Here’s an interesting study by Lawrence Hamilton, published in Climatic Change in December: “Education, politics and opinions about climate change evidence for interaction effects“. He finds that whereas in the past, education level strongly correlated with concern for environmental issues, this correlation has now gone away because of a new interaction effect with political beliefs. (Which is consistent with other recent research, e.g. see the work of Kahan and Braman).

From his survey of people in New Hampshire and Rural Michigan, Hamilton identified strong Democrats and Strong Republicans, and discovered that, with respect to climate change, there is a significant interaction effect between education and party support, and another between how well people believe they understand climate change and party support:

Hamilton 2010, Fig 3: Predicted probability that Upper-Peninsula Michigan residents believe that global warming will pose a serious threat vs. respondent education, for “strong Democrats” and “strong Republicans”

Hamilton 2010, Fig 2: Predicted probability that New Hampshire residents believe that global warming will pose a serious threat vs. self-assessed understanding of the issue, for “strong Democrats” and “strong Republicans”

[I can't help noticing there's no data points for "Strong Republicans" who say they don't understand global warming at all!]

Here’s what Hamilton has to say about the findings:

“The inconsistency marks a social shift away from patterns seen in older research. It reflects the efficacy of media campaigns that provide scientific-sounding arguments against taking climate change seriously, which disproportionately reach educated but ideologically receptive audiences [...]. Among many educated, conservative citizens, it appears that that such arguments have overshadowed the scientific consensus presented by the IPCC reports and other core science sources.”

Hamilton puts a significant part of the explanation for this shift on internet and cable TV as sources of information, which increasingly allow people to tune into only those sources that they find ideologically compatible, and the ability of these sources to propagate politically inspired but scientific-sounding arguments:

“The effective dissemination of contrarian arguments means that many people who have no contact with climate scientists or the primary research literature can nevertheless learn that a scientist says temperatures have risen on Mars (politically spun as evidence that global warming has solar or cosmic origins), or another scientist says it is cooling in East Antarctica (spun as evidence that our planet is not warming after all). They might consider themselves well informed about climate science even while not understanding its basic ideas.”

And he concludes that having more scientists getting involved in blogs and rapid response initiatives is crucial:

“If non-specialists want to find out what scientists really know about temperature trends of Mars and East Antarctica, or other arguments aired in today’s news or last night’s party, they are best served by a relatively small number of active-response Web sites written by climate scientists, such as Unlike journal articles, science meetings or reports, Web sites and blogs have the capability to react quickly (albeit less rigorously), reach broader audiences, and seriously confront arguments that have no scientific merit. Moreover, their online science posts can be passed on from reader to reader, which is difficult to do with journal articles or technical reports.”

[Hat tip to Sol for sending me this]

Update: It appears to also matter whether it’s called “Climate Change” or “Global Warming”, at least to Republicans, who are more likely to believe in the former rather than the latter.


  1. @sme: “an interesting study by Lawrence Hamilton

    Thanks for the open link! (I’ll save my rant about proprietary content in academia, if only because I suspect you can do it better.) Gotta read that.

    Regarding the empirical claim: his results are almost certainly not restricted to NH and the UP. IMHO the intensity of the “Hamilton effect” (i.e., the “interaction … between education[,] party support[,]” and evaluation of environmental threat) appears to vary inversely with latitude :-( One hopes it’s restricted to “faith-based” polities; or can one s/Democrat/Liberal/ and s/Republican/Tory/?

    Regarding the policy prescription: the most charitable assessment I can muster is, it seems merely hopeful, because (as he seems to note) “Strong Republicans” (SRs) will only see (e.g.) when it hits Glenn Beck’s blackboard. SRs simply don’t “want to find out what scientists really know about” global environmental damage, because it offends their notions of God, markets, or both. And it gets worse! per Nyhan and Reifler 2010: correcting SRs makes them worse! And if that’s true in Canada, … don’t tell me, I don’t wanna know …

  2. It’s worse than that. Strong Republicans don’t want to know because it offends their sense of economic truth and ideological limits to the power of the state. If we are truly running off he cliff, are their any solutions to be implemented that do not involve the expansion of government power? If not, they really, really don’t want to hear it.

    I side with Democrats and land up there with the post-grads, but I have to admit that any solution that requires a large expansion of government authority does not sit well with me. It won’t fly in the US with either party because the Democrats actually split a bit and that is enough to stop it. Those solutions simply won’t work here.

  3. @Alfred Differ: “Strong Republicans [(SRs)] don’t want to know because [anthropogenic global climate change (AGCC)] offends their sense of economic truth and ideological limits to the power of the state. [Are there solutions] that do not involve the expansion of government power? If not, they really, really don’t want to hear it.”

    From my American life experience, I must unfortunately agree: SRs refuse to recognize imperfect markets generally, and AGCC externalities particularly. (However I lack empirical cites to that effect: anyone got any?) That being said, attitudes can change, and policies can change even where attitudes do not. E.g., in the past 30 years, we (in the US) have gone from ubiquitous tobacco smoking and criminalization of homosexuality to criminalization of smoking (in all but the most private locations) to the nearly-full legalization (and occasional celebration) of adult consensual sex, even at the intersection of the bible and tobacco belts.

    In both cases demography helped (I believe, though I don’t have empirical cites for either gay rights or tobacco), and it may help with Hamilton’s effect. Table 2 (p6) of Hamilton’s paper claims age-vs-perceived-threat coefficients of −0.031 from the NH survey and -0.023 from the UP survey. 2 things to note (and thanks to Dr Hamilton for clarification) are (1) this is a coefficient not of probabilities (e.g., P(E) of event E occurring) but of odds (i.e., O(E)=P(E)/P(~E)); (2) the coefficent is an exponent, since this is a logit regression. Hence, for each additional year of age, the odds that an NH survey respondent perceives warming as a threat decrease ~3% (exp(-0.031) ~= 0.97), and the odds that a UP survey respondent perceives warming as a threat decrease ~2% (exp(-0.023) ~= 0.98).

    One specific question here is, is threat perception a static or dynamic function of age? If the latter (i.e., folks generally tend toward AGCC skepticism as they age), we’re screwed, presuming the relation between age and voting remains stable (i.e., the propensity to vote increases with age). But if the former (i.e., folks’ perception of AGCC threat is constant with age, and the threat perceptions of youth today are just better, normatively, than those of elders today), we can hope that our opponents just die off, preferably on a time scale compatible with cheap and effective mitigation.

    @Alfred Differ: “any solution that requires a large expansion of government authority does not sit well with me[, and] simply won’t work here.”

    I suspect it’s not that simple. A more general question is, what’s driving the differential threat perception WRT AGCC, and, for that matter, gay sex and tobacco smoking? I suspect (but cannot prove, and note my lack of empirical data for all of the following) that what relates all three are attitudes among US corporate elites, as transmitted through the media they own (e.g., all the major broadcasters and papers) and influence (notably, NPR and PBS). As Marx was among the first to note, capitalists can be quite socially progressive, at least when compared to advocates of prior social systems. In the US (confounding the oddly one-dimensional notion of “left” and “right” dominant in our discourse), corporate elites tend (c.p.) to be social libertarians, and hence have been able to advance equal (legal, not economic) rights for racial and sexual minorities over widespread opposition. (Not sure where the failure of the ERA fits in this analysis.) US corporate elites have also been generally receptive to improvements in public health; I suspect this has played a role in the eventual suppression of tobacco rights, notably over a minority of their own ranks.

    The parallel between the promotions of false controversy by tobacco elites then and fossil-fuel elites now is obvious. The question becomes, will US corporate elites (and therefore their media) see AGCC as a positive threat to their core fiduciary interests (as the insurers seem to be advocating), a non-negligible public health threat (to which they will respond as with tobacco, by advocating threat-suppression policies while seeking to socialize their costs), or as a normative threat to their core ideological interests (to which they will respond implacably, as to, e.g., Cuba and the Soviet Union)? We can watch as we warm :-(

  4. Alfred Differ

    @Tom Roche: Thanks for dealing with my poor homophone usage. There is good evidence that my performance in distinguishing them degrades rapidly after bed-time. 8)

    I can see the parallels with tobacco smoking here in the US, but I think the gay sex one fails to be a good analogy no matter what the numbers might say about it. There is a strong case for defending another person’s choice of sexual partners based entirely upon liberty. One can easily construct an argument for keeping government out of decisions that judge the morality of such acts that libertarians, leftists, and many on the right can tolerate. There is no reason to think that such an argument ould cause an immediate change to our social rules and institutions, so I think it is reasonable to view our growing acceptance of a liberty argument as a cause for our growing tolerance of homosexuality. I don’t have any numbers to back this up and I’m far from qualified to argue this is true, but I suspect it is from a variety of discussions I’ve had with people across the political and faith spectrum. I have lived in California for almost 30 years now and I think this is at the core of our battle over same sex marriage.

    If there is a similar liberty argument to be made for altering market rules to account for the negative externalities we know exist for AGCC, I don’t know it yet. I sure would like to know it too, because without it I personally know a lot of people who will argue against the expansion of US government regulation even at the risk of that approach being the only way we can think of to slow and halt AGCC.

    Laws forcing prices to cover the externalities aren’t enough in our democracy. The public has to believe at best or not care at worst. We don’t have either situation.

  5. Pingback: …My heart’s in Accra » links for 2011-03-09

  6. As a follow-up to this note, readers might be interested to know that we’ve just published a new report that supports the partisan/understanding/climate interaction effect in the second graphic above. The report applies lower-tech methods but looks at more and newer data –9,500 interviews within the past year. Figure 7 in the report suggests a possible weather effect.

    With respect to the partisan/education/climate interaction effect in the first graphic above, I’ve posted elsewhere a note showing its antecedent in a 2008 article (first figure in the linked page), and most recent manifestation in February 2011 data (last figure in the linked page):

    The various studies find broadly similar patterns across a range of different surveys, analytical methods, and climate-related dependent variables.

Join the discussion: