# Phil Jones test 2013

B – Do you agree that from 1995 to the present there has been no statistically-significant global warming
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.

To be clear – that was a question to Professor Phil Jones of the University of East Anglia from the British Broadcasting Corporation. Neither Jones or the BBC are usually taken to be climate “deniers”.

Over the past few years we have been updating that analysis on an annual basis as more data have become available. The complete data for 2012 are now available. So I fired up the trusty Eviews and re-estimated the test using the monthly data from 1995 through 2012.

I’ve replicated this result four years running and each time the t-stat has not met the p = 0.05 test (that is the 95% significance level that Phil Jones refers to). This year the t-stat is 1.35 with a corresponding p = 0.1784.

When I use annual data I get a p-value of 0.1016 – also failing the Phil Jones 95% confidence level requirement.

So what can we conclude? To the BBC question:

B – Do you agree that from 1995 to the present there has been no statistically-significant global warming

The answer to that question remains “Yes”.

This entry was posted in Phil Jones Test. Bookmark the permalink.

### 41 Responses to Phil Jones test 2013

1. Philip Crowley

Someone needs to tell Craig Emerson!

2. Grey

I think Dr Davidson would have made an excellent captain of the Titanic

“Don’t worry, Madame, there is an 18% chance we won’t sink.”

3. Bruce

It’d do better with a non-linear fit. Unfortunately Phil Jones has trouble with even linear regression in Excel, so asking him to fit a sine curve would be futile.

4. Samuel J

Why don’t we hold them at the 5-sigma level of significance that CERN used in testing for the Higgs Boson? That would require a significance of 1 in 3,500,000 (rather than 5 in 100) to demonstrate that there was a statistically significant warming trend.

5. Sinclair, when I tried to read that, my head started to hurt.
Do you do this stuff all the time?

6. Sinclair Davidson

Not so much anymore.

7. Jc

Greys;

It’s over. Jump ship (to use your own analogy), as it’s over. The alarmism and the alarmists are finished.

8. If they didn’t ‘Adjust’ the temperature records there would be no warming at all. Funny how all the records pre 1970’s are cooled through adjustment and the later ones are warmed.

http://joannenova.com.au/tag/australian-temperatures/

Lots more like that if you are prepared to look for it.

9. Eddystone

Isn’t the 95% confidence level a rather low bar?

Although this is climate “science”, so one shouldn’t be surprised!

10. Jim Rose

sinclair, this sample will be denounced as too short, but a hot summer in 2012 is a enough data to prove global warming is real and is happening now.

11. Samuel J

Do you agree that from 1995 to the present there has been no statistically-significant global warming

I take it, Sinc, that to go further and say “there has been no global warming from 1995” would require a test at the 95% confidence level of the opposite question.

So are we really saying: there is insufficient data to know whether the planet has warmed (or cooled or stayed the same) since 1995?

12. JamesK

I remember when I first read that and the squirming and laughable contortions Jones got into to minimise the damage.

That was the time ‘climate change’ really took over ‘man-made global warming., I think

13. WhaleHunt Fun

You and your moon landing denialist loopy friends cannot have your own facts.

The FACTS are that Tim Flannery and Al Gore and Dr Karl and Dr Emerson and dozens of other people at the ABC are all climate experts. Q&A and Tony Jones makes this clear. So you can deny it all you like but you cannot change the rock solid fact that these people are paid for what they do so that makes them experts.

So you can take your NASA data and your UK Met data and shove it. I bet all those nutters at NASA don’t believe in the moon landings either.

14. WhaleHunt Fun

And my mum likes Q&A.

15. Sinclair Davidson

Jim – yes I know. This post causes all sorts of angst every year. That is why I’m now calling the annual post the ‘Phil Jones test’. He devised the test and initially used a smaller time series. I’m simply replicating (and augmenting with the AR(1) and robust p-values) his test. Mind, the unvarnished annual test provides much the same answer.

Samuel – there has been no significant increase in the warming trend. Assuming the data are not corrupted, however, it is still warmer now than it was, say, 20 years ago. The important question is what will happen in the future. The thing is that the planet has not started to cool down – Phil Jones was asked a third question about cooling since 2002. That trend is negative but not statistically significantly different from zero.

16. Grey

Phil Jones was reporting 0.12 C.

I am not quite sure what you have done here, but would it be: 0.00053 * 12 * 10 = 0.064 C?

Using your method do you get the same figure as Phil Jones just suing the 1995-2009 data?

17. jupes

Using your method do you get the same figure as Phil Jones just suing the 1995-2009 data?

And using any method, can you tell us the effect of the carbon tax on global temperature. In degrees celcius please.

18. Sinclair Davidson

Grey – I’ve replicated Jones’ figure before. From memory my replication was slightly smaller than his original figure but small enough to be a rounding issue.

19. John A

“Using your method do you get the same figure as Phil Jones just suing the 1995-2009 data?”

Darn it all, Grey, why do you have to bring the lawyers into this? 🙂

20. Alfonso

“I think Dr Davidson would have made an excellent captain of the Titanic..”

Maybe. Punters, what are you going to do to survive other than use a firearm to acquire a lifeboat assuming the lack of life boat capacity became obvious early…which it did?

SAS operators and airline pilots only need apply……

21. Jim Rose

Sinclair, It was very poor judgement of Jones to answer the question, but was it?

Jones should have pointed out that century long trends are more important as evidence. But that would have been worse because it would have ruled out scare mongering over a few hot summers.

22. Rod Clarke

When I first joined this site some 3 years ago I asked the AGW campaigners for confidence intervals, Rsquared measures and measures of heteroscedasticity and autocolleration in their models. Including from a physicist and mathematician I very much respect Dr Steve Edney.

its good to see that my instincts were right.

23. Sinclair Davidson

Jim – yes. He answered that question very badly – especially given the story he wants to tell. It gets worse – those were written answers.

24. duncanm

Grey,

I think your analogy should be :

“Don’t worry, Madame, there is an 18% chance we won’t hit anything. What we might hit would be something so small as to bounce off the hull. In order to avoid damage to the hull if we do hit something, we could strap passengers from steerage on the bow. A few of them would die due to exposure or drowning, but if it makes you feel better, then its worth it”

25. Grey, if the stats supported warming, it still won’t justify taxing the poor to fund a 10 Billion Green slush fund and 57 Billion a year in dodgy Al Gore carbon derivatives.

If you’re concerned then take cold showers, get a local job within walking distance and insist on being paid the minimum wage, but keep your thriving hands off the wages of the poor!

26. Grey

If you’re concerned then take cold showers, get a local job within walking distance and insist on being paid the minimum wage, but keep your thriving hands off the wages of the poor!

I think I will just massively increase taxes on the wealthy and distribute it in solar panels to civil servants.

That way rich people won’t be tempted to spend it on things that emit carbon.

27. Kaboom

Forester:

“but keep your thriving hands off the wages of the poor!”

I think that THIEVING would be a much better word. The Greys of this world simply don’t “thrive”, except in regard to useless procreation of welfare recipients…

I hate predictive text, and turn it off, just in case I embarrass myself.

28. Johno

I think I will just massively increase taxes on the wealthy and distribute it in solar panels to civil servants.

That way rich people won’t be tempted to spend it on things that emit carbon.

Looks like Grey has outed himself again.

It ain’t about ‘saving the planet’. It’s all about soaking the rich. It ain’t about caring for future generation. It’ just plain old fashion envy.

He is a sad and pathetic little creature.

29. Grant B

Grey, it matters not whether the decadel trend is 0.64 or -0.64. The point is the series t value is less than t (95%)* therefore the null hypothesis cannot be refuted. The null hypothesis being that the slope is zero.

Look up t-values. You might find it interesting especially if you enjoy gamma functions and/or Guiness.

30. Harold

I think it is rising a bit but nothing like they’ve forecast. The skeptic case is favoured by the high late 90s spike which to me looks anomalous. Still no need for alarm.

31. Rosco

This whole thing is really silly.

There is no way that the actual temperature measurements taken are more than 0.5 of a degree accurate.

How the hell can these climate buffoons convert +- 0.5 on a daily basis into trends with one hundredth of a degree accuracy over a decade.

This is scientific fraud !

32. Rosco

You cannot have more precision in the results of a data analysis than the actual data.

A large proportion of the historical temperature data cannot possibly have more precision than a half a degree – I know this to be true for some of the reporting stations I have seen including Amberley air force base from the 60s where my uncle was meteorological officer.

It is not possible to have decadal trends with two decimal points accuracy from original data that is at best +- 0.5 accurate.

33. 2dogs

That AR(1) coefficient is highly significant. I think you should consider taking first differences and work with the residuals.

34. Grey

Grey, it matters not whether the decadel trend is 0.64 or -0.64. The point is the series t value is less than t (95%)* therefore the null hypothesis cannot be refuted. The null hypothesis being that the slope is zero.

Grant, I know you think you understand statistics, but you don’t really.

A linear regression is not the only interpretation to the statement have temperatures increased. There might be a gradual trend increase laid over a 11 year solar oscillation – in which case you would need to adjust your values to the position in the solar cycle.
Or you might have a gradual trend increase underlying a largish variation – which means you need several decades for your trend to establish itself using a linear regression model.

This is just online silliness – not that there is anything wrong with that.

35. JC

Greys

How much of a temp movement would relate to sun spot activity, recalling of course we’re dealing with absolute tiny changes in temp. Go

36. Up The Workers!

How would Craig Emerson know whether temperatures have marginally increased or not?

Did his swallowing both of Juliar’s contact lenses from the glass beside her bed, improve his eyesight? (Maybe somebody ought to tell the good Ph.D that contact lenses are thought by most reputable optometrists to be most efficacious when applied to the outer surface of the eyeball, and are NEVER drunk).

With Einsteins like EMO on the job, how could we possibly go wrong on Global Warming (or is it Gerbil Worming? Or Glow-ball Gurning? – better guzzle another couple of contact lenses.)

37. Bruce

HADCRUT3 says Dec 2012 was .233C warmer than the 61-90 average.

Dec 1939, 1979, 1981 and 1987 were warmer than Dec 2012.

The difference in temperatures is not noticeable by humans and is most likely all UHI.

38. sillyfilly

So, Sinclair, the planet is still warming but not significantly since 1995.

Of course, being your usually totally disingenous self, you completely ignore the measured increase in OCEAN HEAT CONTENT one other signature of global warming

Making a mockery of statistical analysis does not disprove AGW.

39. Grant B

“The planet is still warming but not significantly since 1995”. Thankyou for pointing out that Sinclair is correct. That’s all we need to know. Fullstop.

40. observa

I think Dr Davidson would have made an excellent captain of the Titanic

Well he or the Madame might have to, if like a certain Italian cruise liner captain, the Titanic captain had already jumped overboard because he was overly worried about that 18% chance they were going to hit an iceberg and all perish.

Thanks for your err.. input there captain, but perhaps there’s a moral to the tale in there somewhere for all these budding captains that are so sure they know what’s best for us all, particularly as according to them, there are far too many of us for Gaia’s lifeboats already.

41. Werner Brozek

Here is the latest from:
http://www.skepticalscience.com/trend.php

Here are the times to the nearest month that the warming is not significant for each set. They are as follows: RSS since August 1989; UAH since May 1993; Hadcrut3 since August 1993; Hadcrut4 since July 1994; GISS since June 1995 and NOAA since May 1994.