Skip to main content Accessibility help
×
Home

Information:

  • Access
  • Cited by 5

Actions:

      • Send article to Kindle

        To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        ‘It Isn’t What We Don’t Know that Gives Us Trouble, It’s What We Know that Ain’t So’: Misinformation and Democratic Politics
        Available formats
        ×

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        ‘It Isn’t What We Don’t Know that Gives Us Trouble, It’s What We Know that Ain’t So’: Misinformation and Democratic Politics
        Available formats
        ×

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        ‘It Isn’t What We Don’t Know that Gives Us Trouble, It’s What We Know that Ain’t So’: Misinformation and Democratic Politics
        Available formats
        ×
Export citation

Abstract

This article explores the dangers to the quality of democratic governance of those who are informed but disengaged and, especially, those who are engaged but use false ‘knowledge’. Poll data show the extent of Americans’ misinformation about, or disengagement with, climate change. The main responsibility for these problems lies with politicians, who have partisan incentives to help the disengaged become active, but also partisan incentives to keep the misinformed politically involved. Activity in accord with false ‘knowledge’ can slow needed responses to global warming and lead to concrete harm to individuals, communities and nations.

Footnotes

*

Department of Government, Harvard University (email: hochschild@gov.harvard.edu); Department of Political Science, Boston University (email: kleinst@bu.edu). We thank Robert Goodin and two anonymous readers for their excellent advice on this article.

Ever since political thinkers began to see democratic governance as an aspiration rather than a danger, they have insisted on the importance of a knowledgeable public for a successful democracy. The observation in the title has been variously attributed to Will Rogers, Mark Twain, Satchel Paige, Artemus Ward and others. One can easily find more rhetorically elevated versions of the same point; let the nineteenth century educational reformer Horace Mann speak for legions of philosophers and educators:

In regard to the extent of the education to be provided for all at the public expense,[…] it seems clear that the minimum of this education can never be less than such as is sufficient to qualify each citizen for the civil and social duties he will be called to discharge. [A democracy requires] such an education […] as is necessary for the voter in municipal and in national affairs, and […] as is requisite for the faithful and conscientious discharge of all those duties which devolve upon the inheritor of a portion of the sovereignty of this great republic. 1

Despite these calls for an educated public, the general public frequently falls far short of reformers’ democratic ideal. Again among legions of examples, consider only two:

  • A fifth of British teenagers believe Sir Winston Churchill was a fictional character, and more than a quarter say the same about Florence Nightingale. Conversely, three-fifths think Sherlock Holmes and King Arthur were real, and half say the same about Robin Hood and the Beatles’ Eleanor Rigby. 2

  • In a 2010 poll, Americans, on average, preferred to spend ‘only’ 10 per cent of the federal budget on foreign aid, a dramatic reduction in their view, given the average belief that foreign assistance currently takes up 25 per cent of the budget. The actual amount is considerably less than 5 per cent. 3

Political ignorance and misinformation persist in the face of strong efforts to dispel them. A Lexis-Nexis search shows an average of 1,732 articles per month in 2013 in American newspapers and wire services about the Patient Protection and Affordable Care Act (ACA), more commonly known as Obamacare. 4 In July 2013 there were 2,695 articles. Lexis-Nexis also identified 1,270 ACA stories in the first eight months of 2013 on the ten most prominent American television and radio news shows. And yet, a Washington Post poll found a surprising lack of public knowledge of the ACA. ‘“Never heard”, said Lenard Pringle, 54, a Greenbelt resident who said he was not aware of any changes coming to the health-care system. He did not know that in a few months, he will be eligible for Medicaid, the state-federal health insurance program for the poor that is expanding under the law’. 5 Even many Americans who had heard of the law misunderstood its tenets: ‘“I think if you have a serious illness, you should be entitled to a doctor,” said Joan Fernandez, 62, of Deep Water, Mo.,[…] “I don’t know how Obamacare would affect anything like that”’. 6

Although it is much less frequently studied, citizens’ possession of misinformation – holding apparently factual, but actually false, ‘knowledge’ – is a deeper problem for the quality of democratic governance than is mere ignorance of politically relevant facts. The politically ignorant tend to be politically disengaged and to hold no clear policy views, which is philosophically unfortunate but has little impact on governance. Even some people who hold factually correct information are also relatively disengaged, at least with regard to a particular issue; those people are frustrating both normatively and to political activists who seek their support. But the impact on democracy of both of these groups failing to ‘discharge […] those duties which devolve upon the inheritor of a portion of the sovereignty of this great republic’ is less severe than that of political views held (and actions taken by) people who hold factually incorrect information. Those people endanger the quality of governance in a democratic polity, and can even harm its members and people or countries outside that polity.

One can perhaps blame citizens for not knowing basic politically relevant facts – that Winston Churchill was real and Sherlock Holmes was not, or whether the ACA does or does not help poor people obtain health insurance or health care. But the more appropriate blame accrues to political actors who do not seek to move people into the realm of informed engagement, or who even seek to maintain them in the realm of misinformed engagement. Politicians’ incentives are asymmetrical; encouraging people to stay misinformed and politically active is often easier and produces larger pay-offs compared to encouraging them to obtain correct factual knowledge or persuading the disengaged to use their existing knowledge in the public arena. The combination of politicians’ incentives, misinformation, and political activity in accord with that misinformation can lead to irresponsible governance and bad policies, with potentially devastating consequences. 7

We develop this analysis by examining the American public’s views about global warming; the actions that Americans are (not) willing to take in response to it; and politicians’ behavior in the face of the mix of informed engagement, informed disengagement, and misinformed engagement. We then show how politicians’ asymmetrical incentives interact with citizens’ views when confronting the mammoth but amorphous risk of climate change. We make no claims about its representativeness as a policy issue, but it is intrinsically important and there is no reason to assume that the politics around climate change are unique. 8 We conclude with a brief discussion of a similar case in order to reinforce our claim that misinformed activity is dangerous for people who reside in (or are linked to) a democratic polity.

Global Warming

Over the past few decades, almost all experts have converged on the conclusion that the earth’s temperature is climbing, largely as a consequence of human activity. A variety of ecosystems already shows clear consequences – glaciers are melting and the sea level is rising – and, absent intervention, these and other changes are likely to continue at an increasing rate.

A significant share of Americans agrees that the earth is warming and that increasing temperatures are manmade. Surveys from the Gallup Organization and the Pew Research Center offer the most complete time series on this issue, and their results are remarkably similar. Between 1997 and 2014, roughly half of the respondents to fifteen Gallup polls agreed that ‘the effects of global warming have already begun’. Pew offers more depth through a pair of questions about global warming and its sources. Respondents were asked in eleven polls if there is ‘solid evidence’ that global warming has been occurring over the past few decades. If respondents answered yes, they were then asked if climate change occurs ‘mostly because of human activity such as burning fossil fuels’, or ‘mostly because of natural patterns on the earth’s environment’. From 2006–13, between 40 and 50 per cent of respondents correctly answered both questions. In other words, a large share of Americans possesses one or both essential pieces of information about global warming.

Moreover, a significant proportion of these individuals is, as Horace Mann would hope, engaged with as well as informed about climate change; they not only have correct factual knowledge but also hold policy preferences aligned with it. Gallup respondents were asked in 2012 about seven strategies to reduce energy use. 9 On average, about 70 per cent of informed respondents supported each proffered proposal. Agreement ranged from 53 per cent of the informed (19 per cent of the total sample) supporting more use of nuclear energy, to 79 per cent of the informed (28 per cent of the total sample) endorsing higher emissions standards for automobiles, business and industry. 10 Both Gallup and Newsweek asked a similar battery of policy questions in 2007. While some of the overall sample percentages differ (and are generally higher) as a consequence of different factual and policy questions, 11 the general trends hold: those with correct information largely support public policies designed to reduce energy use.

We contrast these individuals with three distinct categories of respondents. First are those who are simply unaware of the whole topic of climate change (‘never heard’). There are very few of them: an average of about 5 per cent across the two poll series. More politically important are those who are correctly informed about climate change but oppose policies intended to mitigate its effects; they are informed but disengaged from the appropriate policy responses. Their numbers range from 11 per cent of the correctly informed (4 per cent of the total sample) opposing higher emissions and pollution standards for businesses and industry to 55 per cent of the correctly informed (19 per cent of the total sample) opposing the expanded use of nuclear energy. 12 As with the actively informed, the overall sample percentages are higher in 2007 as a consequence of those polls’ looser criterion for factual knowledge. Nonetheless, the proportion of correctly informed respondents opposing environmental policies remains fairly similar across the two years for which there are good polling data.

Most politically important are those who are both misinformed and opposed to policy changes intended to conserve energy. An average of 25 per cent of Gallup respondents from 1997–2008, and 33 per cent since then, agreed either that global warming would not occur within their lifetimes or that it would never occur – and a substantial number of others concurred only with the statement that warming would start sometime in the foreseeable future. Moreover, among the half of the 2012 Gallup respondents that knew global warming was already occurring, 45 per cent incorrectly attributed rising temperatures to ‘natural changes in the environment’ rather than to the ‘effects of pollution from human activity’. Overall, fully three-fifths were wrong about one or both crucial features of global warming.

Among these misinformed individuals, somewhere between a third (on expanding the use of nuclear energy) and a half (on enforcing environmental regulations) opposed a series of climate change policies. In short, up to a third of all 2012 respondents, depending on the policy under discussion, were both misinformed and held policy views consistent with that misinformation. This proportion has risen in comparison with the 2007 Gallup data, although that change may be partly due to differences in question wording.

To understand the impact on democratic governance of both those who are informed but disengaged and those who are engaged but misinformed, we need to know more about the two groups’ characteristics. The crucial point is simple: political partisanship is overwhelmingly the most important demographic predictor of membership in both categories. Consider first those who are informed but disengaged. Again using the 2012 Gallup data, Democrats are 33 percentage points more likely than Republicans to hold correct factual information – but even knowledgeable 13 Republicans were, depending on the policy issue, 10–15 percentage points more likely to oppose policies to mitigate climate change than were knowledgeable Democrats. 14 Holding income, education, age, ideology, gender, race and ethnicity constant does not change these results. 15 (The results do not depend on any distinctive characteristics of 2012 or the survey in that year; partisanship is also the strongest statistical predictor of being informed but disengaged in the 2007 Gallup data).

Party commitments have a similarly strong association with the more problematic category of those who are misinformed but engaged. Three-fourths of Republicans answered at least one of the two factual questions about global warming incorrectly in 2012, compared with 44 per cent of Democrats. The partisan difference persists with controls for income, education, age, ideology, gender, race and ethnicity. And in the final step of the analysis, even among the misinformed, partisanship is associated with holding policy views that conform with incorrect ‘knowledge’: 71 per cent of mistaken Republicans, but only 28 per cent of mistaken Democrats, opposed stronger enforcement of environmental regulations. In what is now sounding like a broken record, the strong predictive power of party identification persists in six of the seven policy arenas in regression analyses using demographic controls, and it remains the same in the 2007 Gallup data.

The finding of a strong association among partisanship, knowledge, and policy views is hardly new. Nor are we making causal claims about these correlations within a single survey. What matters for our purposes is that these patterns of public opinion provide powerful – and different – incentives for public actors to encourage or discourage political activity, and to encourage or discourage abandonment of misinformation in favor of correct knowledge. We turn now to the role of political actors in shaping citizens’ knowledge and its use, and to our core claim that citizens who are engaged but misinformed can harm the quality of democratic governance.

Political Asymmetry

Perhaps political actors care only about the public’s policy preferences and votes, and are indifferent as to whether they hold true or false knowledge about policies and candidates. But politicians claim to want to inform the public, democratic theorists insist on the importance of knowledge, journalists purport to convey facts, and teachers devote their careers to providing information and helping students acquire more – so we proceed on the assumption that facts do (or should) matter.

The condition of political ignorance is fairly stable; people can remain unaware of, or unresponsive to, political cues and politicians’ urging without much stress or strain. It takes a great deal of effort to persuade political abstainers that politics matters, that their participation matters and that there is relevant correct information with appropriate policy consequences. So resource-strapped political actors often leave the politically ignorant alone.

In contrast, being informed and engaged requires, by definition, psychic and cognitive energy. But it can also involve gratification, reward and even enjoyment. So political actors are likely to obtain good results by cultivating their partisan activists; this engagement is essential and takes a good deal of time and energy, but it is not especially difficult. Thus analytically and politically, two of the four categories of information and engagement are relatively unproblematic.

The other two categories are different. A person who is informed but disengaged from the policy views that should follow from that information is in a state of dissonance. To continue with our case study, he knows that the earth is warming dangerously and that humans are largely responsible, but he does not like the options available to address the danger. That implies an internal contradiction that can make one uneasy and therefore available for mobilization that might resolve the unease. This availability, in turn, spurs political actors and parties to expend considerable effort on finding and galvanizing disengaged, but informed, co-partisans. Thus some elites have a strong political incentive to shift individuals from the ‘disengaged informed’ to the ‘engaged informed’ categories, and, because of the category’s inherent dissonance, they have a reasonable chance of accomplishing this goal. But given the likelihood of partisan differences among the disengaged informed, other elites have the opposite incentive of keeping these citizens quiet. In short, in the case of climate change, the Democratic – but not the Republican – party will work to persuade the disengaged informed to move into Horace Mann’s realm of informed engagement.

Finally, a person who is misinformed but engaged is in a state of consonance. She does not believe the earth to be warming dangerously, or perceives global warming only as a danger in the distant future, or does not think human activity is an important cause of global warming – so she feels no compunction about opposing mitigation policies that will, after all, entail costs. In many cases, that person shares views and perceptions with like-minded others. In fact, the role of groups, including but extending beyond a political party, is a crucial element of the dynamics of misinformed engagement. People seek out others with the same views, 16 or people who are already connected develop or reinforce misinformed views together, 17 or people are persuaded into misinformed views by others with whom they identify.

Thus to move a person who is misinformed but engaged into the realm of informed engagement takes a great deal of effort. She must be induced to forsake her group, recognize and renounce false beliefs, learn new facts, come to a new understanding of appropriate policy choices, develop new group ties and partisan loyalties, and then overcome the usual inertia of political inactivity in order to take action. Forging all of the links in this chain is hard, and any one of them is easily broken. As a result, political actors and parties seldom believe it is worthwhile to invest the resources needed to forge the new chain. In contrast, the party that benefits from engaging the misinformed has a strong incentive to keep them loyal, misinformed, and attached, and to reward them for their mistaken policy views or votes. Thus in the case of climate change, the Republican party has more incentive to encourage misinformed activists to remain both misinformed and engaged than the Democratic party has to try to persuade them to learn new facts and act on what they have learned.

There are, in short, three political asymmetries at work that suggest why the active misinformed are especially damaging to the quality of democratic governance. First, inertia is powerful, so people in all four categories are more likely to retain or deepen their current view than to change it. Thus political parties may see it as more rewarding to reinforce their base of informed activists than to try to shift the disengaged or ill informed. Second, a state of dissonance is unpleasant and unstable, so a political party with some slack resources is likely to find it worthwhile to try to move those who are informed but disengaged from the implications of their knowledge into the realm of engagement. That will be more appealing to the party with which a larger share of the disengaged informed is affiliated; the other party would prefer to keep them inactive (or possibly even shift them into the realm of the engaged misinformed by promulgating false information that allows them to retain their policy preferences with a clear conscience). Third, a state of consonance – even if resting on a mistaken base – is pleasant and stable, so a political party will seldom find it worthwhile to try to move the engaged misinformed out of that condition. In fact, it can be very rewarding to keep people closely tied to others in their group who are similarly misinformed and active, so the party that benefits the most from misinformed activity has a strong incentive to encourage people to remain in that category. In concert, the three asymmetrical forces imply that those who are misinformed and engaged will be particularly sticky – they are difficult to change in any case and some political actors have powerful incentives for not changing them.

All three asymmetries are evident in the politics related to global warming. The first seems fairly uncontroversial; it is the social scientists’ version of Newton’s first law, that the ‘innate force of matter, is a power of resisting by which every body, as much as in it lies, endeavours to preserve its present state, whether it be of rest or of moving uniformly forward in a straight line’. Scholars have recently developed a robust literature on motivated reasoning, which shows that people are much better at absorbing and using information that fits their extant beliefs or perceptions of fact than at using new information to change their beliefs. 18 Research similarly shows a strong confirmation bias in the public’s attention to particular media channels, and partisan loyalty is itself a form of inertia.

The second asymmetry – that one party has a strong incentive to resolve citizens’ dissonance resulting from informed disengagement – is well illustrated (though not proven) by an exhortation from President Obama: ‘Someday, our children, and our children’s children, will look at us in the eye, and they’ll ask us, did we do all that we could when we had the chance to deal with this problem [of global warming] and leave them a cleaner, safer, more stable world? And I want to be able to say, yes, we did. Don’t you want that?’ (25 June 2013). The liberal advocacy group Greenpeace is even more explicit in exhorting Americans to demand a federal response to the information included in its sixty-eight reports on global warming:

It is this reality – the physical reality of global climate science – against which the actions of our government will truly be measured. If they don’t adhere to and respect this reality, compromises that seem politically savvy today will prove irrational and irresponsible in the very near future. For both the President and Congress […] the true test of their leadership will be whether they base their actions on political convenience or on scientific facts. 19 (America’s-share-of-the-climate.pdf,: 16).

With regard to the third asymmetry, Democratic leaders exert little effort to change the minds of those who do not believe that the earth is warming: ‘“I don’t have much patience for anyone who denies that this challenge is real’ (Obama, 25 June 2013). Some Republicans and Republican-oriented advocacy groups, by contrast, work strenuously to foster the views and maintain the support of the misinformed who are relevantly engaged. Senator Marco Rubio (R-FL), for example, does ‘not believe that human activity is causing these dramatic changes to our climate the way these scientists are portraying it. I do not believe that the laws that they propose we pass will do anything about it, except it will destroy our economy’. Senator Rand Paul (R-KY) is similarly dismissive: ‘Most of the models [of climate change] have been changed within the last five years. They all predicted, you know, the poor Statue of Liberty was going to drown in 100 years, and the polar bears and all this stuff. And that alarmist kind of stuff really, I think, detracts from the case that we shouldn’t pollute’.

In the case of global warming, Democrats have the partisan incentive to shift disengaged informed individuals into the realm of informed engagement, while Republicans seek to maintain misinformed engagement. This case is not unique; President George W. Bush and his staff sought to maintain Americans’ belief that Iraq harbored weapons of mass destruction for months after the 2003 Coalition invasion of Iraq, and long after they had good reason to believe that no such weapons existed. 20 However, the logic of political asymmetry is itself symmetrical; at other times, or on other issues, the political parties’ roles may be reversed. During the 1950s, Southern Democratic leaders worked strenuously to maintain support for racial segregation on the false grounds that blacks were less intelligent and more aggressive than whites and were happy with their lot; Northern Republicans worked, albeit less strenuously, to promote correct information and some degree of appropriate civic engagement. 21 We conclude with a more contemporary example that is intended to show both the non-partisan logic of our analysis and the genuine danger of misinformation in a democratic polity.

Some Americans (and residents of other countries) refuse to vaccinate their children on religious, personal or medical grounds. Their numbers are small, but sufficiently large and geographically concentrated enough that some communities have lost ‘herd immunity’. As a result, increasing numbers of babies, unvaccinated children and even some vaccinated adults are getting infectious diseases such as measles and whooping cough that had almost been eradicated. Some have died.

A large proportion of vaccination refusers is white, well educated and affluent, and the Americans among them disproportionately live in communities such as Malibu, California; Boulder, Colorado; or Ashland, Oregon. They are, in short, most likely to be Democrats. They received a big boost from a politician with an iconic name – Robert F. Kennedy Jr. – and from the iconic talk show host, Oprah Winfrey. In 2007 Winfrey featured the actress and model Jenny McCarthy; the show promoted McCarthy’s book about curing her son of autism caused by vaccination, and Ms. Winfrey deemed her and others similarly inclined to be ‘mother warriors’. The federal Centers for Disease Control and doctors around the world are pointing with increasing fervor to the ‘unimaginable’ pain of ‘watching a patient die from a preventable cause’. 22

Our purpose is not to embarrass either Democrats or Republicans; the point is that prominent public actors can sometimes find it in their political or personal interest to encourage misinformed activity. People can die and communities or ecosystems can be destroyed – and even when that does not occur, the quality of democratic governance is undermined. The disengaged informed are worrisome and frustrating, but those who act in accord with misinformation can do serious damage despite offering partisan benefits to one or another political party.

References

Berinsky, Adam. 2012. Rumors, Truths, and Reality: A Study of Political Misinformation. Cambridge, MA: MIT. Available from http://web.mit.edu/berinsky/www/les/rumor.pdf, accessed 7 October 2014.
Blee, Kathleen. 2002. Inside Organized Racism: Women in the Hate Movement. Berkeley: University of California Press.
Biel, Steven, and Muffett, Carroll. 2009. America’s Share of the Climate Crisis: A State-by State Carbon Footprint. Washington, D.C.: Greenpeace USA. Available from http://www.greenpeace.org/usa/Global/usa/report/2009/8/america-s-share-of-the-climate.pdf.
Lodge, Milton, and Taber, Charles. 2013. The Rationalizing Voter. New York: Cambridge University Press.
Mann, Horace. 1846. Tenth Annual Report to the Secretary of the Massachusetts State Board of Education. Available from http://rapgenius.com/Horace-mann-tenth-annual-report-to-the-secretary-of-the-massachusetts-state-board-of-education-1846-lyrics, accessed 7 October 2014.
Nyhan, Brendan, Reifler, Jason, Richey, Sean, and Freed, Gary L.. 2014. Effective Messages in Vaccine Promotion: A Randomized Trial. Pediatrics 133 (4):e835e842.
Pearl, Robert. 2014. A Doctor’s Take on the Anti-Vaccine Movement. Forbes, 20 March. Available from http://www.forbes.com/sites/robertpearl/2014/03/20/a-doctors-take-on-the-anti-vaccine-movement/, accessed 7 October 2014.
Roper Center for Public Opinion Research. 2014. iPoll. Available from http://www.ropercenter.uconn.edu/.
Simpson, Aislinn. 2008. Winston Churchill Didn’t Really Exist, Say Teens. Telegraph, 2 April.
Somashekhar, Sandhya, and Craighill, Peyton. 2013. Many Americans Confused About Health-Care Law, Poll Finds. Washington Post, 20 September.
Sunstein, Cass. 2011. Going to Extremes: How Like Minds Unite and Divide. New York: Oxford University Press.
WorldPublicOpinion.Org. 2010. American Public Opinion on Foreign Aid. Available from http://www.worldpublicopinion.org/pipa/pdf/nov10/ForeignAid_Nov10_quaire.pdf, accessed 7 October 2014.

1 Mann 1846.

2 Simpson 2008.

3 WorldPublicOpinion.Org 2010.

4 We searched the Lexis-Nexis Academic database for ‘ACA’, ‘Obamacare’ or ‘health insurance’ in ‘headline’ or ‘lead’, in US Newspapers and Wires. We also searched the same keywords in ‘at least five occurrences’ for TV and Radio Transcripts, of which Lexis-Nexis offers ten.

5 Somashekhar and Craighill 2013. In the nineteen days of September before the Washington Post article was published, there were an additional 2,040 newspaper articles and 278 television or radio stories about the ACA.

6 Somashekhar and Craighill 2013.

7 Our forthcoming book – Do Facts Matter? Information, Misinformation, and Democratic Politics (University of Oklahoma Press, 2015) – explores these arguments in detail. Here we outline the analysis by developing, more fully than the book does, one case study that encompasses public opinion, political elites’ activity, and policy outcomes.

8 One reviewer suggested that global warming is a unique case, since almost all experts line up on one side of the issue while powerful economic interests line up on the other side. We have two responses, beyond the general point that every important and complex policy issue is unique in some ways. First, other issues, such as tariffs or agricultural subsidies for domestic farms, have a similar configuration. Second, to the degree that experts almost universally agree on the nature and impact of global warming, the case of climate change reveals just how powerful the temptations to be misinformed, and to exploit misinformation, can be.

9 We feature analyses from the 2012 Gallup survey because it was the most recent survey with publicly available data on the Roper Center’s iPoll (Roper Center for Public Opinion Research 2014) that asked a battery of policy items in addition to factual questions about global warming. We compare it to 2007 Gallup data because that survey also contained an unusually extensive battery of both factual and policy questions on climate change.

10 The 2012 poll asked two factual questions: whether global warming is already occurring and, if this question was answered correctly, whether its effects were manmade. Respondents had to answer both questions correctly in order to be classified as ‘informed’.

11 Unlike the 2012 Gallup poll’s two-part question, the 2007 Gallup and Newsweek polls asked only whether global warming was already occurring. Thus our criterion for ‘informed’ respondents was necessarily looser in 2007, and the proportion of respondents deemed to be informed was correspondingly higher. Using the same (loose) definition of factual information in 2012 as in 2007 yields very similar sample percentages in the two years. Because we believe the more in-depth battery of questions in the 2012 poll to be a more accurate gauge of factual knowledge about global warming, we report those numbers in the text.

12 Expanding the use of nuclear energy is an outlier in these data; it earned opposition from 55 per cent of respondents. The next-closest policy in terms of generating opposition among the correctly informed is spending more government money to develop alternative sources of fuel for automobiles; 15 per cent of the correctly informed expressed opposition.

13 Knowledgeable respondents are those who answered the two factual information questions about global warming correctly.

14 Expansion of nuclear energy was the only exception among the seven proffered strategies; in that case, correctly informed Democrats were 12 percentage points more likely than their Republican counterparts to oppose the policy.

15 We excluded one of Gallup’s proposed environmental policies, using federal lands for oil exploration, because it was unclear how this initiative linked with climate change. In the logit regression analyses, three of the six policies showed statistically significant relationships between partisanship and disengaged information at the 0.05 level and one at the 0.1 level. The other two policies’ coefficients were in the correct direction but statistically insignificant. Expanded use of nuclear energy was the only policy in which Republican identification did not predict being both informed and disengaged.

16 Sunstein 2011.

17 Blee 2002.

18 Lodge and Taber 2013; Nyhan et al. 2014.

19 Biel and Muffett 2009.

20 A 2012 American National Election Studies survey asked respondents about the veracity of a series of political rumors; Republicans were more likely to agree with statements that made Democrats look bad than the reverse (see Berinsky 2012).

21 We thank an anonymous reviewer for this example.

22 Pearl 2014.