The Economist tells it like it is
This led to a Daily Mail headline reading: "Climategate U-turn as scientist at centre of row admits: There has been no global warming since 1995."
Since I've advocated a more explicit use of the word "lie", I'll go ahead and follow my own advice: that Daily Mail headline is a lie. Phil Jones did not say there had been no global warming since 1995; he said the opposite. He said the world had been warming at 0.12°C per decade since 1995. However, over that time frame, he could not quite rule out at the traditional 95% confidence level that the warming since 1995 had not been a random fluke.
Anyone who has even a passing high-school familiarity with statistics should understand the difference between these two statements. At a longer time interval, say 30 or 50 or 100 years, Mr Jones could obviously demonstrate that global warming is a statistically significant trend. In the interview he stated that the warming since 1975 is statistically significant. Everyone, even climate-change sceptics, agrees that the earth has experienced a warming trend since the late 19th century. But if you take any short sample out of that trend (say, 1930-45 or 1960-75), you might not be able to guarantee that the particular warming observed in those years was not a statistical fluke. This is a simple truth about statistics: if you measure just ten children, the relationship between age and height might be a fluke. But obviously the fact remains that older children tend to be taller than younger ones, and if you measure 100 of them, you'll find the relationship quite statistically significant indeed.
What's truly infuriating about this episode of journalistic malpractice is that, once again, it illustrates the reasons why the East Anglia scientists adopted an adversarial attitude towards information management with regard to outsiders and the media. They were afraid that any data they allowed to be characterised by non-climate scientists would be vulnerable to propagandistic distortion. And they were right.
Meanwhile here in Australia look what IPA executive director John Roskam said on national TV:
You had, the other day, one of the leading climate change scientists in the world say the world hasn't warmed since 1995.
That is a lie.
- Log in to post comments
It would surely be too difficult to actually link to the original BBC source
http://news.bbc.co.uk/2/hi/8511670.stm
and mention what Phil Jones actually said, right?
BBC - Do you agree that from 1995 to the present there has been no statistically-significant global warming?
Phil Jones - Yes, but only just. I also calculated the trend for the period 1995 to 2009. This trend (0.12C per decade) is positive, but not significant at the 95% significance level. The positive trend is quite close to the significance level. Achieving statistical significance in scientific terms is much more likely for longer periods, and much less likely for shorter periods.
Thanks Lubos! That quote very clearly proves that the Daily Mail lied. Or at least that they don't understand statistics, and that they think that if one doesn't understand a term one can just leave it out.
Lubos approves the lie.
i am not surprised.
Lubos - to be fair, everyone and all their dogs will have seen the Phil Jones interview by now. Lars Karlsson is right, though, that it proves the linked article correct.
Put it another way, which of the following statements is true:
a) There has been a rise in temperature between 1995 and the present.
b) There has been no change in temperature between 1995 and the present.
c) There has been a decline in temperature between 1995 and the present.
If you cannot answer either a, b, or c with any degree of confidence, then the wrong question is being asked.
Wow, those are fighting words. Excellent, bravo, here, here! Thank God someone out there with influence and the public ear have gotten it right for once.
Kudos to The Economist. We need to support this journalist. The Economist have actually done a really good job reporting on AGW and the CRU hack since last November-- they have been balanced and clearly done their research.
Can't Jones sue the Mail for libel?
PS: The embedded video is brilliant.
Lubos, what's your point exactly, could you be more explicit?
Meanwhile, Lubos seems [fond](http://motls.blogspot.com/2006/12/2006-probably-coldest-year-in-last.ht…) of [making sense](http://motls.blogspot.com/2006/12/2006-bad-year-for-climate-fearmongers…) out of [noise](http://motls.blogspot.com/2008/07/uah-june-2008-still-negative-anomaly…), yet seems all to ready to [vacuously opine on the irrelevance of statistically significant trends](http://motls.blogspot.com/2006/12/2006-probably-coldest-year-in-last.ht…).
This is a totally irrelevant observation, but if you google(.com) search 'Lubos Motl' with auto complete the fourth suggestion is interesting. I have no opinion on this matter either way.
Daily Mail tries to make it up and overshoots :)
http://www.dailymail.co.uk/sciencetech/article-1156023/Doomsday-How-4C-…
Speaking of distortion and all that, Roger Pielke Jr. translates a bit of a German interview with Mojib Latif where Latif is talking about fraud or something and it is implied in Pielke's post that Latif is accusing the IPPC of fraud; but Pielke does not clarify exactly who Latif thinks is doing the misrepresenting in "misrepresentation of the science of disasters and climate change".
http://rogerpielkejr.blogspot.com/2010/02/mojib-latif-on-zdf-fraud-to-p…
Unfortunately, it isn't just ignorant or tendentious journalists who confuse "not statistically significant" with "not true." Many peer reviewed scientific papers report "no association" when the p value is .06 or higher; whereas the observed association might be of a magnitude which is quite significant clinically or otherwise. This is a common dodge by which industry evades safety regulation, for example.
We really need to get over this habit of dichotomizing "statistical significance" according to an arbitrary p value; instead we should just give the confidence intervals and decide how worried or pleased we ought to be about an observed association based on the particulars of the situation. If it's important to be more certain, go out and get a bigger sample -- or in this case wait a few years -- rather than announcing a negative finding.
There's one subtle point that I think the Economist article misses:
"But if you take any short sample out of that trend (say, 1930-45 or 1960-75), you might not be able to guarantee that the particular warming observed in those years was not a statistical fluke. "
This is essentially true (when using 95% for "statistical fluke") if we only have one warming indicator. Jones was specifically referring to the HadCrut data. We also have GISS and NCDC, which I believe hit the 95% confidence for that interval. We also have independent satellite measurements, which probably have the same confidence level as HadCrut. We also have a significant net depletion in global glaciers. We also have ocean heat content (at all levels) rising during this period. While HadCrut alone might have around a 5-10% of having an actual trend of 0 per decade (or 0.24 per decade) during this period. What are the odds of all of the data being wrong in one direction? It seems when combining all independent measures, it would be virtually certain of warming since 1995.
The last paragraph nails it:
"What's truly infuriating about this episode of journalistic malpractice is that, once again, it illustrates the reasons why the East Anglia scientists adopted an adversarial attitude towards information management with regard to outsiders and the media. They were afraid that any data they allowed to be characterised by non-climate scientists would be vulnerable to propagandistic distortion. And they were right."
Actually, statistically significant or not, warming unequivocally did occur over the period in question. In plain English, the earth was warmer at the end than it was at the beginning. The concept that applies here isn't really "statistical significance" at all -- it's just how often such a difference has been observed over similar periods in the recent past, or to put it another way, how unusual is this?
Whether or not it's "statistically significant" is irrelevant to the question of whether or not it will continue; that all depends on why it happened. It's magnitude or how unusual it is might be evidence that applies to the question of "why" but as a mere isolated fact it doesn't tell us anything.
@cervantes
I'd shy away from that thinking, precisely because it is imprecise - it reminds me of the anti-AGW brigade claiming there had been cooling because 2008 was cooler than 1998 (in the HadCRU set).
Two datapoints doth not a trend make, and there is too much noise to say that warming equivocally did occur - hence the need for statistical significance.
I am not sure whether any of you understands basic science terminology, or whether you want to understand it, but if an observation extracted from the data is "not statistically significant", then it means that it's plausible or likely that it's fundamentally equal to zero and the observation was caused just by noise.
That's the case of the warming since 1995, too. From the viewpoint of statistics, it doesn't exist. Building on a statistically insignificant trend would be exactly as illegitimate as building on a non-existent trend. So the statement that Jones said that there was no global warming since 1995 is not literally accurate, but it is surely more correct than the statement that he didn't say that there was no global warming since 1995.
Funny, those in denial at the same time claim the CRU scientists 'fudged the warming', while also gladly accepting the fact that the warming trend in the CRU data for 1995-2009 is not statistically significant at the 95% level of confidence. WTF!? I wish that they would make up their damn minds.
Also, if they toss CRU, then they are left with the warmer NASA GISS. Oh, and are we to believe that the Japanese Met Agency is also in on the "conspiracy"?
http://ds.data.jma.go.jp/tcc/tcc/products/gwp/gwp.html
Note even Inhofe and his ilk could fabricate such a conspiracy.
Lubos #12 - thats funny, I could have sworn that Tim quoted an article which said
"you might not be able to guarantee that the particular warming observed in those years was not a statistical fluke."
which is effectively the same as your explanation regarding noise. Perhaps you would like to go and learn some science, or better, some reading comprehension?
Why don't you enlighten everyone on the traditional p-value at which it becomes "plausible" that the observation extracted from the data is "fundamentally equal to zero"? Seeing as we don't understand basic scientific terminology.
Once you've done that, maybe you can explain the difference between "plausible", and "likely"?
And maybe once you've done that, you can explain to us why it is appropriate to assert in a headline that a trend is equal or below to zero when the probability of that being the case is below 50%? Or 60%? Or 70%
And finally, you could point to the assertion from any of us that any likely warming between the period 1995 to the present is either statistically significant or meaningful. Because otherwise, your criticism is unfounded.
Lubos,
Jones says that the positive trend is quite close to a 95% significance level. This means that there is only slightly more than a 5% probability that such a trend occurs out of chance.
This is certainly not the same as saying that there is no global warming. I repeat, Jones did not say: "There has been no global warming since 1995".
Have any of you noticed the Institute of Physics response to 'Climategate'?
http://www.publications.parliament.uk/pa/cm200910/cmselect/cmsctech/mem…
Dave Andrews - No. And a document making allegations such as theirs should provide evidence. It provided none.
Disappointing, and raises questions over how they managed to arrive at this judgement.
Bud, please don't feed the trolls.
"I am not sure whether any of you understands basic science terminology, or whether you want to understand it, but if an observation extracted from the data is "not statistically significant", then it means that it's plausible or likely that it's fundamentally equal to zero and the observation was caused just by noise."
No, that is not what it means. I certainly understand basic science terminology, I have a PhD which required me to study 5 semesters of statistics. The term "statistical significance" is properly applied to random sample data taken from a population with unknown parameters. It tests the probability that the sample does not resemble the population from which it is drawn. It is not about "noise" but rather about sampling error. In order for that to be meaningful you need to be testing some sort of association or comparison.
It is generally said that an association or a difference is not "statistically significant" if that probably is greater than .05. However, such a difference is very probably real, it is not at all likely that it is the result of random sampling error. It's just that we have a priori set a high standard of what we need to be convinced.
Computing the significance of changes over time is actually philosophically rather dodgy, because time series observations are probably not properly thought of as samples from a population. Years are ordered, they are not analogous to a bag full of black and white marbles. What we really have here with the earth is what is called a "within subjects design," with an N of 1 -- one earth. It is technically not possible to compute a p value for the difference between time 1 and time 2 in the standard sense.
Mapleleaf - touche! :)
I think troll-feeding is a mood thing. Irritable mood = easily provoked. Mea Culpa.
Bud, no worries. I understand. And I have made that error too many times myself. Ugh.
Bud,
So,according to you, the UK Institute of Physics doesn't know how to respond to a specific inquiry?
Quite breathtaking!
Dang,
'enquiry'
The trolls will have nothing to say on the blatant lies, of which they approve, but seek to move onto newer mendacities.
Dave Andrews, did you even read the submission or did you just cut and paste it from climatefraudit?
From the first sentence in the submission:
Sounds like a real hot bed of scientific endeavour.
I sympathise with Bud. The trolls post up random stuff which they imply is some smoking-gun evidence of wrongdoing, yet when you read it, it generally says nothing of the sort and has been twisted totally out of context.
It's very frustrating, and makes you want to let loose with both barrells.
Most irritating of all - and a work colleague did this to me recently - is when they repeat stuff they obviously don't even understand. It's not the "understanding" as such (some of it is difficult to grasp without further reading and learning), but the fact they never bothered to try to understand it in the first place.
AGW is a sophisticated, yet inconclusive theory about our air con, within subjects design N of 1.
Some characters connected to the IPCC oversimplified the message, the thermostat was on the blink and we needed to fix it fast.
The masses, being extremely vulnerable to propaganda, accepted what was told them. I hear angular momentum is coming back in favor as the answer to everything climatic.
The trolls seem to be more frequent, when something damning comes out against the liars, sorry, I mean, denialists, and that is often nowadays.
Two of the three trolls here have said nothing about The Economist article, and one tried to but had to mangled/distort. The fact that the Economist has called the Daily Mail a liar is significant.
Ignore the trolls, no matter how much they taunt you, b/c once you engage them, then they have a soap box and use you for leverage. I know, I have been baited more times than I wish to recall.
Andrew Bolt made the [very same misrepresentation](http://andrewboltliesdeceptionsonagw.wordpress.com/2010/02/14/not-unusu…).
What a surprise.
El G,
Why do you bother to regurgitate the mindless drivel that you do?
It's clear you have no concept of why the theory is sound given all the available evidence.
The trolls are kept in thrall by "drama".
Nov/Dec it was the fizzled out dud of 'climategate'.
Dec/Jan it was the $50 million dollar "bounty" by that washed-up alleged ex-CIA sad-sack, that disappeared into the ether with nary a ripple.
Jan/Feb it was Glaciergate and Pachuari Porngate.
Feb/March it'll be Inhofe's virtual "indictments", and some other nonsense.
It's no wonder dapielke's choose TV frontmen to present their "case". Infotainment for a TV-scheduled moronosphere is the arena they're engaging in.
My only hope is that as the lies of the denialism boosters are increasingly called for what they are, even the more dense of the trolls like eg and DA will too embarrassed to dribble their inanities.
Nearly everyone that's followed or dealt with Lubos is not trying to trade insults with him by saying he's unstable. Living proof that pseudonymity is not a prerequisite to incivility. However, I have to say that the post today was awesome, for LM, and while a little strange - as the next commenter said, it does make it clear that the headline was a lie, it obeyed all the norms of reasonable blog commenting.
A few observations:
1.Does this same sense of outrage by alarmists apply when the headlines misrepresent in a direction they agree with? It should, bad media is bad media, yet we only see one side of it on this site and elsewhere. For example, headlines such as "U.N. Panel says Global Warming Man Made" would be a "lie", would it not? I recall many such headlines, however, I don't seem to recall any outrage from the likes of those in the "Jakerman Zone".
http://www.reuters.com/article/idUSL0160494020070202
2.In the video, we see a brilliant example of staking wave distances as a simple minded way to explain rising confidence levels with larger sample sizes and time scales. Is the tide coming in or going out?
After many samples, we can all say, with confidence, that the tide is coming in. The leap here, is when it is assumed the tide is coming in solely due to man, and that we can and must change the tides, lest they keep coming in, never to go out again...
It's probably just a coincidence, but temperatures seem to go through the roof in the early 1980's - just about the time they began to get rid of all those pesky stations.
Maybe I'll take this up at proper length in my blog. In the mean time, I'll observe that statistical tests can give quite different answers depending on what prior claims about reality you start with.
First, of course, Jones should not have answered regarding 15 year 'trend'. You need 20-30 years to define a climate trend. The trend over the last 20 (or 30, or 50, or 100) years is positive, and statistically significant versus the null hypothesis of the trend being zero.
That's the second part of the story. In looking at the trend, is 'no trend' a sensible hypothesis to accept as your null? No. The 20 and 30 (and rest long enough to be climatically interesting) of trends ending in 1994 were all positive, and statistically significantly so.
The proper null hypothesis to be looking at -- given that by 1995 there was extensive observation of a climate (20-30 year) warming trend is that the trend was the 0.15/decade (or whatever it was, I don't recall offhand) with error bars of thus and so.
If you start from the null hypothesis that warming has continued as observed in the previous 20, 30, ... years, you will be unable to reject it. Rather than concluding there is no trend, you conclude that prior trend has continued.
Or, again, as I prefer, only bother yourself about trends which are long enough to be physically meaningful.
El Gords @ 37 says:
"It's probably just a coincidence, but temperatures seem to go through the roof in the early 1980's - just about the time they began to get rid of all those pesky stations."
See this post by Tamino about the effect of the dropouts: .
He's pretty passionate about proper analysis of the data, so much so that he's planning to publish an expansion of his work in this post:
Grumbine,
"If you start from the null hypothesis that warming has continued as observed in the previous 20, 30, ... years, you will be unable to reject it. Rather than concluding there is no trend, you conclude that prior trend has continued."
But that's not a null hypothesis.
Jones didn't conclude that there's no trend, he concluded that there's a 0.12C/dec trend. He also said there's a small chance it might be an artefact of the noise - but that's from a purely statistical perspective. That we know that there was a previous trend, and that there's a physical reason the trend should be upwards, somewhat negates the fact that it's not statistically significant... if you see what I mean.
Lubos Motl:
If nothing else, at least we know Lubos thinks 6% probability means likely.
Thax Stu, there's a lot of competing definitions of reality.
While the Doltoids sweat the small stuff they could also check the whoppers:
http://climateaudit.org/2010/02/26/mcintyre-submission-with-figures/
It's also important to remember that the year 1995 was chosen precisely because it's the furthest back you can go without having enough data to find statistical significance.
That is, whoever proposed this question is a conscious fraud as is anyone with statistical training (presumably including Lubos) who repeats it.
But the rest of the "sceptics" aren't much better. These are people who presume to "make up their own minds" on complex scientific issues, but can't be bothered informing themsleves about basic statistics.
Thanks to Mapleleaf and Mike for their understanding. I'm not going to lie, I have not bothered with Dave Andrews since coming into contact nearly a year ago with his extraordinary... well, use your own imaginations as to what I was dealing with. But since in a moment of weakness I rose to his bait, I am honour bound to continue.
Dave Andrews, I'd love some of what you're smoking, but in the absence of it I'm afraid I have no idea how you were making the assertion you made in 24. I'm sure 'they' know very well how to respond to a specific 'inquiry'. However, as to this particular situation, they provide no evidence, have no wider report, and are completely at odds with all reasonable opinion on the matter. So, why should we give a toss?
Betula,
You keep mentioning the 'Jakerman Zone' without providng explanation or links. I can help [direct people to that which plagues you](http://scienceblogs.com/deltoid/2010/02/open_thread_39.php#comment-2277…).
To be fair, we don't. Lubos still hasn't responded to my questions from #16, so his advanced knowledge of significance is as yet unrevealed to us.
"If nothing else, at least we know Lubos thinks 6% probability means likely."
Perhaps our contrarian friend interprets "quite close" to the 95% significance level as being less than 50%. Good stuff.
But again, keep in mind we're only talking about one dataset. With all the other indicators (many independent), it's extraordinarily unlikely that all would be off in the same direction.
[Off topic comment deleted -- Tim]
Stu at #40, sure RG's is a good null hypothesis. It says "nothing's changed" from what's previously been going on. You and those who put those questions to Jones would make a more rash assumption for your null. You'd be supposing something that, given the state of scientific knowledge, would be less likely. To start with an hypothesis that there might be no warming trend for the last 15 years of a time series that shows such a trend over every longer period (from earlier start dates) would be to start with more assumptions - less parsimoniously - than Robert has in his comment.
If you don't know Robert already you might like to check out his publication record or something, but please let's not go into Bayesian/frequentist or other esoteric arguments here in this thread about the Economist's good work.
Lubos your comment at lucky #13 is ridiculous, but mind-boggling coming from a physicist. Even if you removed the revealed cluelessness about statistics from it it'd still have been a stupid comment to write here - what is the matter with you?
There is another aspect of the 'Climategate' emails that is often overlooked. These emails are personal emails between collaborating scientists. They are not part of the peer-reviewed IPCC reports which are intended for publication and use in policy development. There are many steps between the 'working' CRU emails and the published IPCC reports, but is there any evidence that any of the 'questionable' wording or frustrations expressed in the emails actually influenced the content of the peer-reviewed IPCC Reports?
The climate change 'deniers' need to establish such a connection, (needless to say, with a high degree of confidence), before they start slagging off dedicated people like Phil Jones.
I see a lot of denialism about AGW in the world today. I think this is part of the evolution of how we will deal with it. First we simply ignored it, hoping it would go away. It didn't. Now we deny it, hoping it will go away. It won't. Next we'll get all angry about it and blame each other for it. It still won't go away. Then there's the bargaining stage - hey if you use a cloth grocery bag in the year 2030, will the problem go away? No it won't.
Finally, we'll start working on the problem. But it will be too little, too late. The earth won't be destroyed, we won't become extinct. But we could be decimated quite a bit, and surely the biodiversity of the planet will take a major hit.
There is such a thing as reality separate from our beliefs about it. You can't "believe away" a truth, nor can your beliefs create reality.
So far, on average, the earth is warming, and a little more rapidly every decade. There's different lines of evidence - satellite data, ice cores, weather stations, etc.
The main cause is increased Co2 and, if the permafrost melts, a big sudden burst of methane, a far more powerful GHG than Co2.
You can argue about it. You can even win the argument. You an spend billions of dollars winning the argument.
But the victory will be hollow, because the reality itself will not have changed. A strong belief in something that is false can never be strong enough to make it true.
Frankis:
"Stu at #40, sure RG's is a good null hypothesis. It says "nothing's changed" from what's previously been going on. You and those who put those questions to Jones would make a more rash assumption for your null. You'd be supposing something that, given the state of scientific knowledge, would be less likely"
What's the definition of a null hypothesis? In hypothesis testing, the hypothesis we wish to falsify on the basis of the data. The null hypothesis is typically that something is not present, that there is no effect, or that there is no difference between treatment and control.
So we expect the data to falsify the null hypothesis. OTOH, RG's hypothesis is someting we'd expect the data to confirm. It's not a null hypothesis. I don't think this is anything controversial (actually I feel like we're nitpicking).
Did you mean to lump me in with 'those who put those questions to Jones'?
People are a touch sensitive these days, understandably .... of course Stu I wasn't "lumping you in with" those d*ckheads in any way, except to say that you are making a less suitable choice of null hypothesis than Robert has proposed. We all know the definitions of the statistical terms, I presume you also are aware of ongoing excitement between Bayesians and frequentists, and I hope too that if not earlier then by now you've found out who Robert Grumbine is.
Let's not nitpick (there wouldn't be a delusionalist reading this who understood what you were talking about anyway), rather let's applaud with Tim the lucidity of somebody at The Economist and snicker as usual at what passes for "the talent" at Australia's IPA asylum. Seriously, I'm laughing at the thought of the IPA as I type this now; cheers!
It's also important to remember that the year 1995 was chosen precisely because it's the furthest back you can go without having enough data to find statistical significance.
Has anyone checked to see that if you start from more recent years, it stays below 95% confidence? In other words, is 1995 merely the most recent date for which the error bars straddle zero, or is it an isolated cherry-picked island?
Eyeballing it, the trend 1998-2008 does look stronger than 1997-2008.
I'm going to feed Betula the Troll here, because she did attempt to make a point. Betula, you ask how we respond when the media gets something wrong that favors our point of view. But the example you give doesn't support your claim:
headlines such as "U.N. Panel says Global Warming Man Made" would be a "lie", would it not?
The fact is, the IPCC *did* say that global warming is man made. So the headline is not a lie -- it's the absolute truth. The fact that you disagree with the statement "global warming is man made" does not mean that the statement "The IPCC says that global warming is man made." It would be just as correct for them to print the headline "Betula says that global warming is not man made", although I doubt anybody would bother.
"Yes, but.." is still yes. If Phil jones was in the dock, he would have to answer yes.
"Yes, but.." is still yes. If Phil jones was in the dock, he would have to answer yes.
No, that would be perjory. He is required to give the most truthful answer he can. You are attempting to cast a probabilistic issue in boolean terms. Here's an analogy that will help you understand: suppose that you witnessed a gray car fleeing a murder scene. The lawyer asks you, "What the car black or white?" You are not required to answer "black" or "white"; you are required to answer "a shade of gray". You could add that it was closer to black or closer to white, but your most accurate answer would be "a shade of gray". Forcing the witness to describe it as either black or white is a form of deceit. In the same way, forcing Mr. Jones to answer "yes" or "no" to a question whose answer is fundamentally statistical in nature is a form of deceit.
Erasmussimo... not that Betula needs or probably wants my stepping in, but he is male.
Lubos,
How about since 1994? Cherry-picking much?
Dear Deep Climate,
your point is somewhat shallow. In this thread, I discussed 1995 as opposed to 1994 because that's the year that BBC asked Phil Jones about, and for a good reason. 1995 is the earliest year when the statistical significance of the trend from that year to 2009 safely fails. Since 1994, you could get a technically significant trend. It would still not be a robust result because a small change of the beginning year would destroy the statistical significance. Only statistically significant results that persist when small changes are made are really worth talking about as evidence for anything.
Because there have always been low-frequency cycles and signals in the climate data, it's clear that over a sufficiently long interval, one will find "some" trends. But this result is interesting because it shows that 15 years of climate data - and that's a pretty long time (recall how many things have changed since 1995) - is just noise. So even if there were an underlying trend, it isn't urgent in any sense. Waiting 15 extra years will bring no statistically detectable problem.
BBC asked a very good and highly relevant question in a technically accurate fashion and Jones had to answer it - the answer was clearly "Yes". He was also allowed to add refinements. It's just the religious bigots like most of the readers above who dislike the question and the answer - because the answer, "Yes, there's been no statistically significant trend for 15 years", shows that even if there's anything such as "global warming", it's not really worth talking about. And you don't like this fact for the same reason why an Islamic bigot doesn't like the picture of Mohammed as just another obsessed warrior and/or terrorist: because it's an inconvenient truth.
Cheers
LM
Deep Climate @61 makes a good point. If the trend since 1995 (or 1996) is significant after 2010, will Lubos and the like all change their minds?
Shorter Luboš Motl:
The warming trend for 1994--2009 is statistically significant, but that doesn't count because you can make the significance go away by replacing real data with fake data.
Taking 15 years of data, we find there's no statistically detectable warming. This shows that the next 15 years of data will show no statistically detectable warming either, because we'll have to magically forget all the data we collected in the past 15 years.
Therefore, you should be quarantined and euthanized, which makes you all terrorists.
Stepanovic,
No, this is what Lubos actually [said](http://motls.blogspot.com/2010/02/sir-john-houghton-is-liar.html#commen…): "99% of what Lubos has ever said is irrelevant crap."
This quote may be inaccurate in its wording but certainly not in the content.
And I should add to #65: That is statistically significant.
By "safely", don't you mean "only just, by the slimmest whisker"?
The trend only fails the test if you take the arbitrary 95% level as your benchmark. It would pass a test at 91% or 92% or 93%.
Oh, and has anyone pointed out that the "failure" is really a result of aggregating the monthly data into annual averages, meaning you get only 15 data points instead of 180.
Getting into the 90s is pretty solid for only 15 data points IMHO.
You could take 5 year averages and have only 3 data points and an even slimmer chance of passing tests of statistical significance. I predict this is the next tactic of those suffering deniarreah.
Anyone want to fire up their spreadsheet and tell us what the significance of the trend derived from the monthly anomalies might be?
Lubos says: "It's just the religious bigots like most of the readers above who dislike the question and the answer - because the answer, "Yes, there's been no statistically significant trend for 15 years", shows that even if there's anything such as "global warming", it's not really worth talking about."
Questions: 1) do you really think it's constructive to refer to people as 'religious bigots'? If you don't think it's constructive, we'll have to assume you're trying to be disruptive.
I'm religious but I don't think I'm bigoted. Please tell me if you feel otherwise!
2) How short a time period would the trend have to be siginificant over in order to convince you that, even if there's anything such as "global warming", it really is worth talking about? After all, a period with lower than average annual variability would increase the significance even for the same trend. But it would mean bupkis anyway, because those of us interested in the science rather than the soundbites know that the magnitude and sign of the long term trend is what's important, and this quibbling over statistical significance of a small portion of the temperature record is merely a sideshow.
PS if you don't think it's worth talking about, feel free not to ;-)
"The Jakerman Zone" show has been moved @46 to a different station....Open thread 43.
Thank you for watching.
Erasmussimo @57...
Try to back up anything you just said with an example. Anything.
Shorter Betula:
I'm ignoring you, therefore you didn't say anything.
Does anyone else wonder whether Lubos has made up his qualifications?
Ian Forrester,
Did you go on to read past the first half sentence?
"It has a worldwide membership of over 36,000 and is a leading communicator of physics-related science to all audiences...Its publishing company,IOP Publishing, is a world leader in scientific publishing and the electronic dissemination of physics"
Given how often we are told 'its all about the physics' I really think you uoght to pay some attention to what IOP say.
#72 also Lubos seems to be, well, [abusive] (http://backreaction.blogspot.com/2007/08/lubo-motl.html)
#74 Oops didn't preview that link
Even shorter Lubos: * *hand wave* *
Shorter Dave Andrews:
If the IOP says it's a "world leader in scientific publishing", then it's obviously a world leader in scientific publishing.
Batula (thanks for correcting me on the gender, Bud) asks:
Erasmussimo @57...
Try to back up anything you just said with an example. Anything.
Perhaps you might rephrase your challenge. I was describing a specific instance, not a general principle. You're asking me to provide an example of a specific event. The event is its own example! I suspect that you intend something that makes sense, so please expand on your concept.
Lubos Motl:
Thank you for admitting that 1995 is a blatant cherry-pick.
He never gives up, does he? No doubt if we started from 1993, he'd come up with some other pathetic excuse why it's not REALLY statistically significant even though applying his current excuse wouldn't make a diffence. Then he'd make yet another excuse for 1992, and 1991, and 1990 and ad infinitum, every time getting more and more pathetic.
I'm already aware that 15 years is too short to measure climate with a reasonably high signal-to-noise ratio. This result doesn't tell us anything new at all.
Ah yes, the boiled frog excuse. The frog says: "It hasn't warmed significantly in the last 15 seconds therefore the next 15 seconds will bring no statistically detectable problem."
As in, highly cherry-picked question as Lubos helped point out above.
Au contraire, we have no objection to Jones' answer. Our objection is to the lie about Jones' answer in the Daily Mail. If you had paid attention to Tim's posting, you would have noticed the statement:
"that Daily Mail headline is a lie."
The problem is a lie told by someone else, not Jones.
What? Are you no longer saying that we are living "on a planet that has been cooling since 1998"? Please tell me you still do.
So he admitted that there has been NO SIGNIFICANT warming since 1995. If it has NOT been significant why are we handing over so much money and control to global bodies like the UN, EU, IMF. And why am I paying extra UK car tax because I drive a 4x4? The cost that is being imposed is significant, yet the warming is not. http://www.climatejournal.org/
I propose chaining the denialist trolls to a post at low tide in Portsmouth and waiting. It'll be some time before there is a statistically significant change in water depth but they'll eventually get the idea.
-bcw
Take a chill pill and calm down Climate Journal. Climate science isn't a conspiracy to screw you over.
Climate Journal writes:
So he admitted that there has been NO SIGNIFICANT warming since 1995
You greatly misunderstand what he actually said. We have indeed measured warming since 1995. We are only 94% certain that this measurement is not caused by random fluctuations in the data. We normally say that it must be 95% or greater to be "statistically significant", but that's an arbitrary line. The more truthful statement is that our confidence level is 94%.
Climate Gerbil said:
Why shouldn't you pay an extra car tax? Those 4*4s
a) contribute more to pollution than average vehicles
b) subtract more from road safety than average vehicles
c) people who can afford to operate cars, especially the expensive ones, are probably bettter placed to pay taxes than those who can't and are probably evading tax
It's an imperfect world. In one closer to the perfect than this one all of the costs to the commons of driving motor vehicles would be settled on those who operated them. In the meantime, pinging your 4*4 for extra tax would seem to me a reasonable start, even if I were one of those who thought AGW was a load of bollocks.
"So he admitted that there has been NO SIGNIFICANT warming since 1995."
No, he said that the confidence interval for the statistical significance of the warming was not narrow enough with such a small sample. An introductory course in statistics would help you a lot with this.
Without wishing to comment on your understanding of where UK government money goes, I can at least aid your understanding of the reason. No notice is taken of the trend from the period 1995-present because it is far too long to draw any meaningful conclusion from. Instead, people are worried about climate change because there has been a statistically significant increase in GMST between midnight and now.
Oh... wait.
"So he admitted that there has been NO SIGNIFICANT warming since 1995."
Get a clue, Climate journal.
If there was a 94% chance you'd be hit by a bus if you crossed the road, would you do it any way on the basis that it was less than 95% and that therefore there was "no signifcant" chance you'd get run over?
The 95% confidence interval is only a traditional yardstick, there's nothing magic about it.
Climate Journal:
No, Jones said "Yes, but only just" and "This trend (0.12C per decade) is positive, but not significant at the 95% significance level" to the quesion "Do you agree that from 1995 to the present there has been no STATISTICALLY-SIGNIFICANT global warming". Please stop lying about what he said. Didn't you learn anything from the Daily Mail lying about what he said?
Don't you love the unselfconscious irony of taking on the moniker "Climate Journal" and then posting your deep ignorance of climate science? Poe's Law is thriving in hockey stick fashion.
Erasimodo,
If I need to explain it to you, then you're a moron. So I'll give you a hint....
You're most-likely a moron.
Betula, perhaps I am a moron. Perhaps you are inarticulate. Who knows for sure?
Tim, I am pleased to see that you have raised the matter of John Roskam's grubby mendacity. I had hoped to do it last week, but after the time I wasted on drongo, I had none left to focus on Roskam's grubby point-scoring on Q&A.
Luboš Motl:
Permit me to ask you this - given the noise inherent in the temperature signal over the twentieth century, after any long-term warming trend is removed, how rapid would future warming have to be in order for a statisitcally "significant" signal to be observed in less than a 15 year period?
What are the implications for the planet for a rate greater in magnitude than your answer? More interesting, what are the implications for the planet for any of a range of rates less than your answer, down to than currently reflected in the trajectory of the last 100 years?
Anyone who has even a passing high-school familiarity with statistics should understand the difference between these two statements.
If everyone had "a passing high-school familiarity with statistics", Las Vegas would be home mainly to tumbleweeds and the states would have to raise revenue through taxes rather than soaking the poor with lotteries.
Dear Stu,
when it comes to your "constructions", indeed, I was surely not trying to be constructive. I was writing the truth, and I completely agree with you that the truth is completely disruptive, rather than constructive, for all of you. But that's not my fault. You can only build upon lies, obfuscations, distortions, censorship, corruption, and intimidation. That's what you've been doing for quite some time. And you are still doing it today, even 3 months aftedr your house of lies has collapsed.
I think that the few religious bigots who gather e.g. on this blog are dangerous individuals for the society and for the human rights, and as another reader has correctly observed, the responsible officials should think about the most human ways possible how to protect the society - and individuals - against the likes of yours.
Wow, did anyone just get a whiff of naked eliminationism?
If only you'd said that at the top of the thread rather than here, we could have all saved ourselves a hell of a lot of vain efforts asking questions of you and requesting you justify your position.
Lotharsson, Motl is starting to remind me of [Matthias Rath](http://www.badscience.net/2009/04/matthias-rath-steal-this-chapter/), another idealogue whose wrongness is only matched by his moral hatred for those who point it out.
Lubos Motl:
Yes, collapsing like global temperature.
Bud:
I think it must be Lubos' religion that prevents him from answering questions. He must think he's The Pope.
Luboš Motl.
Perhaps you missed my [posting previously](http://scienceblogs.com/deltoid/2010/02/the_economist_calls_a_lie_a_li…).
In case you did I will repeat, for your benefit, the question I asked:
Given the noise inherent in the temperature signal over the twentieth century after any long-term warming trend is removed, how rapid would future warming have to be in order for a statistically "significant" signal to be observed in less than a 15 year period?
What are the implications for the planet, for a rate greater in magnitude than your answer? More interesting, what are the implications for the planet for any of a range of rates less than your answer, down to that currently reflected in the trajectory of the last 100 years?
Bernard:
I wouldn't expect an answer from Lubos Motl anytime soon. He's a pontif troll, i.e. his objective is to pontificate and you won't get any question answered unless he can answer it in such a way as to allow further pontification.
Well, I say that you are dangerous because you are dangerous.
For example, one week ago, two global warming believers in Argentina - people obsessed just like you - shot their two kids and themselves because of fears of "global warming",
http://motls.blogspot.com/2010/03/global-warming-alarmed-parents-murder…
The daughter miraculously survived, swimming in blood and Gore for three days, even though she was shot in the chest. However, the son (and the parents) died. No one asked the children whether they agreed with their parents' beliefs that there is a dangerous global warming underway. They just shot them.
In the very same way, you are attacking me and others just for the very fact that we say the truth and the truth is inconvenient for your wholly irrational cult. You're ready to impose a new tax or decimate the world's economy, or anything like that, in the name of your insane beliefs. It's just a matter of your so-far reduced courage that you don't act as the people in Argentina, or the Islamic terrorists, for that matter.
The logic is surely the same.
Motl, I suggest you stop PWOC.
(posting whilst on crack)
No doubt, Lotharsson. In correspondence with me on his blog, Lubos once characterized eugenics as "impressive". Lubos is an Asperger-suffering fascist. The good news is that, like all high functioning autistics, Lubos possesses a special talent. The bad news is that his is being an Olympic champion asshole.
Ahh, Lubos, you are truly a disturbed individual.
Meanwhile, global temperatures are still trending upwards. And you never present any evidence that you are correct.
My goodness, that is some grade-A bile! I particularly like the argument that those with whom he disagrees would be mass murderers and worse if only they weren't such abject cowards. I'll have to work that one into my conflict management repertoire.
Lubos, good man, YOU have some titanic issues. Nevertheless, your illuminating bonmots are much appreciated. They go a long way towards showing the true mindset of many a denialist. "Dangerous individuals of society" indeed.
Lubos Motl at #101 is clearly very sick at the moment. I don't see any way that someone this unwell could be contributing meaningfully to theoretical physics as he used to claim he did. Here in blogworld there's no possibility of engagement with one who writes as he does. Medication? Asperger's maybe Majorajam but schizophrenia also seems a possibility; if he's religious there are nasty demons tormenting him.
Seek professional help Lubos.
This is very encouraging. The media has generally been so vulnerable to manipulation by people trying to create confusion about the nature and seriousness of climate change, so it is important that forceful counterarguments be used when they are available.
95 Lubos,
You are highly delusional, possibly borderline psychotic. I suggest professional help.
Let's thank the Pope Lubos troll for directing us to those quality tabloids The Telegraph, Daily Mail etc. We wouldn't get the opportunity to believe everything they say without his example.
Ah, so he does think he is a Pope!
;-)
I have read that Lubos is nice in person.
Lubos, what is it about the internet that turns you into such a bitter, twisted individual? If you met us in the street, would you say the same things?
I agree that you really do seem to need help. On the upside, we do now have a stash of Lubos quotes that we could helpfully update his wikipedia entry with... even if they likely wouldn't stay there for long.
Gee, Lubos, if the murder-suicide of three people provides some proof that AGW scientists are wrong, then what does the death of 30,000 people in the USA alone by firearms prove about firearms? Whatever it proves, it 10,000 times more powerfully than the case you make.
Memo to Motl:
http://en.wikipedia.org/wiki/Scientific_opinion_on_climate_change
You're on the fringe, still clinging to flat Earth theory. "How can the Earth not be flat? Should I believe the scientists or my own lying eyes?"
Odd that you preach about religious cultism after claiming "I was writing the truth". Religious zealots always believe they are speaking the truth.
And no - gradaully reducing carbon emissions won't "decimate the world's economy". Such alarmism is quite common among global warming deniers, although one wonders how many of them actually believe such fear-mongering.
Motl is part of a large congregation of zealots, gaining inspiration and reinforcement from each other.
http://deepclimate.org/2010/03/02/round-and-round-we-go-with-lindzen-mo…
The history books won't be kind to them. It hasn't been so far.
http://www.aip.org/history/climate/
Amidst the mass of other nonsense Lubos Motl utters, this does put the finger of the obsessive hatred their lot have about Al Gore.
I've long wondered about this. Personally, I've never felt all that strongly about Al Gore one way or the other. He seems a fairly articulate middle-of-the-road to slightly liberal US politican. Yet for some reason, the right around the world seem utterly obsessed with him.
[Fran](http://scienceblogs.com/deltoid/2010/02/the_economist_calls_a_lie_a_li…).
I suspect that it has something to do with the fact that deep down they know that Gore was the legitimately-elected president of the USA, and that their numpty shrub stole the presidency from him. The only way to legitimise (in their eyes, at least) what was a travesty of democracy, and of subsequent government, is to attempt to convince themselves that he would have been worse than Bush Jnr.
Talk about screaming cognitive dissonance.
There's probably a lot to what you say Bernard, but I also suspect it's partly to do with plebeian angst towards perceived elites. In the mind of the parochial trailer trash fringe, things from outside your immediate physical space, and even more so, from outside your cultural and intellectual space lose political authenticity. Gore probably presses their buttons because he hails from South of Mason-Dixon but sounds educated and is wealthy. This helps explain the gun nut culture -- there's nothing quite so authentic as a single guy with a rifle defending his verandah.
Interestingly and paradoxically, their hatred for him is so great that they miss how their advocacy helps other elites who really are their enemies.
Fran - I'm not american, but know a few of them, and feel that your characterisation of "parochial trailer trash fringe" understates the problem quite severely, in that we aren't dealing just with some unhinged undereducated nutters. Instead, we've got a mix of well educated ideologues, plenty of people who would be called middle class over here but are also rabidly right wing/ christian types, and of course a large number of think tanks and rich people who may or may not believe their own propaganda.
I take your point Guthrie, but it's the trailer trash fringe who make up the foot soldiers, the zombie and bot armies that turn up at tea party meetings with birther slogans and their guns.
These are the people who are the Fox fodder and lend authenticity to the upper middle class/elite creeps at the top. It's the trailer trash who bear the sedan chairs for people like Beck and Palin and Inhofe and the Senator for Lawnmowers ...
115 Fran,
I'm sure we all noticed the capital G.
If there's any one thing that gives away a denidiot, it's any mention of algore, or any variation on that name.
I suppose that Lubos thought he would impress us with his shining wit? ;)
Aha! I think I spot a spoonerism.
117 Fran,
Of course, the Bush family is trailer trash. Isn't it great how Dubya triumphed against such huge odds?
Heck, Dubya is the 21st century Abraham Lincoln!
112 Stu,
(Godwin looms) My sisters Diana and Unity said that Adolf was *very* charming in person. Take that as you will.
It is quite common for sociopaths to have well developed social skills (EQ) so this might well be the case Decca.
This of course is mere affectation rather than anything that goes to the worth of the person in question in a social setting.
Luboš Motl.
You are [missing in action](http://scienceblogs.com/deltoid/2010/02/the_economist_calls_a_lie_a_li…).
What's the problem - was the question too challenging for you?