Slashdot posted this story on Sunday. The criticisms posted there matched the ones in this discussion. They included cherry picking the starting points for events and ignoring fluctuations that didn’t match a world event.
This reminds me of the bible code from a few years ago. Just wait for something to happen then go find the “prediction” ofter the fact.
I hate to point out that James Randi is not the most objective person on the planet. He has a million dollar reason to prove everyone who submits an application wrong. You might also considering that they don’t need a million dollars.
I had considered this only briefly. When I just read the original article (before I went to the website), I thought that perhaps they did just have fast computers seet in an infinite loop of floor(rand*2), and wrote a paragraph on this, but deleted it when I saw that they were using harder REGs. I kind of pictured in my head a little machine with a gap along its wires that the current has an equal chance of either jumping or not jumping, and I was OK with that setup. But if there’s a documented tendency towards non-randomness in one of the machines, this makes the whole work invalid.
I hate to be a pest, but can I get a cite for them doing this?
I think it’s becoming clear that what we have is an experiment which is unfalsifiable. I should clarify…
We have an experiment in which the experimenter claims that something is happening. The fact that he is uncertain of what this something is is currently immaterial. The fact that the experiment is designed to measure and quantify this something is immaterial. He has repeatedly claimed that there is SOMETHING anomalous occurring. Yet he has not stated under what conditions we should accept that something is happening and under what circumstances it could be shown that something is not happening. This is by definition an unscientific claim.
It’s similar to some people claiming that they can sense people’s auras through walls, and then stating that the color of the aura represents the kind of person they are seeing. If we cannot prove first that these people can detect whether there is an aura there, we should not be making correlations between aura color and, say, astrological sign. This is a severe case of putting the cart before the horse.
When I’m on a decent computer at home, I will be emailing Roger Nelson one question.
Under what quantifiable results would your claim that something anomalous is occuring be shown to be false?
I’ll let him know that I plan to post my answer publically, and if I get a response, I’ll be back here.
Wrong. It appears you have not read the Challenge. The money is offered for a VERY OBJECTIVE outcome to almost any paranormal claim.
If that is the way you feel, perhaps you would like to join us in this thread.
They don’t? Then why deprive the world’s poor of it? How selfish, to not win the money and give it to charity!
No, he’s not. However, the money is not his - it’s in an account that can only be used for awarding the prize. I can identify with Randi, and there’s nothing I would like more than to see proof, or even good evidence, for something that is not within our current understanding of the way the world works. I would bet my own money that Randi feels the same. And the researchers don’t need a million dollars? I admit that Princeton has a ton a money, but the individual researchers themselves are almost certainly not wealthy, and wouldn’t turn down an easy million.
No, not really. As Humanists explains above if you can make it through his post, given lots of data, it can be tortured until “statistically significant” correlations are found. I put that in quotes because someone with a good understanding of science recognizes that what appears significant at first blush isn’t really so if you wade through a large amount of data to find it. If I get a million people to flip coins all day long, I will be able to find several who get a highly improbable result. Is that significant? Not with a million data sets to choose from. And that’s what the PEAR researchers are doing.
Let me know when they use their data to actually predict something. Or even find an event while it’s happening. Put a computer on it with an alarm, and have it ring the alarm only when it detects that the EGGs are responding to a significant event in the world. THAT would be a good demonstration.
Sure they are, they just learned how to self-publish their data - check out the numerous Journal of Scientific Exploration cites on that page. Now, let’s look at the list of the staff for the journal and the PEAR staff.
Hey, what a coincidence - Robert Jahn is both the PEAR program director and the vice-president for the society that publishes the JSE. Brenda Dunne is the PEAR lab manager and also the education officer for that same society. But hey, I’m sure those publication credits are just as good as legitimately peer-reviewed articles in respectable scientific journals.[/sarcasm]
I don’t trust that Nobel comittee. They’ve got a multi-million dollar reason to prove everyone who makes a breakthrough or does something that benefits humanity wrong or non-benevolent.
Fortunately, I don’t have to trust them, as they are under a contractual obligation to award the money to whomever meets their requirements. So is the JREF. It’s contractually bound as a non-profit organization to seek out such claims and validate the ones that pass testing. But this sidenote really belongs in another thread.
That would be a good demonstration, if there were a clear quantifiable definition of the type of events that these EGGs are supposed to respond to. Otherwise, we’re left with “the machine went off, let’s go see what’s happening in the world right now, and if nothing’s happening yet, wait a few hours, because it’s predicting something.”
Dear PEAR team,
Sitting at my desk here at work, I’ve noticed that about 1-2 seconds BEFORE my cell phone rings, the CRT monitor for my computer starts acting incredibly erratic and there is clear evidence of (otherwise absent) static and noise. I’ve managed to use this amazing predictive power to answer my cell phone even before it rings so as not to disturb my co-workers.
Is this data useful to you at all?
I have, as of yet, been unable to cause my phone to ring simply by thinking about it. I’ve tried quite a few times. As soon as it does ring when I’m thinking about it, I’ll be sure to let you know.
Regards,
Jayrot
No one is forcing him to approve any conditions.
I’ve pooked around in that forum a few times. To me it seems completely possible to have a discussion about a general question without taking part in the flaming that goes on in great debates.
Or: “The machine always works. We just have to be smart enough to match it to events. If nothing is happening right now, it will, or it did, somewhere in the world (or universe). The machine is never wrong, we just need to interpret it.”
From such philosophies gods are created, religions made, and wars fought.
alterego, there is currently a thread open in Great Debates about the JREF challenge and it’s validity and/or application to the paranormal. If it is your wish to debate the merits of Randi, the JREF, or the Challenge with myself and several other reasoned posters on both sides, I suggest that the thread started for that purpose would be more appropriate. As such, I will post my response to your statement there.
Perhaps a little re-examination of forum concepts is in order.[list][]General Questions: for topics with a FACTUAL answer, not flaming[]The BBQ Pit: flaming allowed, within reason[*]Great Debates: For debates, natch. Like what you seem to be involved in?
Perhaps a little re-examination of forum concepts is in order.[ul][li]General Questions: for topics with a FACTUAL answer, not flaming[]The BBQ Pit: flaming allowed, within reason[]Great Debates: For debates, natch.[/ul] Like what you seem to be involved in?[/li]Sorry, fixed coding.
Not even close to being a fair comparison, if for no other reason than this: the Nobel comittee has actually awarded Nobel prizes to people. It’s really happened. We know it’s actually possible to win one. We know that there’s some point at which the Nobel comittee will actually recognize one’s work and deem it worthy of encouragement.
Sorry, maybe if the Nobel comittee had never awarded a prize. Maybe if Nobel had called himself “The Amazing Nobel.” How is it that James Randi has become the Ultimate Arbiter of Truth for every skeptic in the world?
I don’t think it means anything other than what I presented it as: an example of specifics and publications. It exactly doesn’t surprise me that the folks at *Nature *weren’t jumping all over this. And there’s definitely more than a little bit of active self-promotion going on here, no doubt. But is that really so unforgivable, given the “far-fetchedness” of the idea?
And it is Princeton, you know. It’s a pretty good school from what they tell me. Now it may not be as good as the All-Knowing Institute of Randi, but it’s not exactly the Center For Hollow Earth Studies, either.
Here’s the exact text of an e-mail just sent off to Roger Nelson.
That’s hard for me to understand. You’re saying there really is no point at which variance in this type of experiment would be statistically significant? None? I know you can torture data to say whatever you want, but I thought that’s one of the big reasons we have this field called statistics in the first place: to provide a mathematical basis for evaluating these types of claims.
So say we have a whole team of people flipping coins all day long (I think someone actually did this, looking for a reference…). Most people get within the expected range. Some people get, say 55% heads. No big shock. But there’s guy who’s getting heads 99% of the time. Easy explanations are eliminated. 99%. You’re telling me that it’s simply a highly improbable result, but not unexpected given all the people flipping coins?
Of course this is at the OPPOSITE end of the spectrum of significance than what the PEAR people are claiming, but that wasn’t the question. The question was, at some point, don’t the results become inexplicably significant? And if I’m understanding you, you’re saying they won’t. You’re saying that unless the current experiments end up producing a magical machine or something, the results will never be demonstrably signficant no matter what they are?
How about if they invent a microwave where you put dirt in and get hard currency out? 
This makes sense. So could one generalize to say that in order to demonstrate statistical significance, one has to be able to define a class of results that not only support the hypothesis but also a class of results that could show it is false?
Given an sufficiently large number of ways to interpret raw data, there will be at least one way to interpret the data that produces a stastical correlation as far oustide of the standard deviance for such a correlation as you care to search for.
A simpler example than my previous one. Let’s take one million coin flips. This leads to 2^1,000,000 possible outcomes, of which we will see only one. Thus, this answer is already stastically significant if viewed from the angle of “How likely was this particular order of heads and tails to come out” after the fact. After all, the odds against it were 2^1,000,000:1.
If we are allowed to define what sorts of correlations we are looking for post facto, we are no longer seeking patterns in randomness. We are seeking patterns in an already established larger pattern. If, in our long search for patterns, we come across one that seems to fit our data very well, we might be tempted to think we have found a result. This is jumping to a conclusion that’s not warranted. Our 1,000,000 coinflips that we have already done are no longer defined by chance. We have done them, and they are now in the past (which we will assume is unchanging and immutable). They represent a specific pattern of events that was extremly unlikely before the event, but is now the only result that exists. If in this pattern we find other, smaller patterns, we have not proven anything about the nature of randomness. We have only proven something about the pattern of flips we have already completed.
Now, if, on the other hand, the next 1,000,000 coin flips confirmed our same already-found pattern, we would rightly conclude that something odd is going on. If they instead can be manipulated to form yet another mathematical pattern, this still doesn’t say anything about the nature of randomness, only about the nature of the 2,000,000 flips we have already done.
This, in a nutshell, is why the scientific method works this way:
Observation -> Research -> Hypothesis -> Experimentation -> Conclusion -> Peer Review
In short, we do not start with an experiment and then pore over that experimental data in order to make a conclusion that our hypothesis agrees with. We say what we’re trying to find, then we try to find it. If we went the other way, no experiment would ever be judged a failure.