When do we have a scientific fact? • Hacker News Board

Re: When do we have a scientific fact?

#111

Post by timr » Fri, Dec 30, 2022, 5:32 PM UTC

nobody9999 wrote: post #68

timr wrote: post #54

Earlier quoted context omitted.

There was no significant survival benefit. Just because you find and treat cancer doesn't mean you've made someone's life better or longer. > The risk of death from colorectal cancer was 0.28% in the invited group and 0.31% in the usual-care group (risk ratio, 0.90; 95% CI, 0.64 to 1.16). The number needed to invite to undergo screening to prevent one case of colorectal cancer was 455 (95% CI, 270 to 1429). The risk…

>This isn't the final word on the subject, but it overturns the conventional wisdom that colonoscopy is an obvious win and worth all of the money, time, effort and pain we spend on it. I'd posit that the one in 455 people who would die without a colonoscopy would disagree that colonoscopy isn't worth "all the money, time, effort and pain we spend on it." I would suggest that you ask my uncle about it, as he didn't ha…

> I'd posit that the one in 455 people who would die without a colonoscopy would disagree that colonoscopy isn't worth "all the money, time, effort and pain we spend on it."

Your post is a fantastic illustration of how and why we can't rely on "consensus" to determine fact. People routinely fall in love with narratives based on what they "know", and refuse to accept new data that challenges their narratives.

For literally anything we do, some people will die afterward. You could set up a study where you punch people in the shoulder at random, and follow them to see if getting punched implies a survival benefit. Eventually, everyone in both groups will die -- we'd expect about the same number in each group at any given time, because punching people in the shoulder probably doesn't affect lifespan. But still, people will die after getting punched!

Here we have a large, randomized study where approximately the same number of people died with colon cancer after getting colonoscopy as did without the screening. The screening had ~no effect on outcome.

Also, you've misunderstood: the 455 people is not the number of people who would die with colon cancer -- it's the number of perfectly healthy people you'd need to screen to prevent one case.

> I would suggest that you ask my uncle about it, as he didn't have a colonoscopy and was eventually diagnosed with stage IV colorectal cancer.

While I'm sorry for your loss, the data from this RCT suggests that colonoscopy on perfectly healthy people confers little to no survival benefit, and trumps any anecdote.

Re: When do we have a scientific fact?

#112

Post by kwhitefoot » Fri, Dec 30, 2022, 5:38 PM UTC

Waterluvian wrote: post #2

I think that the mechanism of the universe is a black box. We can provide inputs and get outputs to build confidence of how we think it works, but we can’t ever know conclusively. That is, unless we escape the simulation and evaluate the mechanism’s implementation. I think anything else is just semantics.

off topic, perhaps. Can someone tell me what Americans really mean by this fashionable slur:

> I think anything else is just semantics.

Semantics is the science of meaning. So what do people really mean when they say that something is "just semantics"? Do they really mean sophistry?

Re: When do we have a scientific fact?

#113

Post by marginalia_nu » Fri, Dec 30, 2022, 9:51 PM UTC

TEP_Kim_Il_Sung wrote: post #106

marginalia_nu wrote: post #26

Earlier quoted context omitted.

Huh? Copernicus died in his home when he was in his 70s. Galileo died of old age as well. Tycho Brahe died from politeness. Bruno was burned because he was a hermetic mystic and didn't really prove anything significant about the planets.

Well I'm glad Bruno died. Those hermetic mystics had it coming, and not proving anything significant is basically high treason to science, fully deserving immediate immolation!

My point is that although a lot of people want him to be because it fits the narrative, he is a remarkably poorly martyr for science.

Re: When do we have a scientific fact?

#114

Post by matheusd » Fri, Dec 30, 2022, 10:13 PM UTC

kuhewa wrote: post #94

matheusd wrote: post #88

Earlier quoted context omitted.

This is an excellent comment! And this is a *very* strong threshold to reach: > No matter what your threshold for evidence needed is, it can be met. I'm fearful of mentioning any specific current scientific controversy, but it seems to me that many of the current politically charged topics fail to meet this very strong threshold. I guess I'm thanking you because while I had an intuitive sense for this, I didn't know…

I think the relevant wording here is "problem of induction". You can generate evidence at will and meetn any threshold of evidence for the statement "all swans are white". Anyone can perform the experiment by traveling around the northern hemisphere and find millions of white swans. It would still be silly to call that statement a fact, because it only requires a single swan sighting in Australia to falsify it.

You cannot "generate evidence at will to meet any threshold of evidence" for this specific example ("all swans are white") because, unless you posit that there are swans in Kepler-422B, I can set my threshold of evidence to "we have turned the earth into computronium and have not found any non-white swans", which is a finite (albeit large) threshold.

Once we agree on whatever definition of "swan" we care about (an animal belonging to a specific phylogenetic branch, with such and such qualities) we can agree on a more reasonable threshold, one that should include a visit to Australia to be crossed.

Re: When do we have a scientific fact?

#115

Post by kuhewa » Fri, Dec 30, 2022, 10:19 PM UTC

matheusd wrote: post #114

kuhewa wrote: post #94

Earlier quoted context omitted.

I think the relevant wording here is "problem of induction". You can generate evidence at will and meetn any threshold of evidence for the statement "all swans are white". Anyone can perform the experiment by traveling around the northern hemisphere and find millions of white swans. It would still be silly to call that statement a fact, because it only requires a single swan sighting in Australia to falsify it.

You cannot "generate evidence at will to meet any threshold of evidence" for this specific example ("all swans are white") because, unless you posit that there are swans in Kepler-422B, I can set my threshold of evidence to "we have turned the earth into computronium and have not found any non-white swans", which is a finite (albeit large) threshold. Once we agree on whatever definition of "swan" we care about (an an…

The problem of induction always applies. You also can't examine every bit of matter in the universe to ensure it is made of atoms.

Re: When do we have a scientific fact?

#116

Post by nobody9999 » Fri, Dec 30, 2022, 11:07 PM UTC

timr wrote: post #111

nobody9999 wrote: post #68

Earlier quoted context omitted.

>This isn't the final word on the subject, but it overturns the conventional wisdom that colonoscopy is an obvious win and worth all of the money, time, effort and pain we spend on it. I'd posit that the one in 455 people who would die without a colonoscopy would disagree that colonoscopy isn't worth "all the money, time, effort and pain we spend on it." I would suggest that you ask my uncle about it, as he didn't ha…

> I'd posit that the one in 455 people who would die without a colonoscopy would disagree that colonoscopy isn't worth "all the money, time, effort and pain we spend on it." Your post is a fantastic illustration of how and why we can't rely on "consensus" to determine fact. People routinely fall in love with narratives based on what they "know", and refuse to accept new data that challenges their narratives. For lite…

>Also, you've misunderstood: the 455 people is not the number of people who would die with colon cancer -- it's the number of perfectly healthy people you'd need to screen to prevent one case.

No. I haven't. If we don't do any colonoscopies, that 1 in 455 who develops colorectal cancer will die (as my uncle did) if the cancer isn't diagnosed early enough.

Don't want to go through the "trouble" of a colonoscopy after 50 years of age? Then don't. If so, I hope that you or anyone you love isn't that 456th person.

Yes, Anecdotes aren't data. However, that 1 in 455 is a statistic and what are statistics? Data based on aggregations of anecdotes.

I mentioned my uncle not because I want to pull your heartstrings. Rather, it's because while it's a simple thing to look at data and statistics, it's easy to forget that such data (at least in this case) breaks down in to anecdotes just like that one -- e.g., dead and/or very sick people -- who wouldn't be sick or dead with appropriate screenings.

If you don't care that some people will die without colonoscopies, that's your prerogative. I hope that attitude doesn't keep you or those you love from identifying potential colorectal cancer before it's too late.

That you don't care about others getting sick and dying[0] from the fourth leading cause of cancer deaths in the US, says something. Not sure what though and I won't speculate.

I'm sure that the your full-throated support for not doing colonoscopies is cold comfort to the (~36-56/100,000 Americans) ~118,000-185,000 actual, real people who develop colorectal cancer per year.

I'll say it one more time -- there are real people represented in that data. People who suffer and die every day. And if we did fewer colonoscopies, more of those people would, in fact, die, because their cancer wasn't diagnosed at all/soon enough. And so I have to put on my Quincy, M.E.[1] mask and ask: People are dying, doesn't anyone care? Is it correct to say that you don't care, or am I missing something?

It would be easy for me to just wish colorectal cancer on you and then ignore you, but I don't want you to get colorectal cancer (or cancer of any kind for that matter). Dying of cancer is a slow, agonizing, incredibly painful way to die and should be avoided as much as possible.

If you have a family history of colorectal cancer, please, please start having colonoscopies every two-five years starting at age 45.

If you don't, please do so every five years or so starting at age 50.

Or not. It's your life and (eventually) your funeral. That's up to you.

[0] https://gis.cdc.gov/Cancer/USCS/#/Trends/

[1] https://en.wikipedia.org/wiki/Quincy,_M.E.

Edit: Fixed references and closed emphasis asterisk.

Re: When do we have a scientific fact?

#117

Post by raattgift » Tue, Jan 03, 2023, 1:05 PM UTC

kuhewa wrote: post #92

n4r9 wrote: post #89

Earlier quoted context omitted.

The atomic model is certainly beyond a significant level of doubt. Surely it is not totally beyond question however. There remains a very small chance that an alternative model of the universe is consistent with the experiments you have in mind, and better explains the outcomes of other (as yet unperformed) experiments.

A good but not perfect example would be Newtonian gravitation. Pre-relativity, if it wasn't for the weird precession of the perihelion of Mercury, one could have easily used a form of this argument to claim Newtonian gravity is a fact because you can do an experiment at any moment to demonstrate it by watching an apple fall from a tree. It's been done millions of times, and the acceleration always follows an elegant…

From the early modern discovery of the finite speed of light in the 17th century (postdating and motivated by the discovery of the Galilean moons). From the beginning (about a decade after Rømer and Huygens published their measurements) Newtonian gravitation was on difficult mathematical grounds because the optical image of a source of gravitation (e.g. the Jovian system and its elements) must lag behind its (Newtonian) gravitational source. Indeed these problems drove corpuscular theories of light, including Newton's own, in which perfectly elastic, rigid, and weightless projectiles travel with finite speed from a source to a target.

Early 18th century work on the aberration of light evidenced some of the theoretical problems in Newtonian gravitation (and relatedly interference and diffraction exposed problems in corpuscular theories of light), but almost no practical ones. The practically-measurable non-Newtonian nature of the perihelion precession of Mercury wasn't discovered until at least a couple of decades after Newtonian gravitation exposed the existence of Neptune. For all practical purposes light as a source of gravitational potential (much less the self-gravitation of light) still isn't something we can demonstrate in an Earthbound laboratory (there is ample astrophysical and cosmological support for the active gravitational mass of light, though).

Tl;dr Newtonian gravitation was unphysical at its inception, and Newton understood this in correspondence exploring mechanical explanations for the inverse square law.

Penultimately, a direct laboratory test of Newtonian gravitation wasn't made until Cavendish's experiment in 1798. Was it "fact" or "effective theory" prior to that test? Did it become "fact" at the time of the test, or did it remain an effective theory? I'm not sure the distinction matters in any practical sense.

Finally,

> demonstrate [Newtonian gravitation] by watching an apple fall from a tree.

Try that in a strong wind!

Re: When do we have a scientific fact?

#118

Post by satisfice » Fri, Jan 06, 2023, 3:33 AM UTC

pdonis wrote: post #43

satisfice wrote: post #33

Earlier quoted context omitted.

Seems to me we can identify a scientific fact: it's an empirical proposition for which we have sufficient scientific evidence. "Sufficient" is a community standard, and different communities may have different standards. That's not an ideal state of affairs, but it turns out to be a liveable one. This doesn't solve the problem of "future-proofing" or universalizing facts, because THAT is not a solvable problem. But i…

> it's an empirical proposition for which we have sufficient scientific evidence. "Sufficient" is a community standard It doesn't have to be. We can use predictive power to judge how strong our evidence is.

Make a specific proposal and I will show you how that is a community standard. The only reason I can't right now is because what you have suggested is not yet specific enough to be ANY kind of standard.

Re: When do we have a scientific fact?

#119

Post by pdonis » Fri, Jan 06, 2023, 10:48 PM UTC

satisfice wrote: post #118

pdonis wrote: post #43

Earlier quoted context omitted.

> it's an empirical proposition for which we have sufficient scientific evidence. "Sufficient" is a community standard It doesn't have to be. We can use predictive power to judge how strong our evidence is.

Make a specific proposal and I will show you how that is a community standard. The only reason I can't right now is because what you have suggested is not yet specific enough to be ANY kind of standard.

My point is that what is required by an actual real world application is not up to the community. It's up to the universe.

For example, if you want to judge the predictions of astronomers about close approaches of asteroids to Earth, the standard is not what the community thinks is ok, but whether their predictions are accurate enough to say for sure whether an asteroid will hit or miss. That standard is imposed by the universe, the laws of physics, and our desire to not get killed by an asteroid impact.