Wednesday, December 9, 2009

The survey, conducted by Dennis Bray (and with some help by me), among climate scientists has now be done 3 times - the statistics of the latest survey from 2008 are now online. Dennis Bray will in future contribute regularly to this blog and highlite results from this and the earlier surveys.

20 comments:

MikeR said...

Thanks very much, this is absolutely fascinating.

One comment: It would be much easier for the reader if the box plots were exactly aligned beneath the bar graphs, with horizontal values in the same places. I kept having to check that they weren't weirdly wrong.

Anonymous said...

There appears to be a missing word or two in the text relating to Question 58. This should be fixed.

The charge has been levelled at climate scientists that their concern is driven by pragmatic factors related to accessing funding. It would have been interesting indeed to establish whether the responders' salaries are being paid from public funds or from public company funds, or from self-funded sources - eg retired folk. My hypothesis is that if you were to sort the responses according to this information, you might find alignment of opinion correlates with funding source.

It would also be interesting to pose the same questions to different target groups in the community - journalists, politicians, ordinary folk - to see to what degree the range of opinion expressed by these climate scientists is matched or not matched in the wider community.

Nanonymous said...

This is great, thank you! Illuminating. Paints a picture of "normal" science - with a lot more optimists who think their field is doing generally great (at least that's the normal science I know in my field). As an outsider, things I found of particular note:

1. Most believe that models today can do a good or great job predicting most of the climate, yet at the same time most believe that models do a very poor job accounting for clouds. Question: Isn't it somewhat obvious that clouds must have very major effects? And if the answer is "yes" where does the considerable optimism come from?

2. There really does exist a near consensus that the warming is real *and* that it is anthropogenic.

3. Most believe that the current models do rather poor job of predicting what lies 50 years ahead, yet most also believe that the warming is very dangerous to humanity and that 50 years ahead we *will* experience catastroph. What does this belief is based on then?

4. Most scientists think they are not inclined to make predictions based on models. Clearly though, as evidenced from many answers, they do make a lot of predictions. Based on what then? (The apparent contradiction may simply reflect a total confusion on "prediction" vs "projection").

5. Only 24% believe in Popperian falsification. To me, that's an unpleasant surprise...

Anonymous said...

Nano: You say: "2. There really does exist a near consensus that the warming is real *and* that it is anthropogenic."

That is, among this selected population of 'climate scientists', many of whom are likely to have a vested interest in the world believing that AGW is real.

Joe said...

The most interesting thing to me is that, on average, the respondents are split down the middle on the question of how well the climate science community understands the climate mechanisms, not an unreasonable situation in any developing field of science, and yet a much larger majority believe that the models are accurate and that the climate will change for the worse in the future.

Anonymous said...

Am I right in my conclusion that only 18.2 percent of the scientists who got the questions actually answered them?

If this is the case, I do think this survey has a way too low response rate to draw any conclusions on the consensus of the climate scientists.

The things discussed by "Joe" in the previous comment are interesting though!

(From the survey:
Response Rate % 27.2 10.39 19.1 18.2)

richardtol said...

The answers to the questions in the mid-30s are scary. The interviewed researchers seem to think that the world should be run by technocrats, that is, themselves.

Dennis Bray said...

My response to all comments to date

To Mike R

First, thanks for the accolade. As for the alignment of the box plots with the histograms, I agree that the alignment of the values would be an improvement but I know of no software that can produce such a diagram. To attempt to compensate for this I included the value labels on the box plots.


To Anonymous

Thanks for pointing out the omission in question 58. This will be rectified ASAP. While the funding question would be interesting I doubt many respondents would disclose such information. As for multiple target groups, this would indeed, also be a series of interesting projects. A number of years ago a colleague and myself did survey German policy makes concerning global warming issues. Results are available upon request in the form of the GKSS report C. Kruek and D. Bray, GKSS Report 2000/6 ‘Wie schaetzt die deutsche Exekutive die Gefahr eines globalen Klimawandels ein’ (not available electronically) See also: 2001 Bray, Dennis, Krück, Carsten’Some Patterns of Interaction Between Science and Policy: Germany and Climate Change’ Climate Research. 2001 Nov 22. 19(1).P.69-90. But, as you say, there is much interesting work still to be done. Studies of ‘Ordinary folk’ and their perceptions of climate change, do though, abound.

To nano

Another accolade - thanks! Glad you picked up on the significance of Popperian falsification. I hope to have a paper on that sometime next year. You might want to note that I received a number of comments regarding that set of questions and post-survey many similar comments were made on blogs. All were asking why such questions were included. Glad you spotted it. Regarding your other questions, these are best left to climate scientists to answer.

To anonymous 2

The scientists weren’t selected based on their beliefs or values. Efforts were made to keep the sample relatively random. Oddly enough, in the 2003 survey, the exact opposite charges were made - that the sample reflected skeptic tendencies. See: http://scienceblogs.com/deltoid/2005/05/bray.php. If you would like to see the results of previous surveys, along with sampling procedures, they are available at http://coast/staff/bray/surveyintro.html.


To Joe:

What can I say?


To anonymous again:

That is correct. The response rate is approximately 18%. However, given the nature of the target groups this is not a disappointing rate. For a full discussion of on line survey response rates se http://www.supersurvey.com/papers/supersurvey_white_paper_response_rates.pdf

Dennis Bray said...

Richard

Intersting times!

Anonymous said...

About the response rate of 18 %: how did you do to ensure anonymity to the scientists in the study?

To "come out" as a sceptic, or even as somebody who questions the "consensus", seems to be really tough in the climate science community, and this might bias the survey severely.

You have no possibility to know the opinions of the
82 % who did not answer and this is a serious problem in your study.

Anonymous said...

The fact that your last survey reflected skeptic tendencies is another factor pointing at that your response rate is too low.

Dennis Bray said...

To anonymous

The anonymity of the respondents is ensured by use of an email invitation. Each invitation is assigned a random respondent id number by the bulk email package. The responses are then checked against this assigned list of numbers to avoid duplicates. I suppose if someone is so inclined, and has enough time, he or she might be able to match these numbers with the emails sent out. Also, the survey states that respondents will remain anonymous. Perhaps integrity is a foreign concept to some but in the case of the survey, those respondents who did reply obviously accepted these terms and/or had nothing to hide.

As for the survey reflecting skeptic terms - I am not sure where you see this. Also, that would mean that the sample was biased towards skeptics when indeed it was not.

As for the 'other 82%' - that is 82% of the sample, not of the entire climate science community. First, there is never any expectation of a 100% response rate and nor was there ever any hope of surveying the population (as opossed to a sample). I do not have the time nor inclination here to go into the details of sampling theory or power analysis and would suggest you seek further help on this matter elsewhere.

Anonymous said...

Sorry, I misunderstood this part of your answer: "Oddly enough, in the 2003 survey, the exact opposite charges were made - that the sample reflected skeptic tendencies".

MikeR said...

Relax, folks. All surveys have low response rates, and this was quite a long one. I'm pleased they got 18%

Dennis Bray said...

Ref last Aonymous

You said

'To "come out" as a sceptic, or even as somebody who questions the "consensus", seems to be really tough in the climate science community, and this might bias the survey severely.'

Sorry if I misinterpret you, but it seems to me that this implies that it would bias the sample towards non-skeptics. In the 2003 survey, it was suggested that the sample was biased towards skpetics.

Anonymous said...

Hi Dennis! Yes, you did misinterpret me.

I do mean that the bias probably is towards non-skeptics.

Since I misinterpreted your answer "Oddly enough, in the 2003 survey, the exact opposite charges were made - that the sample reflected skeptic tendencies" as meaning that the outcome of that survey was that there were more skeptics than non-skeptics, I meant that the different outcomes to the surveys was a problem that probably originated in too few people answering the questions, rather than people actually changing their minds between the surveys (I do understand that it wasnt the exact same persons who participated in the 2 surveys by the way).

Hm, do you understand what Im trying to say? This got rather complicated. I think we should just drop it since the reason for this particular discussion is misinterpretations.

Anonymous said...

What would consitute "Popperian falsification" in the context of AGW?

If for example, AGW predicted temperature rise in response to a rise in CO2, at what point would no temperature rise consitute Popperian falsification?

For example, would 10 years of stable and/or decreasing temperatures be enough? Would 30 years be enough? At what point would we be able to conclude that belief in AGW was false?

Nanonymous said...

To the Anonymous (Dec 16, 2009 4:24 AM):
What would consitute "Popperian falsification" in the context of AGW?

:-) Popper's very own criterion for distinction between science and non-science is falsifiability. It does not have to be immediately feasible (say, technical, monetary or temporal limitations) but it requires that in principle there has to be a test that can produce falsifying result. His favorite examples of non-science are Marxism and psychoanalysis - theories that explain everything.

So in a strict Popperian sense it is up to proponents of AGW to specify the falsifying experiment/observation if they insist that their hypothesis is in the realm of science.

Now, of course real life does not work according to the Popperian ideals! Real world is intrinsically messy and, as Feyerabend aptly documented, "everything goes" in it. But the thing is - and even Feyerabend did not seriously dispute it - that more frequently than not it is "good" for science to strive to conform to Popperian ideals. The mess (and all potentially good that comers with it) will come naturally - no worries.

So the point in the context of this thread is that when a good >75% of scientists dealing with AGW are not in any shape or formed concerned with falsifiability of their theories, in the eyes of "proper" scientists it puts them in the bad light. Because in practice it means that they are not concerned with predictive power of their theories and are perfectly willing to explain everything that is to be observed.

Anonymous said...

Is there a mathematical test that can be applied to the predictions of AGW?

For example, there is pretty much a 50-50 chance of the temperature going up or down year to year by chance. So, if the temperature goes up 5 times, and down 5 times in 10 years, while CO2 levels kept going up, that would tell me the odds of AGW being right are maybe 1/32.

So, if the cost of correcting the problem is anything more than 1/32 of the cost of living with the problem, you live with the problem rather than correct it.

TCO said...

Any interesting cross-cuts? Correlations of answering one way with another or correlations of demography with response?