Available via license: CC BY 4.0
Content may be subject to copyright.
Page 1/16
Public support for natural gas bans in the United
States
Elena McLean
SUNY Buffalo https://orcid.org/0000-0002-0512-7495
Taehee Whang
Yonsei University
Joonseok Yang ( jsyang01@skku.edu )
Sungkyunkwan University https://orcid.org/0000-0003-2895-6365
Article
Keywords:
Posted Date: May 3rd, 2022
DOI: https://doi.org/10.21203/rs.3.rs-1573035/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Page 2/16
Abstract
Policies to reduce greenhouse gas emissions in the US have largely treated natural gas as a clean fuel
due to its lower dioxide-to-energy content than other fuels. However, recent local decarbonization
initiatives seek to ban residential uses of natural gas. Public support for such policies will determine
whether other localities adopt natural gas bans and whether state and federal policies may follow. At the
same time, several states adopted policies prohibiting natural gas bans. In this study, we show that public
support for gas bans depends on policy design. Our survey experiment shows that health effects of
proposed policies are associated with public support, whereas monthly energy costs, industry support,
and information about climate effects and political support for the proposed policy do not inuence
public opinion.
Full Text
The United States remains one of the most signicant greenhouse gas polluters globally, in terms of
aggregate carbon dioxide emissions on a per capita basis1. At the same time, the country has become
less dependent on coal for its energy needs, as natural gas replaced coal to a signicant extent. In the
1970s, natural gas began to emerge as a critical source of cleaner energy: its resource base was
extensive, the environmental advantages were well acknowledged, and the gas-using products and
technologies turned out to be markedly ecient2. In 2015, natural gas accounted for 32 percent of total
electricity production, up from 9 percent in the late 1980s, whereas the share of coal in electricity
production declined from 58 to 34 percent during the same period3-4. Natural gas acquired a reputation
as a clean fuel due to its lower carbon dioxide-to-energy content compared to coal5 and became an
acceptable replacement for more-polluting coal and liquid fuels in the energy system. As a result, many
economic sectors, including electric generation, transportation, and commercial cooling, switched to
natural gas to a signicant degree.
Despite this perception, gas combustion generates greenhouse gases and contributes to climate change.
Scholars, policymakers, and environmental activists criticize natural gas for its environmental risks from
the global climate perspective6-10. In particular, recent studies point to problems of aring, venting, and
methane leaks along the natural gas supply chain11. These problems have led to the conclusion that no
sustainable energy mix can include fossil fuels, thereby questioning continued natural gas use. In
addition, natural gas tends to exacerbate public health problems as gas development often contaminates
air and water, increases industrial noise and trac, and leads to residential community changes12.
Finally, the role of natural gas as a transition fuel has come under criticism after 196 countries adopted
the Paris Agreement in 2015 in an effort to limit greenhouse gas emissions as soon as possible. The
agreement necessitates the gas industry to provide a reliable and long-term decarbonization strategy
(e.g., through the production and use of biomethane and low-carbon hydrogen).
One of the recent strategies to address climate change seeks to reduce dependence on natural gas in the
residential sector. In 2019, Berkeley, CA, became the rst municipality to ban natural gas hookups in new
Page 3/16
construction13. Now, over 40 local energy ordinances in California either ban or discourage natural gas
connections in new buildings14. Other areas in the US, including Denver, CO, Brookline, MA, and New York,
NY, are turning to similar measures to eliminate natural gas use in new homes and buildings, thereby
reducing greenhouse gas emissions. At the same time, these efforts to ban natural gas in new
construction have caused a backlash from the gas industry and gas-dependent utilities. Since 2019, 19
US states, i.e., Alabama, Arizona, Arkansas, Florida, Georgia, Indiana, Iowa, Kansas, Kentucky, Louisiana,
Mississippi, Missouri, Ohio, Oklahoma, Tennessee, Texas, Utah, West Virginia, and Wyoming, have passed
state-level restrictions that prohibit local governments from banning or placing limits on natural gas.
While other states seem to be taking a middle path, it is clear that the future of natural gas remains a
contested political issue.
Adoption of measures restricting natural gas in the residential sector depends on public support. Previous
research suggests that public support is critical for implementing renewable energy and climate change
policies15-20. However, no studies currently investigate factors that inuence public opinion on natural
gas bans. We aim to close this research gap and seek to understand whether and under what conditions
the public may support or oppose natural gas bans. Our study contributes to the formulation of clean and
ecient energy policies by focusing on individuals’ preferences over the design and framing of gas bans,
which can directly impact households and provide measurable benets for individuals. The 2021
Morning Consult survey underscores the importance of taking public opinion into account when
developing policies to ban natural gas in new construction. While 44% of adults would support such a
ban and 37% would not, a signicant share – 20% of respondents – do not have an opinion on this
policy. The survey shows minimal regional variation in the share of respondents without an opinion: from
21% in the Northeast and Midwest to 20% in the South and 17% in the West, despite regional differences
in adopted and proposed policies to ban natural gas or prohibit such bans. These statistics suggest that
there is signicant room for expanding the support base for this policy; therefore, policy design should
take into account factors that are associated with greater support.
We conduct a survey experiment to identify factors that explain public support for natural gas bans. A
growing number of municipalities in different parts of the US have adopted measures to ban natural gas
hookups in newly constructed buildings. These ordinances require or encourage all-electric construction,
which means that new homes will have only electric appliances. Our survey experiment aims to evaluate
which factors affect participants’ support for this energy policy. We draw on previous ndings in the area
of environmental and energy policymaking and public opinion. Specically, existing studies show that
economic considerations, including energy costs, affect individual-level support for various energy
policies21-22. Public health benets also appear to determine public support for renewable energy
policies23. Together, these insights inform our study in identifying political, economic, and public health
motivations that can explain individual responses to a proposed policy banning natural gas in new
residential construction.
Page 4/16
Effect Of Information About The Expected Effects Of Natural Gas Ban
Policy
To identify which types of information or attributes of the natural gan ban policy affect public support,
we conducted an online survey experiment. Following the existing studies on public support for
renewable policies23, we utilized a single conjoint experiment design, in which we provided respondents
with randomized sets of information about the effects of a proposed natural gas ban policy. The conjoint
design allows researchers to identify nonparametrically and estimate causal effects of various types of
treatment information on respondents’ multidimensional preferences simultaneously and, hence, mitigate
social desirability biases that often mask respondents’ true preferences24–25. Also, in the conjoint design,
treatment effects of each attribute value are estimated relative to the reference category. In our
experiment, those respondents who were not exposed to any information about a given treatment
attribute serve as the reference category (i.e., control group).
To describe the effects of a natural gas ban policy for our respondents, we include (1) costs of the policy
for consumers; (2) interest groups’ opinions regarding the policy; (3) the policy’s effects on health; (4) the
policy’s effects on climate change; and (5) political support for policy adoption. After presenting a
randomly chosen set of policy effects, we asked how strongly the respondent supports or opposes the
policy using a 5-point scale (strongly support, moderately support, neither support nor oppose, moderately
oppose and strongly oppose). We elded our survey using the Quatrics online platform and recruited a
total of 2,623 US respondents through Dynata in September 2021. To evaluate the causal effect of each
treatment component on support for policy, we estimate linear regression models where the dependent
variable is the policy support level (on a 5-point scale) and binary indicators of each treatment
component serve as regressors (see Methods section for details).
Findings
In this section, we report the results of a survey experiment that examines the effects of various
treatments on public support for the natural gas ban policy. We summarize our results in Figs.1 and 2.
Our survey experiment uses a national sample of 2,623 respondents to investigate which of the ve
included factors, i.e., monthly costs, industry position, health effects, climate change, and political
support, explain individual-level preferences for a policy eliminating natural gas from new residential
construction.
Figure 1 presents the main ndings from the survey experiment. Among the ve dimensions of policy
design and framing, health concerns are the only signicant determinant that explains public support for
the natural gas ban policy. Respondents exposed to the information about adverse effects of air pollution
on their and their children’s health are signicantly more likely to support the natural gas ban policy, than
those who were not exposed to such information. This implies that citizens are most sensitive to non-
economic health frames when evaluating the policy restricting natural gas use.
Page 5/16
Next, we nd only suggestive evidence that consumer costs matter in explaining public support for the
natural gas ban policy. Respondents tend to oppose the gas ban when the policy increases their energy
expenditures, which is consistent with existing studies on public support for renewable or environmental
policies23,26, but these treatments failed to reach the conventional threshold of statistical signicance.
We also conclude that the information describing the policy stance of energy companies, i.e., natural gas
industry opposition or renewable energy industry support, does not affect public preferences toward the
natural gas ban policy. This implies that economic frames such as consumer costs and industry support
are not as important as we would expect, compared with non-economic frames, in explaining the public
support for restrictions on natural gas use.
In addition, we nd that public support does not have a signicant association with climate change
frames, echoing the ndings of previous studies on public opinion toward renewable policies23,27. This
may suggest the limited effectiveness of climate change frames, which have been viewed by
environmental groups as an important tool for inuencing public opinion. Similarly, our analysis does not
yield any evidence of government ocials’ sway over political support. Neither federal nor local
politicians appear to have any inuence on public opinion regarding the policy banning natural gas use.
Next, we investigate whether the effect of health-related information depends on survey participants’
characteristics. Figure2a points to the importance of gender: female respondents are likely to support a
natural gas ban when they see the health treatment, whereas male respondents’ support does not
increase when we present them with this information. Similarly, Fig.2b indicates that individuals’
exposure to harmful products of natural gas combustion due to regular cooking is linked to increased
support for the policy in the group that sees health-related information. In contrast, respondents who do
not cook at home are not swayed by adverse health impacts. Together, these two sets of ndings are
consistent with a traditional gender-based division of household responsibilities, when women tend to do
most of the cooking. Given the amount of time women spend close to the source of pollution, they have
incentives to pay closer attention to information about the negative health impacts of this type of air
pollution.
Another individual-level characteristic that is associated with varying levels of responsiveness to health
information is partisanship. Independents are more likely to express their support for the policy if we
present them with information regarding harmful effects for child health (Fig.2e). At the same time, we
do not nd signicant heterogeneity across parties for our main health treatment.
In addition, we explore differences that can be attributed to characteristics of our respondents’ states of
residence. Our results show that policy support among residents of gas-producing states does not
change when we highlight the health consequences of cooking with gas. However, individuals who live in
other states are more likely to increase their support for a gas ban when they receive information about
harm to their health or their children’s health (Fig.2d). This discrepancy could stem from economic
motivations: those who reside in gas-producing states might worry about the loss of jobs and tax
Page 6/16
revenues from gas production if a natural gas ban goes into force, and that may cancel out any health
concerns.
We also nd a divide between states that moved towards banning natural gas, that made such bans
illegal, and that have not adopted any policies regarding natural gas use in residences. Figure2c
indicates that information about adverse effects of gas on child health is associated with a greater
likelihood of increased support for this policy in states without any existing policies. This means that
residents of these states may be relatively less informed about a natural gas ban (or a counter-ban) and
possess weaker prior beliefs regarding the utility of this policy. Given this nding, one strategy for
increasing support in states that are yet to adopt any policy should center on informing residents about
the harms of indoor air pollution and, by extension, the benets of natural gas bans.
Conclusion
What explains Americans’ support for policies that restrict natural gas use by banning gas from new
residential constructions? While the carbon dioxide emissions resulting from natural gas consumption
are lower than from other fossil fuels, scholars and policymakers have become concerned with the
deleterious consequences of burning natural gas for human health and the environment. Some U.S.
states moved to adopt policies to ban natural gas in new construction, whereas other states are going in
the opposite direction by prohibiting local authorities from passing similar bans. At the same time, a
number of states are yet to adopt any policies in this policy area. Thus, it is crucial to understand the
formation of public opinion on the natural gas ban policy while policymakers attempt to advance new
measures to combat climate change and protect human health. This paper seeks to understand which
factors inuence public support for natural gas bans by conducting a survey experiment.
Our results show that the U.S. public signicantly supports natural gas bans when members of the public
are informed that natural gas is harmful to their (or their children’s) health. Health concerns play a key
role in increasing public support for the natural gas ban policy. Other treatments, such as information
regarding consumer costs, industry position, climate change, and political support, fail to explain public
support for the bans. We also nd that the effects of health-related information vary with survey
participants’ characteristics, such as household responsibilities and economic incentives. For example,
respondents with more household responsibilities support the natural gas ban when they respondents
receive information about health hazards associated with natural gas use. However, respondents who
might expect the loss of jobs and revenues due to the ban do not change their opinions when we expose
them to the same information.
These experimental ndings offer important policy implications. As the energy system is undergoing a
transition away from all fossil fuels, public support is essential for policymakers who seek to advance
policies restricting the use of natural gas. Our study suggests that policymakers can increase public
support for this policy by tailoring its design and framing to focus on health consequences. Future
research may further explore policy design implications by gauging the public’s willingness to absorb the
Page 7/16
costs of natural gas bans or cover the costs of transition away from fossil fuels and towards renewable
energy.
Methods
Survey Experiment Sample
We elded an online survey experiment exposing U.S. adults to information regarding a proposed natural
gas ban policy that varied by expected effects in different areas, policy content, and political support. A
total of 2,623 U.S. adults were recruited by Dynata to participate in an online survey in September 2021.
Employing various recruitment channels (e.g., open enrollment and parnerships with thousands of
websites, and aliate networks including schools and communities), Dynata recruits hard-to-reach
groups, such as ethnic minorities and seniors. As a result, Dynata possesses a diverse group of panels.
Our survey sample is drawn from Dynata’s U.S. panel, which aims to achieve the closest match to the U.S.
Census and social benchmarks. In order to have a diverse sample that is close to the nationally
representative sample, we employed soft quotas with regard to gender, age, regions, and education
(based on the adult population, as reected in the Census, i.e., 18 years of age and older). See Table A1
for the summary statistics of survey participants in comparison to the Census averages. Though our
sample includes a slightly larger proportion of female respondents, a somewhat less educated and
younger group, and underrepresents West, the differences with the Census averages are only marginal.
Table A1
Summary statistics of survey participants in comparison to the
Census averages.
Our sample Census averages
Female 51.1% 50.8%
High school degree or less 40.5% 38.9%
Age less than 44 49.5% 46.1%
Northeast 19.1% 17.3%
Midwest 21.8% 20.7%
South 42.9% 38.2%
West 16.1% 23.8%
To ensure reliable, accurate responses, we rst rely on Dynata’s in-house screening processes based on
“Total Research Quality® system,” which monitor data quality employing various quality check methods.
In addition, we screened out responses that did not pass our own validation check embedded in the
survey. In the middle of the survey, we included ‘the skip response check,’ asking respondents to skip the
Page 8/16
question without choosing any answers. Those who did not read the instruction and randomly chose
responses were considered as inattentive respondents and removed from the sample.
Survey Experiment Design
To assess how the information about the expected outcomes or attributes of the natural gas ban policy
affects public support, we employ a single conjoint design, in which we simultaneously randomize key
attributes (or “treatments”) of a hypothetical proposed policy banning natural gas; this allows us to derive
reliable estimates of each attribute as well as potential interactions between attributes. In our experiment,
each respondent receives a series of statements about the natural gas ban policy. Our experiment begins
with the following short description of a proposed natural gas ban policy:
Since 2019, over 40 cities in different parts of the US have adopted measures to ban natural gas hookups
in newly constructed buildings. These ordinances require or encourage all-electric construction, which
means that new homes will have only electric appliances. In the next year or two, a similar measure may
be considered in your area, which would eliminate natural gas from all new homes. Consider the
following effects of the proposed natural gas ban in your area:
We focus on the following ve pieces of information about the natural gas ban policy that scholars and
practitioners have previously identied as potentially inuential factors, namely (1) economic costs for
households, (2) the effects of the policy on the gas and renewables industries, (3) the effects of natural
gas use on health, (4) the implications of natural gas use for climate change, and (5) political support for
the policy at the local or federal levels.
(1) Economic costs (for households)
First, we present information about the economic costs for households (i.e., changes in consumers’
energy bills). Studies provide some mixed results regarding the public’s willingness to pay higher costs
for cleaner energy. On the one hand, scholars nd that individuals prefer energy at a lower cost28.
Therefore, exposure to cost information reduces support for renewable energy policies29. In contrast, an
average U.S. consumer shows support for a national clean energy standard even if that leads to higher
electricity bills30.
(2) Industry (gas vs. renewables)
Second, the effects of the policy on the natural gas and renewable energy industries can be another
important factor shaping public support for the natural gas ban policy. Natural gas industry has been
effective in establishing this fuel as a transitional solution to the problem of greenhouse gas
emissions31. Therefore, support for this industry can lead to opposition to measures reducing reliance on
natural gas. At the same time, renewable energy industry has grown in importance in terms of its size and
a broad range of economic benets provided to consumers and local economies, including new jobs and
Page 9/16
business opportunities for local suppliers32. This suggests that this industry’s interests may shape
individuals’ support for renewable energy policies.
(3) Health
Concerns about air pollution and resulting adverse health effects increase public support for cleaner
energy generation alternatives33. Such benets t with individuals’ perception of localized and more
tangible social benets connected to reduced reliance on fossil fuels23,28. However, these studies do not
differentiate between ambient (or outdoor) air pollution, which has the most direct link with traditional,
fossil-fuel energy generation, and household (or indoor) air pollution, which is the focus of this study.
(4) Climate change (GHG)
Overall, environmental concerns inuence individual energy preferences. However, these concerns do not
focus on the global environment; they tend to be more localized28,34. Hence, clean energy policies with
global environmental benets receive less support than policies producing local benets23.
The use of natural gas in clean energy transitions is a two-edged sword. The need to stop using coal is
clear, which positions natural gas as an interim fuel on the path of decarbonization35. At the same time,
this temporary solution is risky: reliance on natural gas could result in carbon lock-in, thereby delaying the
process of decarbonization36. Therefore, the public support for natural gas reects this duality: the use of
natural gas to produce electricity receives support as an environmentally friendly solution, but the
recognition of its contribution to global warming reduces support for continued gas use37.
(5) Political support (local vs. federal)
Lastly, the presence of political support may also be an important factor underlying public support for the
natural gas ban policy. Individuals tend to experience different levels of trust for local, state and federal
governments and consequently show different support for policies adopted by these governments38.
Citizens also believe that different levels of government should specialize in certain issue areas39, and in
the U.S. context specically, the public views energy policy as the policy domain of the national (or
central) government40. However, support for narrowly-focused energy extraction policies show variation
based on individuals’ feelings toward the federal and local governments41.
Table A2 provides a summary of these attributes and their levels. For each attribute, the control group,
which serves as a reference category in estimating treatment effects, received no information about a
given attribute. Thus, it is possible that a few respondents viewed no information for any of the ve
treatments.
Table A2: Attributes and Levels in the Conjoint Experiment
Page 10/16
Attributes Control Value 1 Value 2 Value 3
Costs [No
information
is given]
Cost estimates suggest
that all-electric new
construction homes
would likely see
consumer bill savings.
Cost estimates
suggest that all-
electric new
construction homes
would likely see a
consumer bill
increase of less than
$100 per year.
Cost estimates
suggest that all-
electric new
construction homes
would likely see a
consumer bill
increase of more
than $100 per year.
Industry [No
information
is given]
The expected effect of
the ban on the natural
gas industry is negative,
which leads to
opposition from this
industry.
The expected effect
of the ban on the
renewable energy
industry is positive,
which leads to
support from this
industry.
Health [No
information
is given]
Regarding public health
effects, experts expect
that the ban will reduce
indoor air pollution from
gas combustion,
including toxins such as
nitrogen dioxide, which
damages lung and
cardiovascular health in
exposed individuals.
Regarding public
health effects,
experts expect that
the ban will reduce
indoor air pollution
from gas
combustion,
including toxins such
as nitrogen dioxide,
which damages lung
and cardiovascular
health in exposed
children.
Climate
Change [No
information
is given]
When it comes to
climate change effects,
estimates show that an
all-electric single-family
home would reduce
annual greenhouse gas
emissions by 76–88%
compared to a natural
gas-fueled home,
because burning natural
gas creates greenhouse
gas emissions, which
cause climate change.
When it comes to
climate change
effects, estimates
show that natural
gas produces
approximately 50%
less emissions per
unit of energy
compared with coal,
so burning natural
gas instead helps to
cut greenhouse gas
emissions, which
cause climate
change.
When it comes to
climate change
effects, some
experts state that
natural gas
produces less
greenhouse gas
pollution than coal,
while other experts
argue that burning
natural gas still
creates greenhouse
gas emissions,
which cause climate
change.
Politician
Support [No
information
is given]
Politically, this type of
gas ban receives
signicant political
support at the local
level.
Politically, this type
of gas ban receives
signicant political
support at the federal
level.
After reading the statements with information about the natural gas ban policy, respondents were asked
how much they support or oppose the hypothetical natural gas ban policy. The answers to this question
Page 11/16
are on a 1–5 scale provided to respondents with 1 marked as ‘strongly oppose’ and 5 as ‘strongly
support.’ In the survey, respondents were also asked basic demographic and socio-economic questions,
including their 7-point partisan ID, age, gender, race, education, employment status, and the state of
residence. Additionally, we asked how often respondents or their family members cook at home, given
that those who cook at home (and cook more often) may have a greater level of sensitivity to certain
types of information (e.g., health hazards) related to the natural gas ban policy.
Regression and Subgroup Analysis
In order to evaluate the causal effects of provided information about the policy on public support for the
proposed ban, we estimate an ordinary least-squares regression, which is asymptotically equivalent to
average marginal component effect (AMCE) estimators24. The AMCE, which does not rely on functional
form assumptions about the choice probabilities, captures causal average effects of each attribute value
of treatment information over all possible combinations of other attributes on the probability that the
natural gas ban policy will be supported. Note that in the conjoint experiment respondents receive
multiple treatments simultaneously and combinations of treatments, their values and orders are
randomly chosen; thus, estimating regression models with all of the treatments simultaneously does not
lead to biased estimates. To be specic, we estimate the following linear regression model:
Yi=
α
+
β
1Costs +
β
2Industry +
β
3Health +
β
4Climate +
β
5PoliticalSupport +
i
(1)
where i indexes each respondent, and Y denotes the respondents’ support for the natural gas ban policy
on a 5-point scale. Costs, Industry, Health, Climate and PoliticalSupport represent a vector of binary
indicators for specic values in each treatment attribute.
As robustness checks, we estimate OLS models with basic demographic covariates such as gender,
income, education, partisanship, and whether the respondent has a child or not, and ordered probit
models. The results, presented in Supplementary Tables A1 and A2, show that our main ndings remain
essentially unchanged.
We also examine the possibly heterogeneous effects in terms of (1) the respondent’s gender (male or
female), (2) whether the respondent frequently cooks home, (3) whether the respondent resides in a gas
producing state, (4) whether the respondent’s state has already taken policy action on natural gas, and (5)
the party identication (Republicans, Democrats, or independents). We estimate the same OLS models
used in the main analysis, but after splitting the samples for each category to explore heterogeneity in
treatment effects.
To code whether the respondent cooks at home frequently, we asked “How often do you or your family
cook at home?” on a 5-point scale (Never / Rarely / Sometimes / Often / Most of the time). We treat those
who chose either “Often” or “Most of the time” as the individuals who cook at home frequently. We expect
that the effects of health-related information are systematically different for those who cook at home
Page 12/16
often, because they may be more directly exposed to adverse health effects of burning natural gas in the
kitchen.
In addition, we examine potential heterogeneity in the effects of health-related information across
respondents’ states of residence. Specically, we investigate if the effects vary depending on whether the
respondent resides in a gas producing state, and whether the respondent’s state has already adopting a
policy banning natural gas or prohibiting such bans. To identify whether the respondent’s state of
residence produces natural gas, we utilize the state-level natural gas production data from the U.S. Energy
Information Administration (EIA). We coded the respondent as a resident of a state producing natural gas
if her state has produced any natural gas in 2020. Also, based on the S&P Global Market Ingelligence
report on gas ban policies in the U.S. from November 202142, we categorized states into 3 groups—states
advancing natural gas ban policies, states prohibiting natural gas ban policies, and states without any
policy actions in this area. Those states that introduced related policies but have not formally adopted
them are coded as states without policy actions.
For party identication, we asked, “Generally speaking, do you usually think of yourself as a Republican,
Democrat, or as an independent (check the option that best applies)?” on a 7-point scale. We coded a
respondent as a Republican (Democrat), if she answered “Strong Republican (Democrat) or Republican
(Democrat)”. We coded a respondent as independent if she chose either of “Independent, but lean
Republican”, “Independent”, or “Independent, but lean Democrat.”
Declarations
Ethics statement
The University of Buffalo, State University of New York Institutional Review Board (IRB) approved the
survey experiment described in this article. We have also registered a pre-analysis plan (PAP) on a
commonly used repository, but are unable to provide the details due to condentiality in the blind review
process.
References
1. CO2 emissions, metric tons per capita (World Bank, 2018);
https://data.worldbank.org/indicator/EN.ATM.CO2E.PC?
locations=US&most_recent_value_desc=true&year_high_desc=true
2. Burnet, W.M. & Ban, S.D. Changing prospects for natural gas in the United States. Science 244(4902),
305–310 (1989)
3. Electricity production from natural gas sources, % of total (World Bank, 2015);
https://data.worldbank.org/indicator/EG.ELC.NGAS.ZS?
locations=US&most_recent_value_desc=true&year_high_desc=true
Page 13/16
4. Electricity production from coal sources, % of total (World Bank, 2015b);
https://data.worldbank.org/indicator/EG.ELC.COAL.ZS?
locations=US&most_recent_value_desc=true&year_high_desc=true
5. The US Energy Information Administration, How much carbon dioxide is produced when different
fuels are burned? (2021); https://www.eia.gov/tools/faqs/faq.php?id=73&t=11
. Burney, J.A. The downstream air pollution impacts of the transition from coal to natural gas in the
United States. Nat Sustain 3, 152–160 (2020). https://doi.org/10.1038/s41893-019-0453-5
7. Howarth, R. W., A bridge to nowhere: methane emissions and the greenhouse gas footprint of natural
gas. Energy Science & Engineering, 2(2), 47–60 (2014)
. McJeon, H., Edmonds, J., Bauer, N. et al. Limited impact on decadal-scale climate change from
increased use of natural gas. Nature 514, 482–485 (2014). https://doi.org/10.1038/nature13837
9. Nace, T., Plant, L., & Browning, J. The new gas boom: Tracking global LNG infrastructure. Global
Energy Monitor, (2019)
10. Earthworks, “Fracking, methane and climate,” accessed on 16 September 2019,
https://earthworks.org/issues/fracking_methane_and_climate/
11. Losz, A. & Elkind, J. The Role of Natural Gas in the Energy Transition. Columbia SIPA Center of Global
Energy Policy Working Paper (2019).
https://www.energypolicy.columbia.edu/research/commentary/role-natural-gas-energy-transition
12. Witter, R.Z., McKenzie, L., Stinson, K.E., Scott, K., Newman, L.S. and Adgate, J. The use of health
impact assessment for a community undergoing natural gas development. American Journal of
Public Health, 103(6), 1002–1010 (2013)
13. Davis, L.W. What matters for electrication? Evidence from 70 years of US home heating choices.
National Bureau of Economic Research Working Paper 28324 (2021)
http://www.nber.org/papers/w28324
14. California Energy Commission, Local Ordinance Exceeding the 2019 Energy Code, (2021)
https://www.energy.ca.gov/programs-and-topics/programs/building-energy-eciency-
standards/2019-building-energy-eciency-3
15. Bernauer, T. Climate change politics. Annu. Rev. Polit. Sci. 16, 421–448 (2013).
1. Bergquist, M., Nilsson, A., Harring, N. et al. Meta-analyses of fteen determinants of public opinion
about climate change taxes and laws. Nat. Clim. Chang. 12, 235–240 (2022).
https://doi.org/10.1038/s41558-022-01297-6
17. Boudet, H.S. Public perceptions of and responses to new energy technologies. Nat Energy 4, 446–
455 (2019). https://doi.org/10.1038/s41560-019-0399-x
1. Dokshin, F.A. Variation of public discourse about the impacts of fracking with geographic scale and
proximity to proposed development. Nat Energy 6, 961–969 (2021). https://doi.org/10.1038/s41560-
021-00886-7
Page 14/16
19. Hanemann, M. Public support for clean energy. Nature Clim Change 2, 573–574 (2012).
https://doi.org/10.1038/nclimate1640
20. Jacobsson, S. & Lauber, V. The politics and policy of energy system transformation - explaining the
German diffusion of renewable energy technology. Energy Policy 34, 256–276 (2006).
21. Bolsen, T. & Cook, F. L. The polls-trends: public opinion on energy policy: 1974–2006. Public Opin. Q.
72, 364–388 (2008).
22. Farhar, B. Trends: public opinion about energy. Public Opin. Q. 58, 603–632 (1994).
23. Stokes, L., & Warshaw, C. Renewable energy policy design and framing inuence public support in
the United States. Nat Energy 2, 17107 (2017).
24. Hainmueller, J., Hopkins, D. J., & Yamamoto, T, Causal inference in conjoint analysis: Understanding
multidimensional choices via stated preference experiments. Political analysis,
22
(1), 1–30 (2014)
25. Horiuchi, Y., Markovich, Z., & Yamamoto, T., Does conjoint analysis mitigate social desirability bias?.
Political Analysis, 1–15 (2021)
2. Bechtel, M.M. & Scheve, K.F. Mass support for global climate agreements depends on institutional
design.
Proceedings of the National Academy of Sciences
, 110(34), 13763–13768 (2013)
27. Bernauer, T., McGrath, L. Simple reframing unlikely to boost public support for climate policy. Nature
Clim Change 6, 680–683 (2016). https://doi.org/10.1038/nclimate2948
2. Ansolabehere, S. & Konisky, D. M.
Clean and Cheap: How Americans Think About Energy in the Age
of Global Warming
(MIT Press, Cambridge, MA, 2016).
29. Bergquist, P., Konisky, D. M., & Kotcher, J. Energy policy and public opinion: patterns, trends and
future directions. Progress in Energy,
2
(3), 032003. (2020)
30. Aldy, J., Kotchen, M. & Leiserowitz, A. Willingness to pay and political support for a US national clean
energy standard. Nature Clim Change 2, 596–599 (2012).
31. Szabo, J. Energy transition or transformation? Power and politics in the European natural gas
industry’s trasformismo. Energy Research & Social Science,
84
, 102391, (2022)
32. Bayulgen, O. & Benegal, S. Green Priorities: How economic frames affect perceptions of renewable
energy in the United States. Energy Research & Social Science. 47, 28–36, (2019)
33. Sergi, B., Davis, A., & Azevedo, I. The effect of providing climate and health information on support
for alternative electricity portfolios. Environmental Research Letters,
13
(2), 024026. (2018)
34. Howe, P., Mildenberger, M., Marlon, J. et al. Geographic variation in opinions on climate change at
state and local scales in the USA. Nature Clim Change 5, 596–603 (2015).
https://doi.org/10.1038/nclimate2583
35. Janzwood, A., & Millar, H. (2022). Bridge fuel feuds: The competing interpretive politics of natural gas
in Canada.
Energy Research & Social Science
,
88
, 102526. (2022)
3. Bernstein, S., Hoffmann, M. Climate politics, metaphors and the fractal carbon trap. Nat. Clim.
Chang. 9, 919–925 (2019).
Page 15/16
37. Hazboun, S. O., & Boudet, H. S. (2021). Natural gas–friend or foe of the environment? Evaluating the
framing contest over natural gas through a public opinion survey in the Pacic Northwest.
Environmental Sociology
,
7
(4), 368–381. (2021)
3. Sniderman, P. M., Brody, R. A., & Tetlock, P. E.
Reasoning and choice: Explorations in political
psychology
, Cambridge University Press, (1993).
39. Schneider, S. K., & Jacoby, W. G. Public attitudes toward the policy responsibilities of the national and
state governments: Evidence from South Carolina. State Politics & Policy Quarterly,
3
(3), 246–269.
(2003)
40. Bolsen, T., Druckman, J.N. & Cook, F.L. The Inuence of Partisan Motivated Reasoning on Public
Opinion. Polit Behav 36, 235–262 (2014).
41. Ceccoli, S. Explaining Attitudes Toward US Energy Extraction: Offshore Drilling, the Keystone XL
Pipeline, and Hydraulic Fracturing. Social Science Quarterly,
99
(2), 644–664. (2018)
42. S&P Global Market Intelligence. Gas Ban Monitor: Calif. count reaches 50 as West Coast movement
grows. (2021). https://www.spglobal.com/marketintelligence/en/news-insights/latest-news-
headlines/gas-ban-monitor-calif-count-reaches-50-as-west-coast-movement-grows-67732585
Figures
Page 16/16
Figure 1
The effects of each treatment on public support for the natural gas ban policy. Estimates are plotted
relative to the control group (i.e., those who received no information for a given treatment). This plot
presents estimates of the effect of a randomly assigned treatment on a 5-point scale of policy support.
The lines refer to 90% (thicker lines) and 95% (thinner lines) condence intervals based on robust
standard errors clustered by respondents.
Figure 2
The heterogeneous effects of each treatment on public support for the natural gas ban policy by (a)
gender, (b) cooking frequency, (c) residency in gas-producing states, (d) the type of natural gas policy
adopted in the respondent’s state, and (e) partisanship. Estimates are plotted relative to the control group
(i.e., those who received no information for a given treatment). This plot presents estimates of the effect
of randomly assigned treatment on a 5-point scale of policy support. The lines refer to 90% (thicker lines)
and 95% (thinner lines) condence intervals based on robust standard errors clustered by respondents.
Supplementary Figures A1-A5 provide results for other treatment attributes.
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.
Appendixanon.docx
PAPanonNENERGY22040761.pdf
