He gets peoples vaginas all sandy which I find of great entertainment.
So who are these oh-so-pure folks who can’t be tainted by the work of Randi?
I have to say that when it comes to events or skills that are merely above chance I don’t think anyone is going to bring up the JREF challenge as a counterpoint. You are basically inventing an issue that does not exist or is not worth concerning oneself with.
No one statistical test can prove/disprove that - generally tests need to be replicated. But if you want a meaningful test within that constraint, all you need to do is come up with a measurement of what above the expected average by a statistically significant margin.
Off the top of my head, 3 out 10 right where 1 in 10 is pure chance is statistically significant by ordinary standards. (And the more trials you add, the less you have to exceed chance for it it to be significant.)
Truth is that Randi has a problem here, in that ordinary standards consider a 5% or 1% chance of random chance to be significant, and I don’t think Randi can afford to risk $1M on a 1% chance of random failure. So he would have to go higher.
But that’s not relevant. Whether or not Randi has practical obstacles to designing more meaningful tests doesn’t change whether the tests that he does use are meaningful or not.
What’s your point?
I disagree.
Buck Godot, the person’s name is James Randi, not Randy. His original name was Randall James Hamilton Zwinge, and he has used The Amazing Randi on stage.
For everyone else who is arguing if 10,20,50 or 95 % is the number to use, please remember that most claims are by people who claim 100% for their particular skill or talent. It’s only when the test is being designed that some fudging occurs, and if the test is allowed to accept less than 100% in a performance, it is only from the lenient graces of the testors. It gives the claimants an added advantage that they probably do not deserve.
Also note that 100% of the time, the claimants are able to perform when not blinded. Not 95%, not 50%, not 20%, but 100%. That is why the test is designed to NOT be subjective, but have an either/or kind of outcome. The water is in the jug or it isn’t.
Think about this for a minute. I will make a claim right here, that I can detect which of 10 translucent milk jugs contain water and which contain sand if you place them in front of me and allow me to touch them. 100% of the time. Perfect. Absolutely. Every time, no matter how many tests you desire, with my rod in my hand. Do you doubt me?
Now all I would be claiming, if I were applying for the MDC, is that I can do exactly the same thing if the jugs are hidden from view and I cannot touch them. If I truly have the power, and I have done this many times before with 100% success, and I can do it when everything is out in the open, and buckets over jugs have no effect on my incredible powers as I claim, why would I need to accept 90% or lower rate?
Sure, if I can achieve a 60% rate, it bears investigating (probably into cheating), but I have failed the Challenge. Not partly, not barely, but failed. I was unable to perform as I had claimed to do.
Absolutely.
No true paranormalist would ever be debunked in that fashion. Their reputations remain unscathed.
I don’t know if you are being serious or sarcastic here. But either way. I don’t have a lot of experience of the posters you mention but from what I have seen they may be a bit silly or dogmatic but are they utterly obsessed, are they continually abusive, can they engage in logical debate and are they completely dishonest?
PM had an entire website devoted to proving Randi wrong. There are message board and (I’m told) usenet postings by PM under various pseudonyms all over the internet where he tries to find geologists to back his views so he can selectively quote their comments while blanking from his mind what most tell him.
He gets upset and haughty and self righteous beyond measure if you are mean to him in the slightest way. Yet at the same time he continually calls other people liars and frauds in the most outlandish fashion but is completely unable to see the disconnect and hypocrisy.
As to his debating style, in that thread yesterday he spent several posts angrily demanding that Czarcasm say what he meant by something that Czarcasm had written in absolutely plain english. It would have gone on like that for pages if mods hadn’t stopped him. When cornered he starts to lie even to the extent of denying that he has posted things he posted the page before! (how mentally disconnected is that?)
I haven’t bothered posting anything about this in yesterday’s GD thread because, as I say, I think PM is ill and I won’t bother, but his cite of a page showing aquifers in the four state area is a classic of the genre. He linked to one map showing a particular aquifer and said that in areas not marked on the map you won’t find water within reasonable distance. The map in question was a figure from a USGS site on aquifers in the four state region. From reading that site it is utterly obvious that the region has five or six different aquifers all of which provide good water. The cite even says that 95% of the region are covered by these aquifers.
I find it impossible to believe that PM hasn’t read the whole site. He’s not stupid and it’s hard to see how he could even have got to the map without coming through the main site. Yet PM links to one map showing one aquifer. Now yes that is dishonest and jerkish but much more than that: it’s the actions of a mentally ill man. He is capable of just blanking from his mind anything that would upset his carefully nurtured obsession with Randi Being Wrong.
You know, in science usually a 95% confidence is required before a result can be considered significant.
So if you are testing a weak effect you need a large number of trials to get that level of confidence.
If you are testing a coin to see whether it is fair, if the coin has a 100% bias (it always lands heads), you don’t need many trials to determine with 95% confidence that the coin is unfair. If the coin has a 1% bias you’ll need a lot of trials.
But if you’re putting up a million dollars, a 95% confidence level isn’t fair. After all, at that level of confidence you’ll have to give away your million dollars on average 1 in 20 times, even if the subjects perform no better than chance. So it certainly isn’t fair to complain about the level of confidence needed to win the prize. It seems to me that a 99.99999 level of confidence is appropriate here–that is, you’d only expect 1 person in a million to win the prize by sheer chance. Anything else would be unreasonable.
And so the tests aren’t designed to find very small statistically significant differences that are only 95% or 99% likely to be due to chance. They are designed to find differences that are about 1 in a million due to chance.
If the subjects claim to be able to detect water 2 times in 10, when 1 time in 10 is expected by chance, they would need to conduct a very large number of trials to establish this at a confidence level that would be appropriate for winning the prize. It’s been 20 years since I took statistics, but I’m sure one of our resident stats experts could show the chi-square equation that would tell us the number of trials we’d need in that case. But it’s gonna be a lot higher than the number to get a 99% confidence.
But see, the thing about dowsing is that nobody who believes in dowsing thinks it has an effect slightly greater than chance. They think dowsing is damn accurate, 9 times out of 10. But suppose we say, well, that’s obviously false, but what if dowsing has a much smaller effect, like 2 times in 10 or 1.1 time in 10, rather than 9 times in 10. But then, why do no actual dowsers believe this? How did they become convinced that dowsing is 90% accurate? If they are so mistaken about the actual accuracy of dowsing, isn’t it much more likely that the whole thing is bogus than that they are really experiencing some real phenomenon, just much more subtly than they actually believe?
If I claim that I can read with my eyes closed, and then it turns out that I can’t, but I score slightly better than chance at guessing the next letter in a block of text, doesn’t that pretty much completely demolish my claim?
And at that level, scoring better than chance can often be explained by unrecognized systematic biases. Say, that you thought the letters in the block of text were well randomized, but weren’t.
And if we’re testing “find the one bucket out of ten that has the water”, well, there are lots of potential subtle clues that could allow the subjects to score slightly better than chance. Like, you know, scuff marks on the floor. Or non-random distribution of water droplets on the floor. Or a slightly different smell from the bucket. Or maybe one bucket is subtly warmer or colder than the others. Or all kinds of things that I can’t even think of.
And so, if we imagine that dowsing could be real, yet is only slightly better than chance, we’d have to really tighten up on potential biases. It’s easy to remove glaring biases, and debunk claims of 90% accuracy. It’s harder to remove subtle biases and debunk more modest claims. So to even start on this, you’d have to run multiple iterations of the trials, where you progressively try to eliminate these potential errors, and then see if the effect is reproducible in the next trial.
One simple trial where the subject scored better than chance isn’t good enough, because there are multiple explanations for the better than chance score. If the effect is profound you don’t need this. If the test is whether I can read a newspaper using normal vision, eliminating subtle bias is unnecessary, since getting every letter correct in a block of text containing 1000 characters would only be possible one time in 26^1000, if the characters were randomly distributed. But if I can read a large block of non-random characters extremely consistently we don’t even need to make sure those characters are randomly distributed, normal text will do. If I can open a random book to a random page and start reading aloud accurately, that’s a damn good illustration of by ability to distinguish letters by sight. If instead I’ve got an ability sniff a letter, and then by smell guess whether it is a vowel or a consonant, at a rate 1% better than chance, well, that’s going to require a lot better controls as well as a lot more trials.
How many out of 10 do they think pure chance would yield?
This is a neat encapsulation of where you are going wrong. There are an infinite number of things that could be true but which we have no reason to think are true. From invisible pink elephants under my bed to teapots in orbit around Saturn. There is no reason to even begin to look for them. The only reason we have to look for dowsing etc is because of specific outlandish claims. If when those outlandish claims are tested they are found to have no basis whatsoever, why would you think there might be some actual effect at a much lower order of probability?
If some guy on a streetcorner said there were gigantic fully visible spaceships shaped like rocking chairs floating over our heads, and someone looked up and said “actually there are none” would you jump in and call them a fraud because there might actually be gigantic invisible spaceships shaped like rocking chairs floating over our heads?
Your position makes no sense.
Basically what you are describing here is the inherent problem in proving a negative. If I have a psychic power that predicts a coin flip 50.001% of the time, even if its a legitimate power that is going to be hard to prove, and I don’t think you can fault Randi for not trying. 30% vs 5% should be doable, it will probably take 8 times as long as the test for a 85% vs 5% accuracy (37 successes required out of 170 vs 14 out of 21), but not impossible. If you have evidence of such a test being rejected by Randi I would like to see it.
As for the rest of us, extraordinary claims require extraordinary proof as it should be. All that Randi is saying is that thus far the extraordinary evidence has failed to materialize. I fail to see the damage that is done by rejecting those who have powers that are so subtle that it is impossible to test them.
What do you mean here by a statistically significant percentage. Without fixed test and p-value cut-off this seems meaningless to me.
I’ve made this point before, in general terms. But that just means that Randi has practical reasons for doing what he does. It doesn’t mean that his tests show more than they do.
It’s hard to know that, because if there are such people, they are less likely to be the ones making a living as dowsers, and they certainly are not going to be the ones taking Randi’s tests.
It’s human nature to think you have more power and ability than you actually have. Considering that dowsing (and most other paranormal phenomena) are hard to measure in any event, it would be surprising if most dowsers didn’t have an overinflated view of their own abilities.
Not to mention that if you are going to try to get people to pay you big bucks for your paranormal abilities it’s very much in your interests to make your claims as big as possible.
See above.
In addition, it’s frequently the case that folk notions that have persisted over many many years have some basis, although generally not to the extent thought by the masses and not for the reasons that they assumed. It’s even more frequently the case that folklore that persisted over many many years has absolutely no basis whatsoever. That’s why you would test it. But to the extent that you do test it, it’s a mistake to insist on limiting it to the specific levels of potency claimed by proponents, and even more so, to select for extreme claims by filtering out more moderate claimants by the structure of your tests.
Eeew you’re gross.
No, Glutton just doesn’t like being proven wrong. Again and again and again. Glutton, and people like him, are the reason why it’s so hard to engage in rational, factual debates on the I/P subject.
This, is, I think, the best example of how Glutton operates. Here, based on the state run Egyptian news’ second hand reporting on an Ha’aretz article, Glutton posted this, as if it were fact.
The only problem is, every single bit of the report in Al-Ahram was fictional, and Glutton never took the five seconds required to google the primary source and check for himself.
He’d much prefer it if I didn’t point out his errors and, I’m sure, would appreciate me being banned from commenting on such errors in the future.
Since the moderators of GD have requested that folks not bring up Randi in that thread, I am going to answer Peter’ latest psychotic outburst here:
[QUOTE=Peter Morris]
I thought I’d made that plain. I’ll make it simple for you. Basically it comes down to this:
- Randi is a total attention whore. He’s desperate to have people tell him how wonderful he is. He needs constant praise. The central message of everything he writes is "look at ME. What a wonderful thing I am doing. I deserve your praise for this. "
[/QUOTE]
This is at best a subjective observation and at worst an example of your taking your own psychotic opinions as ‘proof’. Randi has showmanship, certainly, he was an entertainer in his day and he uses that skill quite effectively. However I have yet to see anything resembling what you say is his ‘central messaging’.
Oh really?! So he was just getting attention when his work against Geller almost bankrupted him with legal costs? That was ‘just for attention’? When he worked with the victims of Faith Healers (described in painful detail in his book of the same name) that was just for attention?
How the fuck do you know what he is thinking beyond your own psychotic view of the man?!
So what?
Its hard to beat anyone in the game Randi is playing. In fact, it is a problem with all skepticism. You can win battles but it is very tough to win wars. The Straight Dope has known this for ages.
The distortions are Peter’s delusions. Randi certainly does discuss the battles he wins, but he also laments to wars being lost. The return of Peter Popoff is an example of that.
Really? Because you’ve had a lot of trouble proving a single one of these ‘lies people see right through’.
Blah, blah blah.
Gotta love how I am singled out here even though in the GD thread and other recent threads I have taken pains to avoid debating Peter about Randi directly. But the truth is, beyond attending a couple of the TAMs. I haven’t given any money to Randi, and frankly I did not attend TAM for him but others.
yawn. In other words it is just Peter’s delusions. But we all knew that.
As above, what I’ve said here is based on things I’ve seen a while back, IIRC. I don’t know that he has ever rejected such tests, but the ones for which I’ve seen the details have been along the lines of the dowsing one. And I’m sure it’s not the testees who insist on having such high levels of success - not much in it for them - so it has to be Randi, and if he requires such a high level for one test then he most likely requires it for all of them.
[Some people have trotted out the notion of “mutual agreement” to the tests as if it meant that both sides independently decided to use this level. But obviously the testees wouldn’t be insisting on a high level, even if they do believe they can achieve it.]
Understood. In any given test you need to calculate the p-value and select the confidence level etc. I’m not sure what you mean here. My original point in the section you quoted was about the difference between “the phenomenon is real” and “this specific guy’s claims about the phenomenon are true”.
Of course not. Everyone would like to set the bar as low as they think they can get away with. The fact is they agreed to the terms. They did not agree under duress. Randi did not trick them into thinking they were agreeing to different terms than were written down. “Can you do this?” “Yes.”
Well, that’s for sure. The tests don’t show more than they do. What they show is that every person who claims to have an ability to dowse was unable to demonstrate that ability. No more, and no less.
Yes, and this is true whether such paranormal powers exist or not. Except the existence of people who claim to have very easily demonstrated abilities, who then fail to demonstrate those abilities, is not evidence that there are people who have difficult to demonstrate abilities. In fact, rather the opposite.
You don’t understand the point about how someone could come to believe they have a very reliable ability, when in actuality if they do have such an ability it’s reliability must be very very low. There is no evidence for any such ability. The fact that you can form such a strong belief is evidence that your belief is entirely unjustified, not that it is weakly justified.
If some player tells you he can score with any woman, and then proceeds to hit on the next ten women at the bar, and strikes out with all of them, you are justified in rejecting his claim that he can score with any woman. Not only that, you have no reason to believe he has any ability to score with women.
No it isn’t. You don’t have to filter out the moderate claimants because there are no moderate claimants. When strong claimants fail miserably, that’s good evidence that the entire claim is bullshit.
If you think otherwise, well, go right ahead and start testing dowsers. Knock yourself out. I feel I’m justified in rejecting dowsing completely. If you disagree, set up your own fucking testing organization. Or, go to Vegas, and start dowsing cards. There are millions of people who visit Vegas every day, and lots of them have strong intuitions about what number or card is going to come up next. The experience of casino owners is that all these people are wrong. Take a look at the opulent Las Vegas skyline, which is a testament to the inability of human beings to predict random events at a rate greater than chance.
I just don’t get why one has to be a “scientist” to set up a test and a standard by which the test is passed or failed. Is this something only scientists get to do?
PM seems to think so. In reality, scientists may be properly equipped to analyze experiments with lab rats, who don’t cheat, but not people, who do. It may take a magician – someone who knows how humans fool humans, even themselves – to properly conduct a test. Cf. Jacques Benveniste.
[/QUOTE]
Sure he is. He won a “Genius” award in 1986.
So there 