UFOs: The Research, the Evidence

[Rramjet]

Let's start with your reversal of the order in which the hypothesis and the null hypothesis are posited, because that's very important. The null hypothesis should never be posited first, because the null hypothesis is created in order to be disproved in support of the hypothesis.

That’s neither here nor there, as long as the null and alternative hypotheses are clearly delineated, then all that really matters is that:

”We give special consideration to the null hypothesis. This is due to the fact that the null hypothesis relates to the statement being tested, whereas the alternative hypothesis relates to the statement to be accepted if / when the null is rejected.” (http://www.stats.gla.ac.uk/steps/glossary/hypothesis_testing.html#h0)

In this case the hypothesis (your hypothesis, which you have continually offered in one form or another)…

I was merely working off the hypothesis RoboT first proposed and that you subsequently quoted in the post of yours that you stated that I needed to address – so lets not try and get too cute wollery…

…can be stated as, "Some UFO cases are not mundane in origin" or if you prefer "Not all UFO cases are mundane in origin". …

Sure, I have already explicitly acknowledged that substitution is valid.

The wording is important because it covers all mundane possibilities, not just error in identifying extant objects, and this is important because some UFOs are known to be the result of hoaxes, delusions, lucid dreams, hallucinations and optical illusions where no physical object ever existed to be misperceived.

Sure but given independent studies have shown the incidence of hoax/psychological is 1-2%, then lets not get too obsessed with those alternative – the vast bulk of reports (according to the UFO debunkers) arise from a misidentification of mundane objects.

The hypothesis should never include any unwarranted assumptions, and your formulation does i.e. it assumes that there is a physical object (or objects) at the heart of every UFO report.

It is not my formulation – it was RoboT’s, then yours… So let’s not try to misrepresent here wollery, there is no need.

My formulation of the hypothesis includes such reports, and is thus the simplest and therefore easiest to test.

I have acknowledged you can substitute in your terminology…

So, having formulated the simplest hypothesis with no assumption…

I would not go that far (but we’ll come to that soon enough I guess)…

…we now take the inverse as the null hypothesis. This then is "NOT {not all UFO cases are mundane in origin}", and removing the double negative gives us, "All UFO cases are mundane in origin. This then is a testable null hypothesis, since all we need to do is show one case that's not mundane in origin, or show statistically that identified and unidentified UFOs are not drawn from a single population, but are instead drawn from two, i.e. mundane and non-mundane origins.

The general approach is valid.

This leads us to the next problem, which is how to test the null hypothesis. The first way to test it, as noted above is to show one UFO report which is definitively non-mundane in origin. This is of course problematic, because UFOs are by their nature transient events, so such proof would be almost impossible to come by.

Sure…

The second method is to demonstrate statistically that identified and unidentified UFOs are not drawn from a single population (i.e. mundane origins). However, this is also problematic, because identified UFOs are drawn from a huge range of populations, some with common properties, some without common properties, which may be more or less identifiable than others, and some of which (hoax, delusion, lucid dream and hallucination) have the potential to be totally dissimilar from any identified UFO, which will skew any attempt at statistical analysis.

Sure but as hoax etc (according to the independent research of the Air Force, Hynek and Hendry) is between 1-2%, then it is not significant. We can account for such in our analyses. (I’ll get the references for you later – but you can be sure I do have them and they are valid…).

You have correctly pointed out (or rather, picked up repeated my point from a few pages back)…

Again, let’s not get too cute wollery…

…that the quality of the reports may affect whether or not they can be identified, and that this may skew the statistics.

Sure…

This leads to the question, how can we account for the reliability of the report? The problem here is that we are now into a fairly subjective area. As has been shown in this thread over and over, the reliability of witnesses and data are often not clear. If you have any idea how this subjectiveness can be overcome then I'm sure we'd all be happy to hear it.

You assess the reliability of the reports on a number of factors. For example:

Experience of observer (occupation, training, age, etc)
Internal consistency of report
General quality and completeness of report
Observer's fact reporting ability (manner of describing the sighting)

But there is still one confounding factor that you haven't mentioned, which is that there may be UFOs which have mundane origins that are more likely to result in a report remaining unexplained, even if the report is of high quality, and there may be mundane origin UFOs which are often attributed to the wrong mundane cause. These would both skew the statistics and could lead to a false conclusion.

There will be many more false positives (non-mundane explained as mundane) than false negatives (mundane not explained) - and this is simply because there are an overwhelming majority of explained reports. This acts to wash out any potential differences between the known and unknown category reports and actually works in favour of the null hypothesis. I am willing to cop that handicap.

Any statistical study must attempt to take all confounding factors into account, and correct for them simultaneously. It would be a very brave man indeed who took that analysis on.

I am willing to risk it if we both are willing to accept the results. Let the chips fall where they may.

 

[RoboTimbo]

That’s neither here nor there, as long as the null and alternative hypotheses are clearly delineated, then all that really matters is that:

”We give special consideration to the null hypothesis. This is due to the fact that the null hypothesis relates to the statement being tested, whereas the alternative hypothesis relates to the statement to be accepted if / when the null is rejected.” (http://www.stats.gla.ac.uk/steps/glossary/hypothesis_testing.html#h0)


I was merely working off the hypothesis RoboT first proposed and that you subsequently quoted in the post of yours that you stated that I needed to address – so lets not try and get too cute wollery…

Mine was still better than yours but I'm not the one who claims to be a scientist.

Sure, I have already explicitly acknowledged that substitution is valid.

Your substitution is invalid. wollery's original is valid. Don't be too cute in allowing the correct wording to be substituted for yours.

Sure but given independent studies have shown the incidence of hoax/psychological is 1-2%, then lets not get too obsessed with those alternative – the vast bulk of reports (according to the UFO debunkers) arise from a misidentification of mundane objects.

Your lackadaisical attitude is why we must account for those also. A real scientist wouldn't play fast and loose with a null hypothesis.

It is not my formulation – it was RoboT’s, then yours… So let’s not try to misrepresent here wollery, there is no need.

Again, don't be too cute. Your formulation was the one with the unwarranted assumptions and that's the one wollery was referring to. There is no need for you to lie about it.

I have acknowledged you can substitute in your terminology…

I've acknowledged that we cannot substitute yours.

I would not go that far (but we’ll come to that soon enough I guess)…


The general approach is valid.


Sure…


Sure but as hoax etc (according to the independent research of the Air Force, Hynek and Hendry) is between 1-2%, then it is not significant. We can account for such in our analyses. (I’ll get the references for you later – but you can be sure I do have them and they are valid…).

No, I'm not taking your word for anything and the percentage doesn't matter anyway. In the correct wording of the null hypothesis, they are accounted for. In yours they weren't. Nothing more, nothing less.

What percentage were oil well flares?

Again, let’s not get too cute wollery…

Follow your own adivice.

Sure…


You assess the reliability of the reports on a number of factors. For example:

Experience of observer (occupation, training, age, etc)
Internal consistency of report
General quality and completeness of report
Observer's fact reporting ability (manner of describing the sighting)


There will be many more false positives (non-mundane explained as mundane) than false negatives (mundane not explained) - and this is simply because there are an overwhelming majority of explained reports. This acts to wash out any potential differences between the known and unknown category reports and actually works in favour of the null hypothesis. I am willing to cop that handicap.


I am willing to risk it if we both are willing to accept the results. Let the chips fall where they may.

I think you missed reading part of his post.

 

[Astrophotographer]

I am willing to risk it if we both are willing to accept the results. Let the chips fall where they may.

What a bunch of nonsense. Rramjet is trying to corner everyone into thinking he is going to do some sort of independent study but all he is doing is reciting BBSR 14. Once he gets somebody to agrree to this, he will just flash the preselected BBSR14 results and say it proves his point. He would have to do absolutely nothing and already knows the conclusion. There is no risk involved contrary to what he is stating.

When I challenged him to do something similar using the NUFORC data, where he would have to do the work, he refused to do so (impying it would be a waste of time). So, he really is not interested in taking risks at all.

Come on Rramjet. If you are going to do this statistical study, you have to do it with fresh data. Use the entire year of 2010 from NUFORC (4468 reports by my count) and perform this analysis. That would be taking a risk, where you don't know the conclusion. You would also have to have your work presented to JREF for review, where we can determine if you correctly assessed the "reliability" of the report. Are you up for such a challenge or are we going to hear some more "clucking"? Are you really willing to take that risk? Inquiring minds want to know.

 

[John Albert]

The statement merely conceptualises “weight of evidence” for us and allows us to assess any evidence we might come across (including anecdotal) in light of its principles.

Except that the article you cited is not talking about stories, but quantifiable and measurable physical evidence.

The article the quote was drawn from is actually irrelevant to the principles outlined in that quote.

This assertion makes no sense. If the article contains principles irrelevant to its basic premise, then it's a nonsensical article and unworthy of consideration.

The truth is, the article isn't nonsensical. It is a very well-written piece that clearly illuminates the subject of computerized analysis and pattern-matching of quantifiable, physical evidence in forensic science. Your argument that "The article the quote was drawn from is actually irrelevant to the principles outlined in that quote" is wrong.

Your use of that article to support the idea that claims equal evidence is a false comparison. You mined those quotes entirely out of context from an article that has nothing to do with your argument. It's a false analogy. Carefully measured pieces of fragmentary physical evidence is not the same thing as a bunch of incredible stories.

That is, the quote constitutes a general statement of principle – applicable to all evidence – including anecdotal.

Wrong. Anecdotes are claims, not evidence.

This, perhaps unsurprisingly, seems a somewhat elitist approach to science and logic. Does the JREF not purport to promote scientific and critical thought among the general population?

I can't speak for anyone else here, but I'm certainly not elitist. Social standing, economic income, professional occupation, formal education and the like mean very little to me as long as somebody can discuss matters in an honest and rational way, with basic respect and consideration. I just don't suffer fools and liars very graciously. When somebody tries to screw me in a discussion by misrepresenting or ignoring my input, I'm not going to just let it slide.

And if that message is received by a member of the public and they note a concern about an alleged scientific methodology or a flaw in some logic or themselves promote a statement of principle – only to then be told their opinion is invalid because they are not a “professional scientist” …even if that were true, it kind of defeats JREF’s whole reason for being doesn’t it?

Nobody's using their professional position to bully anyone or level any criticism that isn't 100% valid.

Many people who are not professional career scientists have pointed out the gaping, cavernous holes in your rationale and methodology. Yet you simply ignore them and trundle on, tossing around the same foolish assertions over and over again, page after page.

Then, when a professional scientist does happen to weigh in on your flawed approach to research, you get all butthurt and cry persecution. Blaming others for your own failures is cowardly and dishonest. It's another hallmark of that "pseudoscience" thing we were talking about in the other thread. Remember?

Personalization of issues

• Tight social groups and authoritarian personality, suppression of dissent, and groupthink can enhance the adoption of beliefs that have no rational basis. In attempting to confirm their beliefs, the group tends to identify their critics as enemies.
• Assertion of claims of a conspiracy on the part of the scientific community to suppress the results.
• Attacking the motives or character of anyone who questions the claims (see Ad hominem fallacy).

 

[Access Denied]

In an infinite universe anything is, of course, possible – but merely being possible does not equate to probable, likely or even plausible.

Congratulations on debunking yourself for the umpteenth time, by that reasoning you must now concede “aliens” aren’t plausible.

That or make a fool out of yourself again trying to backpedal…

[grabs another bag of popcorn]

 

[corbin]

How does something at "great height" or satellite height (essentially something that is above the atmosphere), not get affected by the exact same atmospheric effect that makes the stars twinkle?

I'm not sure on the technical reasons as to why, but if you have ever seen a satellite you know that it doesn't twinkle. It sometimes gets steadily brighter or dimmer like dimming a light, but it doesn't twinkle.

Even objects as far away as PLANETS generally do not twinkle unless the atmosphere is particularly turbulent.

 

[John Albert]

I'm not sure on the technical reasons as to why, but if you have ever seen a satellite you know that it doesn't twinkle. It sometimes gets steadily brighter or dimmer like dimming a light, but it doesn't twinkle.

Even objects as far away as PLANETS generally do not twinkle unless the atmosphere is particularly turbulent.

Because of the parallax effect, planets are reflecting a larger area of light onto Earth than the tiny points of light we receive from distant stars. Because the beam of light we receive from a star is so narrow and its rays are aligned almost parallel, it is easily deflected by dust ant other particles in Earth's atmosphere.

A planet in our own Solar System, however, presents a relatively much larger area to the Earth, so a lesser percentage of the light is susceptible to being diverted by atmospheric particles at any point in time.

I don't know if that makes sense, but it's the best way I can think of to describe it.

 

[Access Denied]

Sure but as hoax etc (according to the independent research of the Air Force, Hynek and Hendry) is between 1-2%, then it is not significant.

So by that reasoning do you not agree the final 5% unidentified figure reached by the Air Force and Hynek is insignificant too?

[not to mention Hendry concluded only 1.5% had no conceivable conventional explanation which must be considered insignificant by your own standards]

“Scientist” debunk thyself!

What a bunch of nonsense. Rramjet is trying to corner everyone into thinking he is going to do some sort of independent study but all he is doing is reciting BBSR 14. Once he gets somebody to agrree to this, he will just flash the preselected BBSR14 results and say it proves his point. He would have to do absolutely nothing and already knows the conclusion. There is no risk involved contrary to what he is stating.

Indeed. Anyone who did not smell this stinky bait from a mile away the moment Rramjet first proposed his incredibly flawed “null” hypothesis and didn’t know those unrefined intermediate statistical results (the final figure reached in the study was 0% “aliens”) from the 50s that’s been presented completely out of context by every UFOlogist since were coming again if they bit please raise your hand…

[looks around the room for a show of hands]

Nobody?

OK that settles it then, it’s unanimous.

 

[corbin]

So by that reasoning do you not agree the final 5% unidentified figure reached by the Air Force and Hynek is insignificant too?

[not to mention Hendry concluded only 1.5% had no conceivable conventional explanation which must be considered insignificant by your own standards]

“Scientist” debunk thyself!

Actually, Rramjet's response was a response to wollery's point that a number of cases being the results of hoaxes (or hallucinations, dreams, etc.) would mess up the statistics comparing the circumstances in nonmundane cases to the circumstances in mundane cases. Rramjet was saying that the relatively small number of such hoax etc. cases in comparison to physical misidentification cases would not skew the statistics in any significant manner. He further offered that there are ways to control for skewing of statistics. Really, its a very pedantic argument they were having that you can't be blamed for misunderstanding.

The argument was NOT about whether the results are significant by themselves as you imply. He wasn't saying that you can just dismiss the hoaxes because the percentage is low.

The only real meaning in those statistics is between those that have no mundane explanation (5%) and those that have mundane explanation (95%). The 95% is just subdivided for more detail.

 

[Access Denied]

Really, its a very pedantic argument they were having that you can't be blamed for misunderstanding.

Especially if I didn’t…

My understanding is Rramjet claims the 1-2% that were able to be identified as “hoax/psychological” is insignificant therefore there’s no reason to believe a significant percentage of the other 5% that haven’t been identified could be “hoax/psychological” as well.

Do you agree with Rramjet's claim and if so, on what basis?

By the same token, do you believe Hendry’s conclusion that only 1.5% had no conceivable conventional explanation is insignificant?

 

[corbin]

Especially if I didn’t…

My understanding is Rramjet claims the 1-2% that were able to be identified as “hoax/psychological” is insignificant therefore there’s no reason to believe a significant percentage of the other 5% that haven’t been identified could be “hoax/psychological” as well.

Do you agree with Rramjet's claim and if so, on what basis?

It is not Rramjet's claim that hoax cases are insignificant. The percentages are not even in dispute. They weren't arguing whether the percentages are valid or not. They were arguing about what would happen if you subjected the contents of the percentages to a specific statistical analysis. Rramjet's claim of the hoax cases being insignificant was made in the context of wooley's claim that the contents of the hoax cases would mess up the statistical analysis. He wasn't saying that you can just ignore the hoax cases because the percentage is small. He is saying that when put into the same set as comparatively large categories for the purpose of statistical analysis, it has negligible effect on the conclusion of the statistical analysis. That doesn't mean that the category can be erased from whole percentage breakdown.

Maybe Rramjet can explain this better than I...

What you are claiming Rramjet's claim is implies that the Air Force/Hynek study has a margin of error large enough for something under a certain percentage number to be disregarded. Nobody is disputing or discussing whether there is or is not a margin of error in the study. Read the quote that Rramjet was responding to.

The hoax cases being "insignificant" is quite ambiguous. It is undeniable, however, that in the context of the discussion, he is talking about insignificant for the statistical analysis proposed by wooley. A more exact statement would be "The hoax cases are so small in number as to be insignificant contributors to the statistical analysis of cases."

By the same token, do you believe Hendry’s conclusion that only 1.5% had no conceivable conventional explanation is insignificant?

Not by the same token at all. This is question about the margin of error in the Air Force/Hynek study.

 

[John Albert]

Actually, Rramjet's response was a response to wollery's point that a number of cases being the results of hoaxes (or hallucinations, dreams, etc.) would mess up the statistics comparing the circumstances in nonmundane cases to the circumstances in mundane cases.

Wollery is right.

Rramjet's faulty "null hypothesis" is predicated on a number of incorrect assumptions. One major incorrect assumption he makes is that all the reports are the result of actual physical objects being seen. He deliberately discards the possibility of hoaxes, lying, confabulation, dreams, hallucinations, mental defects or other human errors that might cause people to report things that never actually happened.

"All UFO sightings are the result of a misidentification of mundane objects"

See the word "objects" in there? That's his dishonesty showing.

Rramjet handwaves away all the very real and plausible possibilities of hoaxes, lying, confabulation, etc., solely on the claims of only two UFO studies without the benefit of adequate documentation on their statistical methods. Both of these studies trivialize the incidence of hoaxes, lies, hallucinations and other "psychological" causes, estimating them all at only 1-2%. However, neither study provides any specific details on how the researchers arrived at that conclusion or tabulated the results.

This is an extremely important point, because it is impossible to know for certain exactly how many of those cases were actually hoaxed. Before the researchers could classify any given report as a hoax, lie or confabulation, they would first have to discover the hoax, otherwise they'd have no basis for reporting it as such. If they fail to discover a hoax, then that case would need to be put down to some other cause, most likely "unexplained."

The fact that some UFO claimants were successfully caught in lies strongly implies that others must have lied too without getting caught. Therefore, the statistic for "hoaxes," "lies," etc. is almost certainly under-representative.

This is a fundamental problem with relying on anecdotes as your sole source of evidence: people lie, and people get things wrong all the time. Ask any cop, lawyer, judge, teacher, insurance adjuster... You simply can't trust that everyone's word is credible.

Most extraordinary stories are not true. So when you go looking for extraordinary stories like those of paranormal encounters, it just follows that statistically speaking, you're going to get a large percentage of lies. The fact that these ufology reports rank lies, confabulations and hoaxes at such a low percentage indicates that the researchers were more than likely a bit too credulous of the reports they were receiving.

The fact that hoaxes, lies and confabulations are underrepresented in these studies at only 1-2% and the total percentage of unexplained cases is only 5%, I don't think it's much of a stretch to make an educated guess that the unexplained 5% are probably the results of lies, hoaxes and confabulations that slipped past the researchers, with perhaps a few undiscovered mundane causes mixed in.

Rramjet was saying that the relatively small number of such hoax etc. cases in comparison to physical misidentification cases would not skew the statistics in any significant manner.

He's very wrong about that. Especially considering the percentage of actual hoax cases is a wildcard that cannot be adequately established, and the percentage of known hoax cases (at 1-2%) is so close to the percentage of the number of unexplained cases (5%). Even by the woefully diminished standards of those two studies, the percentage of hoaxes amounts to at least proportional to 15-40% of the number of unexplained cases. The margins for error are too great. Hoax, lies, confabulation, etc. are undeniably a significant factor.

He further offered that there are ways to control for skewing of statistics. Really, its a very pedantic argument they were having that you can't be blamed for misunderstanding.

Rramjet is wrong about that, like he's wrong about so many other things. It's an obtuse argument, willfully ignorant of the facts of the matter.

We're all quite familiar with Rramjet's shenanigans where UFO research is concerned. He gets these harebrained notions that make no sense, and he obstinately argues them over and over despite everyone around him telling him they're nonsense. It appears that he assumes he's right about everything and will argue all day without ever admitting he's wrong or putting forth the least effort to try to really understand anything.

The argument was NOT about whether the results are significant by themselves as you imply. He wasn't saying that you can just dismiss the hoaxes because the percentage is low.

You're wrong about this. That's exactly what he said, several times. Scroll up. I quoted him saying it on three different times.

The only real meaning in those statistics is between those that have no mundane explanation (5%) and those that have mundane explanation (95%). The 95% is just subdivided for more detail.

And there we come to another major flaw in Rramjet's notorious "null hypothesis": what will it prove, to find characteristic similarities between the explained and the unexplained sightings? What's the practical upshot of all this? Will it prove aliens? No. Will it prove that all sightings are the result of "mundane" causes? No. It's an exercise in futility.

 

[corbin]

Wollery is absolutely correct about that.

Rramjet's faulty "null hypothesis" is predicated on a number of incorrect assumptions. One major incorrect assumption he makes is that all the reports are the result of actual physical objects being seen. He deliberately discards the possibility of hoaxes, lying, confabulation, dreams, hallucinations, mental defects or other human errors that might cause people to report things that never actually happened.

See the word "objects" in there? That's his dishonesty showing.

As far as I can tell, Rramjet has repeatedly stated that he is willing to accept a null hypothesis that includes all mundane explanations. The null hypothesis serves the same function in either case.

Rramjet handwaves away all the very real and plausible possibilities of hoaxes, lying, confabulation, etc., solely on the claims of only two UFO studies conducted under questionable circumstances without providing adequate documentation on their statistical methods. Both of these studies trivialize the incidence of hoaxes, lies, hallucinations and other "psychological" causes, estimating them all at only 1-2%. However, neither study provides any specific details on how the researchers arrived at that conclusion or tabulated the results...

These are mere guesses that the methodology is probably wrong. You're playing pin the tail on the donkey at this point. The reference to the study has been provided, it is up to you to point out specific errors in methodology rather than assert that the results are "probably the result of lies..." and expect somebody else to spell it out for you.

You're wrong about this. That's exactly what he said, several times. Scroll up. I quoted him saying it on three different times.

Those quotes are all in the context of the statistical analysis. No claim was ever made that by the same logic would disregard the 5% unexplained cases. For Rramjet to make such a claim would mean he is doubting the very same study that he cited. You'll probably chalk this up to "just another example of stupidity," but if you actually read the context of the discussion you'll see that the 1.5% being "insignificant" does not mean what you think it means.

This is quite a belaboring of a very small point...

 

[John Albert]

It may seem like a very small point, but believe me it's not.

If the % of unexplained cases = 5%, and the % of hoaxed cases = 1.5%

5% ÷ 1.5% = 3.333... or 1/3

...then the number of hoaxed cases is proportional to one third of the number of unexplained cases. By what measure is a margin of error that large considered "insignificant?"

If you've been following this discussion for the past few weeks, you'd realize what it is that Rramjet is trying to assert with this absurd manipulation of language and statistics.

He's trying to prove that if the 5% of unexplained cases are really due to mundane causes, then they will consequently exhibit the same characteristics of the other 95% of cases with known causes, otherwise the causes of those 5% must be "something inexplicable."

That's a glaring example of affirming the consequent (an "either-or" fallacy in other words). He's assuming a condition that only two possibilities exist—"mundane physical objects" or "extraordinary physical objects"—and ignoring the significant possibility that no physical objects at all may be behind some or all of the unexplained 5%.

He's dismissing the possibility of hoaxes, lies, psychological issues, etc. out of hand as "insignificant," when they most definitely are quite significant.

All this semantic silliness has resulted from Rramjet trying to dishonestly shoehorn "objects" into the null hypothesis in order to facilitate a jump to the conclusion of ET or other paranormal causes.

 

[corbin]

The claim was never made by Rramjet that the study's margin of error (or methodological dysfunction) was enough to make the 1.5% insignificant. In fact, Rramjet never brought the study's soundness into question at all, which makes perfect sense because it is HIS evidence.

(Let me say that the only reason I'm talking about margin of error or the study being dysfunctional is because this is the only thing that is implied if you take the ambiguous statement "1.5% is insignificant..." by itself, as Access Denied did)

He made the statement that the 1.5% was insignificant in a very specific context. Not "insignificant" in the sense that the study's margin of error (or methodological dysfunction) precludes any results below a certain percentage.

Rramjet's original claim, paraphrased, was this: "The hoax cases are so small in number as to be insignificant contributors to the statistical analysis of cases." He also stated that besides this, even the small number could be adjusted for. This is beside the point, though.

Access Denied saw Rramjet's statement "1.5% is insignificant..." and assumed that he was talking about the numbers of the study itself and reasonably analogized the implications of such a claim (which Rramjet didn't make, if you look at the context) as extending to the also small, 5% of unexplained cases.

The critical point is that Rramjet was never calling into question the 1.5% itself. He was instead saying that 1.5% is so small that it wouldn't matter when compared with other, bigger portions in the case of a statistical analysis.

These are different meanings and have different implications. The analogy is only allowed by one of the meanings, which is clearly not the meaning of Rramjet's post.

 

[corbin]

Repost?

 

[corbin]

If you've been following this discussion for the past few weeks, you'd realize what it is that Rramjet is trying to assert with this absurd manipulation of language and statistics.

He's trying to prove that if the 5% of unexplained cases are really due to mundane causes, then they will consequently exhibit the same characteristics of the other 95% of cases with known causes, otherwise the causes of those 5% must be "something inexplicable."

That's a glaring example of affirming the consequent (an "either-or" fallacy in other words). He's assuming a condition that only two possibilities exist—"mundane physical objects" or "extraordinary physical objects"—and ignoring the significant possibility that no physical objects at all may be behind some or all of the unexplained 5%.

He's dismissing the possibility of hoaxes, lies, psychological issues, etc. out of hand as "insignificant," when they most definitely are quite significant.

Wouldn't the non-physical explanations just fit right into the rest of the 95% mundane explanations, thus including characteristics of even these cases to be compared against the 5%?

I don't see what the problem is. It seems to me that if the 5% of unexplained cases really are due to mundane causes, then by definition they will indeed exhibit the same characteristics as the other 95% of cases with mundane causes. If you don't force him to take out the non-physical category, then there are no other characteristics that the mundane unexplained cases could exhibit other than the very same ones in the mundane explained category, as long as the sample size is big enough and there is no other criterion by which the cases are sorted other than their explainability. If the 95% is representative of the whole, there is nothing that can "escape," if you get my meaning. All sufficient conditions for mundane causes are within the 95%.

 

[John Albert]

Yeah, I reposted to add some context to our discussion.

The claim was never made by Rramjet that the study's margin of error (or methodological dysfunction) was enough to make the 1.5% insignificant. In fact, Rramjet never brought the study's soundness into question at all, which makes perfect sense because it is HIS evidence.

The fundamental unsoundness of the study, as well as the premise it is based upon, is the matter of contention here.

He made the statement that the 1.5% was insignificant in a very specific context. Not "insignificant" in the sense that the study's margin of error (or methodological dysfunction) precludes any results below a certain percentage.

Presumably, he never thought it through that thoroughly. Yet he resists every attempt to correct his abysmally flawed methodology.

Rramjet's original claim, paraphrased, was this: "The hoax cases are so small in number as to be insignificant contributors to the statistical analysis of cases."

Yes, and I have showed that, considering the proportions of the sample sets he's comparing, 1.5% is a very large discrepancy.

Access Denied saw Rramjet's statement "1.5% is insignificant..." and assumed that he was talking about the numbers of the study itself and reasonably analogized the implications of such a claim (which Rramjet didn't make, if you look at the context) as extending to the also small, 5% of unexplained cases.

Rramjet didn't consider those implications, presumably because, as I mentioned before, he just pulled the notion out of his ass and assumed it was correct without even bothering to think it through.

The critical point is that Rramjet was never calling into question the 1.5% itself. He was instead saying that 1.5% is so small that it wouldn't matter when compared with other, bigger portions in the case of a statistical analysis.

Presumably so, but as I said before, he obviously hasn't thought it through before arguing tooth-and-nail that it is a correct and reasonable way to proceed in proving that UFOs are ET (or otherwise inexplicable).

These are different meanings and have different implications. The analogy is only allowed by one of the meanings, which is clearly not the meaning of Rramjet's post.

I'm not sure I understand what "analogy" you're referring to.

 

[John Albert]

Wouldn't the non-physical explanations just fit right into the rest of the 95% mundane explanations, thus including characteristics of even these cases to be compared against the 5%?

Sure, that would be fine if you're honest in your approach and you clearly indicate that's what you're doing and the methodology is consistent with the research you're trying to accomplish. But that's not what Rramjet is trying to do. He's trying to conflate the data on physical and non-physical results and then misrepresent them as if they are all the result of physical phenomena.

I don't see what the problem is.

There are several.

First of all, as I pointed out, the false assumption that all these sightings resulted from actual physical objects. Rramjet's "null hypothesis" is constructed from the explicit viewpoint that all UFO reports are the result of some physical object or other with recognizable characteristics, which is simply not true.

Even the ufology studies he's using for his meta-analysis acknowledged the possibility of lying, hoaxes, psychological instability and other non-objective factors even if they did tend to diminish the incidence of that kind of origin. Rramjet is blatantly, dishonestly dismissing that information out of hand on the basis that it is "statistically insignificant" when in fact it isn't.

That's why he keeps advocating for the inclusion of the word "objects" in the null hypothesis, even though RoboTimbo, Wollery, myself, and others have pointed out the proper way to formulate a null hypothesis.

Clearly,

All UFO sightings are the result of a misidentification of mundane objects

is very different from

All UFO sightings are mundane in origin.

Rramjet's null hypothesis assumes the existence of objects even in cases where the absence of such objects has been proven. As I stated in my post above, Rramjet's null hypothesis makes an unsupported and even disproven assumption.

Then there's the issue of the "characteristics" of the sighting reports, and how those relate to real-world "mundane" objects. Saying that all UFO reports that resulted from misidentified mundane objects should look similar to all other UFO reports that resulted from mundane objects is a huge jump to conclusion. All "mundane objects" are not the same and do not have the same visual characteristics, and many different kinds of mundane objects have been reported as UFOs. There may be a few correlations to be drawn, but it doesn't stand to reason that every object ever mistaken for a UFO will share the same characteristics with every other object that could possibly be mistaken for a UFO.

As I've also pointed out, when you look at the numbers in the studies he's citing, the percentage of cases attributed to hoax, etc. is very close to the percentage that are unexplained. Dismissing the hoax cases as "statistically insignificant" raises the fudge factor to the point where it would totally throw off the error margin with regard to the unexplained cases.

Most sociological studies that deal with case studies or field reports tend to assume a margin of error in the neighborhood of around 5%. Yet we see that the studies Rramjet is working with cite the percentage of "unexplained" (that's synoymous with "ET" in ufology parlance) as right around 5%. Even the percentage attributed to hoaxes is 1-2% so that puts both figures squarely within the standard margin of error.

This study cannot possibly prove that UFOs are extraterrestrial, paranormal, or even unexplainable, as Rramjet keeps insisting.

Please correct me if I'm wrong about any of this.

 

[corbin]

Sure, that would be fine if you're honest in your approach and you clearly indicate that's what you're doing and the methodology is consistent with the research you're trying to accomplish.

So you admit that we would see the same characteristics in both cases if the 5% are explained by mundane causes? (don't answer that right away... you seem to have valid points about the analysis to which I respond further down)

But that's not what Rramjet is trying to do. He's trying to conflate the data on physical and non-physical results and then misrepresent them as if they are all the result of physical phenomena.

There are several.

First of all, as I pointed out, the false assumption that all these sightings resulted from actual physical objects. Rramjet's "null hypothesis" is constructed from the explicit viewpoint that all UFO reports are the result of some physical object or other with recognizable characteristics, which is simply not true.

Even the ufology studies he's using for his meta-analysis acknowledged the possibility of lying, hoaxes, psychological instability and other non-objective factors even if they did tend to diminish the incidence of that kind of origin. Rramjet is blatantly, dishonestly dismissing that information out of hand on the basis that it is "statistically insignificant" when in fact it isn't.

That's why he keeps advocating for the inclusion of the word "objects" in the null hypothesis, even though RoboTimbo, Wollery, myself, and others have pointed out the proper way to formulate a null hypothesis.

Rramjet's null hypothesis assumes the existence of objects even in cases where the absence of such objects has been proven. As I stated in my post above, Rramjet's null hypothesis makes an unsupported and even disproven assumption.

I agree with you that the null hypothesis that excludes "non-object" mundane explanations is indeed lacking. However, modifying the hypotheses to include these explanations doesn't change the logic behind the meta-analysis - that is, that given a big enough sample space, with overlapping and consistent characteristics exhibited by cases in the mundane explanation section, the characteristics exhibited by the cases in the unexplained section should be, to a statistically significant degree, the same characteristics exhibited by the cases in the mundane explanation section if the unexplained section cases indeed have mundane explanations. There would be a margin of error, but you could still come up with statistically significant numbers.

I'm sure Rramjet would be happy to modify his null hypothesis to include object and non-object explanations, as he has repeatedly offered to do:

You may substitute "are the result of a misidentification of mundane objects" with "are of mundane origin" if you wish. It changes nothing about the points I made.

I have... explicitly acknowledged that substitution is valid.

Then there's the issue of the "characteristics" of the sighting reports, and how those relate to real-world "mundane" objects. Saying that all UFO reports that resulted from misidentified mundane objects should look similar to all other UFO reports that resulted from mundane objects is a huge jump to conclusion. All "mundane objects" are not the same and do not have the same visual characteristics, and many different kinds of mundane objects have been reported as UFOs. There may be a few correlations to be drawn, but it doesn't stand to reason that every object ever mistaken for a UFO will share the same characteristics with every other object that could possibly be mistaken for a UFO.

That's why you compare a sufficiently large number of cases such that the margin of error incurred by there being "new" mundane characteristics in the 5% category is statistically insignificant. This would not be difficult to do with a sufficient number of cases. The more you have, the less the chance of one of the 5% characteristics to be unexplained.

As I've also pointed out, when you look at the numbers in the studies he's citing, the percentage of cases attributed to hoax, etc. is very close to the percentage that are unexplained. Dismissing the hoax cases as "statistically insignificant" raises the fudge factor to the point where it would totally throw off the error margin with regard to the unexplained cases.

*Sigh* The 1.5% number itself is not being dismissed on account of the study (the Air Force study, not the statistical analysis) being flawed. It is being recognized as comparatively insignificant when compared, in a statistical analysis, to much bigger numbers. Its a pesky case where "1.5% is insignificant" has a totally different meaning in two contexts. By itself it implies the study is flawed. In response to a specific point about statistical analysis, it is a statement about its bearing on a statistical analysis. "Insignificant" to the analysis, not "insignificant" by itself, next to the other percentages in the study.

It was originally claimed by Access Denied that Rramjet was claiming the 1.5% number is insignificant BY ITSELF on account of the study being flawed. Access Denied then said that by extension, the 5% would also be insignificant by Rramjet's presumed logic. The only problem is that Rramjet never claimed that the 1.5% number was insignificant by itself on account of the study being flawed, therefore, you can't extend this logic to the 5%... because he never claimed it. In this case, the argument remains internally consistent.

Most sociological studies that deal with case studies or field reports tend to assume a margin of error in the neighborhood of around 5%. Yet we see that the studies Rramjet is working with cite the percentage of "unexplained" (that's synoymous with "ET" in ufology parlance) as right around 5%. Even the percentage attributed to hoaxes is 1-2% so that puts both figures squarely within the standard margin of error.

This study cannot possibly prove that UFOs are extraterrestrial, paranormal, or even unexplainable, as Rramjet keeps insisting.

Please correct me if I'm wrong about any of this.

Don't pigeonhole yourself into thinking that UFO's must be ET. You already seem to think it is more likely than paranormal explanations, which means you are not just looking at the evidence, but looking at it and treating it with the more great and difficult expectation of evidence for ET's.