Forum Settings
       
1 2 Next »
Reply To Thread

Remember that hockey stick?Follow

#52 Mar 12 2013 at 1:21 PM Rating: Excellent
Meat Popsicle
Avatar
*****
11,446 posts
His Excellency Aethien wrote:
someproteinguy wrote:
But you very likely can't catch someone falsifying data in a review of a publication unless they really suck at lying. You'll almost need someone from the lab to come out and admit to it.
If I recall correctly, his data was always too perfect to be true. Being exactly as expected each time on top of refusing to let anyone else see any raw data.

Yeah, all tell tale signs for sure. It's really hard to pull the trigger on that kind of thing though, especially prior to publication. It's one thing to think the data looks suspiciously good, and another to hold up publication and ask the submitter to go collect more data. Especially since anything like that will likely be taken to an additional reviewer, and they'll likely have to think the same thing. More likely it gets passed though with reservations, and they wait for others to try and replicate the experiment and confirm their suspicions. Whether or not that happens though...

The whole "I don't think you're innocent, but I can't prove you're guilty" thing, basically.
____________________________
That monster in the mirror, he just might be you. -Grover
#53 Mar 12 2013 at 1:34 PM Rating: Excellent
Liberal Conspiracy
*******
TILT
His Excellency Aethien wrote:
If I recall correctly, his data was always too perfect to be true. Being exactly as expected each time on top of refusing to let anyone else see any raw data.

Coincidentally, if only because I mentioned polling before, this was how Nate Silver exposed... Opinion Strategies(?) or whatever the name of that outfit who used to poll for DailyKos was called. All their numbers stayed within a safe range of everyone else's and they never seemed to have any outliers. Further review showed some weird streaks in their fractions. Like someone saying 45... poiiiiintttt.... 8? Eight sounds good. Had it been natural, you'd see a more even distribution in the numbers but it was as though someone was picking numbers that "sounded right".

Which, as it turns out, they were.
____________________________
Belkira wrote:
Wow. Regular ol' Joph fan club in here.
#54 Mar 12 2013 at 1:49 PM Rating: Good
Avatar
******
26,737 posts
someproteinguy wrote:
His Excellency Aethien wrote:
someproteinguy wrote:
But you very likely can't catch someone falsifying data in a review of a publication unless they really suck at lying. You'll almost need someone from the lab to come out and admit to it.
If I recall correctly, his data was always too perfect to be true. Being exactly as expected each time on top of refusing to let anyone else see any raw data.

Yeah, all tell tale signs for sure. It's really hard to pull the trigger on that kind of thing though, especially prior to publication. It's one thing to think the data looks suspiciously good, and another to hold up publication and ask the submitter to go collect more data. Especially since anything like that will likely be taken to an additional reviewer, and they'll likely have to think the same thing. More likely it gets passed though with reservations, and they wait for others to try and replicate the experiment and confirm their suspicions. Whether or not that happens though...

The whole "I don't think you're innocent, but I can't prove you're guilty" thing, basically.
I guess. Still, after years of this you'd expect people to catch on. But I guess it's different people who reviewed his work every time.
It took the 3 students who caught him months to gather the needed evidence even though they knew he was cheating all that time (and hurting people who used his data).
____________________________
Theophany wrote:
YOU'RE AN ELITIST @#%^ AETHIEN, NO WONDER YOU HAVE NO FRIENDS AND PEOPLE HATE YOU.
someproteinguy wrote:
Aethien you take more terrible pictures than a Japanese tourist.
Astarin wrote:
One day, Maz, you'll learn not to click on anything Aeth links.
#55 Mar 12 2013 at 5:28 PM Rating: Default
Encyclopedia
******
31,372 posts
someproteinguy wrote:
The 73 globally distributed temperature records used in our analysis are based on a variety of paleotemperature proxies and have sampling resolutions ranging from 20 to 500 years, with a median resolution of 120 years (5).


Criminy wrote:
The catch is the rate at which temperatures have went up in the past century has been equivalent to what normally takes four millennium.


Except that the median time frame used to create a data point used in the long term temperature records is longer than the time frame we're comparing to. Comparing those directly is like charting the speed of your car over the last 5,000 miles by averaging 100 mile periods, then comparing that to the speed changes of your car over the last minute. It's just bad science. Period.

Quote:
Amusingly this dramatic speed at which the temperature has increased begins right around the time of the second industrial revolution. Correlation != causation, I know. Just an interesting tid bit of information.


Sure. But we can also say that the dramatic speed at which temperature has increased begins right around the time we actually began measuring such things and comparing them to measurements of past temperatures using a completely different methodology. We cannot possibly know how much temperature variation exists within any given temperature data point in those historical records. We only know that for a given century or two, the average temperature was X. But it could easily have varied from X-5 to x+5 (or more) during that time period and the average would be the same. Simply graphing those averaged data points with a non-averaged running tally just over the last 70 years or so is not giving you a remotely accurate picture.


More correctly, it's giving you a very deliberately constructed false picture. Which is the whole point.
____________________________
King Nobby wrote:
More words please
#56 Mar 12 2013 at 7:11 PM Rating: Good
******
21,717 posts
gbaji wrote:
Sure. But we can also say that the dramatic speed at which temperature has increased begins right around the time we actually began measuring such things and comparing them to measurements of past temperatures using a completely different methodology. We cannot possibly know how much temperature variation exists within any given temperature data point in those historical records. We only know that for a given century or two, the average temperature was X.


Why do you go to such great lengths of written word when a simple "I categorically deny any understanding of the subject whatsoever" would suffice?
____________________________
R.I.P. Jessica M. 5/3/2010
This post brought to you by Carl's Jr.
gbaji wrote:
You guys keep tossing facts out there like they mean something.


#57 Mar 12 2013 at 7:36 PM Rating: Default
Encyclopedia
******
31,372 posts
BrownDuck wrote:
gbaji wrote:
Sure. But we can also say that the dramatic speed at which temperature has increased begins right around the time we actually began measuring such things and comparing them to measurements of past temperatures using a completely different methodology. We cannot possibly know how much temperature variation exists within any given temperature data point in those historical records. We only know that for a given century or two, the average temperature was X.


Why do you go to such great lengths of written word when a simple "I categorically deny any understanding of the subject whatsoever" would suffice?


You don't need to know anything about climatology at all to recognize poor data comparison methods. The methodology being used would be wrong no matter what subject was involved, or what the data represented.
____________________________
King Nobby wrote:
More words please
#58 Mar 12 2013 at 9:26 PM Rating: Excellent
Avatar
******
29,806 posts
Smasharoo wrote:

I'll pass. See, the way science works is that I don't have to demonstrate proof of the existing overwhelming peer reviewed consensus data model. Nor do I have to try to teach a basic statistics course. If you're going to make claims like "no one can know...blank" you have the burden to demonstrate why that's so. Hopefully beyond "that sounds hard".

I've already given you a few examples of areas relied upon by that particular graph which are currently in dispute by climate scientists as to the margin of error for early timeframe analasys. If X+Y=Z, but X+Y are potentially different values than what they were plotted on Graph point Z, that potential variance is called a "margin of error" Statistically, the margin of error they are showing for the 10,000 year old values in particular, is utter crap, because the individual variables themselves presently have statistical uncertanties even within their own climate change scientist community that they are flat out ignoring. They picked a set of assumptions, and made a graph. They didn't include any of the uncertanty factors, and from what I have seen they didn't even use the assumed Median values in their uncertanty calculations. They picked data, made a graph to show a particular viewpoint without giving it a sufficient margin for error, and sent it to press. Thats ignoring the whole quality of sample issue from that far back entirely.

Smasharoo wrote:

It should be obvious that while determining that the average global temperature for March 6, 25000 BCE was 24 C is extraordinarily difficult, that determining the average global temperature between 15000 BCE and 5000 BCE is going to be substantially easier. Accuracy v Precision and all that.

Sure. but only assuming a 0.4 degree +/- margin of error for that data the way they did on the graph based on the underlaying uncertanty and currently disputed variables of the model is being scientifically dishonest if nothing else for the older data they added. The margin of error should quite obviously widen considerably on anything older than 2,000 years as it approaches 10,000 years and our measurements and observable datasets become much more uncertain due to the effects of time and variables we have no historical observational data for. Especially when there is not consensus on early model timeframe climate condition variables which were used to create that model. The basic data and methodology are not at issue here. It's the accuracy of the older datasets, and the currently contradictory theories on various important factors as mentioned previously in the thread that is the problem. Their error bars are crap on the left side of the graph. You seem to think otherwise but provide no evidence to the contrary.

Smasharoo wrote:

Your assertion that it's just too hard reflects a Victorian understanding of analysis at best. Techniques to establish variance and confidence intervals have existed for quite a while now.

It's not "too hard" It's "impossible" to create that particular graph with that narrow a margian of error without ignoring valid, though less popular (note I don't say "correct") studies that show a greater range of uncertanty on the oldest data sets on variables used to make their graph model. Our resolution of data accuracy degrades significantly the further back you go, and our uncertanties rise. There is a degrading confidence interval on the earliest data sets and a larger margin for error that is not reflected on their graph, which has been my point all along.

Would some diagrams with large arrows on them pointing at simple concepts help? Maybe I could toss in an animated dancing frog or in the margins or something if that would make this easier for you to understand.
____________________________
Arch Duke Kaolian Drachensborn, lvl 95 Ranger, Unrest Server
Tech support forum | FAQ (Support) | Mobile Zam: http://m.zam.com (Premium only)
Forum Rules
#59 Mar 13 2013 at 5:09 AM Rating: Decent
Lunatic
******
29,167 posts
Would some diagrams with large arrows on them pointing at simple concepts help? Maybe I could toss in an animated dancing frog or in the margins or something if that would make this easier for you to understand.

Reading your post and feeling like you had any grasp of statistics would help. It seems like you're using the "common sense" method statistical analysis, which isn't know to work very well. Particularly, you don't seem to understand margin of error *at all*.

Honest question: You've never taken a formal statistical analysis course, have you?
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#60 Mar 13 2013 at 5:18 AM Rating: Decent
Lunatic
******
29,167 posts
They picked data, made a graph to show a particular viewpoint without giving it a sufficient margin for error, and sent it to press.

I don't see an evidence for this. Some, anything, could possibly change me point of view. Thusfar, it seems like you want that taken on faith, essentially, or with arguments from "there's no way you can know this".
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#61 Mar 13 2013 at 7:30 AM Rating: Good
******
43,191 posts
His Excellency Aethien wrote:
But I guess it's different people who reviewed his work every time.
I'd imagine that was necessary.
____________________________
George Carlin wrote:
I think it’s the duty of the comedian to find out where the line is drawn and cross it deliberately.
#62 Mar 13 2013 at 8:05 AM Rating: Excellent
Avatar
******
29,806 posts
Smasharoo wrote:


Reading your post and feeling like you had any grasp of statistics would help. It seems like you're using the "common sense" method statistical analysis, which isn't know to work very well. Particularly, you don't seem to understand margin of error *at all*.

Honest question: You've never taken a formal statistical analysis course, have you?


I can sling chi-squares and probability curves around just fine, and I make a Mean Median ala Mode. My Coinfidence Intervals regarding my ability to interperet and generate stochastic statistical analysis models and assorted regression analysis curves remains high. I'm starting to wonder if you failed your course though. Perhaps if you tried presenting an argument other than "I don't think you understand this so you're wrong" might be in order? Because I really do. Honest.

If you come up with a decent argument, let me know. Thanks!
____________________________
Arch Duke Kaolian Drachensborn, lvl 95 Ranger, Unrest Server
Tech support forum | FAQ (Support) | Mobile Zam: http://m.zam.com (Premium only)
Forum Rules
#63 Mar 13 2013 at 8:11 AM Rating: Decent
Lunatic
******
29,167 posts
I can sling chi-squares and probability curves around just fine, and I make a Mean Median ala Mode. My Coinfidence Intervals regarding my ability to interperet and generate stochastic statistical analysis models and assorted regression analysis curves remains high. I'm starting to wonder if you failed your course though. Perhaps if you tried presenting an argument other than "I don't think you understand this so you're wrong" might be in order? Because I really do. Honest.

If you come up with a decent argument, let me know. Thanks!


Wikipedia isn't a statistics course, just to be clear.

Any actual evidence or factual basis for anything yet, or are you sticking with "OMG, this would be tooo hardz..."

Perhaps if you tried presenting an argument other than "I don't think you understand this so you're wrong" might be in order?

Not an argument, just an observation. It's pretty inarguable that you don't understand it. Sorry :(

Edited, Mar 13th 2013 10:12am by Smasharoo
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#64 Mar 13 2013 at 8:35 AM Rating: Excellent
Avatar
******
29,806 posts
Smasharoo wrote:

I don't see an evidence for this. Some, anything, could possibly change me point of view. Thusfar, it seems like you want that taken on faith, essentially, or with arguments from "there's no way you can know this".


Well Lets take a look at what they wrote

teh study wrote:
The 73 globally distributed temperature records used in our analysis are based on a variety of paleotemperature proxies and have sampling resolutions ranging from 20 to 500 years, with a median resolution of 120 years (5). We account for chronologic and proxy calibration uncertainties with a Monte Carlo–based randomization scheme (6). Our data set exhibits several important strengths, as well as limitations, as compared to global and hemispheric reconstructions of the past 1500 years (2, 3, 7, 8). For example, whereas reconstructions of the past millennium rapidly lose data coverage with age, our coverage increases with age (Fig. 1, G and H). Published reconstructions of the past millennium are largely based on tree rings and may underestimate low-frequency (multicentury-to-millennial) variability because of uncertainty in detrending (9) [although progress is being made on this front (10)], whereas our lower-resolution records are well suited for reconstructing longer-term changes. Terrestrial records dominate reconstructions of the past millennium, whereas our stack is largely derived from marine archives (~80%). Unlike the reconstructions of the past millennium, our proxy data are converted quantitatively to temperature before stacking, using independent core-top or laboratory-culture calibrations with no post-hoc adjustments in variability.


So, they started with a random Monte Carlo set to trim off Uncertanties (or skew the data away from dissenting records), and then based their entire graph on 80% marine climate data. That's the whole "number of fossil marine critters per given geological strata layer" and :Carbon dissolved in lime stone per geological strata layer" data for the most part, amongst other metods, and pretty much ignoring any other sources. The problem with marine strata layers is they have a high degree of inherent uncertanty due to the errosive power of water. Terrestrial strata layers are much less prone to having thousands of years of muck suddenly ripped away by a particularily powerful storm surge across a large swath of your sample area for example. The further you go back using such records, particularily marine records, the more uncertanties you encounter as to the accuracy of the observable strata. They attempt to offset that by saying "well, we know that this critter that still exists dies off at this rate today in laboratory cultures, so they must have still died off and deposited at that rate under conditions found at that location 10,000 years ago, because obviously temperature, pressure, acidity, water clarity etc. factors for water bodies don't ever change significantly over time, not to mention genetic drift factors, etc., and using multiplecore samples from independant firms, all of which may or may not have been subject to significant alterations in topology over the last 10,000 years with no way for anyone at all to know. Their first paragraph basically flat out says "our data gets worse the further you go back, but it's ok because our method magically provides better information over long periods of time because we're awsome and you don't need to see our actual math work" They plotted what they think the line should be after "adjusting" their sample set, mixing progressivly lower resolution data with modern higher resolution data with no adjustment for increased uncertanties. The fact that they aren't showing a greater temperature uncertanty for the oldest data and won't release their mathmatical model tells me something isn't right here.
____________________________
Arch Duke Kaolian Drachensborn, lvl 95 Ranger, Unrest Server
Tech support forum | FAQ (Support) | Mobile Zam: http://m.zam.com (Premium only)
Forum Rules
#65 Mar 13 2013 at 8:40 AM Rating: Decent
Lunatic
******
29,167 posts
So, they started with a random Monte Carlo set to trim off Uncertanties (or skew the data away from dissenting records), and then based their entire graph on 80% marine climate data. That's the whole "number of fossil marine critters per given geological strata layer" and :Carbon dissolved in lime stone per geological strata layer" data for the most part, amongst other metods, and pretty much ignoring any other sources.

Like what? I appreciate your response. I *still* don't see a counter-factual argument.

Their first paragraph basically flat out says "our data gets worse the further you go back, but it's ok because our method magically provides better information over long periods of time because we're awsome and you don't need to see our actual math work" They plotted what they think the line should be after "adjusting" their sample set, mixing progressivly lower resolution data with modern higher resolution data with no adjustment for increased uncertanties. The fact that they aren't showing a greater temperature uncertanty for the oldest data and won't release their mathmatical model tells me something isn't right here.

It tells me they want to publish more. Proprietary models aren't unusual or particularly suspicious.
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#66 Mar 13 2013 at 9:49 AM Rating: Excellent
Meat Popsicle
Avatar
*****
11,446 posts
Smasharoo wrote:
Proprietary models aren't unusual or particularly suspicious.

They are really annoying though. Black box software in research makes me want to hurt things.
____________________________
That monster in the mirror, he just might be you. -Grover
#67 Mar 13 2013 at 9:55 AM Rating: Decent
Lunatic
******
29,167 posts
They are really annoying though. Black box software in research makes me want to hurt things.

Well with climate scientists, it's where they keep all the lies they tell for the extremely powerful environmentalism conglomerates who reward them with giant 7 figure bonuses.
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#68 Mar 13 2013 at 10:12 AM Rating: Excellent
Meat Popsicle
Avatar
*****
11,446 posts
Lies are bad. Smiley: disappointed

Also this jumped to mind.
____________________________
That monster in the mirror, he just might be you. -Grover
#69 Mar 13 2013 at 12:58 PM Rating: Good
Skelly Poker Since 2008
*****
15,541 posts
I don't remember much geology. But, I do know that stuff erodes from the land and deposited in the sea. Terrestrial strata, when it even gets preserved is fraught with non-conformity.

____________________________
Alma wrote:
Post and be happy!
#70 Mar 13 2013 at 1:02 PM Rating: Excellent
Lunatic
******
29,167 posts
I don't remember much geology

How about science book or the French you took?
____________________________
Disclaimer:

To make a long story short, I don't take any responsibility for anything I post here. It's not news, it's not truth, it's not serious. It's parody. It's satire. It's bitter. It's angsty. Your mother's a whore. You like to jack off dogs. That's right, you heard me. You like to grab that dog by the bone and rub it like a ski pole. Your dad? Gay. Your priest? Straight. @#%^ off and let me post. It's not true, it's all in good fun. Now go away.

#71 Mar 13 2013 at 1:02 PM Rating: Excellent
Liberal Conspiracy
*******
TILT
Smasharoo wrote:
I don't remember much geology

How about science book or the French you took?

I was just humming that to myself and decided not to post about it Smiley: mad
____________________________
Belkira wrote:
Wow. Regular ol' Joph fan club in here.
1 2 Next »
Reply To Thread

Colors Smileys Quote OriginalQuote Checked Help

 

Recent Visitors: 49 All times are in CDT
Aethien, Allegory, BeanX, Elinda, Poldaran, RedPhoenixxx, Samira, Anonymous Guests (42)