Anecdotes and Scientific Hypotheses

Status
Not open for further replies.
Jabba said:
Ginger,
- I've hilited above what I think is the simple answer to my question. Anecdotes can increase, or decrease, the probability of a scientific hypothesis. To what extent they can depends upon various factors.
- I have a book containing 100 anecdotes in which the subjects claim to have seen or heard events or situations that they could not have heard or seen through normal sensory perception -- but that have been 'verified' by other witnesses to be accurate. Obviously, these other witnesses could be lying, or confused, or whatever, but reading the different stories, the witnesses and stories seem (to me) impressive, and do increase my own estimate of the probability that something paranormal is going on (though I already tended to believe).
- My ultimate claim here is that these anecdotes do support the possibility of an immaterial "self," and life after death. Obviously, they don't prove anything, but they do support those hypotheses -- and added to one's existing knowledge of relevant information, should increase ones estimated probability of immateriality and life after death.
Click to expand...

No they don't they are poorly sourced and hardly credible in any way.
 
I'm aware of a scientific test that was undertaken because the testers involved had no first hand knowledge of the subject matter under examination and wished to debunk something that they consider legend, not fact:

http://powderburns.tripod.com/sharps.html

In a nutshell, they didn't believe that a black powder cartridge rifle was capable of throwing a project out to 1538 yards, the stated distance of the Billy Dixon shot at the battle of Adobe Walls:

In the fall of 1992 the people at Shiloh Sharps were approached by a group of forensic scientists who were going to have a meeting at the Yuma Proving Grounds early in November. The were going to be allowed to use some newly unclassified radar devices to test the performance of various types of ammunition. Shiloh was invited to bring down some rifles and participate in the doings. Especially they wanted was a .50-90 So, Wolfgang Droege, previous Shiloh owner, Kirk Bryan, one of the present Shiloh owners, and Dennis Bardon, Shiloh’s custom gunsmith began making plans to attend. They also asked yours truly if he wanted to go, and I said I wouldn’t miss it.

However, I must admit to being a bit puzzled as to why they wanted to use such new—fangled radar gadgets to test such old guns. Well, when we got there we found out. It seems that one of the forensic scientists wrote an article in their newsletter saying that the Billy Dixon shot at Adobe Walls in 1874 could not possibly have happened. (Remember Billy Dixon knocked an Indian off his horse at a distance later surveyed to be 1,538 yards.)

Anyway, this particular forensic scientist did some calculations and arrived at the conclusion that a .50-90 Sharps (What Billy Dixon said he used could not have a bullet out that far. When I heard what this was all about thought, “That scientist is going to be embarrassed. He must not have fired Sharps Before. We all know they’ll throw a bullet that far.”

The scientist was indeed embarrassed.


From there on it was all fun. We elevated the muzzle to 45 degrees. The bullet again was 650 grains and started at 1,275 fps. It landed at 3,190 yards, but the most amazing thing was that it went up to a few feet shy of 4,000 feet and was in the air a full 30 seconds!

There is no substitute for hands-on knowledge, but that doesn't preclude or substitute for scientific testing.
 
Mojo said:
What if you consider only those anecdotes that seem to support a particular hypothesis?
Click to expand...

This is called confirmation bias. One tries to correct for confirmation bias using a complete set of controls.

Then you aren't using controls. A control is a sample that doesn't satisfy one of your initial hypotheses.

Suppose you are using the hypothesis that people who use brand X toothpaste have whiter and brighter teeth then people who don't use brand X toothpaste. The final hypothesis (whiter and brighter) is about people who satisfy and initial hypothesis (use brand X).

If you choose anecdotes solely about people who use brand X toothpaste, then you have chosen people who satisfy at least one hypothesis. However, you need to compare the results to people who don't use brand X toothpaste.

So you have to tell anecdotes about people who don't satisfy an initial hypothesis. Otherwise, you have confirmation bias.
 
Skeptic Ginger said:
No, that is not the answer. The probability any hypothesis is true is fixed in reality. The evidence supporting the hypothesis doesn't change reality...
Click to expand...
Ginger,
-From JayUtah in the immortality thread
JayUtah said:
Probabilities are pictures of uncertainty. We picture them informally as a percentage of belief that a certain proposition reflects fact. Slightly less informally we picture them as a number between 0 and 1 (just a percentage in proper fractional or decimal form) to which we can apply an algebra and a calculus that allows us to express complex relationships among uncertain values. In the full vista of that realm, the pictures of uncertainty are functions of one or more variables and methods of convolving those functions to produce, in the end, hopefully meaningful advice for how to behave in the face of the uncertainty they represent. We cannot resolve the uncertainty, but we can combine them in ways that let us reason about whether more or less uncertainty results from the combination.

Jabba is correct in trying to tell us that probability is inherently uncertain. The responses that tell him probability is necessarily still based on fact or involve fact are correct. But the dance that's being performed here is not as coordinated as all that.

The Bayes approach to probability is different in that it relaxes the rigor of measurement or simulation that would ordinarily build up the picture of some given uncertainty. Such as in the classic example of Bayesian search, we are allowed to quantify our expertise in a more nebulous fashion in selecting places we believe are most likely to to produce results, and then use actual search results to refine that expert opinion for what to do next. (Or, conversely, use the expertise to evaluate the strength of the search results.) But expertise in this example is still fact -- just not very articulable or otherwise quantifiable fact.

Bayesian inference may be used to reason in the face of uncommon uncertainty, but Jabba is doing it wrong. His approach is entirely uncertain. There are far too many degrees of freedom in the approach he's taken, and he knows it. Which is why he is so desperately trying to beg agreement on one point or another so that he doesn't have to justify the numbers he's simply pulling out of thin air. He wants to constrain his unbridled problem by tricking his opponents into not contesting the constraints he's simply applied arbitrarily.

But the most egregious failure on Jabba's part, in my opinion, is not that he is simply making up numbers. It is that he is proffering an inference as if it were a proof. Jabba is telling us that probability is not absolute as a way of excusing that he has nothing absolute on his side, and that he therefore needs to resort to a handwaving probabilistic argument rather than a factual one, not because it's appropriate but because that's all that's available to him.
Click to expand...
 
sphenisc said:
Not really.

It is possible to make the scientific hypothesis that "People tell anecdotes".
The observation of someone telling an anecdote increasing the probability that the hypothesis is true. Therefore the answer to the OP is "Yes".
Click to expand...
sphenic,
- Me thanking you for your support will only lower your standing in this forum -- but thanks anyway...
 
Skeptic Ginger said:
No, that is not the answer. The probability any hypothesis is true is fixed in reality. The evidence supporting the hypothesis doesn't change reality.

But even if you ignore the semantics of 'probability', the OP has another problem. He is conflating the observation and the conclusion. Conclusions are not evidence.

Person has NDE and believes it to be evidence of life after death. All you have is the person reporting that he saw certain things. It's not evidence he went somewhere out of his body and returned. It's not evidence he left his body.
Click to expand...
Ginger,
- The evidence that he left his body is that others agreed with his claim about what was happening in locations that were not reachable from his body's location.
 
Jabba said:
Ginger,
-From JayUtah in the immortality thread
Click to expand...

Nothing I said in that post in any way disputes Skeptic Ginger. The problem with your misuse of Bayes, and of statistics in general, is that your attempts at quantification are in no way founded in reality. They are your personal beliefs and wishes to which you have attached arbitrary numbers which you insist without evidence must be correct, and then beg others to accept them as mathematically validated models.

Further, it's an incredibly disingenuous move for you to attempt to quote me in your defense when I'm not defending you in the least, and to post this here expecting comment from S. Ginger and others when you yourself never bothered to comment on the original, or have bothered for weeks to address anything I've said challenging your claims. How very rude and selfish of you.

Comment on it now, please.
 
Hokulele said:
How is that "evidence he left his body" as opposed to "evidence someone told him what was happening"?
Click to expand...

It doesn't even have to be that overt. There is a confirmatory dynamic that often happens in those kinds of interviews and interactions where the subject and the witnesses unconsciously try to build a consensus, each "remembering" details that aid the formation of a coherent narrative. A fair amount of the training for conducting forensic interviews revolves around recognizing and avoiding that.
 
Jabba said:
Ginger,
- The evidence that he left his body is that others agreed with his claim about what was happening in locations that were not reachable from his body's location.
Click to expand...

And what other possibilities than 'leaving the body' have been explored and eliminated as possible candidates? ie confounding factors
 
Jabba said:
sphenic,
- Me thanking you for your support will only lower your standing in this forum -- but thanks anyway...
Click to expand...

My, your victim complex is getting out of hand.

You're just wrong, jabba. No one's out to discredit your brilliant ideas or anything. You're just wrong.
 
JayUtah said:
But then it's not an anecdote in the way Jabba intends it. He wants to use uncontrolled, self-reported anecdotes instead of systematically collected and controlled evidence.

Here's where it starts to get the most silly. http://www.internationalskeptics.com/forums/showthread.php?postid=11586163#post11586163
Click to expand...
Thank you for posting the link. I had not followed any of these threads, so I was interested to read that page!



Reading IS blows away any cobwebs of woo and woolly thinking, I find!
 
JayUtah said:
.A fair amount of the training for conducting forensic interviews revolves around recognizing and avoiding that.
Click to expand...


And that's the most important point. Scientificly conducted patient interviews seek to remove bias and (hopefully) any judgment on the part of the interviewer. Sample size and composition are key elements. The use of standardized tests may also help.

There's a huge difference between a properly controlled study of interviews and "anecdotes" as Jabba is using the word.
 
Loss Leader said:
Scientificly conducted patient interviews seek to remove bias and (hopefully) any judgment on the part of the interviewer.
Click to expand...

Or on the part of the interviewee. This is what litigant lawyers do upon cross examination. If the witness testifies she heard an "explosion," the other party may find fruit in questioning how she knew it was an explosion rather than simply a loud concussive noise (e.g., a steel beam snapping or a building collapsing). Forensic interviews also occur when we interview eyewitnesses to an accident. We want to get the information they have, not necessarily their interpretations or attributions.

Sample size and composition are key elements. The use of standardized tests may also help.
Click to expand...

Indeed we expend a great deal of thought in real science upon empirical controls (Are we studying the right variable? Are we studying it in relative isolation?) and statistical controls (How representative is our study of the phenomenon at large? How confident can we be in the strength of our findings?) This is where jt512 and caveman1917 disagreed with me in the other thread when I said that probabilistic arguments are not as persuasive as empirical arguments. They rightly pointed out that any variable we measure empirically is, in fact, an element of statistical probability. They are not wrong, but they are arguing apples to my oranges. What Jabba proposes to do is not a statistical control on a measurement but rather an abstract fairy tale told in the language of statistical probability.
 
Last edited:
SusanB-M1 said:
Thank you for posting the link. I had not followed any of these threads, so I was interested to read that page!

Reading IS blows away any cobwebs of woo and woolly thinking, I find!
Click to expand...

Thank you; I'm glad you found it useful, or at least entertaining. By way of explanation, this particular OP sometimes breaks out one question or another into a separate thread, in a separate part of the forum, to discuss them hopefully in a more abstract way. That's why this is suddenly being discussed here.
 
BStrong said:
I'm aware of a scientific test that was undertaken because the testers involved had no first hand knowledge of the subject matter under examination and wished to debunk something that they consider legend, not fact:

http://powderburns.tripod.com/sharps.html

In a nutshell, they didn't believe that a black powder cartridge rifle was capable of throwing a project out to 1538 yards, the stated distance of the Billy Dixon shot at the battle of Adobe Walls:

In the fall of 1992 the people at Shiloh Sharps were approached by a group of forensic scientists who were going to have a meeting at the Yuma Proving Grounds early in November. The were going to be allowed to use some newly unclassified radar devices to test the performance of various types of ammunition. Shiloh was invited to bring down some rifles and participate in the doings. Especially they wanted was a .50-90 So, Wolfgang Droege, previous Shiloh owner, Kirk Bryan, one of the present Shiloh owners, and Dennis Bardon, Shiloh’s custom gunsmith began making plans to attend. They also asked yours truly if he wanted to go, and I said I wouldn’t miss it.

However, I must admit to being a bit puzzled as to why they wanted to use such new—fangled radar gadgets to test such old guns. Well, when we got there we found out. It seems that one of the forensic scientists wrote an article in their newsletter saying that the Billy Dixon shot at Adobe Walls in 1874 could not possibly have happened. (Remember Billy Dixon knocked an Indian off his horse at a distance later surveyed to be 1,538 yards.)

Anyway, this particular forensic scientist did some calculations and arrived at the conclusion that a .50-90 Sharps (What Billy Dixon said he used could not have a bullet out that far. When I heard what this was all about thought, “That scientist is going to be embarrassed. He must not have fired Sharps Before. We all know they’ll throw a bullet that far.”

The scientist was indeed embarrassed.

From there on it was all fun. We elevated the muzzle to 45 degrees. The bullet again was 650 grains and started at 1,275 fps. It landed at 3,190 yards, but the most amazing thing was that it went up to a few feet shy of 4,000 feet and was in the air a full 30 seconds!

There is no substitute for hands-on knowledge, but that doesn't preclude or substitute for scientific testing.
Click to expand...

(Highlighted), perhaps he was, but the trial really weakened the anecdote. While it showed that a bullet might really have the required range, it also showed that it would be virtually impossible to hit a person at that distance.

At that elevation, at that flying time for the bullet, the gunner would have no chance of willfully hitting a target.

Hans
 
Jabba said:
Ginger,
- The evidence that he left his body is that others agreed with his claim about what was happening in locations that were not reachable from his body's location.
Click to expand...

What exactly was happening?

Hans
 
Jabba said:
Ginger,
- The evidence that he left his body is that others agreed with his claim about what was happening in locations that were not reachable from his body's location.
Click to expand...

That's not evidence. :rolleyes: It's just another uncontrolled, not systematically, retold anecdote.

I'm not sure why you think your arguments are making any sense. They aren't and you don't seem to understand why. But that's not my problem.
 
Last edited:
Jabba said:
- The evidence that he left his body is that others agreed with his claim about what was happening in locations that were not reachable from his body's location.
Click to expand...


Jabba,
- Your 'evidence', as you have described it, is 100 anecdotes out of all the millions of NDEs that have been experienced. What are the odds that 100 people out of all those millions had lucky guesses?
 
Last edited:
JayUtah said:
Or, shall we say, engaged in some shenanigan or another to improve the credibility of their claim, in order to sell books to credulous mystics.
Click to expand...


No shenanigans are necessary for these anecdotes to be worthless as evidence, because Jabba hasn't been systematic in his collection of them. He has no idea how many NDEs included wrong guesses. Or, indeed, how many accurate guesses don't involve NDEs.
 
Last edited:
Plus, how many included easy guesses? Most NDE happen in hospitals, for obvious reasons. Most of us can describe a credible scene in a hospital; one that others would recognize as "real".
Hans
 
MRC_Hans said:
Plus, how many included easy guesses? Most NDE happen in hospitals, for obvious reasons. Most of us can describe a credible scene in a hospital; one that others would recognize as "real".
Hans
Click to expand...


One of Jabba's claims is that:
Jabba said:
...on one plane ride I sat next to a somewhat "famous" neurosurgeon who had a patient with an NDE who was able to tell the surgeon what the surgeon had been doing in the next room.
Click to expand...


What's the betting that he was washing his hands?
 
Jabba said:
- Can anecdotes increase (or decrease) the probability of a scientific hypothesis?
Click to expand...


Anecdotes can increase or decrease the probability of scientific hypotheses, but Bayesian inference is comparative. The probability of a hypothesis is updated by comparing the support for the evidence under the hypothesis with support for the evidence under a competing hypothesis.

Your real question, as you belatedly disclosed, is "Do NDEs increase the probability of the hypothesis that there is life after death?" They do if, and only if, they are better-supported by the hypothesis that there is life after death than they are supported by the hypothesis that they are the result of neurologic changes in the brain under stress.

Which hypothesis, if either, is better-supported by NDEs? Good luck quantifying that, since you have no likelihood function for the data under either hypothesis. It seems to me that this evidence, like all your evidence, does not discriminate between your hypotheses.
 
1. The value of an anecdote as evidence relates entirely to the quality and demonstrated truthfulness of the anecdote. An individual claiming to see a ghost in the dark among a group of bushes at 150 yards is an anecdote that has near zero relevance in terms of evidence even if the person is honest. Another individual claiming to have dined with a ghost in a bright room for an hour may be reporting a better observational opportunity, but then the question becomes if their tale can be proven to be truthful.

2. No amount of poor evidence adds up to become good evidence. Never. Two thousand people providing very weak anecdotal "evidence" of ghosts does not add up to any good evidence of ghosts.

3. Anecdotal reports are useful as suggestions for things to be analyzed in subsequent rigorous, scientific studies; the latter are what produce any real and reliable evidence. The reports themselves are not part of the scientific evidence: they only suggest the type of scientific studies that can generate the scientifically valid evidence. When the anecdotal reports are tested scientifically with controls, etc. and found to not be reproducible then the anecdotes are false clues and become meaningless.

Unfortunately for your hypothesis of immortality/reincarnation, the anecdotal reports of NDEs have failed for these very reasons and when tested rigorously (as explained more than once in detail in your immortality thread and which you appear to acknowledge). Let alone the fact that NDEs do not involve death- they really are periods of unconsciousness from which people who haven't died recover. If someone has well and truly died and they came back to talk about it, I would be impressed. But I know of no proven example.
 
Let me tell you my anecdote. The other day I drove to the supermarket and as I got out of my parked car, I didn't see a ghost, nor anything else out of the ordinary.

'But that's not an anecdote', you complain - and I accept that really.

So anecdotes are self-selecting. There will always be more 'real' anecdotes about people seeing ghosts then there are about people not seeing them, even though for every one incident of reported ghost sighting there are billions of potential reports of non-sighting.

No one could ever report all the non-confirming instances. Note that my story about my supermarket-non-sighting works equally well for Bigfoot, The Loch Ness monster, UFOs, and so on...
 
ceptimus said:
Let me tell you my anecdote. The other day I drove to the supermarket and as I got out of my parked car, I didn't see a ghost, nor anything else out of the ordinary.

'But that's not an anecdote', you complain - and I accept that really.

So anecdotes are self-selecting. There will always be more 'real' anecdotes about people seeing ghosts then there are about people not seeing them, even though for every one incident of reported ghost sighting there are billions of potential reports of non-sighting.

No one could ever report all the non-confirming instances. Note that my story about my supermarket-non-sighting works equally well for Bigfoot, The Loch Ness monster, UFOs, and so on...
Click to expand...

The building I work in is 90+ years old. It was first built as a meeting place for WW1 veterans and contained a bowling ally and pool tables. First building in this town to sport hot and cold running water.
If one is here, alone, after hours, as it cracks and booms occasionally with the brick and wood expanding/contracting at different rates, I admit it can be spooky. I have been here 20 years and while others have said they experienced 'ghosts' I can say I haven't.
 
What I was at school a friend of mine had an after-hours cleaning job at a local firm. One day he told me that the previous night he'd been terrified of a ghost that had been there - it'd made strange noises, moved things around, and caused the lights to flicker on and off. He was terrified until another couple of friends who worked the same job confessed that they'd been doing it to him and explained how they did everything.

Another school friend was once in a church at night, doing some activity for the scouts, when the story of a local witch who had been gutted and burnt alive was told. Strange things started happening and when the kids ran out from the church, they found the ground was covered in entrails and charred bones. Again, he was terrified until at a later point his brother told him that he'd engineered the whole thing with some army buddies, who got the entrails and bones from a local butcher.

The point of these anecdotes isn't to say that all such occurrences are faked by people, but to point out that had the friends/brother never come clean, both of my friends would have a story of the supernatural that they were completely convinced was real. And because we weren't there, we can't possibly know what salient information was missing - in this case, that my first friend worked with another group of teenagers, and that with my other friend was there was a bunch of squaddies on the scene, one of whom he was closely related to.

This is why anecdotes don't count as scientific evidence, and why it's so hard to establish the truth about stories of the paranormal, because you have no way of knowing whether or not you've got the full story, and how important any missing information might be.
 
Jabba said:
- Can anecdotes increase (or decrease) the probability of a scientific hypothesis?
Click to expand...

This is such a poorly-phrased question that it is impossible to intelligently address. Why not try to rephrase is and be more specific?
 
Please take careful note that this thread has a topic that is not immortality, Bayesian inference of immortality, souls, or their definition. You should all be pleased to know we already have such a thread reserved for just that sort of discussion.

This is not that thread.

Discussions of anecdotes and their relationship to scientific hypotheses belongs here. That other stuff, not here.

Thank you for your anticipated cooperation in this matter.
Replying to this modbox in thread will be off topic  Posted By: jsfisher
 
Can anecdotes increase (or decrease) the probability of a scientific hypothesis?
Click to expand...

Jules Galen said:
This is such a poorly-phrased question that it is impossible to intelligently address. Why not try to rephrase is and be more specific?
Click to expand...

Yes. I believe others are responding as if it were phrased properly. Perhaps:
Does a large number of anecdotal testimonies contribute positively to verification of a hypothesis?
 
Last edited:
Jabba said:
- Can anecdotes increase (or decrease) the probability of a scientific hypothesis?
Click to expand...

jsfisher said:
Please take careful note that this thread has a topic that is not immortality, Bayesian inference of immortality, souls, or their definition. You should all be pleased to know we already have such a thread reserved for just that sort of discussion.

This is not that thread.

Discussions of anecdotes and their relationship to scientific hypotheses belongs here. That other stuff, not here.

Thank you for your anticipated cooperation in this matter.
Replying to this modbox in thread will be off topic  Posted By: jsfisher
Click to expand...
js,
- I'm claiming that anecdotes can be used as evidence for immortality, but you're saying that I can only talk about anecdotes in this thread -- that I can't talk about immortality over here...
 
Status
Not open for further replies.

Back
Top Bottom