Thanks for boiling down all that statistical stuff.
Musicat, I think there’s a little more to it than that. He’s explaining why a seeming statistical hit for some validation of the claim is in fact weak, and that the particular test was not designed to evaluate small influences with small increases over chance, but rather large influences with big improvements over chance. Then he provides an analogy to make the distinction clear.
If peter morris needs another lead on Randi, he should check out today’s jref column.
Again, I’m not just looking for claims of unfairness, I want detailed descriptions of what was unfair and exactly how. Perhaps the people in question will give him the details. I note that Randi is being threatened with a lawsuit, and is offering to travel to Denmark to allow them to sue him. Doesn’t sound like the actions of someone who thinks he’s guilty.
You are missing the boat. Badly. I made this point several times to RickJay and thought I had finally gotten through, but I guess not. Here’s the deal, one last time.
There are two entirely separate issues - or potential frauds - we are dealing with here. One is the financial arrangement between Randi and his subjects. Randi signs contracts with these guys promising to pay them if they perform as specified. The issue here is whether Randi is defrauding them in not paying up. The other is fraud of a non-financial nature, and does not directly concern the test participants. It is between Randi and the public at large. Randi publicizes heavily his standing reward offer as a disproof of paranormal phenomena - no one has been able to cash in - surely if paranormal phenomena exist someone would have done so. And Randi apparently publicizes the results of individual tests as being debunkings of the specific phenomena being tested. The suggestion of fraud here is that these tests do not actually show what Randi claims they show, and the fact that no one has been able to win the money is not an indication that paranormal phenomena do not exist.
No one is making the first claim. At least not that I’ve seen. In fact, not many people care a whole lot about whether Randi owes some money to various anonymous test participants. I know I don’t. But what people do care about is whether Randi is correct in putting forward the results of his tests as he does. And this is where the allegations of fraud are being made.
So here’s the crucial point. The test participants are completely irrelevant to this second fraud. Anything that they may believe, not believe, or agree to has absolutely no bearing whatsoever on whether the results of the tests are being misinterpreted. So it is an egregious error to constantly harp on the fact that the subjects agreed to the terms ergo there can be no fraud.
If people cannot understand such a simple concept, there is no hope for any understanding I’m afraid.
We are going to have to disagree here. Randi does not state outright that none of the tests have statistical significance, but he does his best to give that impression. He discusses the overall likelihood, calling it “a figure well within expectation” (actually, 14%, as noted above) but conveniently fails to discuss the water results. He does initially focus, as you say, on the claims of the dowsers, but this is exactly what you would expect - I notice many posters here also prefer to focus on the claims of the dowsers. But Randi goes on to post a lengthy exposition on “Why then, do the instruments show such positive reactions, and what makes them move?”, ending off with his conclusion that “Divining is a delusion, and must be recognized as such.” Clearly the impression being given is that these tests discussed above support the conclusions given - not that they support the exact opposite conclusion.
I can’t force you to believe this, but I think any reasonable person, reading the Randi article without paying too close attention, would come away with the impression that his tests supported his conclusion and not the opposite. It’s even helpfully labeled “Results and Conclusions”, the conclusions being presumably what one can conclude from this test. In this manner, he has misrepresented his results.
Well I said “if”. I don’t know if there are any other instances. Peter Morris is alleging that there are. But I was responding to RickJay’s request for an example. Again, if this is Randi’s MO than he is a fraud.
IzzyR, Peter Morris alleges all sorts of things. Despite heavy challenges to do so, he is able to come up with cites for almost nothing.
By the way, have you heard the comments about IzzyR sexually interfering with small children? If they are true, he is certainly guilty of sexual assualt and rape…
[sub]I hope you take that in the illustrative spirit intended, IzzyR[/sub]
Oh, god. Brucie Dan Kettler and his page of hatin’ Randi. Fresh with Dennis Rawlins rant.
Having gone over that whole sTARBABY nonsense, I have found most of the complaints against Randi to be sourced by an immature and uncompromising source. Rawlin’s has legitimate complaints, but the ones against Randi are not legit.
Klass wrote a reply piece called “CRYBABY”, which unfortunately did not address the scientific issues Rawlins had with CSICOP. However, Klass pretty much skewered Rawlin’s on the emotional issues.
Not long after the sTARBABY mess, Rawlins decided to attack the National Geographic Society over the legitimacy of one of the Pole explorers. Rawlins was demanding surrender on all fronts from the NGS before someone pointed out that Rawlins really screwed up his math.
I tend to take Rawlin’s accounts of events with a very ,very large grain of salt.
Princhester didn’t molest that kid, he just hugged her once when she fell and scraped her knee.
See, distorted accusations.
I agree with you there are two separate issues. The reason we keep mentioning that the testees agreed to the test is because people keep saying that the tests aren’t fair tests of the testee’s abilities. That argument is separate from the second case, but it still keeps being made, so we keep rebutting it.
Now let’s look (again) at the water dowsing vs. the rest of the dowsing that day. Okay, you don’t think it’s fair to lump the numbers, we’ll not lump the numbers. The test was designed to look for a large difference in ability versus chance, not a small one. The small number of iterations per person (5 or 10) reflects the need to distinguish between an 80% performance and 10% performance, not a 22% performance and a 10% performance.
Let’s look more closely. If chance says 10%, then it is within chance for them to hit one out of five, every other time (“on” or “off”). (That’s a distortion of how chance works, but it illustrates the numbers). However, if each of them hit their individually within chance one out of five (everybody had an “on” day, versus some being “on” and some being “off”), the percentage for the total will automatically be 20% hits. That’s a substantial boost in percentage from a reasonable variation in chance happenings - only because of the small sample size.
For the test setup and number of iterations, 22% is not significant. In order to sort out whether the 22% on water that day was significant (albeit at a lower level than the dowsers thought they could perform) or whether it was just an anomaly of the numbers at chance performance, each individual would have had to perform numerous more iterations.
But Randi wasn’t running a test to sort high numbers of statistics, he was looking for a gross difference in performance. For the level of investigation of the test, it did not support dowsing to the level being claimed. Thus Randi’s conclusion.
You run a test, and to the level of sensitivity of your test (i.e. within the accurate measuring zone) you get null results, but there’s noise at the bottom of the scale that looks high, it looks suggestive. The most you can get is the need to run another test with a much finer scale to detect smaller levels. You cannot conclude that the test you ran supported the effect - it didn’t. You can say to the level of our test, the results were negative. You can say there was some noise that was suggestive but not significant for the test conducted.
Come now. Kettler has nothing to do with it. Frankly I never heard of the guy or his website. But I remembered the starbaby article from a previous thread, and as the link in that thread did not work, I googled for Rawlins & starbaby, coming up with that source.
If you are asserting that the linked article is not actually the one written by Rawlins, I’ll try to find another source for it. Otherwise your attack on seems like a sneaky and misleading cheap shot.
Well aparently there seems to be agreement that Rawlins’ dificult personality was a factor in the affair. Still, I don’t know on what basis you would say that his accusations were wrong. (FWIW, here’s the Crybaby article. A quick scan seems to say that 1) Rawlins did not allege cover-up until after the affair was over, and 2) that Rawlins was supressed primarily for being an obnoxious person, not in an attempt to supress his views. I don’t buy the first point, but the second is possible).
It seems to me that this is blatantly false. I don’t think that argument has been made at all in this thread - it is certainly not true that it “keeps being made”. I myself have emphasized that this is not the argument being made, and these “rebuttals” keep coming. Feel free to show me the arguments that these “rebuttals” are responding to.
Actually, I think people prefer to focus on the claims of the dowsers because it superficially seems to support their side. Which is precisely why Randi did the same. But the inability to recognize the distinction when it is repeatedly pointed out suggests a certain amount of general mudleheadedness.
I’m not exactly sure what you are doing here, but it appears as if you are saying that since we are not lumping water in with gold and brass we must look at each individual dowser independently. This is an error. And besides, we don’t know what the individual results were for the various dowsers. You are using the 22% aggregate number and then applying it to the 5 or 10 test per person - another error.
Okay, I reread some stuff. First, I will admit to an error in my statements about the 22% as applied to the group. I was looking at a different way of comparing the numbers, but that is not in fact the way Randi compared them. The 22% was based on the total number of runs (independent of who took each run) and the total number of successful hits.
Second, I looked at what you said about the two cases again, and I think I was on a different point than you. I agree nobody in this thread seems to be saying that Randi owes people money because they passed the test and he refuses to pay anyway. However, peter morris is making the point that the test Randi is using is an unfair test of the dowser’s ability. That has been a primary component of his complaint all along - that Randi takes a strawman of the claimant’s claims and then disproves the strawman. He specifically has commented upon using pipes for the dowsing test versus using natural water sources. So that is specifically a complaint that the test was unfair and not what the dowsers claimed to be able to do. peter morris is the one who asserts dowsers can only search out natural water and not pipes. Thus the need to point out that the dowsers themselves agreed to the test design, and got to actually practice and demonstrate their ability to detect water flowing in pipes when they knew the water was flowing. peter morris has argued that the dowsers must sign a form saying they agree the test is fair even if they don’t think the test is fair.
This is certainly a different idea of fraud than Randi is misrepresenting the test results to his readers to cover up suggestive but weak supporting data. It is also different than Randi is twisting the test results after the fact to not have to pay. This is a third case: that Randi sets up unfair tests that do not actually meet what the claimant says he can do.
However, after peter morris dropped out of this thread, several people made comments about the agreement in response to you. So on that point, you are correct. I was reading both threads and cross referencing to other material over a period of days, and got confused what arguments were in which thread. Sigh