How could such a ability be tested?

[Carn]

I often read in books and even saw on the internet the idea of spontaneous fortune telling. I actually just read about someone being able to do it.

I'm asking for idea's, how such an ability could be tested.
Application to JREF seems to be out of question, as this would require to know beforehand, when the spontaneous fortune telling happens and this is nearly impossible.(Unless one has a spontaneous fortune telling about when one will be able to have a spontaneous fortune telling);)
Therefore saying "apply or shut up" is no good way to deal with such claims.

4 big problems i see in testing such claims:

-constant observation or some other method is necessary to make sure, that prediction happened before any possibility of knowing something

-as it is said that such people experinece it as sensory input and is also accompanied with strong emotions, it would be unfair to require every prediction derived from such an experience to be correct to the detail(e.g. someone predicts a plane crash, gives a date and speicfies "during the night". If the plane crash happens during day, but on a very, very cloudy day, is it a hit or not?)
A example of the problems of interpretation is that painting guy, who claims to have seen 9/11 beforehand, but his picture of the event is pretty lousy and could be any plane crashing in any building anywhere.

-some things do happen anyway(several plane crash per year), so some predictions are pretty difficult to determine how likely a hit by chance would be. In priciple for every prediction, there would have to be a large statistical analysis to be done, to estimate how likely a chance hit is.(e.g. death comes to all of us, so predicting that someone will die until a certain date, would require to check how high the chances for that person to die is for the named time span and this would require the comparison with deaths of people with a similar circumstances, e.g. age, gender,...)

-a lot of people experince experience visions(e.g. dreams), which can also feel special, but are in no way paranormal. Therefore with a "real" spontaneous fortune teller such a vision could be mistaken as fortune telling and thereby lead to a seemingly failure of the ability. This makes it very hard to set a mark, how often predictions have to be right to pass any test, as even the fortune teller himself can often be wrong, in deciding whether his ability was at work or whether he just suffered normal hallucinations(obviously most people thinking, they have such an ability, fall prey to that problem)

Carn

[RamblingOnwards]

Originally posted by Carn
I'm asking for idea's, how such an ability could be tested.


A few ideas:
Every time the person has a vision, they mail it in an approved format to the tester. Once that vision comes true in their opinion, they record the actual details in the same format, together with whatever proof is available. Do this over a long period (say, 10 -12 visions) and take a score.

Now the scoring would be the contentious bit in this arrangement, but I think it could be worked out. Say we have a category 'time of day'. The testee may write
1. unknown
2. dark/light
3. within a specified three hour period.
4. within a specified ten minute period.

If the vision description says 'unknown' or the actual event can not be verified, no score.
If the description matchs
2: 1 point
3: 2 points
4: 3 points
If the vision does not match, -2 points. [If the vision description says between 10:00-10:10 and it happens at 10:15, this is a 'no match' situation.]

Pass level: ask the testee to write descriptions of a typically accurate past vision and event in this format, and score them.

Possible control: get five bored forum attendees to write imaginary descriptions that fit the general category ('plane crash', 'car accident', etc.) and score them using the same system against the actual event. Actually, it might be a good idea to use something like this to develop the score weightings.

Obviously, if the 'pass level' is not sufficiently higher than the control level, inform testee that they do not claim paranormal ability.

This is only possible if the actual events can be confirmed by external sources, of course.

[Traveller]

But, Carn, why would you want to investigate premonitions? As I understand the scientific method it proceeds from formulating an hypothesis to account for an observed phenomena through experimental confirmation of predictions made according to the hypothesis to a theory to explain how the phenomena occurs.

In the case of premonition where is your observed phenomena that you think need explaining?

[Carn]

Originally posted by Traveller
But, Carn, why would you want to investigate premonitions? As I understand the scientific method it proceeds from formulating an hypothesis to account for an observed phenomena through experimental confirmation of predictions made according to the hypothesis to a theory to explain how the phenomena occurs.

In the case of premonition where is your observed phenomena that you think need explaining?

I do not think any observed phenomena there needs explaining, in my eyes, it is all imagination, clever subconcious and luck.

I was musing, if i'm wrong with that, how a test could look like, that really something paranormal is going on.
A lot of believers constantly say "this is beyond your scientific measuring" and frankly with this stuff they might be right(though it is also beyond their abilities to determine whether something paranormal is going on, but that they just ignore).
I'm thinking about, whether with this stuff, it is really beyond testing.

Carn

[DevilsAdvocate]

Good question. This is a tough one (at least for me). I would hypothesize that premonitions are a result of a combination of coincidence, retro-fitting data to match the premonition, forgetting the misses, and changing the remembrance of the details of the premonition to fit the data. I think you summed these up nicely in your original post.

The only way to do this for something like the JREF challenge would be something like RamblingOnwards suggested where you would have to have an agreed upon scoring system for each premonition. So if protocol requires testing of 12 premonitions, then you would also have to have 12 mini-protocol agreements on the scoring. Even then, it would be tough to consider scoring for every possible interpretation of the premonition.

“I see it is night. There is a car crash that kills a famous person in the woods. (more details…)�. Ok. What if the crash was not at night, but the premonition was actually of seeing a news report on TV of the crash and the report was done at night? Or the report was done during the day in a studio but the viewer saw it on the 10:00 news when it was night? Or it wasn’t a car crash, but a truck. A motorcycle? Or the crash was in New York City but a block from Central Park? Or the news report about the crash showed a file photo of the deceased which happened to be taken in a wooded area? Or…It would never work. And the scoring system would be entirely subjective.

Unless the viewing is very precise, I think you will always get stuck with a subjective interpretation of the probability of whether premonition a could come true and whether the premonition fits the actual event.

A possible way to get some somewhat useful data might be to use blinded subjective judges. Have a tester write up, say 20, fake (control) premonitions that range from very probable to improbable and very precise to very vague. Then have the viewer submit any premonitions that they have that they feel may be “real�. Maybe even have the viewer rate the premonitions on how confident they are that it is a “real premonition� and not just a “dream or thought�. The viewer then submits descriptions of any real occurrences that they believe fulfilled the premonition. The tester also writes up descriptions of matching events of the control premonitions. Maybe even do your own investigation and add in any events that seem to match the premonitions for both.

Now have judges review both the control documents and the viewer’s documents and rate each on how much they believe the premonition was “real� (in other words, that the event was not likely to happen and that the hits were not vague enough to retro-fit most any event).

Then you can compare the judges’ rankings of the control documents to the viewer’s documents and see where they fit in. If there really is premonition but the viewer cannot detect the difference between any random thought or dream, then you should see a clear distinction in the rankings such that there is one group that is very high and one that is more lower with nothing in between. You could also compare the viewer’s ranking of whether or not the premonition to the judges’ rankings to see if there is any correlation.

You would probably need a fairly sizable amount of data to really draw any conclusions. The effectiveness of the control documents and the judges would limit the conclusions that could be drawn from the test. If the numbers are all over the place, the test proves nothing. And adding in the text for “real� and “non-real� premonitions obfuscates the data further. If you had a viewer that could always tell that a “vision� is really a premonition, then the conclusions would be easier to draw. If the judges, on average, ranked the control documents in order, then you could take the rankings of the viewer documents and conclude that the viewer’s premonitions are equal to such-and-such control “premonition�. This wouldn’t conclude whether or not there was really a valid premonition, but would match up the viewer’s premonitions to a standard of (perceived) probability and precision. I would expect that the ranking of the viewer’s premonitions would fall into the same rankings as the more vague and imprecise control documents.

Actually, I would expect that the judges would not have truly consistent rankings of the control documents and you would get some mostly meaningless data that could be interpreted to mean just about anything through meta-analysis. But that is the best I can come up with at the moment. :(