Why are climate scientists such bad statisticians?

Question

Michael Mann became infamous because of his statistical errors and unethical behavior.
http://www.uoguelph.ca/~rmckitri/research/Climate_L.pdf
http://greensboro.rhinotimes.com/1editorialbody.lasso?-token.folder=2007-03-08&-token.story=154945.112113&-token.subpub=

Other climate scientists continue to overstate the case for global warming. MIT professor Carl Wunsch thinks global warming may be happening, but it is unhappy with the unscientific statements of his fellow climatologists about specific effects of AGW.
http://www.royalsoc.ac.uk/page.asp?id=4688&tip=1

Wunsch has recently published an online paper trying to teach climatologists the importance of using statistics to keep from jumping to conclusions about global warming.
http://ocean.mit.edu/~cwunsch/papersonline/wunschaha2007.pdf

Climate scientists known for their global warming alarmism are named for their errors.

Why is it necessary for climate scientists to be lectured about the importance of statistics?

Permaculture, an exhaustive study such as you are requesting has not been performed. How would it be calculated? Are you asking the bad statisticians among climate scientists to self-identify? Pretty hard to get someone to hold up their hand on that one. Far better to read the climate literature to see how often statistical errors and incorrect methods are found. I notice the Wunsch paper I linked to is no longer available. I found it on http://climateaudit.org and they probably sent to many readers over there for the server to handle. It's too bad because it is a nice paper on self-deception. Did you read the other links I provided? I would also point you to the Wegman report and fact sheet.
http://energycommerce.house.gov/108/home/07142006_Wegman_Report.pdf
http://energycommerce.house.gov/108/home/07142006_Wegman_fact_sheet.pdf
The key here is that climate scientists are still using some of the statistical methods condemned by Wegman, including decentered data!

Amy, you write: "I just want people to take the emotions and politics out of the mix and look at the facts and projections objectively." That is exactly what Wunsch wants as well. His paper was about self-deception and scientists jumping to conclusions when they should be using statistics to determine if an observation is significant or related to global warming. He points to a number of examples of climate scientists jumping to conclusions instead of using statistics. Did you read it before you posted? Unfortunately, I notice that it is not available now. Hopefully it will become available again soon.

Thomas, my paraphrase of Wunsch had this quote in mind: "Thus at bottom, it is very difficult to separate human induced change from natural change, certainly not with the confidence we all seek. In these circumstances, it is essential to remember that the inability to prove human-induced change is not the same thing as a demonstration of its absence. It is probably true that most scientists would assign a very high probability that human-induced change is already strongly present in the climate system, while at the same time agreeing that clear-cut proof is not now available and may not be available for a long-time to come, if ever. Public policy has to be made on the basis of probabilities, not firm proof."
http://www.royalsoc.ac.uk/page.asp?id=4688&tip=1
Did you have an opportunity to read the paper by Wunsch before it became unavailable? It become available again. I hope you have a change to read it.

Keith, I am sorry but you are incorrect. McIntyre and McKitrick were upheld by the Wegman Report. Even the NAS report (which was much more polite to Mann than Wegman was) agreed with McIntyre on all of the key scientific points, concluding that decentering of data should be avoided and that strip bark bristlecone pine series should not be used. Unfortunately, the ten studies you are referring to all use strip bark proxies of either bristlecone pine or the foxtail series. These studies are exactly the kind of behavior that prompted the question. Why are climatologists still using these proxies when they are statistically known not to be temperature proxies? On the decentering of data issue, Von Storch and Zorita (both are convinced of global warming) investigated this and concluded McIntyre was correct that Mann's method did create an artificial hockey stick even when used with trendless red noise as data. Mann finally admitted at an AGU meeting "I am not a statistician."

Permaculture, the biggest reason I came to the view climate scientists are poor statisticians is because I read http://climateaudit.org The NAS panel that looked into the Hockey Stick controversy agreed with McIntyre on all of the key scientific issues, including that statistics show that the strip bark bristlecone pine series is not a temperature proxy and should not be used in temperature reconstructions. Yet, ten studies have been published after the NAS study using strip bark series from either bristlecone pine of the foxtail series which has the same issues. This shows either an inability to understand the issue or attempt to bias the results. Is this punishable criminally? No, but these actions certainly hurt the reputations of the scientists when this learned. The Divergence Problem is a big issue with the dendro guys. Rob Wilson has attempted to address the issue, but I do not think he has succeeded.

Permaculture, the European Science Foundation recently sponsored a "World Conference on Research Integrity to Foster Responsible Research." Howard Alper makes a claim that is very similar to Wunsch. An ESF press release states:

"Alper also called the misbehaviour found with medical or health science researchers and with climatologists. 'Too many or the former are marketers making exaggerated claims to secure media attention, while the latter peddle biased, ill-informed views as hard facts thereby impacting communication of science to the public.'”
http://www.esf.org/ext-ceo-news-singleview/article/experts-deconstruct-research-misconduct-from-global-and-institutional-perspectives-320.html

If climatologists would use statistics to determine if an event it statistically significant before announcing it is related to global warming, they would have more credibility. The errors in Al Gore's movie were all errors by climatologists first. Evidently, Al never got the memo.

Dana, it is not based on one scientist at all. It is Wunsch, McIntyre, the NAS, Wegman, Alper. The list goes on and on. The fact climatologists are weak in this area is fairly well known. There are exceptions of couse. Pielke, Lindzen, Christy, McIntyre are all good statisticians, but they are all skeptics. Among the more mainstream view, Von Storch, Zorita and Wunsch are good statisticians. I am certain there are others who are good, but the amount of poor statistical work being done in climate science is amazing.

I just noticed the links to the Wegman Report provided earlier are not working. My apologies. These links are working.
Wegman Report
http://img.scoop.co.nz/media/pdfs/0607/07142006_Wegman_Report.pdf
Wegman Report Fact Sheet
http://img.scoop.co.nz/media/pdfs/0607/07142006_Wegman_fact_sheet.pdf

Bob, your answer is exactly the kind of thing I am talking about. Climate scientists try to pretend the NAS did not agree with McIntyre, but if you actually read the report the NAS agreed with McIntyre on every key point of science. They agreed the strip bark bristlecone pine series should not be used and yet climatologists are still using it.
http://www.climateaudit.org/?p=727
The NAS panel did not look into the artificial hockey stick effect from decentered data, but warmers Von Storch and Zorita did and sided with McIntyre. Yet some climatologists are still decentering data! Your blind faith in climatologists is not well-placed.

Dana, Christy did have a small error in his calculations that was found because he was willing to share his data and methods. The size of the error was not significant and did not cause Christy to change his opinion. Christy is a far better statistician than Mann and has made fewer errors than Hanson. Jones is still not providing his data and methods because he doesn't want mistakes found. Did you read any of the links I provided? I cannot tell that anyone who has answered the question has read all of the links yet.

Trevor, of course you are offended by my comment since you claim to be a climate scientist, but you did not even attempt to deal with any of the facts I presented. Never mind any of my supposed errors. Why not deal with the errors you think Wunsch made, or Alper, or McIntyre, or Von Storch, or Zorita, or Wegman. All of these people are talking about errors made by climatologists.

Sometimes it is hard to distinguish between incompetence and unethical behavior. How about addressing the comments by Alper found here?
http://www.esf.org/ext-ceo-news-singleview/article/experts-deconstruct-research-misconduct-from-global-and-institutional-perspectives-320.html

Dana, you wrote: "All scientists share their data and analysis methods." You must not know the facts about climate scientists. Congress had to subpoena Michael Mann to get him to turn over his source code. Congress finally passed a law called America COMPETES Act that requires government scientists to archive and share data. It is a shameful thing a law had to be passed to get Hanson to turn over his code, but that is what happened. The Gary King website on data sharing and replication links to a pdf by McIntyre on data sharing in climate.
http://gking.harvard.edu/replication.shtm
Do some reading.

Answer 1

As your question is about statistics, please can you put your statements about the number of climate scientists being bad statisticians into a statistical context for me?

I am particularly interested to know how many Climate Scientists there are?
Who funds these climate scientists?
How many research papers/journal articles/conference notes your study covers?
What consequences have they suffered? Have these 'inaccurate statistics' been challenged by others and if so, is there a reasonable consensus that these statistics are inaccurate?
Are these statistical 'errors' grave enough to have led to criminal conviction, loss of position and have they had their professional status challenged/registration evoked?
Are these statistics timely?
Is there an accepted margin of error?
In comparative analysis, how do climate scientists compare statistically with other scientists? With other professions?

It is a very interesting question. I should be very grateful for the full research titles/papers if you would paste them please.
Thank you.

Answer 2

The first answer gets to the heart of the issue - you're basing your conclusion on the claims of a couple of scientists. In fact, based mostly on just one scientist - Carl Wunsch. What makes you think he's the ultimate authority on the issue?

As Keith has pointed out, the Mann graph was basically correct. 10 seperate temperature reconstructions look very similar:

http://en.wikipedia.org/wiki/Image:2000_Year_Temperature_Comparison.png

To answer your final question, I don't know why you think it's necessary for climate scientists to be lectured about the importance of statistics. You have very little to base this on.

*edit* Gee, how convenient that you consider all the skeptics to be good statisticians.

Is that why Christy screwed up his atmospheric temperature analysis so badly to conclude that the troposphere was cooling, and later corrected it (after being interviewed for the Swindle)? And why Lindzen thinks smoking doesn't cause lung cancer? Because they're such great statisticians? LOL!

*edit 2* nice try, Ron. The size of his error was not significant? From a cooling troposphere to a warming trend comparable to the surface record and RSS measurements? That's insignificant? Come on, dude.

"The climate trend shown by the UAH satellite data has changed through time, due to corrections in the processing and as the climate has varied. During the first several years of data collection the global trend was downward. That has since changed and the most recent long-term average global climate trend seen in the satellite data is +0.14 C (about 0.25° Fahrenheit) per decade."

http://en.wikipedia.org/wiki/John_Christy

All scientists share their data and analysis methods. No, I did not read your link because I'm not interested in reading a critique of Mann's plot. It's completely irrelevant because the data has been duplicated many times. I glanced over the Wunsch paper and didn't find it very relevant.

Answer 3

You can't make a good case by plunking out 4 sites from a sea of thousands. You can argue that some climatologists statistics can be refuted or argued or put into a more appropriate context, I suppose. And there are plenty of scientists in the mix who are not climatologists but weigh in from other fields of expertise. Slogging through all of this is tough for the laymen, and claims of "bad statisticians" need to be weighed carefully and by those with far better qualifications than I can offer.

Here are some sites I find useful. If your mind is open even a peep, take a look, dig around, and let me know what you think. I'm not out to persuade anyone of anything, I just want people to take the emotions and politics out of the mix and look at the facts and projections objectively.

Answer 4

Climate scientists are fine statisticians. Especially the IPCC, which has some of the best in the world. Mann's work had minor flaws. His conclusions were accurate (per the National Academy of Sciences), and the work has since been redone with better statistics:

http://www.globalwarmingart.com/wiki/Image:2000_Year_Temperature_Comparison_png

It's the climate SKEPTICS who are lousy statisticians. Possibly deliberately so, since there's no other way to get to the conclusion that global warming is not mostly man made.

"Pattern of Strange Errors Plagues Solar Activity and Terrestrial Climate Data", Eos,Vol. 85, No. 39, 28 September 2004

Re Scafetta and West:

"In my physics undergraduate course, we learned that one should stay away from analyses based on the difference between two large but almost equal numbers, especially when their accuracy is not exceptional. And using differences of two large and similar figures in a denominator is asking for trouble."

The bottom line. Note the key word quantitative:

"There's a better scientific consensus on this [climate change] than on any issue I know -
Global warming is almost a no-brainer at this point,You really can't find intelligent, quantitative arguments to make it go away."

Dr. Jerry Mahlman, NOAA

Answer 5

How much do you know about either climate or statistics?

From your previous questions and answers it's quite obvious you know almost nothing about climate and I suspect you know very little about statistics. Let me know any time you want to go head to head on either subject.

I'm not going to waste my time going through your question and correcting all the errors you made, it would take far too long.

I'm a climate scientist and use lots of statistics, I must have quoted many hundreds, possibly thousands in my answers on here. If your claim is correct then go ahead and see how many of my statistics you can prove to be wrong. If you're unable to do so, will you be retracting your statement? No, I didn't think so.

Answer 6

1. The bad statisticians are not Mann et.al, but critics McIntyre & McKitrick, who changed the standardization convention, but incorrectly retained the same number of principal components. Under the M&M standard, at least 5 (perhaps 6) PC's contain significant information, but M&M included only two in their reconstruction. (Using the Mann et.al. standard, two is the correct number.) Had M&M done their statistics correctly, their graph would have been virtually identical to Mann et. al. It is mathematically wrong to retain the same number of PC's if the convention of standardization is changed.
http://www.realclimate.org/index.php/archives/2005/02/dummies-guide-to-the-latest-hockey-stick-controversy/

2. The Mann et.al. 1998 study has been confirmed by no less than TEN further peer-reviewed papers by other authors. Here are all of those on one graph:
http://www.columbusnavigation.com/pictures/paleoclimate1.png

3. The National Academy of Sciences -- the oldest and most prestigious scientific institution in the US -- appointed a special commission to look into the controversy, and determined that the conclusions of Mann et.al. were essentially correct.
http://www8.nationalacademies.org/onpinews/newsitem.aspx?RecordID=11676

Answer 7

Every scientific debate calls into question the opposing view's data. That is expected. Agreeing with Permaculture above, you cannot throw out a blanket statement as you have - site specifics, if you have them.

I suspect the underlying tone you are presenting is a disbelief in global warming with the associated political undertones. I haven't reviewed the data, but based on the photos of glaciers and icecaps disappearing, as well as the freaky weather we're had in this last decade, I've made some changes in my life.

By the by, I'm including a quote from your source below, in which he states that, in his opinion, global warming has a "major human-induced component".

I note that you paraphrase Wunsch as saying it "may be happening" whereas Wunsch is quoted below as saying "I believe it is real." I find that interesting, given the gist of your argument - complaining about the lack of precision in this debate.

"I believe that climate change is real, a major threat, and almost surely has a major human-induced component. But I have tried to stay out of the `climate wars' because all nuance tends to be lost, and the distinction between what we know firmly, as scientists, and what we suspect is happening, is so difficult to maintain in the presence of rhetorical excess. In the long run, our credibility as scientists rests on being very careful of, and protective of, our authority and expertise."
Carl Wunsch, March 11, 2007

update: Thanks for proving my response - the data will be debated ad naseum - there was still "data" that proved tobacco wasn't linked to cancer as little as 8 years ago.

While scientists such as Wunsch, who you quote, believe that human activity is no small factor for this climate change, they are unable to prove it beyond all doubt.

Taking directly from your quote above,” It is probably true that most scientists would assign a very high probability that human induce change is already strongly present…..Public policy has to be made on the basis of probabilities.

Wunsch and Mann are both arguing the same point of view! At best, Wunsch questions some math, but agrees with Mann’s conclusions. If that is your strongest argument against climate change, I claim victory in this debate.

Answer 8

How can you develop accurate statistics with critical information missing? That is why the heated debate and many professionals having problems with the CO2 theory alone, me being one of them.

In order for the globe to be heating because of man made influence, there has to be sources of heat to elevate global temperature.

Surface Temperature Monitoring of the surface of the planet is seriously flawed and doesn't provide an accurate interpretation of urban heat islands. It is assumed that Urban Heat Islands were the buildings and development absorbing the sun's rays and holding the heat. Urban Heat Islands are Urban Heat Generators and then that extreme heat is absorbed by the building as well as radiated atmospherically.

Temperature measurment has evolved and buildings are generating heat close to boiling temperature. We are reacting to the symptoms with ozone depletion, electrical waste, more emissions, toxicity and a domino effect impacting the world.

Scientists need facts to be accurate and there was some critical information missing so the end results were missing data.

When I showed a professor the amount of heat generated right after sunrise, he said "the greenhouse gas theory is seriously flawed, think of the heat generated by the east coast right after sunrise."

Go to http://www.thermoguy.com/globalwarming-heatgain.html to see the heat generation. Simulate that in temperature generation in your environment and it means you have the potential to heat up to close to 200 degrees. Who will live?

Now buildings are being subjected to heat they aren't designed for and it complicates building function including heat loss. Go to http://www.thermoguy.com/globalwarming-heatloss.html and see how were are really using our non renewable resources.

Answer 9

It is because the statistics do not agree with the conclusions that they want.