Psychic mediums really can read your deep secrets

[mummymonkey]

Their methods have been scoffed at and their claims to be in touch with an afterlife taken with more than a pinch of salt – but a five-year test has shown that mediums can indeed discover your deepest secrets.

http://www.sundayherald.com/40843

I could do with a haircut. I'm waiting for a call from the 'psychic barber'.

[Ladewig]

I wanted to view the raw data and see what the hit rate was for the people not targeted by the psychic, but apparently a membership is required to view the Society for Psychical Research online library. (http://www.spr.ac.uk/)

I was rather surprised that the SPR has been publishing a journal since 1884.

[deBergerac]

Originally posted by Ladewig

I was rather surprised that the SPR has been publishing a journal since 1884.

I am more surprised that they are still publishing it.

[shemp]

What a hack job! She sure has a long way to go before she can be called a real reporter. Anyone who wants to email her with opposing information can reach her at Jenifer.Johnston@sundayherald.com

[Lothian]

Originally posted by Ladewig
I wanted to view the raw data and see what the hit rate was for the people not targeted by the psychic, but apparently a membership is required to view the Society for Psychical Research online library. (http://www.spr.ac.uk/)


Yes that seems the key. There is no control.

Gordon Smith is a typical cold reader but with slightly more than the usual amount of "Lovie" and "dearie" thrown in.:(

[Lothian]

This site (http://www.sspr.co.uk/) appears to have more info under Mediumship Information Analysis. Apparently the experiments were quadruple bind !! so that's alright then.

Can someone please calculate the odds of a psychic experiment not producing results which are "a million to one chance ?"

[MRC_Hans]

From the article (bolding by me):
A total of 13 mediums took part in the SSPR study, carried out in Scotland and London. In each test the medium would sit in a different room from the participants and choose seat numbers they wanted to read from the audience. The audience, usually around 30 people, would enter a room out of sight of the medium and on their way in be given a random seat number. After the reading, adjudicators would distribute lists of what the mediums had seen and the audience had to tick which of the mediums’ statements applied to them.

The rules of chance would suggest an accuracy rating of 30%, but the mediums’ average was 70%, with some hitting 80% on some of the participants.

“The results were very surprising,” said Robertson. “I have no idea how mediums can gain this information but the results prove that able mediums can accurately read their subjects. Their chances of guessing this level of information about their subjects is a million to one, statistically.

So, test persons subjectively ticked off which things appled to them?? Howz that for blinding :rolleyes:? Where is the data from the test subjects whose seats were NOT chosen?

How can 70% hits over 30% from pure chance be a million to one?? That would require a quite narrow normal distribution (which you cannot get with yes/no data).

:bs:


Hans

[Ed]

Originally posted by mummymonkey


http://www.sundayherald.com/40843

I could do with a haircut. I'm waiting for a call from the 'psychic barber'.

We seem to have a number of Scots posters. Why not write a nice letter to the editor?

[Drooper]

ALthough this is just a report, alarm bells immediately begin to ring. The report states the following.

The rules of chance would suggest an accuracy rating of 30%, but the mediums’ average was 70%, with some hitting 80% on some of the participants.


Two things require questioning about this remark.


Firstly, even if is true that 30% would be a score representative of chance (which in itself is difficult to assert due to the inherent qualitative nature of this tpye of thing), this says nothing aobut the variance. A 95% confidence interval might include scores of 70%.

Secondly, a proper statistical test would compare the accuracy of predictions as assessed by the targets, versus the accuracy of predictions as assessed by non-targets.

It may not be rigged or fixed (go on sue me:D ), but that leaves open the possiblity of incompetent statistical testing.