“Have Scientists Finally Discovered Evidence for Psychic Phenomena?!”

[Exapno_Mapcase]

The subject was told to open his eyes when a tone was sounded in his ear and a light signal was flashed, then to close them again. The signal last five seconds and was repeated automatically every thirty seconds. Eventually the subject grew accustomed to the noise and fell asleep. But a flurry of anticipatory waves shot through his brains two and a half seconds before each sounding of the tone. Although asleep, the subject knew the tone was coming.

That’s from a description of early EEG experiments made by Alfred Loomis in the late 1930s. From the book Tuxedo Park by Jennet Conant.

Is it possible that Bem simply forgot about or didn’t adequately account for this effect? (I’m not plowing through a 62 page .pdf. but I hope somebody has.)

[Peter_Morris]

@mozchron

I just had to have a conversation with one of my post-docs today about the proper use of control. He had a great looking gel. I said “great! Where’s the positive control?”

He said he ran a negative control to control for contamination, but didn’t run a positive. “Who cares, he said. it worked - we don;t need a positive control”

I had to disabuse him of this. “How do we know that these results are real without a positive control? You need to throw this out and re-do the entire thing.”

He won’t do that again. I guarantee it, since i’ll fire him if he does.

Can you explain what exactly you mean by a positive and a negative control, and what is the difference?

Just as an example, suppose there is a claim that food additive XY123 causes cancer. You wish to test this claim.

You run a test with 5000 lab rats, give them food with XY123 in it. You also have a control group of 5000 rats, who have exactly the same conditions, except that their food does not have XY123 in it. The lab assistants who give out the food are “blind” in the sense that they don’t know which group is which.

Would that control be a positive, or a negative one? And what is the opposite control that you would also need?

[mozchron]

@Peter_Morris

Can you explain what exactly you mean by a positive and a negative control, and what is the difference?

Just as an example, suppose there is a claim that food additive XY123 causes cancer. You wish to test this claim.

You run a test with 5000 lab rats, give them food with XY123 in it. You also have a control group of 5000 rats, who have exactly the same conditions, except that their food does not have XY123 in it. The lab assistants who give out the food are “blind” in the sense that they don’t know which group is which.

Would that control be a positive, or a negative one? And what is the opposite control that you would also need?

Well, our experiments are usually a bit more prosaic than that.

We are performing a test to see if a particular sequence of DNA in in our sample. If it is there, it shows up as a band on the gel. The test id sensitive to contamination and false positives, so the negative control is a test sample with all the reagents but no DNA. If the band appears here we know we had contamination and throw out the results.

However, if bands appear, how do we really know it is what we think it is? We run a positive control that is spiked with the DNA. If this doesn’t come up as positive then we know something is wrong, even if some of the samples come up positive as well.

[mr.jp]

@mozchron

However, if bands appear, how do we really know it is what we think it is? We run a positive control that is spiked with the DNA. If this doesn’t come up as positive then we know something is wrong, even if some of the samples come up positive as well.

But the band appeared at the expected size, right? I don’t really see the importance of the positive control either, when the experiment worked.

[naita]

If you click the link to the follow up blog post at the bottom (PSI research: What do these numbers really mean) you get what I find a convincing put down of the research.

[mozchron]

@mr.jp

But the band appeared at the expected size, right? I don’t really see the importance of the positive control either, when the experiment worked.

Because if the one sample (the positive control) that is supposed to come up positive doesn’t, we can’t trust the assay. If we have some negative samples, how are we to say that they are true negatives or failures?

In the larger sense, we want our assay to be consistent every time we run it. if the controls don’t work then the assay isn’t consistent.

This is how good science is done.

[mr.jp]

@naita

If you click the link to the follow up blog post at the bottom (PSI research: What do these numbers really mean) you get what I find a convincing put down of the research.

That’s it? He didn’t correct for multiple testing? How ridiculous.

[mr.jp]

@mozchron

Because if the one sample (the positive control) that is supposed to come up positive doesn’t, we can’t trust the assay. If we have some negative samples, how are we to say that they are true negatives or failures?

In the larger sense, we want our assay to be consistent every time we run it. if the controls don’t work then the assay isn’t consistent.

This is how good science is done.

Hmm. But the sample is positive, so it’s not an assay that produces only negatives.

I think the positive control guards against a very specific error: That the whole assay doesn’t work, and a different factor by chance happened to make the same band as you had expected. Depending on the setup, this doesn’t seem very likely.

Including the positive control, other errors are still possible. It could be something else than your expected factor that caused the band in your sample, while the positive control has it’s positive band. If you want to be more certain of the result, you could also test it a different way instead.

[mozchron]

@mr.jp

Hmm. But the sample is positive, so it’s not an assay that produces only negatives.

The samples can be positive or negative. Therefore, if the positive control is negative, we can’t trust any samples that test negative.

And we do test and confirm by multiple different methods.

[Jackmannii]

@njtt

Most Published Research Findings are False

Including that one.

[Bad_Astronaut]

@the article

For example, Einstein believed that the mere act of observing something here could affect something there, a phenomenon he called “spooky action at a distance.”

There is so much wrong with this sentence it makes my brain hurt. Not that this discredits the research reported in the article, but it suggests that the author lacks familiarity with general science. This may help to explain why he believes that the correlations found in the experiment are comparable with the correlation between using condoms and not contracting HIV.

[TriPolar]

I postcognated the precognition of that paper’s publication. Postcognition is much more difficult than precognition. It requires the ability to postdict the prediction of something before it’s occurence. Those researchers have nothing until they explain the mystery of postcognitive prediction. Or maybe they just have nothing to start with.

[naita]

@Bad_Astronaut

There is so much wrong with this sentence it makes my brain hurt. Not that this discredits the research reported in the article, but it suggests that the author lacks familiarity with general science. This may help to explain why he believes that the correlations found in the experiment are comparable with the correlation between using condoms and not contracting HIV.

She.

[mr.jp]

This article pisses me off. If you test for 6 things at the 5% confidence limit, most likely one of them will be nominally significant. If the test are independent, you should use

overall confidence limit = individual confidence limit / number of tests

As I understand it, the article just proceeds to ignore this, and talk about stupid psychic babble.