The ol' water going down the drain thing

[DaveC426913]

Split off from here:
https://www.physicsforums.com/threa...the-spinning-of-a-galaxy.974218/#post-6201131Has the idea of Coriolis force affecting the direction of swirl down a drain been compellingly vindicated or debunked?

I don't mean in theory. We all know the theory. I'm looking for experimental confirmation.

The above video makes one comparison, between a setup in each hemisphere, and declares the effect has been demonstrated.

Unless I missed a longer version of this experiment, doing the experiment once in a given hemisphere demonstrates nothing but a lucky first try. For all we know, the next time they do the exact same experiment it will go the opposite way.

@Orodruin

 

[Orodruin]

Unless I missed a longer version of this experiment, doing the experiment once in a given hemisphere demonstrates nothing but a lucky first try. For all we know, the next time they do the exact same experiment it will go the opposite way.

This is wrong. As long as you have your experimental systematics under control. It can be sufficient to perform an experiment once.

 

[DaveC426913]

This is wrong. As long as you have your experimental systematics under control. It can be sufficient to perform an experiment once.

And they don't know if they have them under control.
Nor do they know if their result (one in each hemisphere) is mere chance.

I bet if I flipped a single penny in N.Am and a single penny in S.Am and they happened to come up opposite, that would not be declared sufficient data to draw a conclusion about how pennies land in different parts of the world.

 

[Orodruin]

And they don't know if they have them under control.

How do you know that? It is not the point of an educational video to start talking about what systematics appear. If they did that people would turn it off very quickly. If you want to make an argument, you should argue for why you think they do not have them under control, i.e., you should estimate what kind of size of the swirl you would expect by chance alone.

Nor do they know if their result (one in each hemisphere) is mere chance.

This is not true. If they have the systematics under control, they know exactly whether or not their result can be due to chance or not.

I bet if I flipped a single penny in N.Am and a single penny in S.Am and they happened to come up opposite, that would not be declared sufficient data to draw a conclusion about how pennies land in different parts of the world.

This is not at all the same thing. This example has a single binary outcome. In the case of the swirl you can estimate what kind of swirl speed you would get by chance alone due to other external factors. If that speed is sufficiently smaller than observed swirl, then yes you can make the conclusion.

 

[DaveC426913]

If that speed is sufficiently smaller than observed swirl, then yes you can make the conclusion.

Unless it has occurred by chance, say, by not controlling confounding factors.
There is a list of factors they may not have thought of.
An easy one is simply the shape of the pool. There are surely others.

The only way to confirm the phenomenon is to show a statistically significant result.

 

[Dale]

One test certainly cannot give information on the reproducibility of a phenomenon.

 

[Orodruin]

Unless it has occurred by chance, say, by not controlling confounding factors.
There is a list of factors they may not have thought of.
An easy one is simply the shape of the pool. There are surely others.

The only way to confirm the phenomenon is to show a statistically significant result.

But you have no basis to claim that the result is not statistically significant. Show me the valid reasoning leading to an estimate of systematics that are larger than the signal observed. I think it is rather naive to think that this production would have been made at all if they had not done the maths first to at least make it reasonable that systematics are under control.

One test certainly cannot give information on the reproducibility of a phenomenon.

Without further information. In this case knowledge about the systematics. If the systematics are under control and you believe in your estimate of them, you certainly could claim that the experiment is reproducible. You might not get exactly the same number of sigmas out, but typically enough to claim discovery and that a similar experiment will give you the same result. If this was not the case a large number of experiments would never have been built.

 

[russ_watters]

Uh, why does any of this debate on what it means for an experiment to be reproducible really matter to the main question?

Maybe look at it another way: is it possible to force the whirlpool to rotate in an arbitrarily chosen direction?

 

[PeterDonis]

But you have no basis to claim that the result is not statistically significant.

The burden is not on us to show that it isn't. The burden is on the people who produced the video to show that it is. That's why scientific results are normally written up as peer-reviewed papers. Is there a peer-reviewed paper that supports the result the video claims to show?

I think it is rather naive to think that this production would have been made at all if they had not done the maths first to at least make it reasonable that systematics are under control.

"Naive" or not, PF's rules about valid sources are pretty clear, and they're there for a reason.

 

[Orodruin]

Maybe look at it another way: is it possible to force the whirlpool to rotate in an arbitrarily chosen direction?

This is not the same question. Nobody is discussing this. The question is whether the experiment in question controls systematics (ie, external factors that would give a swirl in an arbitrary direction give a smaller effect than the sought signal) well enough to claim that the effect is established.

 

[DaveC426913]

It is not the point of an educational video to start talking about what systematics appear.

Educational component aside, the consideration of any practical challenges will have to be handled before one could state that the conclusion is valid.

 

[DaveC426913]

The question is whether the experiment in question controls systematics...

No it isn't. The question is: has it been shown experimentally to be confirmed or debunked?

A single result - one that could just as easily have been the result of controlling no variables yet still gets the desired result once - is not a convincing test.

As I said: one (OK, two) data points (under different circumstances, and getting different results) is insufficient.

Is there a peer-reviewed paper that supports the result the video claims to show?

Personally, I don't need it to be a peer-reviewed paper. Multiple tests, varying the setup (such as using a different pool), all showing the expected result would be sufficient for me.

One test merely puts it in the 'plausible' category, where it's always been.

 

[Orodruin]

The burden is not on us to show that it isn't. The burden is on the people who produced the video to show that it is.

For a scientific discovery, I would agree. This is not a scientific paper, it is a demonstration intended to show how large the effect is. If it was an experimental paper, then yes I would expect systematics to be motivated. For a popular science presentation on youtube, I would not. It would completely alienate the intended audience.

"Naive" or not, PF's rules about valid sources are pretty clear, and they're there for a reason.

We are not discussing an original result. We are discussing a demonstration of the Coriolis effect, which is well established.

A single result that could just as easily have been the result of controlling no variables (yet still getting the desired result once) is not a convincing test.

Of course not, but that is not the argument you are presenting. You are essentially namecalling a respected science educator for fraudulent behaviour. I would not expect systematics to be shown in an educational video so I think this is unfair behaviour, in particular as they do not have the possibility to actually provide you with the appropriate information upon request. The appropriate response would be to ask them directly what kind of systematic effects they considered and give them a chance to convince you the experimental conclusion is valid. Your claim in the other thread is that their behaviour is fraudulent on the same level as charlatans at the equator fool tourists.

 

[PeterDonis]

This is not a scientific paper

Then, by PF rules, we should not be using it as a reference to support a scientific claim. The claim under discussion is a scientific claim.

We are discussing a demonstration of the Coriolis effect, which is well established.

The Coriolis effect is well established, but the Coriolis effect is not the only effect that is relevant for water flowing in a toilet. The claim under discussion is whether, under appropriate conditions (such as those shown in the video), the Coriolis effect does indeed dominate the dynamics. That is a scientific claim and should be subject to the usual rules for scientific claims.

You are essentially namecalling a respected science educator for fraudulent behaviour.

No, he isn't. He hasn't said anything to them. He's only made posts here on PF.

 

[russ_watters]

Thread locked pending moderation.

 

[anorlunda]

After much discussion, this thread is re-opened to allow participants to post new source information. It has been a challenge for everyone, but I hope we can come to a consensus.

 

[Orodruin]

That is a scientific claim and should be subject to the usual rules for scientific claims.

I dug out the original reference to the MIT experiment performed by Shapiro. It is published in Nature:
https://www.nature.com/articles/1961080b0The paper is behind a paywall, but I do not think that invalidates it as a scientific reference (if we do not accept Nature, what do we accept?). The introduction reads (after setting up the problem of kitchen sinks and bath-tubs)

Both schools of though are in some sense correct. For the everyday observations of the kitchen sink and bath-tub variety, the detection of the vortex seems to vary in an unpredictable manner with the day and the time of day, and the particular household experimenter. But under well-controlled conditions of experimentation, the observer looking downwards at a drain in the northern hemisphere will always see a counter-clockwise vortex, while one in the southern hemisphere will always see a clockwise vortex.

The paper goes on to discuss what is meant by "well-controlled conditions of experimentation" and to describe the results of the experiment and the magnitude of the effect, including the opposite-direction vortex resulting from pulling the plug to early after the tub is filled in the opposite direction as well as an initial opposite swirl later reversed to a "correct" swirl when pulling the plug in a medium range time.

Based on this, I would say that the Veritasium video falls under the category of physics demonstrations or reproductions rather than original research. Once having read the Shapiro paper, it is obvious that Veritasium has followed the same procedure as the things he points out are essentially pitfalls explicitly mentioned in the paper - such as not introducing any vorticity by pulling the plug through the water.

Coincidentally, this paper was critiqued for not citing an earlier source from 1908:

WITH reference to Prof. A. H. Shapiro's recent communication on the above subject, I venture to call attention to the fact that in 1908 the Austrian physicist, O. Turmlïtz, described careful and effective experiments which demonstrated the effect of the rotation of the Earth on the outflow of water through a central aperture, in a paper entitled “Ein neuer physikalischer Beweis der Achsendrehung der Erde”.

That reference is: Turmlïtz, O., S.B. Akad. Wiss. Wien, Abt. IIa, 117, 819 (1908).
I have unfortunately not been able to locate a copy of this paper, but based on the Nature comment, it should also describe a similar setup and discussion.

I hope this brings some clarity as to the reproducability (Veritaserum reproduced the experiment, he did not do an original experiment but rather a demonstration) as well as to the establishment of the effect in the scientific literature going back at least 57 years (and likely 111 years).

 

[PeterDonis]

The paper is behind a paywall, but I do not think that invalidates it as a scientific reference

No, of course not. It would be nice if scientific research that the public funded was always available freely to the public, but unfortunately we're not there yet. (And arxiv.org unfortunately didn't exist when this paper was written.)

 

[PeterDonis]

Btw, it looks like the replication of Shapiro's experiment in the Southern Hemisphere was also published in Nature:

https://www.nature.com/articles/2071084a0
(Unfortunately it's also paywalled.)

 

[jedishrfu]

This site has the two youtube videos side by side and all you have to do is run them synchronized

https://www.smartereveryday.com/toiletswirl

and you can see the Coriolis effect in the kiddie pools with food coloring showing the flow more clearly. They go on to explain the nature of the effect and why you will get inconsistent results if you only consider toilets or sinks.

There's also a Scientific American article on the same topic here:

https://www.scientificamerican.com/article/can-somebody-finally-sett/

 

[DaveC426913]

The Definitive Verdict Is In

https://www.smartereveryday.com/toiletswirl
At the suggestion of a Mentor, I reached out to the makers of the video, Derek Muller and Destin Sandlin via their Veritasium website. I am pleased to say Derek got back to me within three hours.

Amongst humble appreciation and gushing of their fine work, my core question was this:

< snip >
I know you went to great lengths to eliminate confounding variables, but one thing that surprised me is that you did not repeat the experiments to ensure it gave consistent results. In my humble opinion, that would go a long way toward eliminating any unforeseen variables ... thereby showing a *statistically significant* result.

Am I wrong in thinking that such an experiment would ideally provide more than one data point per setup?
< /snip >

This was the core of Derek's response:

< snip >
We did repeat the experiment - three times in each location. I think you can see in some of the shots, particularly in australia, the pool drainings are not the same.
< /snip >

So that puts the final nail in the coffin of doubt, and provides a definitive answer to this longstanding erstwhile "myth":

With sufficient variables controlled, in a pool of a mere 5 foot diameter, the Coriolis Effect is sufficiently large that it can be - and has been - demonstrated reliably and repeatably. To-wit: Water in a small pool in the northern hemisphere does drain in a counterclockwise direction, the same direction as northern hurricanes. Water in a small pool in the southern hemisphere does drain in a clockwise direction, the same direction as southern cyclones. Status: Confirmed. Case closed.

Thank you all for your patience and contributions.