Short communication

Perceptions of scientific dissent undermine publicsupport for environmental policy

Michael Aklin a,1, Johannes Urpelainen b,1,*aNew York University, New York, NY, USAbColumbia University, New York, NY, USA

e n v i r o n m e n t a l s c i e n c e & p o l i c y 3 8 ( 2 0 1 4 ) 1 7 3 – 1 7 7

a r t i c l e i n f o

Article history:

Received 5 August 2013

Received in revised form

11 October 2013

Accepted 13 October 2013

Available online 25 December 2013

Keywords:

Scientific uncertainty

Survey experiment

Skepticism

Public opinion

Environmental policy

a b s t r a c t

This article shows that even modest amounts of scientific dissent reduce public support for

environmental policy. A survey experiment with 1000 Americans demonstrates that small

skeptical scientific minorities can cast significant doubt among the general public on the

existence of an environmental problem and reduce support for addressing it. Public support

for environmental policy is maximized when the subjects receive no information about the

scientific debate, indicating that the general public’s default assumption is a very high

degree of scientific consensus. Accordingly, a stronger scientific consensus will not generate

public support for environmental policy, unless skeptical voices become almost silent.

# 2013 Elsevier Ltd. All rights reserved.

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/envsci

1. Introduction

Environmental problems are often so complex and technical

that scientific knowledge is necessary to evaluate their

severity. However, studies have noted that instead of resolving

policy controversies, science can become a vehicle of conflict

by strengthening the prior beliefs of different sides (Sarewitz,

2004; Pielke, 2007; Hoffman, 2011). Moreover, the inherent

difficulty of many environmental problems undermines the

effectiveness of communication because there is a mismatch

between their ‘‘usual modes of understanding’’ and what is

needed to grasp the essence of the problem (Oreskes, 2004;

Moser, 2010; Weber et al., 2011). In politically salient cases

such as climate change, there is also the possibility that

* Corresponding author. Tel.: +1 734 757 0161.E-mail addresses: michael.aklin@nyu.edu (M. Aklin), ju2178@colum

1 These authors contributed equally to this study.1462-9011/$ – see front matter # 2013 Elsevier Ltd. All rights reservedhttp://dx.doi.org/10.1016/j.envsci.2013.10.006

opponents of action deliberately mislead the public (Oreskes

and Conway, 2010). Often, science communication fails to

inform the public about the severity of the problem. Instead of

providing the public with information about the state of the

art, science communication is either ineffective or even

worsens the situation.

When does the general public believe the scientific

consensus to be strong enough to warrant action? How much

does public support for environmental policy depend on

scientific consensus? These are the questions this article

addresses. As noted above, several scholars have challenged

the idea that scientific knowledge about a severe environ-

mental problem automatically creates public concern and

prompts rapid policy action by the government. We conduct

an empirical analysis to evaluate the sensitivity of public

bia.edu (J. Urpelainen).

.

e n v i r o n m e n t a l s c i e n c e & p o l i c y 3 8 ( 2 0 1 4 ) 1 7 3 – 1 7 7174

opinion about environmental problems to scientific dissent.

We do this by exploring how much scientific dissent is needed

to significantly undermine the public’s support for environ-

mental regulation. If scientific communication were easy, one

would expect public support for environmental regulation to

remain high, as long as the vast majority of scientists believe

the problem to be a serious one.

Our simple experimental approach to the effect of scientific

dissent on public support for policy fills an important gap in

the literature. Although there are observational studies on

public perceptions of the scientific debate on environmental

problems such as global warming and nuclear power (Johnson

and Scicchitano, 2000; Krosnick et al., 2000; Dunlap McCright,

2008; Ding et al., 2011; Kahan et al., 2012; Leiserowitz et al.,

2013), and some have examined the role of scientists as a

trusted source of information both experimentally (Rabino-

vich et al., 2012) and in a traditional survey (Malka et al., 2009),

there are few experimental studies that shed light on how

people’s beliefs vary with the perceived degree of scientific

consensus. While one experimental study examines how

controversy influences people’s perceptions of scientific

reporting in newspapers, it does not compare different levels

of scientific consensus and the sample only comprises

undergraduate students (Corbett and Durfee, 2004). Another

study analyzes the effect of experimentally manipulated

uncertainty among different types of recipients, finding that

the respondent’s belief about the nature of science condi-

tioned the effect of uncertainty on willingness to engage in

environmental behavior, but the participants were students

and the study did not analyze support for regulation

(Rabinovich and Morton, 2012).

Given that scientists remain among the most trusted public

authorities in the United States (Lang and Hallman, 2005;

Gauchat, 2011, 2012; Leiserowitz et al., 2013), it is important to

understand when and why the public accepts the scientific

consensus as a legitimate basis of policy formulation. If the

origins of such trust were better understood, policy interven-

tions to improve public awareness could be designed. In other

words, understanding the sources of trust could improve

science communication. In this sense, this article contributes

to the literature on the interactions between scientific debates

and public opinion.

To investigate the issue, we conducted a survey experi-

ment on a sample of 1000 American adults between the ages

of 18 and 65 in November 2010. The sample is nationally

representative across standard population characteristics

such as gender, age, race, and education. We presented the

respondents with information about a hypothetical study

concerning changes in the levels of biochemical oxygen

demand (BOD), a commonly used measure for effluent water

pollution, in American lakes and rivers. The control group

was told that such a study had been conducted, while the

treatment groups were given additional information about

the proportion of scientists who believe the study methodol-

ogy is sound: 60, 80, and 98 percent to capture varying degrees

of scientific consensus. In the treatments, the remaining

proportion of scientists was presented as questioning the

merits of the study.

We found that even a relatively small scientific minority

can significantly reduce public support for addressing the

environmental problem in focus. While support levels were

high when 98 percent of scientists agreed that the problem is

real, there was a substantively large and statistically

significant drop with 80 percent of scientists being presented

as skeptical. Since a one-fifth minority is common in the case

of new and complex environmental problem, this means that

the scientific community can only convince the public about

the existence of a problem with a high degree of consensus. In

other words, even a modest amount of scientific dissent

significantly decreases public support for environmental

policy.

2. Research design

In the survey, 1000 English-speaking Americans of age 18–65

were interviewed about a variety of economic and political

issues in November 2010, immediately following a midterm

election. The survey experiment discussed here was part of a

broader survey implemented by the 2010 Cooperative Congres-

sional Election Study (CCES). This is an opt-in, on-line survey

managed by a research team at Harvard University and

implemented by YouGov/Polimetrix (http://projects.iq.harvard.

edu/cces/book/sample-design). The respondents were chosen

so that each matches a randomly drawn individual from the

general population. In other words, the survey respondents are

selected so that they represent the broaderAmerican population

based on official data from the United States census. Although

the survey responses are collected on-line, the pool of

respondents matches well with the broader American popula-

tion along standard covariates such as age, gender, education,

income, and race. Survey weights along these lines are used in

the analysis. The response rate for both of our two outcome

variables was 83.8 percent.

In the survey, the respondents were given hypothetical

information about a study related to water pollution. Each

respondent was randomly assigned to a control group or one

of four treatment groups, each with equal probability (1/4). The

wording of the treatments our outcome variable was as

follows:

According to a recent scientific study of pollution problems,

biochemical oxygen demand has increased in rivers and lakes

throughout the country due to industrial activity. [TEXT A].

Reducing biochemical oxygen demand is technically feasible but

economically costly.

The control group received no additional information: for

them, [TEXT A] was left empty. The three treatments were of

the following format:

About [PERCENTAGE YES] percent of all scientists believe that the

results are credible. However, the other [PERCENTAGE NO]

percent argue that the results are weak and that further research

is required.

The percentages for [PERCENTAGE YES] were 60, 80, and 98;

the percentage for [PERCENTAGE NO] were 40, 20, and 2. As the

number of supportive scientists grows, scientific dissent

decreases. Since the control group received no information,

22.

12.

22.

32.

42.

5Li

kelih

ood

that

BO

D h

as In

crea

sed

Baseline 60% 80% 98%Response

Belief that biochemical oxygen demand has increased (0: very unlikely; 4: very likely).

Belief in Increase of Biochemical Oxygen Demand

Fig. 2 – Mean response to question about belief in increase

in biochemical oxygen demand by treatment.

e n v i r o n m e n t a l s c i e n c e & p o l i c y 3 8 ( 2 0 1 4 ) 1 7 3 – 1 7 7 175

their views represent the default belief of the general public in

the absence of any specific information.

Following the treatments, two questions were asked in

this order:

Do you support or oppose new environmental regulations to

reduce biochemical oxygen demand?

From what you have heard, how likely is it that biochemical

oxygen demand has increased?

The question about policy was given first to avoid the

respondents’ considering the survey a test of their abilities.

The question about beliefs complements the policy question,

and we expect responses to the two to be highly correlated. For

the first question, the respondents were given five options: (1)

strongly support, (2) somewhat support, (3) neutral, (4)

somewhat oppose, (5) strongly oppose. For the second

question, the respondents were also given five options: (1)

very likely, (2) somewhat likely, (3) not sure, (4) somewhat

unlikely, (5) very unlikely. We rescaled and inverted the

responses to create a measure from 0 to 4, where higher values

indicate greater support and belief, respectively.

3. Results

Fig. 1 shows the response to a question about the need for new

environmental regulations to reduce BOD pollution. The

respondents indicated their support for new environmental

regulations on a 0–4 scale, with higher values indicating

support for regulation. We then computed the mean response

by treatment, each respondent being weighted in order to give

a representative sample. As the figure shows, the mean level of

support in the control group was 2.64. This level decreases to

2.11 when 60 percent of scientists endorse the study. Even

with 80 percent scientific endorsement, the support level

remains at 2.15. Both levels are statistically distinguishable

from the control group at the p < 0.01 level. Even when

98 percent of scientists endorse the study, the mean support

22.

22.

42.

62.

8S

uppo

rt fo

r E

nviro

nmen

tal R

egul

atio

n

Baseline 60% 80% 98%Treatment

Support for new environmental regulation to reduce biochemicaloxygen demand (0: strongly oppose; 4: strongly support).See text for treatments. Whiskers represent 95% confidence interval.

As Scientific Consensus ChangesSupport for New Regulations

Fig. 1 – Mean response to question about need for new

environmental policy by treatment.

level is below the baseline case, at 2.49, though this level is

not statistically distinguishable from the control level. These

results are unchanged if we include covariates as control

variables in a multivariate regression, as shown in the

supporting information.

As shown in Fig. 2, the effects are similar for a follow-up

question on belief about the existence of the problem. We

asked this question later in the survey so that the respondents

would not interpret it as a ‘‘test’’ of their abilities. The

respondents were asked to indicate their belief in the BOD

increase on a 0–4 scale, with lower values indicating skepticism

about the existence of the environmental problem. As with the

question about support for environmental regulation, infor-

mation about scientific endorsement created skepticism.

However, the differences are not statistically significant at

conventional levels.

In sum, even a very broad consensus among scientists

barely achieved the levels of policy support and belief that

would have been achieved without any communication about

the scientific consensus. As soon as the respondents were told

that there is some disagreement among scientists, the level of

public belief in the problem and demand for a policy response

decreased. One plausible interpretation of this finding is that

the public considers a scientific study reliable in the absence of

additional information about the scientific debate. If the

public’s prior expectation is that scientific studies and results

are beyond dispute, even small skeptical minorities under-

mine the public’s faith in the study, even as scientists are used

to continual criticism and debates.

The supporting information shows that the findings hold

across different subgroups. While conservatives are less

supportive of environmental policy than liberals, the informa-

tion treatment effects apply to both subgroups. They also

apply to both men and women, to respondents with different

levels of education, and to more and less religious Americans.

Remarkably, the results also hold for subgroups of people who

are for or against international climate cooperation. These

subgroup results indicate that the effect of information about

the scientific consensus is not specific to any given subgroup

of Americans.

e n v i r o n m e n t a l s c i e n c e & p o l i c y 3 8 ( 2 0 1 4 ) 1 7 3 – 1 7 7176

4. Second experiment

To further scrutinize the robustness of our result, we

conducted a second on-line experiment with a convenience

sample of 2031 Americans using Amazon’s MTurk service

(https://www.mturk.com/mturk/). The methods and results

from this experiment are described in detail in the supporting

information. First, we varied our experiment so that the

respondents were given more detailed information about the

consequences of increased biochemical oxygen demand. This

variation allowed us to verify that our results were not biased

by a lack of information about the environmental problem at

hand. As expected, the treatment effects were slightly

stronger.

Second, we included a treatment that explicitly empha-

sized the existence of a unanimous consensus among

scientists. We found that the support levels were not different

from those obtained without any information on scientific

consensus. This suggests that in the absence of information to

the contrary, Americans believe individual scientific studies to

be underpinned by a broad consensus. In other words, the

respondents’ prior belief about the credibility of the scientific

study seems to have been high. On average, respondents tend

to trust scientific studies as credible sources of accurate

information about environmental problems.

Finally, we analyzed the second set of experimental results

separately for people who claimed they trust scientists and

those who said they did not. While the treatment effects were

strong for people who trust scientists, increased scientific

consensus decreased support for environmental regulation

among people who do not trust scientists. This suggests that

people who distrust scientists become more skeptical as

consensus among scientists increases.

5. Conclusion

Even small skeptical minorities can have large effects on the

American public’s beliefs and preferences regarding environ-

mental regulation. Several scholars have warned that due to

the uncertain and continually evolving nature of environ-

mental science, building popular support for public policy

based on scientific reasoning is difficult (Oreskes, 2004;

Sarewitz, 2004; Pielke, 2007). According to these authors, the

common expectation that scientific research produces con-

clusive proof and unambiguous policy recommendations may

impede meaningful public debate about alternative courses of

action. The problem is particularly acute for politically

controversial issues, such as climate change. In these

circumstances, opponents of environmental regulation can

frame the science in different ways to cast doubt on the

scientific findings and to downplay the urgency of mitigation

policies.

Our results also show that even if the consensus among

scientists is unassailable and people are aware of this, public

support for environmental policy does not increase above the

baseline without no information. If even small amounts of

uncertainty plant skepticism in people’s minds, it is not

difficult for groups opposed to environmental regulation to

confuse publics about the state of the scientific debate. This is

exactly how the debate on climate change has played out in

the United States and many other industrialized democracies.

While negative, the result can also inform future efforts to

improve scientific communication about environmental pro-

blems. Given that a very high degree of consensus is needed to

achieve high levels of public trust, it is difficult for scientists to

try to ‘‘win’’ the debate in the media. A more promising

strategy would try to increase public awareness about the

inherent uncertainties that surround science. If the public did

not expect unrealistically high degrees of consensus to accept

scientific findings as a legitimate basis for environmental

policy, then legitimate dissent would not raise an insurmoun-

table obstacle to policy. One way to achieve this goal would be

to follow Jasanoff (1997) advice and enhance public participa-

tion in scientific and regulatory processes. Future research in

the social sciences could examine this hypothesis along the

lines of Rabinovich et al. (2012), who show that people’s

responses to scientific uncertainties depend on their concep-

tion and understanding of science and its progress. As Miles

and Frewer (2003, 281) warn, ‘‘[i]t may be that denying

uncertainty will have a negative impact on trust in informa-

tion sources and regulators’’. Our findings underscore the

importance of the issue and further research on it.

Acknowledgments

We thank Yotam Margalit for his generous help with the

design of the survey experiment. We thank Scott Barrett,

Arthur Lupia, Michael Bechtel, Michael Tomz, Robert O.

Keohane, Maureen Lackner, Ezra Markowitz, Robert Socolow,

Stephen Pacala, Marko Klasnja, as well as seminar audiences

at New York University and Princeton University for helpful

comments on a previous draft. We also thank Martin Beniston

and anonymous reviewers for useful advice.

Appendix A. Supplementary data

Supplementary data associated with this article can be

found, in the online version, at http://dx.doi.org/10.1016/

j.envsci.2013.10.006.

r e f e r e n c e s

Corbett, J.B., Durfee, J.L., 2004. Testing public (un)certainty ofscience: media representations of global warming. ScienceCommunication 26 (2) 129–151.

Ding, D., Maibach, E.W., Zhao, X.Q., Roser-Renouf, C.,Leiserowitz, A., 2011. Support for climate policy and societalaction are linked to perceptions about scientific agreement.Nature Climate Change 1 (12) 462–466.

Dunlap, R.E., McCright, A.M., 2008. A widening gap: republicanand democratic views on climate change. Environment:Science and Policy for Sustainable Development 50 (5) 26–35.

Gauchat, G., 2011. The cultural authority of science: public trustand acceptance of organized science. Public Understandingof Science 20 (6) 751–770.

e n v i r o n m e n t a l s c i e n c e & p o l i c y 3 8 ( 2 0 1 4 ) 1 7 3 – 1 7 7 177

Gauchat, G., 2012. Politicization of science in the public sphere:a study of public trust in the United States, 1974 to 2010.American Sociological Review 77 (2) 167–187.

Hoffman, A.J., 2011. Talking past each other? Cultural framingof skeptical and convinced logics in the climate changedebate. Organization and Environment 24 (1) 3–33.

Jasanoff, S., 1997. Civilization and madness: the greatBSE scare of 1996. Public Understanding of Science 6 (3)221–232.

Johnson, R.J., Scicchitano, M.J., 2000. Uncertainty, risk, trust,and information: public perceptions of environmentalissues and willingness to take action. Policy Studies Journal28 (3) 633–647.

Kahan, D.M., Peters, E., Wittlin, M., Slovic, P., Quellette, L.L.,Braman, D., Mandel, G., 2012. The polarizing impact ofscience literacy and numeracy on perceived climate changerisks. Nature Climate Change 2 (2) 732–735.

Krosnick, J.A., Holbrook, A.L., Visser, P.S., 2000. The impact ofthe fall 1997 debate about global warming on Americanpublic opinion. Public Understanding of Science 9 (3)239–260.

Lang, J.T., Hallman, W.K., 2005. Who does the public trust? Thecase of genetically modified food in the United States. RiskAnalysis 25 (5) 1241–1252.

Leiserowitz, A., Maibach, E.W., Roser-Renouf, C., Smith, N.,Dawson, E., 2013. Climategate, public opinion, and the loss oftrust. American Behavioral Scientist 57 (6) 818–837.

Malka, A., Krosnick, J.A., Langer, G., 2009. The association ofknowledge with concern about global warming: trusted

information sources shape public thinking. Risk Analysis 29(5) 633–647.

Miles, S., Frewer, L.J., 2003. Public perception of scientificuncertainty in relation to food hazards. Journal of RiskResearch 6 (3) 267–283.

Moser, S.C., 2010. Communicating climate change: history,challenge process and future directions. WileyInterdisciplinary Reviews: Climate Change 1 (1) 31–53.

Oreskes, N., 2004. Science and public policy: what’s proof got todo with it? Environmental Science and Policy 7 (5) 369–383.

Oreskes, N., Conway, E.M., 2010. Merchants of Doubt: How aHandful of Scientists Obscured the Truth on Issues fromTobacco Smoke to Global Warming. Bloomsbury Press,New York.

Pielke Jr., R.A., 2007. The Honest Broker: Making Sense ofScience in Policy and Politics. Cambridge University Press,New York.

Rabinovich, A., Morton, T.A., 2012. Unquestioned answers orunanswered questions: beliefs about science guideresponses to uncertainty in climate change riskcommunication. Risk Analysis 32 (6) 992–1002.

Rabinovich, A., Morton, T.A., Birney, M.E., 2012. Communicatingclimate science: the role of perceived communicator’smotives. Journal of Environmental Psychology 32 (1) 11–18.

Sarewitz, D., 2004. How science makes controversies worse.Environmental Science and Policy 7 (5) 385–403.

Weber, E.U., Stern, P.C., 2011. Public understanding of climatechange in the United States. American Psychologist 66 (4)315–328.