Any advice or guidance would be appreciated please

[Eddie B]

I have been corresponding with fellow members of this community, in relation to a hypothesis  posed based on observation. This has unfortunately been closed to further replies, because it does not meet two of the requirements for speculations discussions. The first point relates to the lack of statistical evidence, of which I am sure I will be able to compile, thanks to suggestions and guidance from a fellow member. However, the second point is directed at being able to "falsify" my data. I have been having some problems during my correspondents, with the use of scientific terminology. I am a non academic which I pointed out several times during my correspondents, but this has now been addressed with advice given not to use terminology I do not understand. So I now stand with a hypothesis down graded to "idea" which I can no longer describe using "big words". This has left me a little confused regarding the meaning of falsify. Does this mean what I believe it means e.g. falsifying the results of pharmaceutical trials, so they are passed for general release. Or is there a scientific terminology, which gives different meant to the same word... and I am unaware of this ?

Secondly, if my first suggestion is correct, how would I go about falsifying my data so that it would meet the requirements of speculations discussions ?.

My posts reached the 16 mark, which was 14 more than I had anticipated... so this my be deemed a good statistic for the life expectancy of an "idea" on this site !

Thank you

E B 

[Strange]

6 minutes ago, Eddie B said:

However, the second point is directed at being able to "falsify" my data.

It is not the data that needs to be falsifiable; it is the hypothesis.

So, you need to work out what evidence would show you to be wrong.

A hypothesis can never be proved to be correct. We can gather some evidence that appears to confirm it. Then we can get some more evidence that also confirms it. But we can never be sure that one day we will find some data that contradicts it (see the "black swan" effect).

However, a hypothesis can be proved wrong. A scientific hypothesis must be capable of being proved wrong; I other words falsifiable. If it can't be falsified then you are  always assuming it is true and evidence become irrelevant.

Any hypothesis like yours needs to do two things: produce positive evidence (ie. a correlation between your predictions and earthquakes, and also between non-prediction and absence of quakes); and just as importantly show there is no negative evidence (eg no big earthquake when you predict one; earthquakes when you predict there should not be one). Obviously, in your case, this is going to require a lot of data because you are (I assume) looking for a very small effect that will only be apparent in very large data sets.

The other possibility is to look at the mechanisms. Without wanting to reopen the discussions in the closed thread, this would mean looking for evidence that the mechanisms exist. For example, can the position of the moon and sun significantly affect the magnetic field at the surface of the Earth; can such changes in the magnetic field affect geological processes; can these changes make earthquakes more likely; and so on. (I am guessing that the answer to each of these is "no"; but it is up to you to provide the chain of evidence to prove sceptics wrong).

[Phi for All]

Not falsifying. Falsifiability. It's the concept that an hypothesis needs a way to be shown false. If I claim aliens make it rain but can't be observed doing so due to advanced technology, no evidence can be produced to show I'm wrong. It would be an unfalsifiable claim.

Also, you weren't criticized for using "big words", you were criticized for using made-up terms like "solar threshold" and "UV connection". They make sense only to you, because you made them up. One of the big reasons why science uses shared terminology is that experimental results must be consistent between scientists. You need to be able to share with your peers.

[swansont]

Regarding falsifiability, put a slightly different way, there needs to be a way to show that your idea is wrong, if it is indeed wrong. You need to be able to analyze data or do experiments in such a way that will only work if you are correct, and not work if you are incorrect.

You could predict that more massive objects fell faster than lighter objects, and that experiment is both falsifiable and false. Also, the more precise your predictions are, the easier it is to show that the idea is falsifiable. That why e.g. we don't just say that gravity makes things drop (general prediction), we say it makes things drop at an acceleration of 9.8 m/s^2 near the earth, and it is independent of mass. Showing that any element of that was wrong would falsify the hypothesis, so it's falsifiable, but we get results that agree with it. 

Having statistically significant results is a step in that direction, but that should be coupled with predictions that are more precise. 

 

 

[Eddie B]

Thank you so much, my appreciation goes out to both of you. If could impose further, I would like to present a small calculation and ask if the result could be deemed as unfalsifiable or falsifiable by other means. I believe you are familiar with the concept of my "idea", so I shall not expand on this further.  I have learned on the prediction circuit that the calculated odds of a prediction being successful, works in your favour depending on the region you have selected. Meaning I would have a greater chance of securing a positive result by predicting an event in Indonesia, in preference to Madagascar. Average rate of 6+ occurrence in Indonesia, once every 3 months compared to "no" similar magnitude events in Madagascar from post 2004. So the odds for securing a 6+ result in Madagascar would be extremely high. 

If I now compare this example to my "idea", and relate this too one lunar cycle as suggested by a fellow member, or the one relevant month the lunar calculation relates too. This would be April (just gone), and the calculation relates to the moons perigee position on the 16th of this relevant month. I selected this particular moon event because it had already been represented in my closed thread. My calculations determined by online tools determined that the moon would be at its closest point to Earth between time periods 21:24:22 UTC - 22:50:45 UTC at a distance of 357,829 km's. I have been referring to these times as "prime focus" in reference to my magnifying glass analogy. They represent the first second the moon reaches this distance, and the last second before the moon alters distance. I shall refer to these positions as "stop / start" until I am advised otherwise. As I have just noted, my prediction calculations would have to be more precise. So with this in mind, the subject times here will be Judged as 21:24 UTC - 22:50 UTC = 2 minutes. If I calculate the odds that one of these two times related to stop / start perigee, would be associated with a pre destined seismic location, this should calculate as event-month-stop / start. 

April  43,200 minutes = odds 43,200/2 or 21,600/1

The unaccredited result relates to coordinates 82'49'W - 8'37'N on the Panama-Costa Rica border. This location experienced an M 6.1 event, approximately 44 hours prior to this post. It occurred 19 days after perigee took place, and the subject time related to to this event was "stop"... it was moonrise on this epicentre at 21:24 UTC 16th April. This happens every month related to differing solar / lunar combinations etc.

This result would be included on my "perigee" spreadsheet as contributing to statistical evidence. However, I would naturally assume that I would be asked how I formulated the calculation to secure this result... and would be obliged to direct enquiries to the online tools. If my calculation was verified as being correct, then this "precise" calculation had a 43,200 /  2 chance of determining the correct pre destined location. However it is not that impressive if you consider it was also moonrise at this time over half the world, so you would need to triangulate position. I do have the formula for this, but did not get the chance to demonstrate its abilities.

As I have demonstrated, this calculation is "past-tense", the event has occurred and the calculation relates to the event. Therefore there are years of unaccredited statistical evidence that could be calculated and presented for verification... would calculation "past-tense be regarded as statistical evidence ?. And if I relate back to my original question, what part of this methodology could I attempt to falsify ?

My sincere thanks again...

E B

 

 

 

[Strange]

1 hour ago, Eddie B said:

As I have demonstrated, this calculation is "past-tense", the event has occurred and the calculation relates to the event. Therefore there are years of unaccredited statistical evidence that could be calculated and presented for verification... would calculation "past-tense be regarded as statistical evidence ?

This technique, sometimes called "back-casting", is often used to test statistical models. It can be a good way of validating a model. It is not usually enough by itself to confirm a theory; you also need show the ability to forecast future events. (This is because it is too easy to accidentally "tune" a model to fit existing data.)

You are right that using a location where large earthquakes could make the analysis easier - if you can accurately predict an event that only occurs, on average, every century that is more convincing (everything else being equal) than predicting an even that occurs every week.

As far as falsifiability is concerned, you need to loom at correlations. Basically a list of your predicted dates/locations (and magnitudes, if possible) and a list of actual dates/locations (magnitudes). You then need to show that the number of matches (and mismatches) are better than would occur by chance. You might need the help of a mathematician with some expertise in statistics. 

1 hour ago, Eddie B said:

The unaccredited result ...

You used "unaccredited" a few times. I don't know what you mean by it.

[Eddie B]

To show ability to forecast future events would require an overseer  to confirm a positive or negative result of trialling the "idea", which would be deemed as prediction. I have been a little busy in speculations to look around the site properly, so could you advise me if  you have a section or category for predictive experimentation ?. 

Unaccredited definition - not recognised as having attained an acceptable standard (Oxford Dictionary). I'm sorry, my terminology and communication skills seem to have been questionable lately.  I had to look online to determine what an obtuse cherry-picker posting anecdotal information was... but no harm done, I can live with that !

Thank you my friend

E B