Proof of Immortality, the IV league

[Mojo]

Mojo,
- But the evidence (the claim that she was there in a red dress) was falsifiable.

It was an unsupported claim. It was not evidence of anything beyond an apparent belief that his wife was in the waiting room wearing red.

 

[annnnoid]

this is the 4th time i've brought this up and you've never addressed it. it's a major part of the reason why anecdotal evidence is only good for hypothesis generation. and you disagreeing with this puts you at major odds with the vast majority of the statistics world (i've never seen, nor heard, of a any statistician or stat book saying one can use anecdotal evidence to confirm a claim. it's always to generate hypotheses).

Innumerable scientific hypothesis have been created exploring issues of subjective experience. Since science has absolutely no way of empirically adjudicating subjective experience, every single one of these studies can only be initiated by, based on, and resolved with / through…anecdotal evidence.

Not to mention…that just about everyone everywhere all-the-time engages in just this kind of activity.

If I asked you right now …”did you in fact write that post?”

…you would answer with the anecdote…”yes I did write that post.”

No verification required. All that was necessary to confirm your (implicit) claim of authorship (that you wrote the post)…was anecdotal evidence.

 

[jt512]

In which case the medical profession are using anecdotal in a different sense to the rest of the scientific community.

Evidence for that claim?

 

[Nonpareil]

But the evidence (the claim that she was there in a red dress) was falsifiable.

The claim is not the evidence.

Going into the room and observing a lack of wives in red dresses is the evidence.

If I asked you right now …”did you in fact write that post?”

…you would answer with the anecdote…”yes I did write that post.”

That is still not what "anecdote" means.

No verification required.

And that is still incorrect.

All that was necessary to confirm your (implicit) claim of authorship (that you wrote the post)…was anecdotal evidence.

Edited by jsfisher: 

<snip> Edited for compliance with Rules 0 and 12 of the Membership Agreement.

 

[jt512]

most people will say 'yes' to this question as they don't want to admit they aren't voting

Evidence for that claim?

that's why the USC/LA times polls asked about 'enthusiasm' about their candidate.

Evidence for that claim? That is, evidence that they thought this proxy measure would be better than standard likely voter screens, and that they actually used it, instead of standard screening questions to determine likely voters. (ETA: In fact, they appear not to use enthusiasm to predict likelihood of voting. [1])

enthusiasm turned out to be a better predictor of whether they actually voted or not.

Evidence for that claim?

Also, you do realize that the USC poll was the least accurate of all national polls. The vote margin in their final poll before the election was off by 4.6 percentage points from the current popular vote tally—double the average absolute polling error of 2.3 percentage points—making USC the least accurate among the final results reported by 22 national polling firms.

and this is another reason anecdotes are bad evidence. self-reporting is filled with bias.

Self-reporting of voting intention, but not self-reporting of voting enthusiasm? Interesting.

it's a tough field. and designing even simple questionnaires can cost upwards of $200k.

I worked for more than 10 years in survey research. I'm not the one who needs a lesson in it.

______________

1. http://cesrusc.org/election/

 

[godless dave]

Slowvehicle,
- I can't remember what I've presented so far, and it would take me awhile to look them (or it?) up, but the book I'm slowly reading (The Self Does Not Die) is only $10 by KIndle, and contains 100 anecdotes that I consider evidence of immateriality...

Why should we believe any of the anecdotes? Why do you believe them? Anybody can write a book that says whatever they want.

 

[Mojo]

- I can't remember what I've presented so far, and it would take me awhile to look them (or it?) up, but the book I'm slowly reading (The Self Does Not Die) is only $10 by KIndle, and contains 100 anecdotes that I consider evidence of immateriality...

How many NDEs have there been that aren't included?

 

[Squeegee Beckenheim]

Innumerable scientific hypothesis have been created exploring issues of subjective experience. Since science has absolutely no way of empirically adjudicating subjective experience, every single one of these studies can only be initiated by, based on, and resolved with / through…anecdotal evidence.

These claims seem to tie in to the questions I asked you in post #1596, which you have thus far failed to answer. Since you have berated other people in this thread for not answering your questions I can only assume that you have failed to answer them by accident, rather than on purpose, and I give you the opportunity to do so now.

 

[John Jones]

Argumemnon,
- There is some evidence in favor of immateriality, and this evidence is appropriately considered when estimating the prior probability of immateriality.

Jabba,

Put up or shut up.

 

[Belz...]

Good luck, John. I've been telling him that for four years now.

 

[jt512]

for these reasons, controlled trials can provide evidence that we can generalize to the population. anecdotes can never do this.

Edited by jsfisher: 

<snip> Edited for compliance with Rules 0 and 12 of the Membership Agreement.

the hypothesis that the Zika virus can be sexually transmitted between humans was confirmed on the basis of a single case report.[1] No randomized controlled trial was ever conducted to test the hypothesis, and none ever will be, because it would be unethical and is unnecessary. The question has been settled decisively based on a single case report.


___________

1. http://www.nejm.org/doi/full/10.1056/NEJMc1604449

 

[John Jones]

[...] Since science has absolutely no way of empirically adjudicating subjective experience, every single one of these studies can only be initiated by, based on, and resolved with / through…anecdotal evidence.
.

Uh huh huh. You said adjudicate.

 

[Jedi]

Evidence for that claim?
Evidence for that claim? That is, evidence that they thought this proxy measure would be better than standard likely voter screens, and that they actually used it, instead of standard screening questions to determine likely voters. (ETA: In fact, they appear not to use enthusiasm to predict likelihood of voting. [1])

you may have noticed that i deleted the election post a few minutes after posting it (and before anyone responded). i was concerned that others (i.e., you) would focus on this rather than on the anecdote discussion we were having. so my subjective prior was confirmed by the data (your latest post).

i have no problem continuing the election discussion but how about we finish the anecdote discussion first? specifically:

you avoided addressing anything i wrote in the previous post, and so, for the 5th time now:

a well run controlled clinical trial generally has the following components:
1. Randomization. This helps assure us that we don't have a biased sample of, say, zinc believers only in our sample.
Anecdotes. sample is self-selected (non-random) and is usually comprised of believers. Example: 1000 big foot sightings in the past 100 years. seems like a lot of anecdotal evidence. but this number is never compared to the 100's of millions of visits to forests around the world during the same time who see nothing. anecdotal evidence is self-selected and biased in favor of the claim.

2. Masking. this helps assure that the respondent isn't biased by knowing he took the zinc or a placebo.
Anecdotes. hopelessly biased in this way

3. Placebo controlled. We have a comparison group that we know didn't get the treatment. so if we end up with 90% of the people saying zinc worked we wouldn't get so worked up about it if 90% of the placebo group said the same.
Anecdotes: only looks at zinc group. so if you got 90% of the people in the zinc group saying it worked you would be pretty ecstatic about it. and could be completely wrong. again, anecdotes are hopelessly biased.

4. Safety. controlled trials assess safety. anecdotes don't. does no good to take a drug that's effective if it can kill you.

5. statistically adequate sample size. controlled trials tend to have lot's of people based on the number of people required to confirm the results. Anecdotes are convenience samples with as little as 1 person.

for these reasons, controlled trials can provide evidence that we can generalize to the population. anecdotes can never do this.

this is the 4th 5th time i've brought this up and you've never addressed it. it's a major part of the reason why anecdotal evidence is only good for hypothesis generation. and you disagreeing with this puts you at major odds with the vast majority of the statistics world (i've never seen, nor heard, of a any statistician or stat book saying one can use anecdotal evidence to confirm a claim. it's always to generate hypotheses).

I worked for more than 10 years in survey research. I'm not the one who needs a lesson in it.

you've been telling a lot of anecdotes about your stat background. you work as a statistician, you explain stats to people all day at work, and now you've worked in survey research for 10 years. but all your anecdotes are a biased, self-reported sample of 1, and the data (your posts) don't bore this out. not to mention that these are all ad homonym arguments.

it's not only what you post but what you avoid responding to. and you're avoiding responding to the main thrust of the argument, that anecdotal data is biased for the reasons i wrote above.

if you're as much a statistician as you claim, it should be a piece of cake to answer me. so far, you've not responded to any of it.

 

[Nonpareil]

Edited by jsfisher: 

Moderated content redacted.

the hypothesis that the Zika virus can be sexually transmitted between humans was confirmed on the basis of a single case report.[1]

As has been explained to you repeatedly, this is not the definition of "anecdote" as it applies to this discussion.

Medical case studies are only anecdotal in that they are generally single reports made by a small group in an informal setting, and are therefore - as has been explained - not necessarily reliable. It is possible that the doctor in question was wrong about something, or fabricated data, or something of that type.

But, as they are a doctor, steps have been taken to verify that the data included in the report is accurate. Your own cited source includes supplementary appendices detailing the methods of data collection and figures showing the data itself.

Again, individual report, smaller group. The possibility of falsification or inaccuracy exists, which is why case studies are generally not considered one hundred percent conclusive evidence. However, the results are not entirely unverified, which means that they have some value as evidence, even if it isn't very much.

This entire thing about case studies is just a pointless semantic disagreement. The value of case studies as evidence comes from the fact that they are at least somewhat verifiable. Anecdotes, as we are discussing them, are unverified. If they were verified, they would have value as evidence.

But they aren't, so they don't.

 

[Loss Leader]

LL,
- This is certainly interesting...
- I think you're saying that "falsifiable evidence" is evidence that can prove an hypothesis false -- rather than, "evidence" that can be proven false. Is that what you're saying?

.


I am saying that evidence is anything that is necessarily true or necessarily false if the hypothesis is true.

Now stop dithering about your silly story book and the meaning on "evidence." You have yet to show that anything has ever happened that could only have been true if the immaterial soul existed.

Waste no more time. Confront yourself and honestly ask whether you could ever even entertain the idea that you are wrong.

 

[jt512]

you've been telling a lot of anecdotes about your stat background. you work as a statistician, you explain stats to people all day at work, and now you've worked in survey research for 10 years. but all your anecdotes are a biased, self-reported sample of 1, and the data (your posts) don't bore this out. not to mention that these are all ad homonym arguments.

I'm guessing you didn't intend the irony in that statement.

 

[jt512]

i have no problem continuing the election discussion but how about we finish the anecdote discussion first? specifically:

you avoided addressing anything i wrote in the previous post, and so, for the 5th time now:

I guess you missed my response here.

 

[jt512]

As has been explained to you repeatedly, this is not the definition of "anecdote" as it applies to this discussion.

As I'll explain to you just once, the discussion has been transferred over from another thread where the discussion was more general. It should have been left there, but the moderators, unfortunately stuck it back here.

 

[Jedi]

I'm guessing you didn't intend the irony in the statement.

it's only ironic to you

 

[Jedi]

<snip>

<snip>

<snip> the hypothesis that the Zika virus can be sexually transmitted between humans was confirmed on the basis of a single case report.[1] No randomized controlled trial was ever conducted to test the hypothesis, and none ever will be, because it would be unethical and is unnecessary. The question has been settled decisively based on a single case report.

i can see why you like anecdotes so much.
it was suspected after 1 case and confirmed after 15 more cases were found. it was also already known to exist in semen. and these cases are CONFIRMED cases, i.e., much interviewing and follow-up was done to vet that these are actual cases. but even if the confirmation is wrong, the CDCs mission is to swiftly investigate disease outbreaks. and they make recommendations inline with protecting the public good given the risk. so it's a no brainer to tell people to use condoms even if they are wrong.

of course it's not always ethical to conduct trial. people with terminal illnesses is a common example. but there are lot's of statistically sound studies you can do instead, even if they're not as good as a clinical trial. doesn't change the fact that trials are still the gold standard and anecdotes the worst.

but, as in your earlier zinc example, if condoms had a 1% chance of working, i highly doubt the CDC will encourage their use. would you use a condom with a 1% success rate? how about if i told you we had more anecdotal information that increase the chance to 1.01%? i wouldn't be buying stock in any condom companies just yet.

so, tell us how you get unbiased results with anecdotal data? given you worked as a statistician for years and years you do know that to mathematically show a statistic is unbiased you have to assume that the data used to calculate the statistic come from a random sample (among other things). biased samples lead to biased conclusions, whether you're a Bayesian or a Frequentist (user of classical statistics).

your zika anecdote is not a response to the fact that anecdotes are not random samples, have no control group, are unmasked, etc.. you have yet to respond to this (7th attempt).

Edited by jsfisher: 

Moderated content and response to same redacted.