Frenemies of TalkRational:
Nontheist Nexus |  Rants'n'Raves |  Secular Cafe |  Council of Ex-Muslims |  The Skeptical Zone |  rationalia |  Rational Skepticism |  Atheists Today | 
TalkRational  

FAQ Rules Staff List Calendar RSS
Go Back   TalkRational > Discussion > Life Science Discussions

Life Science Discussions Biology, Natural History, etc.

Closed Thread
 
Thread Tools Display Modes
Old 08-07-2010, 03:37 PM   #1045036  /  #1326
Mike PSS
Senior Member
 
Join Date: Mar 2008
Location: Florida
Posts: 8,149
Mike PSS
Default

Quote:
Originally Posted by Loechelt View Post
Crazalus, I have to correct you on one point here. The prediction was tested and proven to be wrong.



I took the biotite diffusion data from Appendix B of Humphreys' 2003 ICC conference paper. For some strange reason, Humphreys never bothered to plot the data and compare to the original predictions of his 2000 RATE model.
Quote:
Originally Posted by Crazalus View Post
Quote:
Originally Posted by Dave Hawkins View Post
It was risky because the Jemez zircons had not been tested before. Humphreys took Argon in biotite data [Grove and Harrison, 1996], scaled it [Fortier and Gilletti, 1989], plugged the numbers into the diffusion equations [Carslaw and Jaeger, 1959] and came up with the data points in his graph in Fig. 7.
Dave... you've got that wrong. I've got the PDF open in front of me right now... and the notation is as follows...
Quote:
Predictions of yet-future experiments on He diffusion in Biotite, using the observed He retention in Jemez Zircons in two very different theoretical models.
Got that? The prediction was for BIOTITE ONLY!

Don't believe me?

Quote:
And ... voila! ... the lab results corresponded perfectly to his prediction.
Really? They tested Biotite?
Quote:
Highly specific.

Highly risky.
Unconfirmed because he changed his mind and decided they should test something else. That specific prediction has never been tested Dave...
Quote:
Originally Posted by VoxRat View Post
I'm not quite following this...
Quote:
Originally Posted by Loechelt View Post
... Why would Humphreys pay Farley extra to fudge the data to fit his model when it was so much easier to fudge the model to fit the data! He had to change the dimensions and boundary conditions of his model otherwise it would not fit the data! I pointed this out over a year ago in my reply to Humphreys. Please see the following graph.



The red stars show what Humphreys' unfudged model would have looked like. It does not match the measured data at all.
What do you mean "would have looked like"? The red stars represent your "alternative model", no?

I'm looking at the original prediction (RATE 1; p. 348; Fig. 7)right now. I see five points corresponding to "the Creation model" they are (best I can tell reading the graph) at the following coordinates:
  • 1.78, 10-15.0
  • 1.95, 10-15.8
  • 2.14, 10-16.7
  • 2.35, 10-16.9
  • 2.65, 10-17.5

Those were made in 2000 without any reference to subsequent data. They look like they correspond to the yellow triangles in your graph, identified as "RATE model 2003".

Why isn't Dave right, then, when he says the 2000 predictions were borne out by the eventual data (the gray boxes)?
Remember that all of Humphreys work in RATE 1 was based upon his readings and data from biotite (and all the mistaken assumptions).

As you can see in the Crazulus post above, the bottom sentence of Figure 7 is "Thus He diffusion measurements in biotite are likely to reject the evolutionist model and confirm the creationist model."

But, as Dr. Loechelt pointed out above, when you graph the 2003 data FROM BIOTITE with the model BASED UPON BIOTITE then they don't align.

Therefore the prediction in RATE 1 is false.

But THEN Humphreys went on to "correct" his work in RATE 2 and said the following (in Section 7; A New Creation Model; page 51; .pdf page 27)
http://www.icr.org//i/pdf/technical/...f-Zircons.pdf/
Quote:
This new model turns out to be very close to my previous Creation
model—within 0.5% for sample 1 and 0.05% for the others—despite
the different assumptions and equations.
The effect of two changes
(going from cavity in biotite to solid in biotite, and increasing the
effective radius from 22 μm to 30 μm) almost completely canceled
each other out (see Appendix D, Section D4.) Thus my previously
published predictions [Humphreys, 2000, p. 348, Figure 7] of diffusion
coefficients still happen to be numerically valid—no thanks to me! But
the numbers should be re-interpreted to apply to zircon, not biotite
.
So Dr. Loechelt's graph in your post shows the "alternate model" which is the correction of Humphreys work in RATE 2.
Mike PSS is offline   topbottom
Old 08-07-2010, 04:06 PM   #1045060  /  #1327
VoxRat
humble rodent
 
VoxRat's Avatar
 
Join Date: Mar 2008
Posts: 39,710
VoxRat
Default

Quote:
Originally Posted by Mike PSS View Post
...
So Dr. Loechelt's graph in your post shows the "alternate model" which is the correction of Humphreys work in RATE 2.
In other words, you're saying that the red stars represent what Humpy should have predicted, if his model had actually made any sense.

But what he, in fact, did predict, based on the wrong model* and the wrong math, turned out to be remarkably prescient!

Moral of the story: wrong is right!

* This seems to me the critical point. I don't think it is, as Dave would like to believe, just substituting zircon for biotite as the limiting barrier. I think this "bubbles of helium in a matrix" model is fundamentally wrong.
__________________
'The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.' - Bertrand Russell

"I already know the answer of course, being scientific" - F X
VoxRat is offline   topbottom
Old 08-07-2010, 06:25 PM   #1045162  /  #1328
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by VoxRat View Post
I'm a little puzzled by one thing.
In describing how he made his original prediction (the one that had to be rejiggered after seeing the 2003 data so that the prediction would be fulfilled ) Humpy says:
Quote:
... the biotite crystals iin which the Jemez zircons were embedded could "bottle up" the He in the zircons, causing longer retention times. So the real question is: how fast does He diffuse through biotite?
Of course we know now he got that completely wrong, but let's let that go for the meantime.
Quote:
From [Argon diffusion studies] we can very roughly extrapolate the Ar-in-biotite data to He-in-biotite, getting a band of estimated He diffusion rates for biotite.
That's the shaded band you see in his Figure 7 - the figure that constitutes his "specific, risky, prediction"*[though what "evolution" has to do with helium or zircons or anything else in this discussion is beyond me. I suspect it follows from the bizarre creationist notion that real-worlders came up with the age of the earth in the first place to retrofit it to Darwin's theory. But I digress...]
Quote:
Next we can plug each of these models into the well-understood equations for diffusion [Carslaw and Jaeger, 1959, pp. 234-235]. For simplicity, I assume the zircons to be He-filled spherical cavities in the biotite.
This is because he had mistakenly concluded that diffusion through zircons was so much faster than through biotite that we could essentially think of it as bubbles of free He in a biotite matrix.
Quote:
The cavity diameters are 44 microns, to match the surface areas of the zircons. Then for each temperature, I calculate what diffusion coefficient in the biotite is necessary to get the observed percent He retention in the zircons during the time allotted by each of the two models.
Now here's where I'm puzzled. What, exactly, counts as "He retention" in this model? How much He is left after some of it has left the bubble and entered the biotite matrix? I normally think of diffusion equations as telling me how much stuff travels how far how fast through other stuff. I think of them as predicting concentration gradients. It's not immediately obvious to me how I apply them to how much material crosses a phase boundary per unit time, and I find Humphreys's omission of these equations - referring to a 50 year old inaccessible reference which may or may not answer my question - suspiciously facile.

Without seeing those equations - and how the multiple arbitrary assumptions enter in - I can't judge the match of this "prediction" to the eventual rejiggered version of the prediction, or the eventual actual data. [But apparently Dave can! Or - you don't suppose? - maybe he's just taking Humpy's word for it?]

Moreover, this model for the process - He crossing a phase boundary - seems like a pretty poor match for the real situation, where you'll have a complex (time, temperature, diffusion coefficient)-dependent concentration gradient within the zircon.

Did Humphreys' ever flesh out these equations, and justify the equivalence of the two (seemingly, to me anyway) very different models (cavity vs. crystal)? Did someone else?
VoxRat,

The missing piece of information in your reasoning is that Humphreys used the helium retention data from Gentry. That was the starting point for all of his work. In a previous post I discussed how Humphreys used an inverse modeling approach which used a model and the measured helium content of the Fenton Hill zircons to calculate the diffusivity. Then he used the measured diffusivity as the test for his models. Consequently, the calculation of the data points on his graph require no knowledge of the material parameters of the system, a point which Dave Hawkins does not comprehend. Once he had better measured data, he could easily tweak the model to fit it. This is not predictive, physical science but merely curve fitting.

My model may not have been predictive, but it was physical since I tried to set every parameter in the model by either theoretical argument or independent data. Dave obviously has little appreciation for physical modeling. Fitting the curve is all that matters to him, even if the input to the model are non-physical.
Loechelt is offline   topbottom
Old 08-07-2010, 06:29 PM   #1045171  /  #1329
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by Dave Hawkins View Post
Quote:
Originally Posted by Loechelt View Post
Quote:
Originally Posted by Crazalus View Post

Ok... let's have fun...

First, "That graph has 5 data points which represent diffusivities of He through zircon/biotite at the indicated temperatures." is wrong... in that it represented estimated diffusion through Biotites only... it was based upon known rates of retention of He in Zircon, but was all about the predicted rates of Diffusion of He in Biotite. To suggest that the graph represented diffusivities of He in Zircon/Biotite is to misrepresent the graph.

The graph only refers to Biotite Diffusivity... any results of Zircon Diffusivity is irrelevant and does not confirm the prediction.

Second, "So Humphreys prediction was that the lab testing of zircons would result in the data values plotted in Fig. 7." He did not... he predicted that lab testing of BIOTITE would result in the data values in the graph. Testing of Zircon is IRRELEVANT.

Third, "The only thing Humphreys got wrong was that the restriction to flow of He was not the biotite ... it was the zircons themselves." Two different materials Dave...

Fourth, "The testing was done by Farley after Humphreys published his graph and guess what ... the numbers came out almost exactly to what Humphreys had published." Humphreys published a graph predicting values of diffusivity of He in BIOTITE... Farley tested ZIRCON. What numbers did Farley come out with for the diffusivity of He in BIOTITE?


Yes, there is a prediction in there... but it has not been tested, nor has it been proven correct. Dave... you've just shown that you don't know what the prediction was, or if it really was correct.

Now that you've been shown your error, will you acknowledge it? (I'm not holding my breath...)
Crazalus, I have to correct you on one point here. The prediction was tested and proven to be wrong.



I took the biotite diffusion data from Appendix B of Humphreys' 2003 ICC conference paper. For some strange reason, Humphreys never bothered to plot the data and compare to the original predictions of his 2000 RATE model.
You say "The prediction was tested and proven to be wrong" but your graph is not showing the results of the zircon testing. It's showing the results of the biotite testing. Humphreys prediction of diffusion rate was accurate when considering the zircon / biotite complex, which is the important thing anyway. Remember, we are interested in understanding why He is retained in zircons, not biotite.
Dave, you miss the point. Since the original prediction was for biotite, the proper test for the prediction is to compare it to biotite not zircon. Get it?
Loechelt is offline   topbottom
Old 08-07-2010, 06:38 PM   #1045181  /  #1330
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by Dave Hawkins View Post
Quote:
Originally Posted by Loechelt View Post
Quote:
Originally Posted by Dave Hawkins View Post
I think it's like this ... imagine a small bucket inside a large bucket and we pour water in the small one. Humphreys thought that the small bucket had 10 holes in the bottom and the large bucket had 2 holes. He made a prediction of the rate of outflow from the bottom of the large bucket. Turns out he got the rate right, but it was actually the big bucket with the 10 holes and the small bucket with 2 holes. Either case gives the same answer. Now I may be off on my analogy, but it seems right.
For the third time tonight, I have to point out that you are wrong. Your analogy is overly simplified. Let me rephrase it to more accurately depict what really happened. Start with your model with a small bucket with 10 holes and a large bucket with 2 holes. Then, upon physical inspection you discover that the small bucket actually has 2 holes and the large bucket has 10 holes. So, you begin a long discussion about how to model such a case. However, then you "simply the math" by making the model to have only 2 holes in both buckets, despite the fact that you have already seen the big bucket with 10 holes in the bottom. Not satisfied with this, you claim that the holes which were once 1.5" in diameter are now only 1" in diameter, even though direct measurement shows them to actually be 1.5" in diameter. All of these changes to the original model, by the way, were made after having studied preliminary data from experiments on water flowing from buckets. A year later, you repeat the water flowing experiments, find that the results match your preliminary data well, and publish the result as a significant scientific finding.

You may be impressed by this kind of "science", but I am not.
I think you are over-complicating this by putting in extraneous stuff that is not necessary. Why is it so difficult to see that Humphreys prediction was fulfilled (it was), but for a reason different than what he originally thought? Is that so difficult to admit? It's so obvious to me. If you focus simply on Humphreys prediction of diffusion through biotite, then yes, you're right, his prediction was wrong. But what Humphreys prediction really was - was a prediction of how fast Helium diffuses out of zircon encased in biotite. I'm not getting why you disagree with this.
I am glad to see that you finally admit that we are right and you are wrong -- Humphreys prediction was wrong! My "over-complicating" is not an exaggeration but is very indicative of what Humphreys actually did. If you allow someone to continually fudge a model, then you will be able to fit almost anything. I see no value in this kind of fudging, but evidently you do. This is not good science.
Loechelt is offline   topbottom
Old 08-07-2010, 06:42 PM   #1045182  /  #1331
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by Dave Hawkins View Post
Quote:
Originally Posted by Loechelt View Post
Quote:
Originally Posted by Dave Hawkins View Post
Jeepers ... Let's try this again. The prediction was regarding biotite because that's what he thought was the main restriction to He diffusion from zircons encased in biotite. Prior to doing any experiments he realized that this was wrong. Thus he changed the experiment to test zircons instead of biotite. And voila! ... the diffusivities came out exactly as that predicted for biotite.

Specific, risky prediction. Wrong reason. Still got it right. Way cool.
Dave, once again I have to point out that you are wrong. Humphreys realized that he was wrong after performing his first experiments. That is how he know he was wrong. It is all in his 2003 ICC conference paper. Take a look at the appendices. All the diffusion data is there for both biotite and zircon.
OK. I'll take your word for it. I was wrong about the "Prior to doing any experiments" bit.
Please, Dave, don't take my word for it. I gave you the reference. Look for yourself and see whether his conference paper has diffusion data for biotite and zircon in the appendices or not. Report back to us and tell us what you find there.
Loechelt is offline   topbottom
Old 08-07-2010, 07:32 PM   #1045205  /  #1332
Dave Hawkins
Senior Member
 
Dave Hawkins's Avatar
 
Join Date: Mar 2008
Location: Kansas City, MO Area
Posts: 12,960
Dave Hawkins
Default

No, Dr. Loechelt, I think you missed this ... Me (earlier today)
Quote:
If all you are interested in doing is quote mining Humphreys and disregarding the context, then yes, in a convoluted sort of way, you could say his prediction failed.

But if you are interested in reading the context (like you guys always say you are), then you will easily understand that the prediction was really about Helium diffusing -- not just through biotite -- but out of zircons through biotite. And this prediction was confirmed. And yes, it was specific (very) and it was risky (he had no prior data for Jemez zircons).
__________________
“The modern literature is full of funny ideas about the state of the natural sciences in antiquity. Those funny ideas can only come up if somebody does not study very carefully ... " --Dieter Lelgemann, Researcher at the Institute for Geodesy and Geoinformatics, Technical University of Berlin.
Dave Hawkins is offline   topbottom
Old 08-07-2010, 07:39 PM   #1045209  /  #1333
Dr. Nelson C. Armadingo
Senior Member
 
Dr. Nelson C. Armadingo's Avatar
 
Join Date: Mar 2008
Posts: 1,801
Dr. Nelson C. Armadingo
Default

Quote:
Originally Posted by Dave Hawkins View Post
No, Dr. Loechelt, I think you missed this ... Me (earlier today)
Quote:
If all you are interested in doing is quote mining Humphreys and disregarding the context, then yes, in a convoluted sort of way, you could say his prediction failed.

But if you are interested in reading the context (like you guys always say you are), then you will easily understand that the prediction was really about Helium diffusing -- not just through biotite -- but out of zircons through biotite. And this prediction was confirmed. And yes, it was specific (very) and it was risky (he had no prior data for Jemez zircons).
How was it confirmed if zircon/biotite matrix was not tested?
IOW, you are presenting an untruth dave. Again Still.

no hugs for thugs,
Dr. Nelson C. Armadingo

no fall, no flood, no resurrection
Dr. Nelson C. Armadingo is offline   topbottom
Old 08-07-2010, 07:47 PM   #1045216  /  #1334
Crazalus
Senior Member
 
Join Date: Jan 2009
Posts: 3,222
Crazalus
Default

Quote:
Originally Posted by Dave Hawkins View Post
No, Dr. Loechelt, I think you missed this ... Me (earlier today)
Quote:
If all you are interested in doing is quote mining Humphreys and disregarding the context, then yes, in a convoluted sort of way, you could say his prediction failed.

But if you are interested in reading the context (like you guys always say you are), then you will easily understand that the prediction was really about Helium diffusing -- not just through biotite -- but out of zircons through biotite. And this prediction was confirmed. And yes, it was specific (very) and it was risky (he had no prior data for Jemez zircons).
Dave... if he didn't do any testing on diffusion through Biotite, then his prediction hasn't been confirmed.

And as for reading the context... note that I posted a picture of the entire paragraph that explained Fig 7 on page 348, and it specifies that the measurements were to be of He diffusion through Biotite, not Zircons in Biotite.


We are not having to twist a thing here Dave... if you can point to where Humphreys says that the measurements were to be of He deffusion from Zircons in Biotite, AND if you can point to where the testing was done on Zircons in Biotite (which it wasn't) then, and ONLY then... can you claim that his prediction was confirmed.

Unfortunately for you, you can't... so every time you claim the prediction was confirmed, you are LYING!
Crazalus is offline   topbottom
Old 08-07-2010, 09:10 PM   #1045262  /  #1335
Pingu
I did. F. Poste.
Admin; GLaDOS
 
Pingu's Avatar
 
Join Date: Mar 2008
Location: UK
Posts: 46,871
Pingu
Default

Quote:
Originally Posted by Dave Hawkins View Post
No, Dr. Loechelt, I think you missed this ... Me (earlier today)
Quote:
If all you are interested in doing is quote mining Humphreys and disregarding the context, then yes, in a convoluted sort of way, you could say his prediction failed.

But if you are interested in reading the context (like you guys always say you are), then you will easily understand that the prediction was really about Helium diffusing -- not just through biotite -- but out of zircons through biotite. And this prediction was confirmed. And yes, it was specific (very) and it was risky (he had no prior data for Jemez zircons).
Dave, when we test hypotheses, one of the things we have to do is make an adjustment for multiple hypotheses, because the more hypotheses you make, the less "risky" any one is.

So if you are very specific then you can get a way with a higher p value than if you aren't. However, contrary to your assertion, Humphreys wasn't specific at all. I could make a prediction that the answer to a question would three. If I sample enough questions, I will almost certainly find one to which that answer is correct. That's neither specific not risky. Humphreys found, fortuitously, that the answer to a different question to the one he'd originally predicted the answer to was the answer he'd predicted for his original question. That makes him lucky, not right, as he admits. His answer to his original question was wrong. That makes him wrong.

Neither right answer means that the earth is than 6,000 years old.
Pingu is offline   topbottom
Old 08-07-2010, 10:53 PM   #1045340  /  #1336
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by VoxRat View Post
I'm not quite following this...
Quote:
Originally Posted by Loechelt View Post
... Why would Humphreys pay Farley extra to fudge the data to fit his model when it was so much easier to fudge the model to fit the data! He had to change the dimensions and boundary conditions of his model otherwise it would not fit the data! I pointed this out over a year ago in my reply to Humphreys. Please see the following graph.



The red stars show what Humphreys' unfudged model would have looked like. It does not match the measured data at all.
What do you mean "would have looked like"? The red stars represent your "alternative model", no?

I'm looking at the original prediction (RATE 1; p. 348; Fig. 7)right now. I see five points corresponding to "the Creation model" they are (best I can tell reading the graph) at the following coordinates:
  • 1.78, 10-15.0
  • 1.95, 10-15.8
  • 2.14, 10-16.7
  • 2.35, 10-16.9
  • 2.65, 10-17.5

Those were made in 2000 without any reference to subsequent data. They look like they correspond to the yellow triangles in your graph, identified as "RATE model 2003".

Why isn't Dave right, then, when he says the 2000 predictions were borne out by the eventual data (the gray boxes)?
VoxRat, you have to go back to the discussion in my reply to Humphreys. Briefly, the argument goes like this.

Humphreys' Model 2000: Assume Dzircon >> Dbiotite. Then a reasonable geometry is a zircon-sized cavity in a biotite sphere. The radius of this cavity representing the zircon was determined to be 22 um.

Farley's Measurements 2003: Measurement shows that Dzircon << Dbiotite. Therefore, the diffusion model has to be revised.

Alternate Model: Given the two points above, the simplest and most obvious way to change the model is to turn it "inside-out", making the cavity a solid sphere and ignoring the biotite. Since Dzircon << Dbiotite does not affect the original choice for the geometry, the radius of the solid sphere should again be 22 um. This model was used to calculate the red stars. Note that the numbers change from the 2000 model because, when you change a model you usually expect the out-come to be different. Funny how that works in science.

Humphreys' Model 2003: Additional changes were made to the model. Humphreys started by considering a two-material system. However, in the end, he simplified that system so that Dzircon = Dbiotite even though his own data showed that Dzircon << Dbiotite. Furthermore, he increased the radius of the zircon sphere from 22 um to 30 um for apparently no good reason. Why all these additional changes, especially the choice of Dzircon = Dbiotite when measurements showed that Dzircon << Dbiotite? This model was used to calculate the yellow triangles.

Farley's Measurements 2003: This data was available before Humphreys published his new model. If one extrapolates the measured diffusivity of helium in zircon down to temperature of interest, one gets the blue line. Three of the yellow triangles lie almost exactly on the blue line. Is this a remarkable and risky prediction? No, it cannot be called a prediction because the new model was published after the measurements were made. Is it pure coincidence? I will let you decide. Did Humphreys make these apparently unphysical changes to his 2003 model so that it would match the blue line? I cannot rule out this possibility. What I can say is that Humphreys used Dzircon = Dbiotite when his own data showed that Dzircon << Dbiotite, and that he changed the size of his zircon from 22 um to 30 um. When I follow through with the calculations, I find that it moves the red stars to the yellow triangles, 3 of which just "happen" to lie on the blue line. Again, there is no prediction here since all of this information was known at the time Humphreys published his revised model.

Farley's Measurements 2004: These are the gray sqaures in the figure, most of which happen to lie along the blue line from the 2003 measurements. Since 3 of the yellow triangles also lie along the blue line, they also match the gray squares. It is a simple consequence of the mathematical law

If A = B and B = C then A = C (transitive law of equality).

Now Humphreys' publishes his 2003 model and his 2004 measurements on the same plot and declares victory in a remarkable prediction. However, how remarkable is this agreement of the model to the measured data? Two questions need to be answered.

1. How remarkable is it that the 2003 diffusion measurements (blue line) match the 2004 diffusion measurements (gray squares)? It is not remarkable at all. A bulk material property is being measured, and for most zircons this should be a well defined number that can be printed in a handbook. Subsequent measurements in the literature show that the diffusivity of helium in zircon is very reproducible.

2. How accurate is Humphreys' 2003 model (yellow triangles)? Well, I showed by counter-example that other young-earth models are possible (red stars). Therefore, the good agreement between the model and measured diffusivity is not a consequence of a young-earth belief system, but is rather the result of very particular choices in the model that Humphreys made in 2003.

My argument is that in order for the agreement between Humphreys' model and experiment to be considered significant, these particular choices in the model must be defended against all reasonable alternatives. Humphreys has not done this. In contrast, I can make a good case that the alternate model is more physical. In particular, it takes into account Dzircon << Dbiotite. I argue that based upon the known physics, it is a better choice of geometry and boundary conditions. However, it does not agree well with measurement, which suggests that the fundamental flaw in the model is related to its young-earth assumptions.
Loechelt is offline   topbottom
Old 08-07-2010, 11:11 PM   #1045362  /  #1337
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by VoxRat View Post
Quote:
Originally Posted by Mike PSS View Post
...
So Dr. Loechelt's graph in your post shows the "alternate model" which is the correction of Humphreys work in RATE 2.
In other words, you're saying that the red stars represent what Humpy should have predicted, if his model had actually made any sense.

But what he, in fact, did predict, based on the wrong model* and the wrong math, turned out to be remarkably prescient!

Moral of the story: wrong is right!

* This seems to me the critical point. I don't think it is, as Dave would like to believe, just substituting zircon for biotite as the limiting barrier. I think this "bubbles of helium in a matrix" model is fundamentally wrong.
There may be some degree of coincidence in a wrong model giving a numerically right answer, but there are also more subtle manipulations that Humphreys could have done as well. For instance, Martin pointed out that the measurement of diffusivity gives D/a2, not D directly. Therefore, in changing a from 22 um to 30 um, he already has a fudge factor of about 2 to play with. Then you add to that the fact that the defect tail is starting to affect the data, you have another fudge factor to play with. Remember, the RATE team had 4 experiments to determine the diffusivity of helium in zircon, and they only published results from 2 of them. If the data from experiments #2 and #3 (the unpublished ones) had met their objectives, they might not have requested a fourth experiment. Incidentally, it is common practice in science to publish even the bad data from "failed" experiments. If you look at the paper on helium diffusion by Reiners, some of the curves are not that clean and have kinks and anomalies. So, between fudging the model, fudging the data a little, and repeating the experiment multiple times, the chances of "success" probably improve.
Loechelt is offline   topbottom
Old 08-07-2010, 11:17 PM   #1045371  /  #1338
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by Dave Hawkins View Post
No, Dr. Loechelt, I think you missed this ... Me (earlier today)
Quote:
If all you are interested in doing is quote mining Humphreys and disregarding the context, then yes, in a convoluted sort of way, you could say his prediction failed.

But if you are interested in reading the context (like you guys always say you are), then you will easily understand that the prediction was really about Helium diffusing -- not just through biotite -- but out of zircons through biotite. And this prediction was confirmed. And yes, it was specific (very) and it was risky (he had no prior data for Jemez zircons).
I fail to see your point here. Please elaborate.
Loechelt is offline   topbottom
Old 08-07-2010, 11:37 PM   #1045394  /  #1339
Dave Hawkins
Senior Member
 
Dave Hawkins's Avatar
 
Join Date: Mar 2008
Location: Kansas City, MO Area
Posts: 12,960
Dave Hawkins
Default

I don't know how to make it more clear.
__________________
“The modern literature is full of funny ideas about the state of the natural sciences in antiquity. Those funny ideas can only come up if somebody does not study very carefully ... " --Dieter Lelgemann, Researcher at the Institute for Geodesy and Geoinformatics, Technical University of Berlin.
Dave Hawkins is offline   topbottom
Old 08-07-2010, 11:43 PM   #1045405  /  #1340
Steviepinhead
Senior Member
 
Steviepinhead's Avatar
 
Join Date: Mar 2008
Location: Seattle, WA
Posts: 29,985
Steviepinhead
Default

Of course you don't...

That should worry you, Dave.
Steviepinhead is offline   topbottom
Old 08-07-2010, 11:46 PM   #1045407  /  #1341
ChocolateChipCookies
Member
 
ChocolateChipCookies's Avatar
 
Join Date: Jun 2010
Posts: 78
ChocolateChipCookies
Default

Quote:
Originally Posted by Loechelt View Post
Quote:
Originally Posted by VoxRat View Post
Quote:
Originally Posted by Mike PSS View Post
...
So Dr. Loechelt's graph in your post shows the "alternate model" which is the correction of Humphreys work in RATE 2.
In other words, you're saying that the red stars represent what Humpy should have predicted, if his model had actually made any sense.

But what he, in fact, did predict, based on the wrong model* and the wrong math, turned out to be remarkably prescient!

Moral of the story: wrong is right!

* This seems to me the critical point. I don't think it is, as Dave would like to believe, just substituting zircon for biotite as the limiting barrier. I think this "bubbles of helium in a matrix" model is fundamentally wrong.
There may be some degree of coincidence in a wrong model giving a numerically right answer, but there are also more subtle manipulations that Humphreys could have done as well. For instance, Martin pointed out that the measurement of diffusivity gives D/a2, not D directly. Therefore, in changing a from 22 um to 30 um, he already has a fudge factor of about 2 to play with. Then you add to that the fact that the defect tail is starting to affect the data, you have another fudge factor to play with. Remember, the RATE team had 4 experiments to determine the diffusivity of helium in zircon, and they only published results from 2 of them. If the data from experiments #2 and #3 (the unpublished ones) had met their objectives, they might not have requested a fourth experiment. Incidentally, it is common practice in science to publish even the bad data from "failed" experiments. If you look at the paper on helium diffusion by Reiners, some of the curves are not that clean and have kinks and anomalies. So, between fudging the model, fudging the data a little, and repeating the experiment multiple times, the chances of "success" probably improve.
Dave, are these terms simple enough for you?
ChocolateChipCookies is offline   topbottom
Old 08-07-2010, 11:54 PM   #1045418  /  #1342
VoxRat
humble rodent
 
VoxRat's Avatar
 
Join Date: Mar 2008
Posts: 39,710
VoxRat
Default

Quote:
Originally Posted by Loechelt View Post
...
Alternate Model: Given the two points above, the simplest and most obvious way to change the model is to turn it "inside-out", making the cavity a solid sphere and ignoring the biotite. Since Dzircon << Dbiotite does not affect the original choice for the geometry, the radius of the solid sphere should again be 22 um. This model was used to calculate the red stars. ...
Forgive me if this is treading well-trodden territory... Are the calculations for this spelled out in one of your papers or in this thread somewhere?
Quote:
My argument is that in order for the agreement between Humphreys' model and experiment to be considered significant, these particular choices in the model must be defended against all reasonable alternatives. Humphreys has not done this. In contrast, I can make a good case that the alternate model is more physical. In particular, it takes into account Dzircon << Dbiotite. I argue that based upon the known physics, it is a better choice of geometry and boundary conditions. However, it does not agree well with measurement, which suggests that the fundamental flaw in the model is related to its young-earth assumptions.
So Humpy basically got lucky - because even though there's nothing about the "Creation model" that dictates the yellow triangles, as opposed to the red stars, it happens that the 2000 paper did plot the yellow triangles, which did turn out to match the eventual Farley data pretty closely. Right?
Quote:
There may be some degree of coincidence in a wrong model giving a numerically right answer, but there are also more subtle manipulations that Humphreys could have done as well. For instance, Martin pointed out that the measurement of diffusivity gives D/a2, not D directly. Therefore, in changing a from 22 um to 30 um, he already has a fudge factor of about 2 to play with.
However, on a Log(10) scale, a factor of 2 doesn't change things much - only by about the size of the data point symbols used in that graph.

(Not really playing Davil's Advocate here; just trying to be thorough)
__________________
'The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.' - Bertrand Russell

"I already know the answer of course, being scientific" - F X
VoxRat is offline   topbottom
Old 08-08-2010, 01:26 AM   #1045523  /  #1343
Tension
Member
 
Tension's Avatar
 
Join Date: Jan 2010
Posts: 88
Tension
Default

When you say that Humphreys RATE experiment got peer reviewed and published, do you mean it was reviewed and published in a creationist journal or a REAL science journal?
Tension is offline   topbottom
Old 08-08-2010, 01:36 AM   #1045532  /  #1344
VoxRat
humble rodent
 
VoxRat's Avatar
 
Join Date: Mar 2008
Posts: 39,710
VoxRat
Default

Quote:
Originally Posted by Tension View Post
When you say that Humphreys RATE experiment got peer reviewed and published, do you mean it was reviewed and published in a creationist journal or a REAL science journal?
Heh.
Who did say that? Dave Hawkins?
It was a creationist journal. "Peers" being - other creationists.
__________________
'The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.' - Bertrand Russell

"I already know the answer of course, being scientific" - F X
VoxRat is offline   topbottom
Old 08-08-2010, 02:48 AM   #1045578  /  #1345
Dave Hawkins
Senior Member
 
Dave Hawkins's Avatar
 
Join Date: Mar 2008
Location: Kansas City, MO Area
Posts: 12,960
Dave Hawkins
Default

Uh huh ... like an evolutionary biology journal. "Peers" being - other evolutionary biologists.
__________________
“The modern literature is full of funny ideas about the state of the natural sciences in antiquity. Those funny ideas can only come up if somebody does not study very carefully ... " --Dieter Lelgemann, Researcher at the Institute for Geodesy and Geoinformatics, Technical University of Berlin.
Dave Hawkins is offline   topbottom
Old 08-08-2010, 02:58 AM   #1045583  /  #1346
Loechelt
Senior Member
 
Loechelt's Avatar
 
Join Date: May 2010
Location: Tempe, Arizona
Posts: 674
Loechelt
Default

Quote:
Originally Posted by VoxRat View Post
Quote:
Originally Posted by Loechelt View Post
...
Alternate Model: Given the two points above, the simplest and most obvious way to change the model is to turn it "inside-out", making the cavity a solid sphere and ignoring the biotite. Since Dzircon << Dbiotite does not affect the original choice for the geometry, the radius of the solid sphere should again be 22 um. This model was used to calculate the red stars. ...
Forgive me if this is treading well-trodden territory... Are the calculations for this spelled out in one of your papers or in this thread somewhere?
The general mathematical equations are given in my technical paper. Some of the specific choices for parameter values are given in my response to Humphreys. Contact me if you have any further questions, and I can provide additional details.

Quote:
Quote:
My argument is that in order for the agreement between Humphreys' model and experiment to be considered significant, these particular choices in the model must be defended against all reasonable alternatives. Humphreys has not done this. In contrast, I can make a good case that the alternate model is more physical. In particular, it takes into account Dzircon << Dbiotite. I argue that based upon the known physics, it is a better choice of geometry and boundary conditions. However, it does not agree well with measurement, which suggests that the fundamental flaw in the model is related to its young-earth assumptions.
So Humpy basically got lucky - because even though there's nothing about the "Creation model" that dictates the yellow triangles, as opposed to the red stars, it happens that the 2000 paper did plot the yellow triangles, which did turn out to match the eventual Farley data pretty closely. Right?
The calculations from the 2000 paper do match Farley's data pretty well. The two aspects that I have a hard time with is how can a model fundamentally change without the resulting numbers calculated by it changing, and how can a model be considered correct if it has so many errors in its construction. Perhaps Humphreys got lucky. Perhaps there are subtle things that could have been done to manipulate the results. I have seen some evidence of manipulation already.

Quote:
Quote:
There may be some degree of coincidence in a wrong model giving a numerically right answer, but there are also more subtle manipulations that Humphreys could have done as well. For instance, Martin pointed out that the measurement of diffusivity gives D/a2, not D directly. Therefore, in changing a from 22 um to 30 um, he already has a fudge factor of about 2 to play with.
However, on a Log(10) scale, a factor of 2 doesn't change things much - only by about the size of the data point symbols used in that graph.

(Not really playing Davil's Advocate here; just trying to be thorough)
Agreed, a factor of 2 on a log scale is not that much. Perhaps the point I made about data shopping is more relevant. They had to repeat the experiment 4 times before they published their final result, and they never published the results from experiments #2 and #3. They only published #1 and #4. The defect tail creates a bend in the Arrhenius curve. If you find a sample with just the right number of defects at the right temperature, placement of that tail might make a difference. What I am tempted to do is plot Humphreys' model against several different diffusion datasets using just the high temperature data. If you look closely at a previous graph, I think an effect can already be seen.



The high temperature data begins to diverge from Humphreys' model data right around the last data point. The remaining model points all lie well above the trend line. As we have discussed before, the low temperature laboratory data are very problematic, being dependent upon defect density, helium background concentration, thermal history etc. The high temperature data are more robust because they reflect a fundamental material property. Since the defect line can move around from sample to sample, data shopping is a real possibility. Some of the good agreement may actually be more of an illusion and the result of good graphsmanship. The high temperature trend line does not match the model data well. Only the last 2 model points overlap the last 4 measurements. These last 4 measurements are the lowest temperatures in the laboratory experiment and are subject to the most uncertaintly because of the low gas yield. Similarly, the last 2 model points are the deepest samples from the Fenton Hill wells and are again subject to the most uncertainty because of the low gas concentration. So you have a minimal overlap of two datasets at the point where both dataset have the most uncertainty.

When looked at another way, the agreement between the experiment and model is actually quite poor. For instance, the high temperature part of the experiment is the reliable portion, and should be extrapolated down to lower temperatures for making any comparison. Similarly, the low temperature part of the model is the reliable portion because it is based upon samples with higher (and hence more accurately measured) helium content. The comparison of extrapolated high temperature laboratory data with low temperature model data makes the uniformitarian model look better. We have come full circle with the argument, arriving again at the conclusion that Humphreys' incorrect interpretation of the low-temperature laboratory data is the fundamental flaw (among a host of others).

Therefore, having thought about the question a little bit, the agreement between the 2000 model data and the 2004 laboratory experiment really is magic -- the ability of a magician to make you see what he wants you to see rather than see things for what they really are. By focusing the attention entirely upon a small region of overlap between two datasets in the temperature region where both have the greatest uncertaintly, the viewer is distracted from considering how the overall trends of the two datasets diverge in the region where they have the greatest amount of confidence. Very clever. Don't for a moment think that Humphreys is stupid. In a strange sort of way he is a very intelligent man. Given good graphsmanship, data shopping, and a little bit of luck, the agreement may not be as unlikely as it seems.
Loechelt is offline   topbottom
Old 08-08-2010, 03:05 AM   #1045590  /  #1347
BelindaP
Senior Member
 
BelindaP's Avatar
 
Join Date: Mar 2009
Posts: 8,922
BelindaP
Default

Sounds like they repeated till they got the results they wanted. Classic noob scientist thing to do. Usually they beat that out of you pretty early in grad school or your professional career. Apparently, not so much in creationist circles.

Pay attention, Dave. This is why he can't pass peer review in a quality science journal. The reviewers would have wanted to see the results from experiments #2 and #3 with an explanation for why the data were not included.
__________________
We <3 SMS.
BelindaP is offline   topbottom
Old 08-08-2010, 03:25 AM   #1045608  /  #1348
Tension
Member
 
Tension's Avatar
 
Join Date: Jan 2010
Posts: 88
Tension
Default

Quote:
Originally Posted by Dave Hawkins View Post
Uh huh ... like an evolutionary biology journal. "Peers" being - other evolutionary biologists.
Your saying Humphreys experiment was published in an evolutionary biology journal? What the hell do zircons and helium have to do with evolutionary biology? whats the name of the journal?
Tension is offline   topbottom
Old 08-08-2010, 03:30 AM   #1045612  /  #1349
VoxRat
humble rodent
 
VoxRat's Avatar
 
Join Date: Mar 2008
Posts: 39,710
VoxRat
Default

Quote:
Originally Posted by Dave Hawkins View Post
Uh huh ... like an evolutionary biology journal. "Peers" being - other evolutionary biologists.
And you see some symmetry there?
You think there are equivalent bodies of evidence for creationism and for evolutionary biology?

Well? We keep asking for this evidence.
And you keep not providing it.
__________________
'The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.' - Bertrand Russell

"I already know the answer of course, being scientific" - F X
VoxRat is offline   topbottom
Old 08-08-2010, 03:31 AM   #1045615  /  #1350
SomecallmeTim
That's no ordinary rabbit!
 
SomecallmeTim's Avatar
 
Join Date: Mar 2008
Location: Mt. View, CA
Posts: 5,011
SomecallmeTim
Default

Quote:
Originally Posted by Tension View Post
Quote:
Originally Posted by Dave Hawkins View Post
Uh huh ... like an evolutionary biology journal. "Peers" being - other evolutionary biologists.
Your saying Humphreys experiment was published in an evolutionary biology journal? What the hell do zircons and helium have to do with evolutionary biology? whats the name of the journal?
Dave is saying Humphreys' work was peer-reviewed, and in a way it was. The work of one dishonest YEC idiot was reviewed by a few other dishonest YEC idiots and given the Creto seal of approval.

I think Dave is admitting that the Humphreys' work was not scientifically peer-reviewed.
__________________
"We measure heat in watts, dumbshit." - Schneibster

"Furthermore momentum can be converted to heat. Directly." - Schneibster
SomecallmeTim is offline   topbottom
Closed Thread

  TalkRational > Discussion > Life Science Discussions

Tags
"algebras, a cromulent discussion, could sanshou be the one?, dave the king of science, don't mention sanshou, how do they work?", wishful thinking, would sandshoe tea be fun

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT. The time now is 06:05 AM.


Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright ©2008 - 2014, TalkRational.org