52. Skeptic Lynne Kelly on Cold Reading Techniques (Podcast)

[hoggworks]

Originally Posted by Mizzark

Ya know, this question comes up every now and then about how we can distinguish telepathy among humans from mediumship. Has anyone ever thought of bringing an audio recording device along and seeing if you can get voices at the same time that the medium is supposedly communicating with spirits? If the voices match up with what the medium is saying then it should be pretty good evidence that the medium is probably communicating with a ghost.

What makes you think it's an audio communication? If it's audio, wouldn't other people be able to hear it? Wouldn't security cameras occasionally pick up very clear bits of audio from ghosts?

Wouldn't there be a single person who's died since electronic recordings started actually make a deliberate attempt to contact the living, and REALIZE that they could run around the world recording the same message in disparate devices?

Monitoring equipment seems a good idea, though, and I've thought that you just need to monitor a person's brain to see if they're communicating telepathically. If the person is interacting with another person's brain or a ghost, there's got to be a physical mechanism, right? My computer communicates with my router over the air, but that's only by the dint of a transmitter and receiver built into the computer. Similarly, one would have to assume the presence of a psychic antennae in a human mind; we ought to be able to detect the frequency at which psychic behavior occurs, and potentially monitor, record, and even simulate it.

If it's something that we can measure (which, if it happens in our brain, we should be able to), then we can replicate it, right? We should be able to turn us all into psychics, if the phenomenon exists at all.

[alextsakiris]

Originally Posted by hoggworks

It seems surprising that Alex wasn't able to grasp the difference between scaling something subjectively (such as on a scale of 1 to 10, where a judgment would depend entirely on how similar a person feels it is [for example, take "old." If I'm 50, I might think the person dying isn't old if they died at 55, because old is a relatively consideration for many people]), and scaling something objectively (such as a binary scale).

There is a huge difference between me saying "your loved one was 54" and "your loved one was old." The former only encompasses one year, while the second could potentially encompass several decades.

The deceased "targets" are piared to be dissimilar. "Old" will only fit one. Besides it all washes out in the scoring, if it's all random guessing you get random/chance results. Read: http://www.reason9.com/JP71_2007-200...el_Methods.pdf

[hoggworks]

So you're saying that if you and I are scoring the same reading, and old is used, I'd be picked for that reading because my deceased person is young and yours is old? Fair enough (and that seems a greater clarification than was given in the episode, and you did seem cagey about answering the question), but that doesn't seem to sidestep the issue of subjective versus objective. Old still fits with a bunch of ages, while a concrete age, an actual number, would only fit with one. So even if you're limiting it so that two readings can't be considered accurate to the same person, it doesn't sound like you're necessarily preventing that one person from getting a false positive, which seemed to be Lynne's issue in the interview.

Her point also seemed to be that cold reading works in the generalities, and since "old" is a generality, even if the medium can't get a feedback from the person they're giving a reading to, the techniques still work because there are a lot of those generalities available to them.

[hoggworks]

If it's been answered before, I apologize for the repetition and hope that I'll get pointed in the right direction, but what makes anyone think that it's reasonable to assume that a psychic/medium can operate strictly with a person's name?

[Mizzark]

Okay, I'm going to go out on a limb here and say that Mr. hoggworks, you are not open to the possibility that you are wrong. I will personally not address anything that you post on here until you:

1) Apologize for posting a bunch of issues on my thread about bad reasoning that Radford used that had nothing to do with what Radford actually said, like "Overwhelming Evidence" - something that Radford never mentioned in his original essay and thus is something that was irrelevant to my thread and probably used as a red herring - and caricaturing my Susan Blackmore example by introducing analogical aspects that were inaccurate and never in the original Radford example or my Susan Blackmore example, and
2) Apologize for posting your opinion about my post about EVP as it is irresponsible to do so when you clearly (from the contents of your post) know very little about EVP. No disrespect, but I wouldn't be surprised if you don't even know what EVP stands for.

I realize that this may seem closed-minded of me, but I think that there needs to be punishment for skeptics using bad reasoning flagrantly. (Not that I think that he will actually apologize, but I hope I'm pleasently surprised) I just wish I could have gotten more people who claim to be open-minded on this message board to understand. I think too many people on this board who claim to be open-minded are only open-minded about psi...and the kicker is I wasn't even asking anyone to believe the underlying ideas, merely to say that a skeptic was using transparently bad reasoning, when he clearly was.

[alextsakiris]

Originally Posted by hoggworks

So you're saying that if you and I are scoring the same reading, and old is used, I'd be picked for that reading because my deceased person is young and yours is old? Fair enough (and that seems a greater clarification than was given in the episode, and you did seem cagey about answering the question), but that doesn't seem to sidestep the issue of subjective versus objective. Old still fits with a bunch of ages, while a concrete age, an actual number, would only fit with one. So even if you're limiting it so that two readings can't be considered accurate to the same person, it doesn't sound like you're necessarily preventing that one person from getting a false positive, which seemed to be Lynne's issue in the interview.

Her point also seemed to be that cold reading works in the generalities, and since "old" is a generality, even if the medium can't get a feedback from the person they're giving a reading to, the techniques still work because there are a lot of those generalities available to them.

I think Lynne actually came around to the protocol we're using because it completely side-steps (to use your word) rater bias. The rater scoring the reading is completely blind to who the reading is for. So, there isn't any advantage to answering in generalities, or using vague catch-all phrases (e.g. this has happened... or may happen... to someone your close to... or may become close to).

[Mizzark]

I was thinking about something that Baischel said in her interview. (Maybe I should have posted it on that thread, but I only just thought of it recently) Maybe there is an innocent explanation for this, (like me not knowing what I'm talking about) but I thought that I heard her talking about how she was using the readings that the sitters scored low as the control group. It almost sounds like doing such a thing would be like hijacking the control group to try to make it fail as badly as you can and make the successful group look better. I think that it would be a good idea to have a group of regular people as control group #1, a group of cold readers as control group #2, and then the real medium group. That way if the control groups 1 and 2 both get similar results, you're pretty certain that the cold readers are neutralized. It may not be absolutely necessary, as Julie's protocol may also neutralize cold readers, but I think it would be a good idea, nevertheless.

[scomsjw]

Originally Posted by Mizzark

I think that it would be a good idea to have a group of regular people as control group #1, a group of cold readers as control group #2, and then the real medium group. That way if the control groups 1 and 2 both get similar results, you're pretty certain that the cold readers are neutralized.

That would be really interesting. But is it possible to get a group of cold readers? Perhaps they would have to be trained first.

The protocol seems watertight to me. But I have seen a lot of well-designed psychology experiments turn out misleading results that could not be replicated. I used to work in developing online learning tools and became frustrated with the amount of academic research into various learning styles. There were a huge number of experiments done to evaluate how effectively people could learn using different methods. When I did a literature review I found experimental results that contradicted each other. Yet the experimental protocols followed (designed by professional psychologists) seemed flawless. Its hard to see where things go wrong with well-designed experiments but it usually has something to do with the details of how the experiment was conducted. Something that you can't see from reading the research papers. You have to be in the room with the experimenters or have detailed transcripts of everything that happened. I have also seen this in the many attempts psychologists have made to evaluate the usability of interactive software systems. Watertight protocols but then the results can't be replicated.

[alextsakiris]

Originally Posted by Mizzark

I was thinking about something that Baischel said in her interview. (Maybe I should have posted it on that thread, but I only just thought of it recently) Maybe there is an innocent explanation for this, (like me not knowing what I'm talking about) but I thought that I heard her talking about how she was using the readings that the sitters scored low as the control group. It almost sounds like doing such a thing would be like hijacking the control group to try to make it fail as badly as you can and make the successful group look better. I think that it would be a good idea to have a group of regular people as control group #1, a group of cold readers as control group #2, and then the real medium group. That way if the control groups 1 and 2 both get similar results, you're pretty certain that the cold readers are neutralized. It may not be absolutely necessary, as Julie's protocol may also neutralize cold readers, but I think it would be a good idea, nevertheless.

When Dr. Beischel is talking about the "control reading" she referring to the reading not intended for the sitter. So, for every pair of sitters, one acts as a control for the other.

[alextsakiris]

Originally Posted by scomsjw

The protocol seems watertight to me. But I have seen a lot of well-designed psychology experiments turn out misleading results that could not be replicated. I used to work in developing online learning tools and became frustrated with the amount of academic research into various learning styles. There were a huge number of experiments done to evaluate how effectively people could learn using different methods. When I did a literature review I found experimental results that contradicted each other. Yet the experimental protocols followed (designed by professional psychologists) seemed flawless. Its hard to see where things go wrong with well-designed experiments but it usually has something to do with the details of how the experiment was conducted. Something that you can't see from reading the research papers. You have to be in the room with the experimenters or have detailed transcripts of everything that happened. I have also seen this in the many attempts psychologists have made to evaluate the usability of interactive software systems. Watertight protocols but then the results can't be replicated.

I hoping the open source process will allow more eyeballs to catch and fix potential problems.