Telekinesis, telepathy and their impact on science [Absolutely NONE]

[Robittybob1]

That's the joke point.

You need to be more precise. I'm also joking making my point but it is important that you be precise when it comes to statistics (Type 1 Errors must be controlled).

Edited January 29, 2016 by Robittybob1

[Robittybob1]

I am thinking of doing the experiment myself, and I'm just a little concerned that it might be against my religion to practice divination.

But then I remembered the supposed words of Jesus when he said something to the effect "if you had faith as small as a mustard seed you could say to the mountain be thrown into the sea", so that is a TK event like no other - moving a mountain - so moving a PSI wheel sounds like child's play.

[Phi for All]

I am thinking of doing the experiment myself, and I'm just a little concerned that it might be against my religion to practice divination.

But then I remembered the supposed words of Jesus when he said something to the effect "if you had faith as small as a mustard seed you could say to the mountain be thrown into the sea", so that is a TK event like no other - moving a mountain - so moving a PSI wheel sounds like child's play.

 

You should try moving a mustard seed. 2 milligrams.

[Robittybob1]

 

You should try moving a mustard seed. 2 milligrams.

Sounds like a good suggestion.

[imatfaal]

If I get time I'll go back and find it and see who made the suggestion.

 

I made a suggestion about flipping a coin every so often - but two crucial things; it was the same undisturbed apparatus and secondly it was blinded. Which I think was exactly what Arete suggested

 

 

You need to be more precise. I'm also joking making my point but it is important that you be precise when it comes to statistics (Type 1 Errors must be controlled).

 

That's not a type one error - it is spurious or unwarranted accuracy.

 

The null hypothesis is that the wheel is just moving naturally with air currents and/or vibration and that nothing the wannabe psychic is doing makes any difference. We are looking for evidence that the null hypothesis is not supported and that the experiment tends to make one favour a different hypothesis. Type one errors are the incorrect rejection of this null hypothesis - ie someone is too easily convinced that the boring explanation is not the correct one. Occam's razor is a kinda qualitative version

 

We would look for results where the chances of the data being consistent with null hypothesis is less than 5% or less than 1% or less than .5% etc. When we have achieved our desired level of improbability of the null being true but providing those results we can say that we no longer give credence to the null hypothesis.

[Robittybob1]

 

I made a suggestion about flipping a coin every so often - but two crucial things; it was the same undisturbed apparatus and secondly it was blinded. Which I think was exactly what Arete suggested

 

 

 

That's not a type one error - it is spurious or unwarranted accuracy.

 

The null hypothesis is that the wheel is just moving naturally with air currents and/or vibration and that nothing the wannabe psychic is doing makes any difference. We are looking for evidence that the null hypothesis is not supported and that the experiment tends to make one favour a different hypothesis. Type one errors are the incorrect rejection of this null hypothesis - ie someone is too easily convinced that the boring explanation is not the correct one. Occam's razor is a kinda qualitative version

 

We would look for results where the chances of the data being consistent with null hypothesis is less than 5% or less than 1% or less than .5% etc. When we have achieved our desired level of improbability of the null being true but providing those results we can say that we no longer give credence to the null hypothesis.

Sorry - I was playing and continuing the joke. I was not really saying that this was a type 1 error but joking about it. Now I didn't mean to play a joke on you.

I am pleased you have clarified who it was who proposed the high frequency changing experiment.

The only idea I wasn't clear on at present is what you mean by "blinded" in the part sentence "it was the same undisturbed apparatus and secondly it was blinded"?

[Arete]

Sorry - I was playing and continuing the joke. I was not really saying that this was a type 1 error but joking about it. Now I didn't mean to play a joke on you.

I am pleased you have clarified who it was who proposed the high frequency changing experiment.

The only idea I wasn't clear on at present is what you mean by "blinded" in the part sentence "it was the same undisturbed apparatus and secondly it was blinded"?

 

A single blind experiment is an experiment in which the subject is not made aware of the results of the study (e.g. a test drug is administered, but patients do not know if they got a placebo or a drug) . In a double blind, the researcher is also unaware until the experiment is concluded (e.g. a test drug is administered, but neither the researchers or the patients know who received a placebo or a drug).

 

In this case, you would conduct the study in a manner which prevented the person being tested for their TK ability from seeing the wheel when during the course of the experiment. The purpose is to remove subjective bias (a source of type 1 error) from experimental results.

[Robittybob1]

 

A single blind experiment is an experiment in which the subject is not made aware of the results of the study (e.g. a test drug is administered, but patients do not know if they got a placebo or a drug) . In a double blind, the researcher is also unaware until the experiment is concluded (e.g. a test drug is administered, but neither the researchers or the patients know who received a placebo or a drug).

 

In this case, you would conduct the study in a manner which prevented the person being tested for their TK ability from seeing the wheel when during the course of the experiment. The purpose is to remove subjective bias (a source of type 1 error) from experimental results.

We are going to measure degrees (as in 360 degrees = 1 circle) so I can't see how being able to see the PSI wheel will make one iota of difference.

Let Imatfaal answer the question please.

 

[i see someone has the cheek to give me a negative score for being correct and honest. Discuss the issue please instead of being so vindictive.]

Edited February 1, 2016 by Robittybob1

[Strange]

[i see someone has the cheek to give me a negative score for being correct and honest. Discuss the issue please instead of being so vindictive.]

 

I apologise. I hit the down arrow by accident. I actually think you deserve it for the "joke" but your question about blind testing is very important and shouldn't be denigrated.

[Robittybob1]

 

I apologise. I hit the down arrow by accident. I actually think you deserve it for the "joke" but your question about blind testing is very important and shouldn't be denigrated.

I forgive you Strange. I still find it frustrating when Arete talks of Type 1 errors when they are more of an experimental design issue.

As Imfataal pointed out correctly "Type one errors are the incorrect rejection of the null hypothesis".

[Strange]

I still find it frustrating when Arete talks of Type 1 errors when they are more of an experimental design issue.

 

That is equally true of type 2 errors.

[Arete]

We are going to measure degrees (as in 360 degrees = 1 circle) so I can't see how being able to see the PSI wheel will make one iota of difference.

 

The issue is one of cognitive bias. A person being tested has a tendency to record results which lead towards a conclusion which favors their preconceptions - even when they are attempting to remain objective. As a result, having results recorded by a subject, regardless of the experiment, leads to biased results. Given the nature of TK, it involves self reporting (i.e. you cannot objectively measure when a person is attempting to use the "ability" and when they are not) the only true way to eliminate that source of bias is to control for it by blinding the subject from the result. Same applies to the drug studies cited earlier.

 

Let Imatfaal answer the question please.

 

Imatfaal explicitly cited a previous post made by me when commenting about the blinded nature of the experiment. If he meant something different, I'm sure he can clarify.

 

 

I forgive you Strange. I still find it frustrating when Arete talks of Type 1 errors when they are more of an experimental design issue.

As Imfataal pointed out correctly "Type one errors are the incorrect rejection of the null hypothesis".

 

This is confusing, as we are discussing the design of an experiment. Controlling for both types of error is inherently a component of experimental design, analysis and interpretation. What is it exactly that you are considering separate to experimental design?

 

Also, any measurement made during the course of a experiment which is a false positive is a type 1 error, not just rejection of the overall experiment.

[Robittybob1]

 

The issue is one of cognitive bias. A person being tested has a tendency to record results which lead towards a conclusion which favors their preconceptions - even when they are attempting to remain objective. As a result, having results recorded by a subject, regardless of the experiment, leads to biased results. Given the nature of TK, it involves self reporting (i.e. you cannot objectively measure when a person is attempting to use the "ability" and when they are not) the only true way to eliminate that source of bias is to control for it by blinding the subject from the result. Same applies to the drug studies cited earlier.

 

 

Imatfaal explicitly cited a previous post made by me when commenting about the blinded nature of the experiment. If he meant something different, I'm sure he can clarify.

 

 

 

This is confusing, as we are discussing the design of an experiment. Controlling for both types of error is inherently a component of experimental design, analysis and interpretation. What is it exactly that you are considering separate to experimental design?

 

Also, any measurement made during the course of a experiment which is a false positive is a type 1 error, not just rejection of the overall experiment.

As I understood it Type 1 and Type 2 errors involve the analysis of the results and the acceptance or rejection incorrectly of the null result. This has to do with the level certainty you apply ie was it 90% 95% or 99% etc.

The design of the experiment will increase the amount of motion that is available in both the control and the test apparatus. So we can look for ways to tighten that up, but if the operator of TK says he must see the object that is turning you can't design an experiment that blocks his view. In truth it should make no difference for he can see the test and the control apparatus equally.

 

Well there could be a bias if the operator is also making the measurements, granted, but in this experimental case let's say it is all independently verified.

So tell me what other situation would cause a false positive movement to occur please? We must eliminate all possible cheating obviously.

 

That is equally true of type 2 errors.

What was equally true? Type 2 errors are the incorrect acceptance of the null hypothesis. That is opposite to Type 1 hence not really the same.

Edited February 2, 2016 by Robittybob1

[Eldad Eshel]

I have some progress to report. I tried to spin the wheel without covering it with my hands, and I have been successful. My hands are placed casually to the sides of my body, and I am also standing a bit further away from the wheel than usual. So you can rule out the idea of heat from the hands being the cause.

Edited February 2, 2016 by Eldad Eshel

[Robittybob1]

I have some progress to report. I tried to spin the wheel without covering it with my hands, and I have been successful. My hands are placed casually to the sides of my body, and I am also standing a bit further away from the wheel than usual. So you can rule out the idea of heat from the hands being the cause.

Would you say it was essential to be able to see the wheel? Did you accept the idea your view of the wheel should be blocked (blinded)?

 

In the above if your hands or body are not near the wheel, what are you doing to make it any different to being not in the room at all? Are you mentally focusing on the object?

Edited February 2, 2016 by Robittybob1

[Eldad Eshel]

Would you say it was essential to be able to see the wheel? Did you accept the idea your view of the wheel should be blocked?

First I am alone in my apartment, doing the testing alone. Second It is probably much harder to make it spin without looking at it.

[Robittybob1]

First I am alone in my apartment, doing the testing alone. Second It is probably much harder to make it spin without looking at it.

Unless there is some effort why do you say "harder"? What we are more likely to accept is that it spins less if you are not looking at it.

 

For that is what we want to test whether looking at it makes any difference at all. The null hypothesis would be equivalent to saying looking at the wheel makes no difference to the motion of the wheel. During the test the null hypothesis says you would get the same amount of movement more than 5% or 1% (depending on the level of certainty) of the time as when it is unobserved.

Does it ever appear to have moved when unobserved?

Edited February 2, 2016 by Robittybob1

[Eldad Eshel]

Does it ever appear to have moved when unobserved?

I look at it from far and it is completely still, always.

[Strange]

What was equally true?

 

That "they are more of an experimental design issue".

First I am alone in my apartment, doing the testing alone. Second It is probably much harder to make it spin without looking at it.

 

Until you make the effort to do some proper controlled and blind experiments, as suggested by various people here, you are just wasting your (and our) time.

[Robittybob1]

 

That "they are more of an experimental design issue".

.....

Type 1 and 2 errors are not experimental designs issues. Wikipedia:https://en.wikipedia.org/wiki/Type_I_and_type_II_errors#Definition

 

In statistical hypothesis testing, a type I error is the incorrect rejection of a true null hypothesis (a "false positive"), while a type II error is the failure to reject a false null hypothesis (a "false negative"). More simply stated, a type I error is detecting an effect that is not present, while a type II error is failing to detect an effect that is present. The terms "type I error" and "type II error" are often used interchangeably with the general notion of false positives and false negatives in binary classification, such as medical testing, but narrowly speaking refer specifically to statistical hypothesis testing in the Neyman–Pearson framework, as discussed in this article

.They seem to be more related to the degree of sensitivity and sample size. For example we could insist EE would have to repeat the experiment sufficient times so that the results would happen less than 1 time in a thousand by chance.

That is 0.1% chance it was due to random chance.

[imatfaal]

Firstly Rob - it is hubris of the worst sort to be told what a type 1 error is earlier on in a thread and later on start to argue definitions with people who actually know what they are talking about. From Strange's participation here I know he really knows his stuff and from Arete's blog I know he runs a lab at a very prestigious US University - and within a few days of learning a term of art you presume to be able to lecture them on the reality of the meaning.

 

And Second - "blinded" has been explained before. This is stop the bias of the experimenter - and this is not a slur everyone does it unless it is taken into account. If EE can see the psiwheel then he can "decide" that he was trying to move it everytime the breeze moves it and "decide" he wasn't trying when nothing happens. If EE cannot see the results of his effort then a proper comparison of him trying to move it and it moving can be compiled.

[DrP]

QUOTE: EE "I looked at it from afar and it is completely still, always"

 

Well that's that then, it must be TK.... solid proof at last!...

 

....If you can be sure you aren't breathing on it by using the bowl as we suggested, you are using a feather or a seed or something very light to make it easy.... (as we discussed, if you use the BS psiwheel then it's just too unstable and no-one will take you seriously). You have ruled out vibrations by allowing the thing to come to equilibrium... (don't let it settle and walk fast towards it - that will make it spin, even with the bowl the vibrations could set it off... unless you are using the speck instead of the dumb wheel, but I assume you dropped the wheel now or no-one in their right mind will take you seriously)

[Strange]

Type 1 and 2 errors are not experimental designs issues.

 

I was quoting you. Despite imatfaal's kind words, I am not an expert in this area. However, you said that type 1 errors are "more of an experimental design issue". I was simply pointing out that that is equally true (or not) of type 2 errors. Not a big deal.

[Robittybob1]

Firstly Rob - it is hubris of the worst sort to be told what a type 1 error is earlier on in a thread and later on start to argue definitions with people who actually know what they are talking about. From Strange's participation here I know he really knows his stuff and from Arete's blog I know he runs a lab at a very prestigious US University - and within a few days of learning a term of art you presume to be able to lecture them on the reality of the meaning.

 

And Second - "blinded" has been explained before. This is stop the bias of the experimenter - and this is not a slur everyone does it unless it is taken into account. If EE can see the psiwheel then he can "decide" that he was trying to move it everytime the breeze moves it and "decide" he wasn't trying when nothing happens. If EE cannot see the results of his effort then a proper comparison of him trying to move it and it moving can be compiled.

Good point about seeing it move and making it a positive result. That was why I was suggesting having two wheels side by side, one the test subject and the other the control, and both under glass bowls etc. That to me seems so improbable that one would move and not the other. Can you imagine that situation?

Ever since Arete started using the term "type 1 error" (seems like a week or more now) I have been trying to get him to define his term. All my definitions of type 1 error have come straight from sites that were teaching statistics.

 

Is there an alternate use of the term "type 1 error"?

 

I was quoting you. Despite imatfaal's kind words, I am not an expert in this area. However, you said that type 1 errors are "more of an experimental design issue". I was simply pointing out that that is equally true (or not) of type 2 errors. Not a big deal.

I do occasionally make a grammatical error. I have been known to miss the word "not" out of a sentence giving it the complete opposite meaning to what I had intended.

What I understand at this stage is that type 1 and 2 errors are not caused by experimental design issues.

[Strange]

What I understand at this stage is that type 1 and 2 errors are not caused by experimental design issues.

 

If your experiment is getting false positive or false negatives, then it could be that your analysis of the results is wrong, perhaps not taking into account some confounding factor or incorrectly caclulating measurement errors. Or it could be that your experimental setup is not eliminating some other influence.

 

As I say, this is not my area. Are there other sources of these errors? Why wouldn't experiment design not be a factor?

 

 

They seem to be more related to the degree of sensitivity and sample size.

 

Isn't that part of the design of the experiment?

Edited February 2, 2016 by Strange