads/images/IsThereSomeoneInTheHereafter_22.pdf .

 

[Merle]

Here is what I wrote on my blog regarding the Medium study when it was presented to me. I think my case is a little weak, but I could not find much better analysis on the web.

Here we have another study with blinded mediums apparently giving readings, with the readings being at least partially recognizable as from the intended deceased person.

100 sitters had come and asked for readings from a particular deceased person. Mediums were given only the first name of the deceased and came up with readings for that person. Later the sitters were grouped in pairs, where both sitters received both readings. The sitters apparently picked the reading intended for their requested deceased person 65% of the time.

In the past such studies were often plagued with methodological flaws or showed no significant conclusion for the mediums. (Battista et al, The Myth of an Afterlife, p615) This study claims to have a more rigid control and to be clearly positive for mediumship accuracy.

If we assume that nobody altered the readings or biased the study to filter out the results they wanted, we are left with reports from the mediums that had a slight tendency to match the requested person. They were not a perfect match, but they were accurate enough for 65% to choose the "correct" reading. How did these reports come to contain this level of correct information? I can think of three ways it could happen:

1. The mediums got their information from the deceased.
2. The mediums go their information from the living through PSI.
3. The mediums got their information from some physical means.

Option 1 is clearly incredible. I find no evidence that a mind could continue after death. I have seen no post in that thread that gives an answer for anterograde amnesia after brain injury that is reasonably consistent with an afterlife. I think the same applies to retrograde amnesia, loss of consciousness under anesthesia, language difficulties after brain injury, etc. So, I find it hard to believe that the dead are communicating in this study.

Even if the dead could communicate with the living, how did the mediums contact the right person? All they had was the first name. If the medium was told the name was Teresa, how did he contact the right Teresa? There were three different people named Teresa whom the mediums were told to contact (full data is at https://doi.org/10.6084/m9.figshare.13311710). How did they get the right one when they were told the next reading was to come from Teresa?

Option 2 is also clearly incredible, as it makes an unrealistic claim for PSI. But at least the supposed source would be alive, which I would think is more likely compared with getting information from the dead. Dead men tell no tales.

Option 3 is also incredible, as the controls in place should have prevented the mediums from getting information from elsewhere. But I suspect this is the case. For mediums have long used fraud to verify their skills (see https://en.wikipedia.org/wiki/Mediumship). If they had real powers, why did many mediums resort to fraud? In this study, could some of the sitters have found a way to contact the mediums before the study? Could the mediums, who gave their readings over Skype or WhatsApp without anybody supervising them, have had assistants on the Internet gleaning information to use in the readings? I don't know, but I find this possibility more credible than options 1 and 2.

A very similar study is found at:

Explore (NY); 2015 Mar-Apr;11(2):136-42. doi: 10.1016/j.explore.2015.01.001. Epub 2015 Jan 7.
Anomalous information reception by research mediums under blinded conditions II: replication and extension
by Julie Beischel, Mark Boccuzzi, Michael Biuso, Adam J Rock (https://pubmed.ncbi.nlm.nih.gov/25666383/) (#490)

Again, I suspect the mediums somehow got information fraudulently. I would like to see these studies reviewed in more detail.

I would like to see a study like this done where the medium does not even know the questions. Questions could be asked at random to the sitter in another room, with the deceased directed to give the answer to the medium in a nearby room. The deceased would supposedly know the question, but the medium would not. If the answers are just coming from the medium's mind, rather than coming from the deceased, this would reveal the problem. (#603)

I had suggested another study that I would find impressive: Believers in an afterlife could all agree here on a time and place where the first person to die would meet the others to reveal the card held in a dealer's hand to the others who would be sitting in designated rooms. If that's not plausible, revise the experiment such as having trained mediums. Then see if survival believers can do better than others who have no access to the deceased and are simply guessing.

If mediums could really hear from the dead, imagine what we could learn from Socrates, the Apostle Paul, or George Washington by simply holding a conversation with their souls. Imagine what the dead could do for us as spies or at the gambling table. Imagine the books they could write about their experiences beyond. If only they could send emails or even ring bells to answer yes and no questions. Instead, we have only what comes through the mind of a medium. Some of us think we are just hearing what the medium wants to say, not what the dead tell those mediums.

( https://mindsetfree.blog/adventures-in-psienceland/ )

 

[John J.]

By the way, there is a dataset available at: https://figshare.com/articles/dataset/Mediumship/13311710

Hiya Merle,
first thing I found on the webpage you linked to is the researcher's rating system:

It is very heavily biased: the rating "2" states that the reading had insufficient information "...to be certain of real communication with the deceased", i.e. there might have been communication with the dead, but it's not certain.
And that's the second lowest level of certainty.

"3", on a scale of 1-6, means the reading "...has enough information to indicate that there was indeed communication with the deceased."

So of 6 possible ratings, 4 are taken as indicating not only that the claimed medium has provided substantially correct information, but -without consideration of other possibilities- "...that there was indeed communication with the deceased."
One rating ("2") indicates there was not enough information to be certain that the reading resulted from communication with the dead- still an astonishing claim.
Only a rating of "1" does not mention or imply that the reading might have been due to communication with the dead.
There is no rating available for "No correct information".

-Thank you for providing a fresh link for the paper @Merle,
"Is There Someone in the Hereafter? Mediumship Accuracy of 100 Readings Obtained with a Triple Level of Blinding Protocol",
Patrizio Tressoldi, Laura Liberale, Fernando Sinesio, 2022, OMEGA—Journal of Death and Dying 91 (2), 2025
(list of Omega 91 (2) contents here, https://journals.sagepub.com/toc/omea/91/2).
The gap between online publication and journal publication isn't explained.

I won't pretend to have read it particularly closely yet, but there is a major issue with the study participants which I think potentially invalidates the methodology and any results.

Of the three authors, Tressoldi is a psychologist at the University of Padova, Italy.
He has a long interest in psi, and has made several remarkable (and, I strongly suspect, unreplicated) findings, according to this uncritical summary of his work: "Patrizio Tressoldi", Psi Encyclopaedia, written by Michael Duggan, updated 03 May 2025 https://psi-encyclopedia.spr.ac.uk/articles/patrizio-tressoldi

The other two authors are described in the paper as

Dr. Laura Liberale, executive member of the Gruppo di Ricerca Italiano sulla Medianit` a (GRIM, Italian Research Group on Mediumship). Dr. Fernando Sinesio, Head of the Gruppo di Ricerca Italiano sulla Medianit` a (GRIM, Italian Research Group on Mediumship)

The paper refers to the volunteers who participated as sitters.

Sitters
Sitters were recruited by email among people known by the authors' friends and colleagues, offering them free consultations.

For a study like this, this is not a suitable way of recruiting subjects, particularly if any of the friends or colleagues were in any form of dependent or professionally subordinate relationship with the authors.

In addition, Tressoldi clearly has views on "psi" phenomena that are not widely held by his peers; Liberale and Sinesio are identified as being senior members of the Italian Research Group on Mediumship, but no academic qualifications or institutions are stated for either. It is not implausible that some of the friends and colleagues recruited as sitters by the authors might also be interested/ have pre-existing beliefs in mediumship.

We might reasonably expect Tressoldi, as a psychologist, to know that this could be a major criticism; it's hard to understand why he chose to recruit subjects in this manner.

 

[NorCal Dave]

Here is what I wrote on my blog regarding the Medium study when it was presented to me. I think my case is a little weak, but I could not find much better analysis on the web.


( https://mindsetfree.blog/adventures-in-psienceland/ )

From your blog:

1. The mediums got their information from the deceased.
2. The mediums go their information from the living through PSI.
3. The mediums got their information from some physical means.

I think you could add: 4. It's a poorly designed study, that is largely geared toward making a supposedly high positive result.

1. There is no control. As @John J. pointed out above, the "sitters" are NOT randomized. They are all acquaintances or otherwise part of a group of people that believe in Psy, the afterlife, medium ship, psychics. There appears to be NO control group of any kind. No group of "sitters" of that are just random people that may or may not believe in these things. One should always have a good control group.
2. The "judgments" that the "sitters" are asked to choose from are, again as @John J. noted above, heavily biased towards a positive result. Of the 6 "judgments" the already biased "sitters" got to choose from, only 1 was definitively negative, 1 was vague and the other 4 affirmed some sort of communication with the dead.

It's kinda of a wonder the whole study didn't end up at 80-90% positive.

 

[yoshy]

I just wrote a post on p-hacking that was inspired by this thread. The application here is asking whether the statistical analysis was sound. Was there unintentional p-hacking? Was the analysis pre-registered (meaning they released their exact analysis plan before conducting the study)? Are there possible multiple testing problems?

 

[FatPhil]

Try this link: http://www.patriziotressoldi.it/cmssimpled/uploads/images/IsThereSomeoneInTheHereafter_22.pdf .

Thanks.

Terrible study - trivial to get a bogus result, it does't control for the medium wanting to cheat.

That, and "Italian first names did not convey any information about age" is an unfounded assumption which implies there's no such thing as popular culture in Italy, which seems unlikely.

 

[Mendel]

I think you could add: 4. It's a poorly designed study, that is largely geared toward making a supposedly high positive result.

I'm very happy with my previous posts on this thread, which make the case very clearly.

My option 4a is, the first names given to the readers reflected a statistical clue to the deceased's age; this enabled them to tailor their readings to one young and one old deceased (which was how the experiment was set up); and at that, they were statistically successful.

Plus the shenanigans with skewing the averages etc.

 

[Merle]

Thanks for the responses. I have been away from this subject for a while, but this forum looks like a good place to make some progress on evaluating medium studies, so let's see where it leads.

By way of background, I have this habit of going into the thick of things, looking for civil discussion among people who strongly differ with me on the Internet, often alone. In these discussions, I try to listen to the opposition and hone my case to be effective when it is placed along those with strong bias against me. (e.g., see
Source: https://medium.com/@Merle2/are-we-making-america-hate-again-65ee12a43f56
)

With that in mind, I am strongly in your court but may play devil's advocate to see how these arguments would stand in a forum like Psiquest .

Hiya Merle,
first thing I found on the webpage you linked to is the researcher's rating system:

View attachment 82531

It is very heavily biased: the rating "2" states that the reading had insufficient information "...to be certain of real communication with the deceased", i.e. there might have been communication with the dead, but it's not certain.
And that's the second lowest level of certainty.

Understood, but the study put sitters in pairs. Sitters A and B received the readings for both A and B. The rating system you mention may have tended to give high scores for both A and B, but that should balance out by the methodology. The sitters were told to choose between readings A and B. The rating system may have tended to cause high ratings for both A and B. That doesn't change the result. The problem is, we are told that 65% of the time sitter A chose reading A as better that reading B, supposedly when everyone was blinded.

In addition, Tressoldi clearly has views on "psi" phenomena that are not widely held by his peers; Liberale and Sinesio are identified as being senior members of the Italian Research Group on Mediumship, but no academic qualifications or institutions are stated for either.

Thanks for the information regarding the author's bias. Yes, that does give reason to question the study. But having a reason to question a study and being able to explain why the study is wrong are different things.

What I needed at the time of the debate was a good argument explaining why this study should not be trusted. I wrote it up as I copied above but always wished I could have done better.

 

[Merle]

I just wrote a post on p-hacking that was inspired by this thread. The application here is asking whether the statistical analysis was sound. Was there unintentional p-hacking? Was the analysis pre-registered (meaning they released their exact analysis plan before conducting the study)? Are there possible multiple testing problems?

Thanks. I would like to look into what you say about p-hacking. Another person familiar with such studies told me this is the problem: they simply manipulated statistics to make it look significant.

I'm not an expert on statistics, but a simple smell test tells me this is significant. 100 people were divided into pairs and given readings intended for both people, supposedly with everyone blinded. 65% of the time they choose the "correct" reading. Although the effect is weak, it is hard to argue that, if the methodology is good, then this finding is statistically significant.

As I concluded above, I find fraud as the most likely cause of what I read in that study. However, it is a little difficult to go into a room of people that are heavily biased for woo and tell them I think the study they reference must be fraudulent. If I say that, then I can hardly tell them they have no evidence. They will tell me I just ignore the evidence.

When faced with that situation, I need a little stronger argument.

 

[Merle]

Terrible study - trivial to get a bogus result, it does't control for the medium wanting to cheat.

As I wrote above, I think there was some fraud in this study. However, if the study really went down as described, then it is hard for me to say where this result could have been established fraudulently.

Of course, the researchers could have been lying about the methodology. Perhaps there was some path to get information to the medium, or for somebody to fraudulently modify some medium readings to make them more applicable.

Assuming they actually did what they state in the paper, what are possible causes for this 65% success rating besides actual contact with the dead?

That, and "Italian first names did not convey any information about age" is an unfounded assumption which implies there's no such thing as popular culture in Italy, which seems unlikely.

That is a good point. They simply state this is so, but do not verify it. That bias has been demonstrated to be a problem with other such studies and may have crept into this study also.

 

[Merle]

My option 4a is, the first names given to the readers reflected a statistical clue to the deceased's age; this enabled them to tailor their readings to one young and one old deceased (which was how the experiment was set up); and at that, they were statistically successful.

We have a list of the names of the deceased used in the study. I guess the test for this would be to give these names to a number of people and ask them to guess the person's age based on nothing but the name. If there is a strong correlation between the guesses and the ages, this would confirm that the mediums had a clue on the deceased's age before giving the reading.

And perhaps, even though all the names were Italian, there were other clues from the names that the mediums could have used to determine cultural background.

Also, if the range of time of death was known, the mediums could have possibly checked obituaries to find help in making a reading more likely to be selected. The mediums worked over the Internet, with nobody watching what they did (with assistants?) to tailor their readings based on the name given.

 

[Merle]

The whole claim of this study is preposterous. We are told that a sitter came and asked for a reading on, for instance, a particular deceased Teresa. In another city, a medium supposed had access to billions of deceased people, including thousands of Teresas. And yet somehow, in that cacophony of voices, this medium picked out the deceased Teresa this sitter was asking about and got information from her. Even if the dead live on in the spirit world, the claim that the medium can do this is preposterous.

Or perhaps the claim is that the requested deceased Teresa is the only one that responded. There were three different deceased Teresas called in this study. Did the correct Teresa, on seeing her niece as the sitter, assume she was the one being called, and then she alone contacted the medium? This assumes the deceased know a mind-boggling array of information about what is going on and want to communicate with us.


If they have this mind-boggling array of information of what is going on, imagine the service they could be us if we had them ring a bell to yes or no questions about what our enemies are doing, or what cards the blackjack dealer is holding.