by Andy West
Climate culture versus knowing disbelief.
In a previous post at Climate Etc I showed two analyses on US public attitudes to climate change, based upon data from psychologist Dan Kahan’s studies plus some independent surveys. Both of these seek to explain what social / psychological mechanisms are driving the observed attitudes. The first analysis is Kahan’s own, which concludes that identity defense by adherents of particular political views / parties is the chief mechanism explaining the data, of which ‘knowing disbelief’ is the strongest form exhibited by those Conservatives / Republicans who are science aware. The second analysis, mine, demonstrates that the concept of a ‘climate culture’ provides a much better fit to the data, a culture that has adherents in its own right plus asymmetrical alliance with politics; also that Kahan’s conclusion is largely a product of his own bias due to a major influence from this same climate culture.
For some time before the above post was published (Jan 30th), I’d been looking out for a particular kind of survey that ought to provide significant evidence supporting either one or other of the above analyses. Unfortunately there seemed to be no such survey measuring the group I was interested in, so I had to run without this. However I noticed in June that a Gallup poll (G1) had appeared at the end of March, which while not ideal does measure a superset grouping in an appropriate enough manner to provide useful insight. This new poll data matches closely what the ‘climate culture’ hypothesis expects, and I believe it strongly challenges the ‘knowing disbelief’ hypothesis. While G1 covers respondents from the full US political spectrum, the most insightful data comes from the Independents. The next section is thus helpful context for understanding this data. (Note: for new readers wanting a short-cut, or prior readers wanting a refresher, there’s a compressed summary of Kahan’s theory versus mine in the ~700 word Appendix).
A brief characterization of US Independents
In an interesting 2012 Washington Post article (WP1), journalist and political analyst Linda Killian says: ‘In hundreds of interviews with independent voters, I found that they tend to be well informed and care about the political process — even though the two parties have done their best to alienate them through attacks, gridlock and dysfunction.’ And later: ‘Sixty percent of independents say they are not aligned with a party because they agree with the Republicans on some things, such as the economy and national security, and with the Democrats on social issues.’
Another Gallup poll (G2) from 2013 claims that 42% of Americans identified as Independents, of which figure 16% percent generally lean towards the Democrats and also 16% lean towards the Republicans, so have partial predictability. This leaves 10% of ‘centrists’ whose allegiance is still less consistent and will always swing on particular issues. However, the WP1 article claims it is ‘a myth’ that most Independents are ‘leaners’, and that in fact about half are ‘truly independent’. A Pew pole quoted in a CNN article (C1) about Independents, approximately agrees with the figures in G2, yet whether it is a quarter or a half of Independents who are ‘leaners’, everyone seems to be agreed that Independents generally are much less partisan, and are driven away from the main parties in part by negative tribalism:
The CNN article C1 says: ‘The shift away from partisan affiliation has occurred during a sustained period of government distrust and distaste for partisan politics. In the last year, negative impressions of government have displaced the economy atop Gallup’s monthly measure of the nation’s most important problem.’ The article WP1 says: ‘Independents are more turned off than partisan voters by negative campaign ads and are more likely to say they want more substantive discussions from the candidates and the media. Independents take voting seriously but are less moved by partisan appeals.’ Poll G2 says: ‘The rise in political independence is likely an outgrowth of Americans’ record or near-record negative views of the two major U.S. parties, of Congress, and their low level of trust in government more generally.’
The new survey data (from G1)
The first task is to register this data against what we’ve seen before, given that ‘what is said in the news’ about global warming is rather fuzzy, as are other terms. For instance opting for ‘exaggerated’ doesn’t necessarily mean a disbelief in MMGW, although likely means a disbelief in calamitous narrative. And having some college education is a superset characteristic that won’t always mean being science aware. However we are chiefly looking for polarization here, which distances us rather from absolute meanings, and both the Republican and Democrat responses show a strong correspondence to previous findings; the enhanced polarization that Kahan reports for the science aware is clearly reflected for the higher educated, albeit not quite to the same extreme. More education for the Republicans leads to more skepticism, more education for Democrats leads to more belief in MMGW and more belief in dangerous scenarios, in this case a belief that (on average) the news underestimates the seriousness.
So far so good, this is what both Kahan’s model and mine expects. And performing a common sense comparison of the RH column, ‘generally underestimated’, with those who place global warming as a top priority (by political stripe) in this Pew survey cited in the previous post, yields a pretty fair fit that helps orientate us in belief space. (RH column average after rough college attendance profile per wiki ~ R17, I31, D45, Pew ~ R14, I27, D42). Hence we can go on to examine the responses of the Independents with relative confidence, including the attitudes of higher educated Independents, which is the element that I couldn’t find before.
What the higher educated Independents tell us
The attitudes of the Independents also change significantly with education. Overall the effects move in the same direction as for the Republicans, i.e. those with more than high school education are more skeptical. On average 23% more in the ‘exaggerated’ column, and 16% less in the ‘underestimated’ column. I think this modest yet clear movement is a major challenge for the theory of ‘knowing disbelief’, which theory (see the previous post or the Appendix) hinges upon powerful identity defense to explain the very same movement for the Rep/Cons, i.e. a tribal allegiance to the Republican party and conservative principles. But from section 2 we see that these individuals are more flexible or split on principles, and that partisan politics (from Republicans and from Democrats) does not turn them on. They do not actually have a Republican identity to defend!
I can only assume what Kahan would predict for higher educated Independents. Given that they’re the least partisan of all voters and so the least blinded by party loyalty, one would think that their education would guide them more towards the truth. But they certainly do not move towards the ‘truth’ that Kahan has pinned the pole of the science aware Dem/Libs to. They move towards skepticism, i.e. in the opposite direction.
What does the theory of climate culture predict? Given the symmetry of the Independents as noted in section 2, a plausible start is to assume this same symmetry on the particular issue of climate change. I.e. the Independents are formed from two approximately equal populations, one adopting similar attitudes to the Democrats and the other adopting similar attitudes to the Republicans (assuming that on average the ‘centrists’ fall 50:50). So the ‘High school or less’ row for instance, should be about an average of the Dem and Rep rows for this same category, which comes out at GE42, GC23, and GU32. That’s pretty close, it seems a reasonable assumption. But the key test for ‘climate culture’ is a prediction that higher educated Independents will be led into a path of either climate orthodox comprehension or skeptic comprehension depending upon their initial leaning, in exactly the same manner (so the same strength / proportion of attitudes) as for the Reps and Dems. Hence the same simple average should hold for the higher educated Independents too. For instance in the graduate category, this yields GE45, GC25, GU30, once again a good match.
Now this result may seem intuitive to some readers, perhaps even blindingly obvious. Yet if one follows Kahan’s ‘knowing disbelief’ theory it should be completely counter-intuitive. For a survey like this with a centre position, in the absence of both higher education (including science literacy) and any strong party loyalty, there ought to be clustering within that centre, because in Kahan’s model there is no other strong influence here that would cause polarization or indeed any main departure from a centrist view. Yet the centre is lowest in all of the Independent rows; these shallow ‘V’ shapes should be inverted, which would completely break my match of Independents with simple Rep/Dem averages. And where there is higher education in this non-partisan population, as noted above Kahan’s model would surely predict a strong movement towards climate orthodoxy. (Note: the only way Kahan could similarly claim a blended effect from two populations is if each were as just as powerfully partisan as the Reps and Dems, yet we know this is not at all the case. In my own model the combined effect is via biased assimilation from a modest starting stance, with one population led much more into climate culture, the other led away).
This data from the Independents adds to the evidence that attitudes on climate change in the US are not primarily a simple reflection of partisan Republican or Democrat identity. (Note: the text with the G1 poll itself does not concentrate on the Independents. While correctly noting that education for the Republicans ‘leads them in a different direction’ to Democrats, it incorrectly draws the conclusion that the dominant effect overall must therefore be ‘partisanship’).
The critical difference between models
Kahan’s model assumes the simple reflection above because it does not acknowledge a third influence in addition to Rep/Con and Dem/Lib influences. People’s attitudes are governed by a belief in, or cultural alliance with, or lack of belief in, climate culture itself. Their position on the left-right political axis is only a good indicator of where folks will likely sit on this issue, because en-masse the two main political parties in the US have either a largely supportive (Dems) or largely resistive (Reps) relationship with climate culture. So Kahan’s model is two way upon a left / right axis, and mine is three-way (i.e. Dem/Lib culture, Rep/Con culture, and climate culture) on a conceptual map like below, from the previous post:
This model also better explains edge cases that are difficult for Kahan’s model. For instance the minority of higher educated Democrats and Republicans (to a lesser extent this occurs for the science aware too) who are opposite to their respective party trends. While I guess this is not impossible to address within Kahan’s framework, it’s certainly a serious challenge to provide a robust and generic explanation, given that their political allegiance pulls in the opposite direction to their views about dangerous MMGW. Yet climate culture is a strong independent influence that in some folks will overwhelm any political stance; and likewise some who resist this strong culture can come from the political left as well as the right.
The view that Kahan and many others hold regarding attitudes to climate change in the US tends to boil everything down to a left / right tribal political influence, and this view masks what’s really going on. For instance that there are far more Democrats (‘the shifters’ above) who don’t believe the orthodox narrative of dangerous MMGW, than do. They are merely allied to this narrative for the sake of their identity and party, yet never place climate policies high on the agenda. In some surveys these shifters outnumber the full climate culture adherents who are also Democrats, by almost 2 to 1.
Kahan’s model is also challenged in countries outside the US where climate change has not become so mixed up with political identity. For example the UK, where all the main parties support policies to fight man-made climate change, yet nationally there is still significant skepticism. In the diagram above this is represented by shallower left-right lines, yet still a significant block at the bottom of the graph. In other words, the adherents climate culture has claimed are more evenly spread across the political spectrum than is the case for the US, as are those that resist the culture too. This is clearly a problem for a model that doesn’t acknowledge a climate culture in the first place, and essentially places ‘disbelief’ only at the door of right-wing identity defense; it doesn’t seem plausible that fundamentally different mechanisms are driving attitudes to climate change in the US than in the UK (or elsewhere).
The attitudes of non-partisan Americans shed light on an issue that is often perceived as chiefly partisan on the left / right political axis. Especially the subtle yet significant attitude shift for the higher educated Independents. These support the case that the true situation is more complex, and that a third cultural pole is involved.
If ‘knowing disbelief’ is why many millions of science aware Rep/Cons resist the narrative of dangerous MMGW much more even than Rep/Cons generally, a theory founded on a subconscious apprehension of danger plus potent identity defense of right-wing politics / principles, then apparently about half of higher educated Independents must suffer from this very same problem too. This is despite these least partisan of US voters not actually having a Rep/Con identity to defend in the first place. ‘Knowing disbelief’ is not a plausible explanation.
Caveats: One small survey doesn’t make a summer, and these are rough and ready figures to get an idea of what’s going on. Plus a difference effect between two populations of Independents is bound to be modest; it seems not much more than twice the natural slop in the figures (all Independent cells are at most 3 points from the Rep/Dem average bar the middle left, which seems a slightly high outlier). Nevertheless, this survey shows with relative confidence what is very unlikely to be going on. As noted in section 4 the Independents’ responses would have to be very different indeed to match Kahan’s model. A high quality survey of the science aware / climate science aware Independents would be of great value in exploring this angle further.
JC note: As with all guest posts, please keep your comments civil and on topic.