Analyse statistique du consensus

Par Eric Steig (traduit par Pierre Allemand)

Y a-t-il réellement “consensus” dans la communauté scientifique sur la réalité du changement climatique anthropogénique ? Comme N. Oreskes le fait remarquer dans un récent article de Science, c’est une question qui peut être elle-même traitée scientifiquement. Oreskes a pris un échantillon de 928 articles sur le changement climatique , objectivement choisis (utilisation de la phrase clé “changement climatique”) dans la littérature scientifique relue par des pairs. Oreskes en a conclu que parmi les articles (environ 75 % du total) qui traitent de la question 100 % (tous) partagent la vue consensuelle selon laquelle une part significative du changement climatique récent est due à l’activité humaine.


Evidemment, il existe sans doute certains articles qui ont été publiés après relecture par des pairs, qui donnent un avis différent, et qui n’ont pas été décomptés dans l’étude d’Oreskes, mais le fait que son mode d’échantillonnage ne les ait pas trouvés indique que leur nombre est très, très petit. On pourrait débattre pour savoir si l’existence d’un très large consensus constitue une base pour une action sur le changement climatique ; en revanche il n’existe pas de base de discussion sur le fait de savoir si ce consensus existe ou non.

Mise à jour du 26/12/2004: Naomi Oreskes a écrit un éditorial qui prolonge le sujet dans le Washington Post -mike

[NB : il sera plus approprié de discuter du sujet dans le nouvel article de RealClimate En quoi consiste donc ce consensus?]

42 comments on this post.
  1. Peter Thejll:

    Is it possible to see what fraction of these papers also said that the Sun could have an influence? Don’t we all hedge our bets by keeping away from one-sided statements?

  2. Eli Rabett:

    Peter: Hopefully all of them, as the sun has a major influence. It is simply that after accounting as best one can for the influence of the sun you still can’t match the observed climate over the past ~30-50 years without including the effects of additional greenhouse gas forcing.

    It is not solar OR anthropic greenhouse gas forcing. It IS solar AND anthropic greenhouse gas forcingS.

  3. mike:

    David Appell has a nice discussion of this. One piece of information he uncovered was that the search was actually done on the phrase “global climate change” and not “climate change”. The point would still seem to stand, however.

  4. Jeffrey Davis:

    The problem is climate change. One of the issues is the degree to which our behavior contributes to it. Why? Because our contributions would be the easiest to change. What if it were 0%? Should we remain fatalistically idle? “Oh, well. The solar wind is going to make the planet uninhabitable. Pass me the heroin.” No, we’d figure out what to do about it. After all, we wear clothes, we build houses, we grow food, etc. We already change our environment to suit our needs.

    We need to deal with changing climate regardless of the degree of our involvement. Determine the causes. Assess what looks like the extent of the change. Figure out what can be done to stabilize it. (Agriculture works best when farmers can be sure that next year will be a lot like this year).

  5. Jay F:

    Surely the phrase “in part” must be defined rigorously to justify any sort of “action”, otherwise the statement “climate changes [...] due to human activities” has little value and is practically a tautology.

  6. Woody Emanuel:

    I’ve seen no actual links to Oreskes’survey data and methodology. I’m hesitant to accept the conclusion of 100% support without access to that data.

  7. Slade:

    Since she’s polling on the notion that human activity is involved, why doesn’t she go on
    to poll for other factors? If she were to ask the exact same question substituting ‘solar
    activity’ for ‘human activity’, might she not also have gotten 100%?

  8. eric:

    In response to some of the comments above: the conclusion as stated by Oreskes is stronger than my paraphrasing of it, and I’ve edited my original post to make this clear. But you really ought to read the original article in any case.

    In specific response to the comments on solar variability, comment #2 is exactly correct. Note that the relevant question to ask — in response to #7 — is “what fraction of papers support the view that solar variability is the dominant cause of the warming of the last century?” The answer will not be anywhere near 100% (or even 1% I suspect), though it would certaintly be greater than zero, as some peer-reviewed papers have certainly tried to make this case. Those papers are widely viewed as flawed by the scientific community, however. I won’t say more here as there is already a post on the purported sun-earth connection on this site.

  9. James B. Shearer:

    The inference that the failure of a random sample of 928 papers to find a skeptical paper indicates that the proportion of skeptical papers must be “significantly less than .1%” is unjustified. Suppose for example that the exactly .1% of all papers are skeptical. Then (assuming an infinite number of papers) the probability that a random sample of 928 will not pick up a skeptical paper is .999**928=.395 so there is no strong reason to believe the proportion is less than .1%.

    Editor’s note: I had written that the number of skeptical papers must be significantly less than 0.1%, and Shearer correctly pointed out my error.
    In the interest of readability and accuracy, I changed the original post to read “very very very small”. Please also note my comment below giving the correct statitics.
    Note also that .999**928 in the comment above means .999 raised to the power of 928.

  10. Andrew Boucher:

    “a significant fraction of recent climate change is due to human activities”

    Well I probably believe that. But there’s a long road from that to knowing which fraction and what human activities.

  11. eric:

    NOTE: James (#9) commented that my statement that the number of
    papers that concluded the opposite from the “consensus” view was
    less than 0.1% was incorrect. He is right and I have changed the
    statement to “very very small”, which is the important point. We can
    be more precise about this than either of us were, however.

    Using a Poisson approximation (which is the appropriate limit of the binomial
    distribution for N large and p small), the probability for any
    discrete number of “successes” n is p(n) as follows for the case
    N=981, p=0.001 (.1% rate of occurrence):

    p(0) = 0.395
    p(1) = 0.367
    p(2) = 0.170
    p(3) = 0.053
    p(n>3) = essentially zero.

    So indeed, there is a roughly 40% probability that one might
    observe no occurrences given a true occurence rate of 0.1%.

    However, consider the numbers for a 0.3% rate (still, very, very
    small). Then we have:

    p(0) = 0.06
    p(1) = 0.172
    p(2) = 0.239
    p(3) = 0.222
    p(4) = 0.155
    p(5) = 0.086
    p(6) = 0.040
    p(n>7) approx 0

    Thus, we can state with high (~95%) confidence that the true rate
    of peer-reviewed articles that go against the consensus view is
    almost certainly less than 0.3% (though it may be greater than 0.1%)


  12. Ferdinand Engelbeen:

    There is another bias, as the “peer-review” process itself forms a bias, because peer-reviewers mostly agree with the majority that makes the consensus. Thus any article that may hint to other causes are filtered out or need some “adherence” to the general accepted theory (even if it contradicts that!), or your article will not be published… This process may be real or imaginary, but certainly influences the wordings used in “difficult” articles.

    Editors note: I deleted much of this long comment because it is off topic. The point of the Oreskes article is not to address whether there is a “conspiracy” or even whether the consensus view is correct. It was simply to point out that there is a consensus view. — eric

  13. John Wrobel:

    What if there were 10001 articles and all were wrong except one that was scientifically right? I thought you were a science oriented site.

    John Wrobel

  14. John Fleck:

    John Wrobel -

    That’s not a particularly useful way of phrasing the question, because it assumes you know know what’s “right” and “wrong.” But the whole point in this sort of debate is that you don’t, except by looking at the 10,002 papers in question.

    What if there are 10,001 papers that say “A” and one that says “B,” and let’s say you’re a policymaker trying to formulate an action that depends in some fashion on which is correct. It’s not unreasonable to go with “A.” That’s the way we make these sorts of decisions all the time. If we used your approach, we would never take any science-based public policy action at all, because in any interesting science there are always mavericks and outliers.

  15. Benton Maples:

    There are two core issues here that are far different, IMO.

    1. Is human emissions of C02 likely to be an additive factor in global temeratures?

    2. Is most of the warming of the past 50 years attributable to the greenhouse effect?

    I think it’s pretty obvious that there’s a concensus regarding #1.

    But I don’t think it’s so clear that there’s a concensus regarding #2.

    Oreskes seems to blur the distinction between these, and suggests the IPCC statement: “Most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations.” is not only a consensus view, but a unanimous view among around 700 randomly selected peer-reviewed studies that address the issue.

    Since the methodology is not discussed, I am highly skeptical of this study.

    If the methodology in fact tests whether the authors agree with #1, then Osrekes paper would appear to go beyond bias and into the realm of dishonest.

  16. tom:

    Re: the last post.

    “100% (all of them) support the consensus view that a significant fraction of recent climate change is due to human activities”

    Did the original post misstate Oreskes’ view?

    Response No. In the original post I had written “recent climate change is due in part to human activities” whereas Oreskes’ statement was stronger. In the revision I tried to more accurately reflect Oreskes’ emphasis — but as always, going to the original source is valuable!

  17. John Fleck:

    Benton -

    Oreskes’ paper (I think it’s on line free here, though subscription may be required), defines the “consensus” using this direct quote from the IPCC: “Most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations.” So she’s arguing that her literature review found a clear consensus around your point #2.

  18. GeniusNZ:

    I wonder how many studies directly (as in in a scientifically valid manner) address #2 as opposed to assuming it from theory.
    I think it is reasonably likely to be true but it is easy to confuse the obvious #1 with and #1.5 “the earth is warming” with #2 “CO2 explains most of the earth warming” or even the possible “the earth should be cooling but it is warming instead due to CO2″

    Response: That is a perfectly valid question. You could try to answer it yourself with some work with databases, but it would be quite a lot of work.

  19. vernaculo:

    Anthropogenic climate change is, for all its gravity, only one negative in the list of changes caused by the personal automobile.
    The single leading cause of death for people under 30 in the US is traffic accidents. The war in Iraq, the threat of war in Iran, the bloodshed in Chechnya and Ossetia – are all petroleum-based misadventures.
    And it needs saying, and repeating, that the automobile industry is the largest industry in the US, with oil hovering right behind it. The US burns around 380 million gallons of gasoline each day; co-incidentally the world is dumping 380 million gallons of petroleum into the sea every year.
    I understand that science must hold to standards that preclude rhetoric and moral distinction, but the social context of the debate is what’s preventing the information, the virtually unanimous accord that anthropogenic climate change is here, from being acted on, now.

  20. Benton Maples:

    tom – I would argue that “significant fraction” could be construed quite differently than “most” which clearly implies a majority.

    John – In the first paragraph, Oreskes states: “Such statements suggest that there might be substantive disagreement in the scientific community about the reality of anthropogenic climate change. This is not the case.”

    Thus the thesis framed out in the opening paragraph is on issue #1, not issue #2.

    The impression is definitely left that the research showed consensus on #2, but it doesn’t unambiguously state that…Which makes me wonder if this is misleading, and the research actually focused on the claim she sets out in the opening paragraph.

  21. Benton Maples:

    Maybe someone with more knowledgable here can help me out with this…

    Oreskes states:

    “In its most recent assessment, IPCC states unequivocally that the consensus of scientific opinion is that Earth’s climate is being affected by human activities: “Human activities … are modifying the concentration of atmospheric constituents … that absorb or scatter radiant energy. …[M]ost of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations” [p. 21 in (4)].”

    Going directly to the IPCC report she quotes on page 21 I find:

    “Human activities (primarily burning of fossil fuels and changes in land cover) are modifying the concentration of atmospheric constituents or properties of the surface that absorb or scatter radiant energy. The WGI contribution to the TAR (Climate Change 2001: The Scientific Basis) found, “In the light of new evidence and taking into account the remaining uncertainties, most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations.”

    I don’t get it…Where does the IPCC ever “unequivocally” claim that this is the consensus of scientific opinion? It seems to me that they never make this claim. The claim #2 appears to be a consensus opinion among WG1, but I can’t find any suggestion that the IPCC asserts it to be a universal consensus.

    Can anyone?

  22. J. Sperry:

    The following is copied from Dr. Roy Spencer’s article on the subject:

    In her Science editorial, Ms. Oreskes also makes a curious claim about past research on “climate change”: that of 928 climate research paper abstracts published from 1993-2003, none rejected the consensus view on climate change. While I doubt that I’ve read this many climate change papers, I do have several in my office that specifically state that quantitative estimates of global warming are not possible without further knowledge of certain elements of the climate system (e.g. Renno, Emanuel, and Stone, 1994; Grabowski, 2000) or that current climate models are overly sensitive (e.g. Hu, Oglesby, and Saltzman, 2000). And remember, the consensus view Oreskes refers to is so qualitative and innocuous that few scientists would dispute it anyway.

    [Response: Oreskes says that she searched on the keywords "climate change" (subsequently corrected (see Quark Soup) to "global climate change"). Spencers papers don't include those. See also here for more detail - William]

    “Consensus” among scientists is not definitive, and some have even argued that in science it is meaningless or counterproductive. After all, even scientific “laws” have been disproved in the past (e.g. the Law of Parity in nuclear physics). Global warming is a process that can not be measured in controlled lab experiments, and so in many respects it can not be tested or falsified in the traditional scientific sense. Nevertheless, I’m willing to admit that in the policymakers’ realm, scientific consensus might have some limited value. But let’s be honest about what that consensus refers to: that “humans influence the climate”. Not that “global warming is a serious threat to mankind”.

  23. Bill Hicks:

    Eric, Thanks for another great post. I am a plant ecologist who studies carbon and nitrogen cycling and I have been yearning for a site/blog this. Kudos to you all. Your arguments are simple, informed and persuasive. Now to sit back and giggle at the grasping at straws of the uninformed. Time for some popcorn.

  24. Emmanuel M:

    In my opinion, the question is not whether human industry increases atmospheric CO2 or not (it does), but whether this will increase the earth’s temperature significantly, knowing that fossil fuels reserves are limited and our infatuation with them will end in about 50 years. We should be more concerned about adapting to climate change, natural or anthropogenic, than about its causes.

  25. Michael Tobis:

    #24 is entirely off topic to the present question of existence/non-existence of a scientific consensus. However it seems to represent two common misperceptions which I think can be addressed briefly here.

    First, while our reliance on petroleum is significantly supply-limited, in fact the dominant component of the problem going forward is the coal reserves, which can supply the world’s energy demands at the present rate for some centuries. If coal usage is never constrained by environmental policy nor limited by emergence of less expensive energy sources, the resulting climate disruption can be expected to reach catastrophic levels. While gas and oil contribute significantly to contemporary greenhouse forcing (as, for some reason, does manufacture of cement, as well as agricultural methane emissions from livestock and rice) the biggest part of the foreseen problem comes from coal.

    Also, as for the conclusion that we should be indifferent to the source of climate change, that presumes that such change is sufficiently small. There is, I can argue, nothing on the horizon other than vigorous international policy response that will prevent rate of change of climate from continuing to increase until it eventually overwhelms our adaptive capacity. To reach the conclusion that we can be indifferent to the source of the change, you must first understand the quantitative arguments driving the concern over anthropogenic climate change and then make a convincing risk-weighted argument that this expectation of accelerating change is not a serious concern.

    Since the contributor of comment #24 seems unaware of the size and nature of fossil fuel reserves, it seems unlikely that such a quantitative argument has been formulated as yet. I therefore hope the poster of #24 will reconsider the stated opinion.

  26. Stephen Berg:

    J. Sperry, you cite Roy Spencer in your message, saying that the view that the current episode of global warming is caused by human activities is not the consensus view. There could have been many better sources of information than Dr. Spencer.

    Dr. Spencer is a member of the Heartland Institute, which has received money from the fossil fuel industry (ex: ExxonMobil, to the tune of almost $350,000 from 1998 to 2003, source: (A further description of Spencer and the Heartland Institute can be found at the site given above.)

    Most of you realize how much the fossil fuel industry has been obstructing governments from actually getting something done to fight climate change. Such influence (especially on the current administration in the US) is causing the problem to become more and more impossible to solve.

  27. John Laumer:

    Consensus has two “frames”: one, the accuracy and precision of ‘climate science’; the other, the reasonableness of impacts projected. Mass media get these confused and interwoven to the point where meaning is lost, the public disengages, and wingnuts are given an opening in which to mock serious thinkers. That may well be an outcome that the masters of denial are avidly pursuing. ALthough it seems that a tipping point is rapidly being approached for demonstrting a consensus of peer reviewed science, we must keep in mind that corporate interests control all mass media. My main point is to encourage you to carefully frame and give context as often as possible, something that does seem to happen so far. Keep up the great work.

  28. Benton Maples:

    There’s also two separate issues – the science, and the societal take on the science.

    Unforunately, this debate has gotten so intertwined with politics that it has become a polarizing issue with preposterous propaganda from both sides.

    As a layperson, when I read a peice like Oreskes’, and try to verify what she is claiming, I’m forced to conclude that her piece is not only misleading, but more frankly a propaganda piece masquerading a science. I cannot understand how she claims that the IPCC asserts “uneuqivocally” that it’s summary reports represent a consensus view of scientists at large, or why she extrapolates the consensus beyond the one she claims in the thesis (and the one Science refers to in it’s summary of her study.) While it is certainly believable that these papers unanimously assert an anthropogenic factor to warming, it seems silly to me that they unanimously explicitly or implicitly (using _any_ subjective criteria Oreseks chose to apply) support the idea that a majority of warming is do to greenhouse gases.

    Further, I don’t find the attacks on every dissenting scientist because of loose ties to foundations which have been supported by oil interests to be particularly helpful or enlightening.

    It’s become a polarized debate with loads of propaganda from both sides, and attempts by special interests to tarnish reputations of those that don’t present the view of science they want.

    I do think there is large consnsus that burning fossil fuels increases greenhouse gases, and that this increase has led to temperature rise lately – I even think a big majority would agree that it is mostly due to greenhouse gas increases.

    In my mind, the real issues center on how to improve existing models (and underlying assumptions) of both projected temperature rise and projected impact. I think there’s a lot of good work being done on this.

    Unfortunately, I think the science has entered a political arena where fame can be obtained by being an outlier on either side, and grants from like-minded interests will be forthcoming. The net result of this could easily be a loss of faith in science by the public – not a loss of faith in the scientific method – but a loss of faith in the practicioners.

  29. John Humphreys:

    Stephen Berg, the source of funding doesn’t make somebody right or wrong. Otherwise, we should simply ignore every person working for the government (including academics). :)

    It seems that there is a consensus that something is happening, and no consensus on exactly what. So long as both sides continue to talk at slightly cross purposes, both sides can continue insisting they are correct and their adversary is lying.

  30. Tom Rees:

    I’ve found a paper that disagrees with the consensus that the recent warming is mostly anthropogenic: “The simulation implies that the solar part of the forcing, alone, would account for 71% of the global mean temperature variance, compared to 51% for the greenhouse gases part, alone.”

    OK so it’s only one out of several hundred, but I’m playing Devil’s Advocate here :) Does anyone know of a critique of this paper?

    [Response: first, note that this paper wouldn't have shown up in Oreskes survey: it doesn't feature the keywords "(global) climate change". Second, there has been a lot of research since 1996 and techniques rather more sophisticated than simply correlating one series with another. See IPCC, ch 6, section11 for more - William]

  31. Robert Coté:

    I did a quick Google search on the phrase “flat earth conspiracy” and got 179 hits. It appears the the consensus is that there is a flat earth conspiracy. But unlike Oreskes I did contra-test. “Round earth conspiracy” yielded 17 hits. Those however supported the same consensus that indeed the world is flat.

    What we do know (consensus) is 6.2 Gigatons of carbon are introduced into the atmosphere by the burning of fossil fuels (and production of cement) annually. This represents 0.8% of the annual 778.8 Gigaton global total from all sources. It is also consensus that 0.8% is well within measurement error and far below natural variability.

    [Response: 0.8% is well within measurement error, but thats irrelevant. Your argument (human emissions are much less than natural and therefore must be irrelevant) is an argument from the wilder shores of skepticism. The basic answer is, natural fluxes are in balance and human emissions disturb that balance. The recent post by Eric demonstrates why the rise in CO2 levels is human-caused. If you want more, try looking at a graph of CO2 levels, e.g. here. Do you really think that huge recent spike is natural? And is it just coincidence that it matches recent human emissions? - William]

    Point being that recent literature searches for keywords are not probative regarding endogenous climate influence. A similar time specific search of the mid 1970′s would have revealed a scientific consensus that global cooling was occuring and that it was human activity that was causing it.

    [Response: this is wrong, and the link in the next post demonstrates it - William]

    All this exercise reveals is that the methodology of literature searches needs more work. Oreskes should have been able to easily find several papers with contrary opinions and then should have investigated why the original search revealed none of them. Rather than the term â��global climate change” how about “anthropogenic uncertainty?”

    The existence of climate change is a non-issue. We used to call it weather but you cannot get a peer reviewed article published for the assertion that there has been weather and will be weather and it will change so instead we get these incredible leaps of logic that since we agree there is weather then we can predict whether it will rain next Tuesday.

  32. caerbannog:

    A similar time specific search of the mid 1970�s would have revealed a scientific consensus that global cooling was occuring and that it was human activity that was causing it.

    If I were a professional climatologist, my reply to you would be, “been there, done that, worn out the t-shirt”. See for more.

  33. Dennis Bray:

    In an article recently submitted but not accepted by Science I reported the results of survey of climate scientists conducted in 2003. One question on the survey asked “To what extent do you agree or disagree that climate change is mostly the result of anthropogenic causes? A value of 1 indicates “strongly agree” and a value of 7 indicates “strongly disagree”. Such a measure excludes the possibility of editor bias in the selection of publications (as it asks scientists directly) as suggested in previous comments on Oreskes’ article. Countries, and number of responses from each country are as follows:

    USA n = 372
    Canada n = 14
    Germany n = 56
    Italy n = 14
    Denmark n = 5
    Netherlands n = 4
    Sweden n = 5
    France n = 5
    U.K. n = 18
    Australia n = 21
    Norway n = 3
    Finland n = 3
    New Zealand n = 6
    Austria n = 3
    Ethiopia n = 1
    South Africa n = 3
    Poland n = 1
    Switzerland n = 7
    Mexico n = 3
    Russia n = 1
    Argentina n = 1
    India n = 3
    Spain n = 2
    Japan n = 3
    Brazil n = 1
    Taiwan n = 1
    Bulgaria n = 1

    To the question posed above there were 530 valid responses. Descriptive statistics are as follows:

    Mean = 3.62
    Std. Error of mean = .080
    Median = 3.00
    Std. deviation = 1.84
    Variance = 3.386

    1 strongly agree 50 (9.4% of valid responses)
    2 134 (25.3% of valid responses)
    3 112 (21.1% of valid responses)
    4 75 (14.2% of valid responses)
    5 45 (8.5% of valid responses)
    6 60 (10.8% valid responses)
    7 strongly disagree 54 (9.7% of valid responses)

    These results, i.e. the mean of 3.62, seem to suggest that consensus is not all that strong and only 9.4% of the respondents “strongly agree” that climate change is mostly the result of anthropogenic causes. This is however, a slight rise in consensus of the same survey conducted in 1996 that resulted in a mean of 4.1683 to the same question (Five countries – USA, Canada, Germany, Italy, and Denmark only in 1996 survey, N = 511). In the 1996 survey only 5.7% of the valid responses “strongly agreed” that climate change is mostly the result of anthropogenic causes.

    [Response: this is an interesting survey, and I think I took part... there are three concerns though: (1) self selection bias and (2) possibility of multiple returns from the particularly... avid (3) possibility of returns from non-(climate-)scientists. Do you address those? - William]

  34. Dennis Bray:

    Response to comments made by William

    The 1996 survey was a mail out survey to only climate scientists in the respective countries. Lists of scientists were obtained from various organizations i.e. American Meteorological Association, the German equivalent, etc. Original copies were returned so there was no possibility of duplicate submission. The 2003 survey was conducted on line with password protection so as to limit access to only those who received notification. Notification of the survey was made though various mail groups i.e. Climlist, BAMS, etc. I am not certain if self selection refers to bias in sample selection but if so, in the 1996 survey the selection was randomly drawn from membership lists after the affiliation and activity of the respondent had been determined to represent the climate science community. In the 2003 on-line survey, this, of course, was not possible. It would have been possible should data protection laws allow mailing lists to be distributed but in many countries this is not the case. Multiple returns were not possible in the case of the mail out survey as the original brochures were returned in prepaid self addressed envelopes. It would be possible however, should one wish, to submit a duplicate return in the 2003 survey. 2003 surveys were checked for identical response patterns and none existed. Of course this does not mean that a single person could not fill out two surveys using different responses. However, this applies to both alarmists and sceptics. Given the patterns of change in the mean responses between the two surveys, all indications would be that should there have been attempts to sway the result by multiple submission (which I doubt) it would have been the alarmists that eventually won in the number of submission sent as the data indicates consensus has indeed risen. Of course, it might have risen more. As the sceptic – alarmist battle seems to rage the most fierce in the US I could simply exclude the US sample from the analysis or select a random sample from the US responses, neither process of which I have yet attempted. As for non – climate scientists, the survey does contain some demographics pertaining to disciplinary and sub-disciplinary areas the respondent works in, what training the respondent received etc. and on this account I can only assume the responses to mostly represent the truth. One other check was employed in the construction of the survey (to the annoyance of some respondents) and that was to fluctuate the order of the responses, for example, a response of 1 indicating a positive response for some questions and a response of 1 indicating a negative response in other questions. This is standard procedure in questionnaire construction and attempts to identify those cases that simply checks all right hand response options on the assumption that they are positive (or negative) without giving consideration to the question being posed. Fortunately, no completed surveys followed this pattern, suggesting at least that all respondents acted in a responsible manner when submitting their responses. As with any measurement operation in any scientific endeavour, measurement is subject to error. We do the best we can.

    [Dennis: thanks for your response. By "self selection" I meant that certain groups might preferentially respond: though it would be hard to know if this would produce a pro/anti bias. You said that the 2003 survey was conducted on line with password protection so as to limit access to only those who received notification but there was nothing (I think) to prevent someone passing out the password to interested friends, so it would have been possible to "farm out" responses - William]

  35. eric:

    The comment by Dennis Bray above is interesting and the greater detail than in Oreskes’s survey is useful. However, the question asked by Brey is a very ambiguous one (which may be one of reasons the paper was rejected from Science), because the answer depends greatly on what timescale one is talking about. Indeed, my own inclination would be to answer this question in the negative. Oreskes’s essay reports the IPCC statement that “most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations”. I expect Bray would have obtained different results had his survey asked for agreement or disagreement on this.

  36. B.W. Emanuel:

    In reference to #32 above, this from the popular literature, Newsweek, April 28, 1975, “The Cooling World.”

    There are ominous signs that the Earthâ��s weather patterns have begun to change dramatically and that these changes may portend a drastic decline in food productionâ��…

    The evidence in support of these predictions has now begun to accumulate so massively that meteorologists are hard-pressed to keep up with it.

    Meteorologists disagree about the cause and extent of the trend, as well as over its specific impact on local weather conditions. But they are almost unanimous in the view that the trend will reduce agricultural productivity for the rest of the century. If the climatic change is as profound as some of the pessimists fear, the resulting famines could be catastrophic. �A major climatic change would force economic and social adjustments on a worldwide scale,� warns a recent report by the National Academy of Sciences, �because the global patterns of food production and population that have evolved are implicitly dependent on the climate of the present century.�

    Climatologists are pessimistic that political leaders will take any positive action to compensate for the climatic change, or even to allay its effects.

    [Response: globalclimate have got their transcription of the newsweek article somewhat wrong (nothing serious, just enough to allow tracking) - see here for a correct transcription and some comments. The main point though is that newsweek is not a science journal. I don't know if you noticed that. Newsweek *did* manage to reference the 1975 NAS study (see here) which says, correctly for the time, we do not have a good quantitative understanding of our climate machine and what determines its course. Without the fundamental understanding, it does not seem possible to predict climate (from the foreword).

    And the conclusions of the NAS report are worth reading. In summary they are:

    1. Establish National climatic research program
    2. Establish Climatic data analysis program, and new facilities, and studies of impact of climate on man
    3. Develope Climatic index monitoring program
    4. Establish Climatic modelling and applications program, and exploration of possible future climates using coupled GCMs
    5. Adoption and development of International climatic research program
    6. Development of International Palaeoclimatic data network

    Which was all very sensible (again for the time): more research was needed - William]

  37. B.W. Emanuel:

    Of course I know Newsweek is not a science journal. That is why I was specific to characterize it as the popular literature. The link to the article was the first to come up when I did a Google search of “global cooling.”

    What I think interesting as a layman and non-scientist is what the article was actually conveying to the general public at the time: an impression that there was consensus in 1975 that the climate was cooling, a period in which the communication of such things to the general public was chiefly limited to newspapers and newsmagazines.

    [Response: err, well, OK. But please don't assume that Newsweek actually was reporting the "scientific consensus" of the time. If you want the sci opinion, you're better off with the NAS report. And since you were responding to comment #32, you should have followed the link therein. The wiki article is useful, too. Its #4 on the google search. Happily for you, nowadays the IPCC report is online and you don't have to rely on Newsweek - William]

  38. B.W. Emanuel:

    I assure you, William, I don’t rely on Newsweek for my science, otherwise I wouldn’t be here.

    [Fair enough. The reason this blog exists is because we think the news media don't do a good job of explaining the real story - William]

  39. William:

    The Washington Times has a good piece about Oreskes work:

  40. Jeff Guinn:

    It seems Oreskes, and many others, missed what appears to me to be two thorough papers on the subject:

    Conclusion: 1) Yes, there is antrhopogenic forcing. 2) But so small as to be nearly inseparable from the background noise.

    [Response: I can't see any papers on the link you point out. You're not mistaking the dubious web articles Dr Dewpoint wrote for "papers" in the scientific sense, are you? If you're referring to his post on the UHI, then you need to read the RealClimate article here - the Peterson paper (no UHI effect on the T record in the USA) totally supercedes Dr D - William]

  41. B.W. Emanuel:

    [Fair enough. The reason this blog exists is because we think the news media don�t do a good job of explaining the real story - William]

    Then you agree with my point.

    [Response: I'm not sure what point you think I'm agreeing with... if you're arguing that the page you mention comes up first in a google search, then I agree. If you're arguing that the Newsweek article is useful to assess the state of science at the time, then I disagree - the NAS report proves that - William]

    Plus ca change, plus c’est la meme chose

  42. George E. Smith:

    Well ALL scientific advances are the result of disagreement with the ‘concensus’ viewpoint.

    Science is about what you can prove to be true; not about opinion polls which is what concensus really is.

    According to a mini-bio article in my local newspaper (San Jose Mercury News) about a certain NASA climate scientist; she gained peer notice, when she “upset 60 years of climate science” (concensus) when she proved that a liquid rain drop freezes from the outside in, rather than from the inside out (the prevailing concensus view). So much for concensus.

    [Response: if you provide her name and/or a working link, I can check. Until then: "don't believe everything you read in the newspapers, especially if it supports your prejudices" - William]

    But what is particularly troubling about this concensus view, is the simple fact that any eighth grade science student could tell you that the second law of thermodynamics would prohibit a rain drop from freezing from the inside out. Evidently this famous young lady didn’t even know that either, since her proof is quite arcane.