Not exactly. YOu keep making up models that don't match reality. and then ignoring things like the 0.3 deg C added merely by editing.
Glenn, I've been downloading your data and plotting it. The fact you keep acting as if I'm not doesn't speak well of your veracity. Sure I've generated models
along with this. But do me the great honor of at least presenting what I have done accurately.
I most assuredly do
not appreciate someone clipping my data, Trusty.
Prove you wrong? Life doesn't consist of mathematical proofs.
If you can't use math to prove a mathematical concept wrong then it says more about your ability with the math than it does about "life".
And I didn't come here to debate you even though you seem to be under that misapprehension. I came here to show the data that I have been showing.
Too bad for you you ran into a PhD scientist with about years of experience in industrial/academic/governmental research science and who has some statistical chops.
I showed that the cold end of the data does not have the slope of 1. You can see that in my graph. I also pointed out to you from your own graph that when one town is 40 degrees the other can be anywhere from something like 15 to 70 degrees and these towns are only 20 miles apart or so.
All you have to do is replot the data as you originally had and which you had somehow lost 29,900 data points. I didn't ask you to show me
another graph of unrelated format.
I will stand by the math and the graph until you prove mathematically otherwise.
A slope is a really, really easy thing to generate, Glenn.
This part is frustrating. I can't get excel to open after last night but I did plot the data and you are lazy in not going to find it. If you can't remember it then maybe you should pay more attention to what I write.
I actually would look quite forward to that. I don't mean to be a complete tool on this but I've had plenty of training from a pro these past 500 posts.
All one has to do is look at the temperatures in the picture I posted of Atlanta to see that the red, which is in the city is representative of a significantly higher temperature than the pale/whitish yellow out in the surrounding countryside, to see that urbanization does indeed have a huge impact on the temperature. Often times, Thau, one merely needs to look at the temperature rather than using obfuscation.
Wow, Glenn. So when someone
speaks against your confirmation bias you call it obfuscation.
I see how this works.
Your link was broken on the page, although I was able to find it and did look at it. Thau, just because someone publishes something doesn't automatically make it correct, nor does it make it incapable of being criticized.
Well, unless you can prove his numbers to be lies or incorrectly processed you'll have to keep your yap shut on it won't you?
Remember just yelling at someone, while it may work for you "managers" doesn't change what the numbers say.
I've heard plenty of desk-jockies talk a big game about science, but until I see them prove the point with either hard data or accurate assessment of said data I have little patience for it.
Anyone whose rolled their sleeves up and gotten dirty with the hard data can tell you the "declarations" of a manager or desk-jockey matters not one whit.
And lets remind people of your own howlers. The SD of the temperature difference between almost any two towns I have ever looked at is 3-4 degrees. To be 95% sure that the world is warming, the data must be precise to within 3 standard deviations.
Er...95%
of the population is within about
2 standard deviations of the mean
* (this is, indeed, within 3 standard deviations, but the rule is generally called the 68-95-99.7 Rule in stats...3 sd contains 99.7% of the population).
But that is not the same as a 95% confidence interval on the mean
As I've repeatedly pointed out the 95% confidence interval on the mean is equal to a band that is
+ [FONT="]1.96*s/sqrt(N)[/FONT]
YEs, Thau, I messed up the graph. I will admit it as many times as you wish me to. You, on the other hand have yet to acknowledge that you made that howling statistical misunderstanding about the error bars around warming.
the 95% confidence interval on the mean is equal to:
mean+1.96*s/sqrt(N)
where:
s = standard deviation
sqrt(N) = square root of the number of samples
(1.96 is the z-statistic value applicable for populations, but the t-statistic value can be used based on the number of degrees of freedom)
What I'm curious about, Glenn, is is it the
fraction part of the math that is throwing you off or is it the color of the font?
I can change the color for you, or I can help you with understanding what a
denominator is if you like.
Let me know what your issue is.
95% ci on the mean =mean+1.96*s/sqrt(N)
where:
s = standard deviation
sqrt(N) = square root of the number of samples
(1.96 is the z-statistic value applicable for populations, but the t-statistic value can be used based on the number of degrees of freedom)
You also made a physics howler trying to figure out if temperature was additive.
Was I incorrect in the
conclusion from that post?
Please show me how the conclusion was incorrect.
(I posed the question in a rhetorical fashion in case there was something missing, but I don't recall making a factual error. I do recall
asking if it would be additive and
then quite quickly in the same post ultimately building out the hypothesis that indeed it should not be and would bias the temperature high by placing a floor on temperatures.)
And you made a howler out of explaining how heat is transferred.
Ahhh, but sadly I was also
technically correct there as well, Glenn. While you don't
need statistical mechanics to explain convective heat transfer
YOU CAN USE STATISTICAL MECHANICS FOR IT.
Now of course at the time I clearly stated this was not my area.
At least I didn't tell you you were WRONG. The reason your stats errors are so disgusting is
not because you were so wrong but rather because you were such a SNOT about it and told me repeatedly how bad I was with stats.
That's the difference, Glenn.
I don't care that you never got past high school stats or that you don't understand stats. But please don't insult me
and be bad with stats.
People who don't understand physics shouldn't be trying to use statistics.
Uhhh...why?
I told you right up front Thau, that I am not getting into a discuss like we did last time. I am not going to have you and I speaking only to each other with no one else understanding.
So you can't talk science at the big-boy level? Got it.
Frankly I'm not here to talk try to convince the gullible. I'm not here to spin the data or "cartoonify" it. You came to a gunfight with a butterknife, Glenn.
Given the current state of my computer, I am forced to agree. I just found out that Word won't work.
So now Word is involved in processing data? Your computer is very messed up!
I recommend you buy a new one. Might try an Apple this time?
And you can't seem to understand the connection between noise level and SD which we debated before and you hilariously think that a large sd doesn't mean a large noise level when one subtracts the two nearby towns,
Oh my. It's almost like you are incapable of understanding the
t-test.
Let me trot it out for you again:
t = (Mean[sub]1[/sub] - Mean[sub]2[/sub])/s[sub]mean1-mean2[/sub]
where
s[sub]mean1-mean2[/sub] is the standard error of the difference between the means. It itself is:
s[sub]mean1-mean2[/sub] = sqrt{((N[sub]1[/sub]s[sub]1[/sub][sup]2[/sup] + N[sub]2[/sub]s[sub]2[/sub][sup]2[/sup])/(N[sub]1[/sub]+N[sub]2[/sub]-2))*((N[sub]1[/sub]+N[sub]2[/sub])/N[sub]1[/sub]N[sub]2[/sub]))}
N= number of samples
[FONT="]s= standard deviation
[/FONT]
There's a pairwise version also available
t = d/(s[sub]d[/sub]/sqrt(N))
where
d= (1/N)*sum(differences between individual pairs)
s[sub]d[/sub] = sqrt{(1/(N-1)*sum((d[sub]i[/sub]-d)^2))
But maybe I'm missing something. You seem to freely throw around sd and error. Which error specifically are you talking about?
The standard error of the mean is actually the standard deviation divided by the square root of the number of samples
The standard deviation is a measure of dispersion of the data about the mean.
You know, given what I am paid, I would cite this as evidence that I am pretty smart to get someone to pay me like that for that kind of skill.
Actions speak far louder than paychecks in the world of science.
No, actually that is the chart I posted BEFORE my issue with Excell.
Well, I sure couldn't find your
corrected Okmulgee vs Okemah post. And the fact you can't seem to repost it or do anything other than say you already did and you aren't going to show me just sounds weird.
If you posted it after post #414
just show me where it is. I'll gladly apologize!
I posted it again in that post but that isn't the correction Apparently you can't follow the temporal order of a discussion.
IF ANYONE CAN SHOW ME WHERE GLENN POSTED THIS CORRECTION< PLEASE DO SO. I WILL GLADLY APOLOGIZE TO GLENN.
I'm guessing he can't apparently find it either. (maybe Microsoft Internet Explorer is getting in on the action and messing up stuff too! Along with Excel and Word!)
It isn't from the post where I did replot it. But, as with the claim that I didn't acknowledge the error, you seem incapable of actually finding a post.
Apparently you can't either.
I'm noting this about you Glen:
YOU TALK BIG.
Maybe after we go through another couple of rounds of you mis-claiming things I will point you to it.
I looked. But if it allows you to
talk big some more I'll wait.
And like last time, you won't acknowlegde that you were wrong in your claim about what I did or didn't do.
I will give you blessings and a reputation bump if you do that very thing.
And
I WILL PUBLICLY APOLOGIZE IF YOU SHOW ME THE POST.
You have this post as proof of that. If you show me where you reposted the Okmulgee vs Okemah x vs y plot along with the slope
I will apologize and you can repeatedly take this post as a bat to hit me with repeatedly.
Please just do something!
STOP TALKING talking talking. That's all you desk-jockies/managers do!
