Sampling and processing of climate change information and disinformation across three diverse countries

Abstract

In popular media, accurate climate information and climate disinformation often coexist and present competing narratives about climate change. As climate disinformation can undermine public support of climate policies and trust in climate science, it is crucial to understand what leads to exposure and acceptance of climate disinformation. Whereas previous research examined the effects of disinformation on climate beliefs, little is known about how people seek climate-related content (Pro-or Anti-climate) and how this varies between cross-cultural contexts. In a preregistered experiment, we studied how individuals sequentially sample and process climate-related information and disinformation. Participants from the U.S., China and Germany (Ntotal = 2226) freely sampled real-world climate related statements. Across 15 rounds, participants decided between two boxes containing Pro-climate or Anti-climate statements , respectively. Overall, reading a statement influenced climate concern in all countries. Participants preferred the box that was better aligned with their initial climate beliefs, and this confirmatory tendency intensified in later rounds. While climate concern was mostly stable, in the U.S., climate concern levels and box choices mutually reinforced each other, leading to greater polarization within the sample over time. The paradigm offers new perspectives on how people process and navigate conflicting narratives about climate change.

Full text

This manuscript is currently under review.
Sampling and processing of climate change information and
disinformation across three diverse countries
Zahra Rahmani Azad1, Tobia Spampatti2,3,4, Sebastian Gluth5, Kim-Pong Tam6, and Ulf J.J. Hahnel1,7
1Faculty of Psychology, University of Basel, Switzerland
2Department of Psychology University of Geneva, Switzerland
3Swiss Center for Affective Sciences, University of Geneva, Switzerland
4Center for Conflict and Cooperation, New York University, United States of America
5Department of Psychology, University of Hamburg, Germany
6The Hong Kong University of Science and Technology, Hong Kong
7Institute for Sustainability Psychology, Leuphana University Lüneburg, Germany
In popular media, accurate climate information and climate disinformation often coexist and
present competing narratives about climate change. As climate disinformation can undermine
public support of climate policies and trust in climate science, it is crucial to understand
what leads to exposure and acceptance of climate disinformation. Whereas previous research
examined the effects of disinformation on climate beliefs, little is known about how people
seek climate-related content (Pro- or Anti-climate) and how this varies between cross-cultural
contexts. In a preregistered experiment, we studied how individuals sequentially sample and
process climate-related information and disinformation. Participants from the U.S., China and
Germany (Ntotal = 2226) freely sampled real-world climate related statements. Across 15
rounds, participants decided between two boxes containing Pro-climate or Anti-climate state-
ments, respectively. Overall, reading a statement influenced climate concern in all countries.
Participants preferred the box that was better aligned with their initial climate beliefs, and this
confirmatory tendency intensified in later rounds. While climate concern was mostly stable,
in the U.S., climate concern levels and box choices mutually reinforced each other, leading to
greater polarization within the sample over time. The paradigm offers new perspectives on how
people process and navigate conflicting narratives about climate change.
Keywords: disinformation, climate action, climate policies, information sampling,
confirmation bias
The scientific consensus on anthropogenic climate change
has been documented by the Intergovernmental Panel on Cli-
mate Change (IPCC) since its establishment in 1988 (Cook
et al., 2016; Lynas et al., 2021). In an extensive global
collaboration, the IPCC presents the best available evidence
about climate change, its risks for humankind, and mitigation
pathways (IPCC, 2023). While the factual evidence seems
indisputable to experts, for non-scientists, navigating a po-
larized information landscape around climate change as a
layperson can create a different experience. Available infor-
mation on climate change on popular or social media out-
lets is more noisy and conflicting, which may make climate
change appear like a contentious or unsettled issue (Bolsen
Correspondence regarding this article should be addressed to
Zahra Rahmani Azad, zahra.rahmani@unibas.ch, Missionsstrasse
62, 54055 Basel. Tobia Spampatti Sebastian Gluth Kim-
Pong Tam Ulf Hahnel
& Shapiro, 2018; Falkenberg et al., 2022; Lewandowsky,
2021; Petersen et al., 2019). Exposure to contrarian claims,
including deliberate disinformation campaigns, can create a
false perception of scientific controversy and leave a person
uncertain about what to believe (Brüggemann & Engesser,
2017; Goldberg & Gustafson, 2022).
Climate disinformation has been around since the pioneer-
ing discoveries of climate change (Farrell, 2016; Oreskes
& Conway, 2011). Fossil fuel industries have deliberately
spread false claims about climate change to protect their
profitable businesses (Cook et al., 2019; Supran et al., 2023;
Supran & Oreskes, 2017). Fossil fuel corporations and other
stakeholders invest heavily in campaigns designed to enhance
or ‘greenwash’ their image and engage in public relations
and lobbying efforts to obstruct climate action (Brulle, 2018;
Brulle & Downie, 2022; Brulle et al., 2020). The disinfor-
mation tactics employed are versatile, ranging from outright
climate denial to the promotion of ineffective climate mea-
sures (Brulle, 2014; Coan et al., 2021). Climate disinfor-
mation can be broadly categorized as either climate science

2RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
denial or climate solution delay (Coan et al., 2021; Lamb
et al., 2020; see also Supran and Oreskes, 2021). Climate
denial questions the existence of global warming, disputes
the scientific basis of anthropogenic climate change, or the
integrity of climate scientists (Dunlap & McCright, 2010).
They also sow doubt about the scientific consensus, portray-
ing climate change as controversial or uncertain (Coan et al.,
2021; Oreskes & Conway, 2011). On the other hand, climate
delay discourses, also known as ‘response skepticism’ try to
delay or thwart adequate climate action. Beyond legitimate
discussions about the costs and challenges of a sustainable
transition, their rhetoric depicts climate action as a pretense
to torpedo a society’s prosperity or freedom (Boussalis &
Coan, 2016; Lamb et al., 2020; Supran & Oreskes, 2021).
Climate contrarians and the climate countermovement strate-
gically employ delay discourses with the aim to obstruct the
effective, large-scale transformations (Doyle, 2011; Pringle
& Robbins, 2022; Supran & Oreskes, 2021). Although these
discourses are more subtle, they are increasingly recognized
as disinformation as they misrepresent robust scientific evi-
dence about appropriate responses to climate change (Lamb
et al., 2020; Painter et al., 2023; Roper et al., 2016).
The spread of climate disinformation presents a barrier to
effective climate action (IPCC, 2023; Meng & Rode, 2019).
Climate disinformation influences public discourses, dimin-
ishes public support of climate policies and erodes trust in
climate science (Judge et al., 2023; Tam & Chan, 2023; Treen
et al., 2020). Disinformation has a detrimental influence at
the societal level as it reduces beliefs in anthropogenic climate
change beliefs and support for climate action (McCright et
al., 2016). Climate disinformation shifts the public discourse
away from important societal debates how to integrate the
best available scientific evidence into an action plan. For
example, discussions on how to decarbonize the energy sec-
tor can get sidetracked by spreading false fears about wind
turbines, inflated expectations about the future availability of
hydrogen power, or claims that using peat-fired heating is
a basic right (Winter et al., 2024). As it targets and better
resonates with specific population segments, climate disin-
formation can contribute to an increasing polarization of the
issue (McCright et al., 2016; Pennycook & Rand, 2021). Po-
larization, as defined here, refers to the divergence of attitudes
toward opposing extremes within a society, resulting in dis-
tinct opinion clusters (Fischer & Frey, 2023; van Eck, 2024).
Ultimately, strong ideological opposition on climate issues
impacts broader societal discourses and creates barriers to
policy implementation (Freelon & Wells, 2020; Kahan et al.,
2017; Oreskes & Conway, 2011; Pennycook & Rand, 2021).
Disinformation also affects people at the psychological
level. In a cross-national randomized controlled trial, Spam-
patti et al. (2024) found that climate disinformation exerted
detrimental effects on affective, cognitive, and behavioral
responses to climate change. Their registered report also
revealed that common interventions to combat disinforma-
tion (Lewandowsky, 2021; Lewandowsky & van der Linden,
2021) could not prevent the negative effects of repeated expo-
sure to climate disinformation (Spampatti, 2024). McCright
et al. (2016) found that statements promoting the benefits
of climate action are less effective when presented alongside
climate denial messages (see also van der Linden et al., 2017).
The findings on the impact of climate disinformation un-
derscore the need to examine how individuals actively seek
climate information. Understanding what predicts selection
and acceptance of either accurate information and disinfor-
mation can serve as a starting point for effective strategies
to curb exposure to disinformation. In this preregistered
study, we employ a novel sequential sampling paradigm to
investigate how individuals select and process contradictory
climate-related statements. Over multiple rounds, partici-
pants freely chose between two boxes featuring contrasting
climate content. After every box choice, participants were
asked to rate how much they agreed with the statement fea-
tured by the box and their momentary climate change con-
cern. The "Anti-climate box" contained climate disinforma-
tion, including statements that questioned climate science or
argued against climate action, reflecting climate denial or
climate delay narratives as described above. In contrast, the
"Pro-climate box" featured statements advocating for decisive
climate action and presenting findings from climate science.
All statements were empirically pre-tested, real-world social
media posts collected from the platform X/Twitter, sourced
from a diverse array of accounts across more than 19 countries
(Spampatti et al., 2023).
Factors leading to Disinformation Exposure
For a person navigating the media landscape filled with
contradictory messages about climate change, discerning
trustworthy sources of information can be difficult. One
way to go about that is to consume content that confirms
one’s established beliefs. Empirical studies show that peo-
ple prefer articles that confirm their existing beliefs (Jonas
et al., 2001; Kerschreiter et al., 2008; Knobloch-Westerwick
& Meng, 2009) and favor news outlets that are aligned with
those beliefs (Chopra et al., 2024; Garrett, 2009; Iyengar
& Hahn, 2009). This phenomenon, known as confirmation
bias, can influence not only content selection but also infor-
mation processing: individuals tend to pay more attention
to confirming evidence, and interpret ambiguities in a way
that favors existing beliefs (Amasino et al., 2024; Corner
et al., 2012; Gluth et al., 2024; Lewandowsky et al., 2012;
Sunstein et al., 2016–2017; Talluri et al., 2018). Social and
cultural contexts may increase tendencies to consume belief-
confirming disinformation. If views on climate change are
divided between different groups within a society, consum-
ing news sources from one’s in-group can serve a sense of
belonging and social connection (Fielding & Hornsey, 2016).

SAMPLING CLIMATE (DIS)INFORMATION 3
Content shared by the out-group can be perceived as threaten-
ing one’s identity and therefore be avoided (West & Iyengar,
2022; Wojcieszak & Garrett, 2018). Confirmatory behavior
can occur both on the level of source selection, and on the
level of content evaluation, such that they agree more with
messages that confirm their pre-existing beliefs.
Our experimental paradigm allows to test for confirmatory
behavior and we expected that climate change beliefs mea-
sured at the beginning of the study to predict subsequent box
choices. Specifically, we expected people with higher initial
climate beliefs to select the Pro-climate box more often and
to agree with its statements more compared to people with
lower initial beliefs (Hypothesis 1). Our second focus was on
the effects of Pro- or Anti-climate statements. We expected
momentary climate change concern to increase after choosing
a statement from the Pro-climate box and to decrease after
reading an Anti-climate statement (Hypothesis 2). Taken to-
gether, the preference for belief-confirming content and its in-
tegration into one’s worldviews can lead to reciprocal effects
that mutually reinforce each other. Through the sampling,
existing attitudes might be bolstered, with climate skeptics’
and believers’ views becoming more extreme over time, po-
tentially leading to a divergence of their attitudes. The rein-
forcing spirals model posits that under some circumstances,
positive feedback loops can occur, where views become more
polarized (Slater, 2007; Spohr, 2017). Utilizing the sequen-
tial nature of our experimental paradigm allowed to test for
such reciprocal effects. We expected that climate change
concern increases during the sampling task in participants
who selected the Pro-climate box more often and decreases
in participants who selected it less (Hypothesis 3).
Comparisons Across National Contexts
Most research on disinformation was conducted in single
Western countries, but patterns of information search and pro-
cessing might vary across national and cultural contexts ( Tam
et al., 2021; but see Ballew et al., 2025; Ballew et al., 2025;
Većkalov et al., 2024; Vlasceanu et al., 2024). Discourses
around climate change and climate mitigation strategies vary
between countries (Hornsey et al., 2018; Lewis et al., 2019;
Smith & Mayer, 2019). Perceived or actual political polar-
ization of climate issues can affect sampling and processing
of climate statements (Judge et al., 2023). A more polar-
ized context might lead to stronger confirmatory tendencies
and a more pronounced acceptance or rejection of statements
compared to one where the topic is less politically charged.
Political polarization around climate-related issues has been
found to strongly increase in the United States over the last
decades (Carmichael et al., 2017; Dunlap et al., 2016; Egan
& Mullin, 2017; McCright & Dunlap, 2011). In the U.S. and
other Western countries, climate skepticism has evolved as
a characteristic stance of political conservativism (Hornsey,
2021; Kahan et al., 2011). Political polarization of climate
issues may be exist in multiple countries, but outside of the
U.S. and Europe, there is less empirical evidence available
(Berkebile-Weinberg et al., 2024; Hornsey et al., 2018).
Compared to climate communication in the Global North,
which is often characterized by pessimistic future scenar-
ios, in China, optimism about climate solutions is prevalent
(Graaf et al., 2024; Guenther et al., 2022; Zeng, 2022). China
presents itself as a global leader in climate action and renew-
able technologies, which contributes to national pride and
support for governmental climate action (Guo et al., 2023).
Generally, there is a high awareness of anthropogenic cli-
mate change concern in China (Lewis et al., 2019; B. Wang
& Zhou, 2020), but conspiracy beliefs against climate ac-
tion also seem to exist (Du et al., 2024; J. C.-E. Liu, 2015,
2023). For instance, some media outlets proclaimed that
climate change was a Western invention to stifle China’s eco-
nomic growth: a narrative that has been similarly proliferated
by Western countermovement (J. C.-E. Liu, 2015). Some
disinformation has been found to promote over-optimistic
assessments of technological solutions (Chu et al., 2023;
Rodenburg, 2024). Given the diversity of national climate
discourses, we hypothesized differences in the sampling pat-
terns between countries. Specifically, we expected to find
larger effects of confirmatory box choices for the U.S. than
in China and Germany. As polarization can lead to hostility
towards opposing views (Slater, 2007), we expected confir-
matory statement evaluation, meaning agreement to belief-
congruent statements, to be more pronounced in the U.S.
than in Germany and China. Comparing information search
of individuals from countries with various climate discourses
can contribute to a better understanding of how aggregate
beliefs form within a population and help to improve climate
communication.
Method
Participants
The preregistration, data, and code are available on OSF
(https://osf.io/hr4pb). In the United States and Germany, data
was collected via a market research institute as part of a larger
panel study. The panel study included two other experimen-
tal tasks for separate research projects that are not reported
here. The order of the three tasks was randomized and we
controlled for the order of the experimental sampling tasks in
the analyses (i.e., whether the sampling task was completed
in first, second or third position). In China, we collected data
separately with a different market research institute. We pre-
registered a sample size of N=3000 participants (n=1000
per country).
We aimed to obtain nationally representative samples, with
simple quotas for age, gender, income and residential area
(urban/rural) in the U.S. and Germany. In China, we could
apply simple quotas for age and gender only. Due to difficul-

4RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
ties with data collection and budget constraints, particularly
in the larger panel study, the actual sample sizes deviated from
the preregistered target sizes. We obtained data from n=880
participants in the United States, n=883 in Germany and
n=1136 in China. We also decided to apply two post-
hoc non-preregistered exclusion criteria to ensure high data
quality and to reduce the likelihood of data points based on
insincere participation (Goodrich et al., 2023; Storozuk et al.,
2020). These exclusion criteria were: 1) following instruc-
tions to move the range sliders to the leftmost position after
round 7, and 2) sampling from either of the two boxes at least
once. Since participants received no prior information about
the boxes, they could only learn about them through sam-
pling. Participants who failed one of these attention checks
were excluded. Analyses conducted on the full sample (see
Supplemental Material) yielded consistent results, confirm-
ing the findings reported in the main manuscript. This proce-
dure differs from the preregistration where we stated to report
the results based on all participants in the main manuscript
and the robustness analysis with attentive participants in the
Supplemental Material. The final sample sizes were n=674
participants from the United States, n=795 from Germany
and n=757 from China. We report the demographic com-
position of the samples in the Supplemental Material.
Material
Sampling Task
Over 15 rounds, participants repeatedly chose between two
boxes (Figure 1). One box consistently revealed a Pro-climate
statement, while the other box consistently revealed an Anti-
climate statement. Participants then rated how much they
agreed with the statement and indicated their momentary
climate change concern on range sliders. Whether Box A
(left-hand side) or Box B (right-hand side) was the Pro- or
Anti-climate box, was counterbalanced between participants.
There was no a-priori information about the content of the
boxes and participants could only learn about them through
selecting them. A pretest (n=500 from the United King-
dom) compared informative box labels ("Pro-climate" box
and "Anti-climate" box) with neutral box Labels ("Box A"
and "Box B"). In this pretest, no differences in the distribu-
tion of box choices were found between the label conditions.
In every round, participants could freely choose between one
of the two boxes to retrieve a randomly drawn statement from
the respective Pro-climate or Anti-climate database. Every
fifth round, participants were asked to rate the boxes. They
indicated their liking for Box A and Box B on one-item mea-
sures, and whether they would follow an account with similar
content on social media (see Supplemental Material for more
detail).
Real-world Stimuli used for the Sampling task
The statements used in the sampling task were derived
from two databases with real-world climate related statements
from public Twitter (now X) posts. Every statement had pre-
viously been manually coded and empirically tested in vali-
dation studies. The "Anti-climate" database featured 75 dis-
information statements downplaying and contesting climate
change (Spampatti et al., 2023). To retrieve eligible state-
ments, a list with Twitter accounts from conservative think-
tanks, well-known members of the climate-countermovement
and climate-denying institutions was curated. All public
posts by these accounts (until October 2022) were down-
loaded and a random subset of 20’000 posts were manually
coded. Manual coding identified climate-relevant posts that
were understandable with minimal background knowledge.
All eligible statements were classified among a taxonomy for
Anti-climate arguments based on Lamb et al. (2020), and
Coan et al. (2021). The remaining 78 statements were em-
pirically tested with a representative sample from the UK
(with N=500). In the validation study, participants rated
the statements on various dimensions (e.g., whether they were
familiar with it, its persuasiveness, political slant or emotional
content). Every statement was rated by 50 to 70 participants.
For the present study, we excluded three items because they
contained slurs against China.
The "Pro-climate" database contained 146 statements that
promoted facts about the severity of climate change and the
urgency of climate action. It was developed similarly to the
Anti-climate database. A list of Twitter accounts from cli-
mate scientists, NGOs, climate activists, science journalists
from more than 19 different countries was curated. All their
public posts in the English language (until January 2023)
were scraped from Twitter, totaling 4.3 million posts. A ran-
domly drawn subset of 20’000 posts was manually coded and,
if applicable, classified among a taxonomy for Pro-climate
communication. Similarly to the Anti-climate database, the
resulting 146 statements were empirically tested in a valida-
tion study with a representative sample from the UK with
N=500.
For both databases, two independent raters carefully fact-
checked every statement against the best available evidence.
All statements from the Anti-climate database were identi-
fied as inaccurate contrarian claims, reflecting climate denial
and delay narratives. While the Pro-climate statements were
largely accurate information about the risks of climate change
and the benefits of climate mitigation measures, we identified
9 inaccurate ones. For every inaccurate statement, debunking
corrections were formulated with scientific references. At the
end of the study, all participants read a note about the sci-
entific consensus on climate change (with country-specific
links to additional official information about climate change).
Additionally, participants received corrections for every false
or misleading statement they had encountered during the sam-

SAMPLING CLIMATE (DIS)INFORMATION 5
pling task. This in-depth debriefing was employed to mitigate
the harm of disinformation exposure. The procedure was ap-
proved by the Faculty of Psychology’s Ethics Committee at
the University of Basel (Application 028-23-1).
Climate Belief Questionnaire
To assess climate change beliefs, we used a six-item cli-
mate belief scale from van Valkengoed et al. (2021). We
used two subscales (belief in anthropogenic climate change
and the severity of the consequences of climate change) with
three items each (e.g., "Climate change will bring about seri-
ous negative consequences."; see Supplemental Material for
the full scale). Agreement to these statements was measured
on 10-point Likert scales (from "not at all" to "fully"). Partic-
ipants scoring low on these scales can be classified as climate
skeptics who are not convinced that human activities cause
climate change with serious negative consequences. High
scorers are often referred to as climate believers, implying
that they acknowledge the human causation and severity of
climate change. Internal consistency the joint scale was Cron-
bach’s alpha = .88 which was higher than for either subscale
alone (α=.79 and α=.83). We therefore aggregated all
items to a score assessing initial beliefs about climate change
before the sampling task.
Procedure
Experimental Procedure
An illustration of the procedure can be found in Figure 1.
First, participants rated their climate change concern on a con-
tinuous range slider (from not at all = 0 to very much = 100)
and filled out the climate beliefs questionnaire as a measure
of baseline attitudes about climate change. Participants then
proceeded to the instructions explaining that they could freely
sample statements as they would when browsing information
online. We intentionally did not provide more specific in-
structions about how they should sample statements because
we were interested in their spontaneous sampling behavior.
Participants completed 15 rounds of sampling climate-related
statements from two boxes as described above. In the end,
they were debriefed with general information about the sci-
entific consensus on climate change and received corrections
for every misinformation they had seen.
Data Analysis
Statistical analyses were conducted using R version 4.3.3
(R Core Team, 2024) and several software packages: lme4
(Bates et al., 2015), lmerTest (Kuznetsova et al., 2017), lavaan
(Rosseel, 2012) and tidyverse (Wickham et al., 2019). When-
ever we complemented the preregistered analyses (https:
//osf.io/hr4pb) with additional analyses, we label them as
exploratory. In all regression models, wherever applicable,
we controlled for the effects of order of the task, country and
round number. We tested the following hypotheses:
H1 (confirmatory box choices): We hypothesized that
participants would prefer the belief-congruent box. Specifi-
cally, we expected participants with stronger climate beliefs to
choose the Pro-climate box and those with climate-skeptical
beliefs to choose the Anti-climate box more often.
H1b: We hypothesized that this relationship would be
stronger in the U.S. than in China or Germany.
H1c (explorative, not preregistered): We tested if the
preference for the belief-congruent box would be more pro-
nounced in later rounds.
For Hypothesis 1, we computed a mixed effects logis-
tic regression model with random intercepts for participants
and box choice as a dependent variable. Deviating from
the preregistration, we did not include random intercepts for
statement, as their inclusion prevented model convergence.
Fixed effects were initial climate beliefs. For H1b, in a sec-
ond step, we computed a model that additionally included
the interaction of initial climate beliefs by country to test
for differences between countries. For H1c, we added the
interaction between initial climate beliefs and round number
to the logistic regression.
H1d (confirmatory statement evaluation): We expected
that higher [lower] initial climate beliefs would predict higher
agreement with Pro-climate statements [Anti-climate state-
ments].
H1e (cross-country differences in confirmatory state-
ment evaluation): We expected that the relationship between
initial climate beliefs and statement agreement by statement
type would be stronger in the U.S. compared to China and
Germany. To test H1d, we computed a linear mixed effect re-
gression model with agreement to statement as the dependent
variable. As fixed effects we included initial climate beliefs,
statement type (Pro- vs. Anti-climate statement) and their
interaction. As random effects, we included participant-ID
and statement-ID. Exploratively, we also included the three-
way interaction with round number to test for temporal effects
in statement evaluation. For H1e, we added the three-way in-
teraction with country to test for country-specific differences
in the hypothesized relationship.
H2 (influences of statements on climate concern): We
expected that box choices impact individual climate change
concern levels, with higher [lower] climate concern immedi-
ately after reading a Pro [Anti] climate action statement.
H2b: Further, we expected that this effect would be moder-
ated by agreement with the statement. This means, the more
a participant agrees with a statement, the more influential the
statement is on the participant’s climate change concern level.
H2c (explorative, non-preregistered): We tested if the ef-
fect of box choice on momentary climate change concern
differed between countries.
We tested Hypothesis 2 with a mixed effects linear regres-

6RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
Sampling and
Rating Statements
Round 1 – 5
Timeline of the Experiment
Baseline Climate
Concern and Beliefs
Box
Rating 1
Sampling and
Rating Statements
Round 6 – 10
Sampling and
Rating Statements
Round 11 – 15
Debriefing with
personalized debunking
of all seen disinformation
Exemplary Sampling Screen.
Box Choice (Screen 1)
Exemplary Sampling Screen.
Statement Rating (Screen 2)
Box
Rating 2
Box
Rating 3
Figure 1
Procedure of the Experiment With 15 Rounds of Statement Sampling and Rating
sion model with random intercepts for participants for state-
ment, and climate change concern as a dependent variable.
Fixed effects were selected box type and initial climate change
concern. To further examine and visualize this effect, we
conducted a robustness check using within-person centered
climate change concern. Within-person centering captured
round-wise variation of a participant’s average concern over
all 16 measurement points. We then tested whether these
round-wise variations in climate change concern differed af-
ter reading a Pro- or Anti-climate statement. For Hypothesis
2B, we added the interaction with statement agreement. For
Hypothesis 2C, we added the interaction with country to test
if the effects of box choice on climate concern varied between
countries.
H3 (reinforcing spiral): We expected the number of Pro-
climate box choices to be predictive for individual changes
in climate concern over the course of 15 rounds. To test
Hypothesis 3, we computed a mixed effects model with ran-
dom intercepts per participant and climate change concern as
dependent variable. As fixed effects, we included the total
number of Pro-climate box choices, round number and the
interaction between the two as well as initial climate change
concern.
H3b (explorative, non-preregistered): To test whether the
reinforcing effects of box choices and climate change concern
differed between countries, we added the three-way interac-
tion between country, number of Pro-climate box choices
and round number to the regression model. Country was
dummy-coded (with China as the reference category). As an
additional exploratory robustness check for Hypothesis 3, we
also fitted a latent growth curve model to test whether the
total number of box choices could significantly predict the
change in climate concern over the course of the 15 rounds.
Additionally, we preregistered a number of secondary hy-
potheses that are reported in detail in the Supplemental Ma-
terial. We had also preregistered additional analyses that in-
clude political worldviews as a predictor variable. Since we
were not able to assess these worldview measures in China,
we report these analyses for the Western countries in the
Supplemental Material.
Results
Distribution of the Main Dependent Variables
Climate Change Concern
Initial climate change concern was above the midpoint in
all countries. In China, average initial climate change con-
cern (M =75.24; 95% CI [73.96,76.52]) was significantly
higher than in the U.S. (M =67.34; 95% CI [64.91,69.78])
and Germany (M =64.21; 95% CI [62.08,66.35]). Initial
climate change concern in China was almost uniformly high,
whereas responses in Western countries were more varied,

SAMPLING CLIMATE (DIS)INFORMATION 7
with a greater concentration at the extremes. This was also
reflected in the bimodality coefficients, a measure to assess
whether a distribution has two rather than one peaks and
which is also used as an indication for polarization of a sam-
ple (Fischer & Frey, 2023). Bimodality coefficients for initial
climate change concern were above the threshold of 0.555
(Freeman & Dale, 2013) in the Western countries (U.S. = .65,
Germany = .71) suggesting a bimodal distribution and below
in China (0.41), consistent with a unimodal distribution. Plots
visualizing the distribution of initial climate change concern
can be found in the Supplemental Material.
Box Choices
The distribution of box choices across all participants and
rounds was relatively even, with 53.1% of all selections made
for the Pro-climate box. Between countries, there was lit-
tle variation in aggregated box choices with 52.4% of Pro-
climate box choices in China, 53.2% in the U.S. and 53.6% in
Germany. The between-participant standard deviation in box
choices was greater in China (SD = 3.21) than in the U.S. (SD
= 2.21) and Germany (SD = 2.39), as indicated by Levene’s
test for equality of variances, F(1,2224) =112.9,p< .001.
This indicates that in China more people showed a stronger
preference for one box than in the Western countries, where
many participants sampled more evenly from the boxes. Dis-
tribution plots can be found in the Supplemental Material.
Predicting Box Selection from Initial Climate Beliefs
We hypothesized (H1) that participants select the belief-
congruent box more often, meaning that initial climate beliefs
can predict box choices in the sampling task. The logistic
mixed effects regression model revealed a significant main
effect of round number on box choices, indicating that in
later rounds, the likelihood to choose the Pro-climate box
increased (ORroundNumber =1.09,95% CI [1.07,1.12],p<
.001). Moreover, there was a main effect of initial beliefs on
box choice (ORinitialBeliefs =1.14,95% CI [1.10,1.18],p<
.001). In line with preregistered Hypothesis 1, higher ini-
tial climate beliefs were associated with a greater propen-
sity to select the Pro-climate box (see Figure 2). Table
1 shows the results of the logistic mixed effects models
computed separately for each country. Descriptively, the
confirmation bias effect was stronger in the Western coun-
tries than in China. Contrary to H1b, we did not find
statistically significant differences in the strength of con-
firmation bias between countries, as indicated by a non-
significant interaction of country and initial climate beliefs
(ORChinaVSGermany =1.03,95% CI [0.94,1.13],p=.565,
ORChinaVsUSA =1.00,95% CI [0.91,1.10],p=.954). To test
explorative Hypothesis 1c, we added the interaction between
initial climate beliefs and round number, which was sig-
nificant (ORinteraction =1.07,95% CI [1.04,1.09],p< .001).
This indicated that the effect of climate beliefs on box choices
was stronger in later rounds (see Figure 3). As one might
contend that this interaction can be attributed to the fact that
participants needed to explore the boxes first before develop-
ing a preference, we tested if the interaction effect remained
significant after removing the first five rounds. A logistic
mixed effects model with the same specification including
only rounds 6 to 15, revealed a smaller, but still significant
effect for the interaction of round number and initial beliefs
(ORinteraction =1.04,95% CI [1.02,1.07],p=.002). This
suggests that even after participants formed initial impres-
sions of the boxes, participants preferred the more belief-
congruent box the more they had sampled.
Predicting Agreement to Statement by Initial Beliefs
Using mixed effects regression models, we tested preregis-
tered Hypothesis 1d that agreement to a statement type (Pro-
vs. Anti-Climate) could be predicted from initial climate be-
liefs. Indeed, the interaction between statement type and ini-
tial climate beliefs was significant (βAnti-ClimateX InitialBeliefs =
−23.29,95% CI [−23.81,−22.76],p< .001). This indicates
that participants with higher climate beliefs agreed with
Anti-climate statements substantially less than participants
with lower initial climate beliefs. Looking at the temporal
component, we found a significant three-way interac-
tion with round number (βAnti-ClimateX InitialBeliefsXRound =
−1.78,95% CI [−2.30,−1.27],p< .001). This indicates
that over the course of the sampling task, the evalua-
tions of the statements became more slanted towards
preferring confirmatory statements. Again, testing for
differences between countries, we added the three-way-
interaction with country to the regression model. This
interaction was significant, suggesting that confirmatory
statement evaluation, while significant in all countries,
was higher in the U.S. (βAnti-ClimateX InitialBeliefsChinaVsUS =
−8.61,95% CI [−10.22,−7.00],p< .001) and in
Germany (βAnti-ClimateX InitialBeliefsChinaVsGer =
−10.64,95% CI [−12.25,−9.03],p< .001) compared
to China.
Influences of Climate Information and Disinformation on
Climate Concern
Next, we tested preregistered Hypothesis 2 that reading a
Pro- or Anti-climate statement affected momentary climate
change concern. As expected, we found a significant effect
of box choice on climate change concern on the trial level
(see Table2) indicating that climate concern levels decreased
after reading an Anti-climate statement and increased after
reading a Pro-climate statement compared to a participant’s
initial level of concern. Supporting preregistered Hypothesis
2B, there was a significant interaction between agreement to
the statement and statement type on climate change concern
(β=−1.84,95% CI [−2.10,−1.59],p< .001) , indicating
that higher agreement to a statement was associated with

8RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
OR = 1.10, p = .005
0%
25%
50%
75%
100%
2 4 6 8 10
Predicted Probability to Choose Pro Climate Box
China
OR = 1.16, p < .001
2 4 6 8 10
Germany
OR = 1.15, p < .001
2 4 6 8 10
U.S.
Figure 2
The Relationship Between Initial Climate Beliefs and Box Choices in the Three Countries.
Note. The slopes show the predicted probabilities to choose the Pro-climate box from the mixed effect logistic regression
models, the points show the actual percentage of Pro-climate box choices by participant. Shaded areas show the 95%-
confidence interval for the predicted values of the regression model.
the statement having more influence on climate change con-
cern. In other words, the content of the statement (Pro-
or Anti-climate) influenced a participant’s climate concern
and more so, the more they agreed with the statement. To
test the explorative H2c, if the influence of box choice on
climate concern differed between countries, we examined
the interaction with country. Indeed, the influence of box
choice was stronger in China than in the Western countries
as indicated in significant interactions effects (China vs. US:
β=0.88,95% CI [0.31,1.46],p=.002; China vs. Germany:
β=0.89,95% CI [0.34,1.45],p=.002). In the individual
country models, the effect remained significant in all three
countries at the p< .001 level; China, however, showed the
strongest effect. In our robustness check for Hypothesis 2,
we examined the deviations in climate change concern from
participants’ average levels of concern after each round. The
results by country are depicted in Figure 4: After reading
an Anti-climate statement, within-person centered climate
change concern was significantly below zero in all countries
(i.e., lower than the average climate concern of a partici-
pant). Similarly, climate change concern was significantly
above participant’s average levels after reading a Pro-climate
statement in all countries. This suggests that individual vari-
ations in climate concern can be predicted from whether a
participant was exposed to a Pro- or Anti-climate statement.

SAMPLING CLIMATE (DIS)INFORMATION 9
Table 1
Logistic mixed effect models predicting box choice from initial beliefs and round number
China U.S. Germany
Predictors Odds
Ratios
95% CI p Odds
Ratios
95% CI p Odds
Ratios
95% CI p
(Intercept) 1.52 1.38 – 1.67 < .001 1.20 1.07 – 1.35 < .001 1.18 1.03 – 1.34 < .001
Initial Beliefs 1.10 1.03 – 1.17 .005 1.15 1.10 – 1.22 < .001 1.16 1.11 – 1.22 < .001
Round number 1.13 1.08 – 1.17 < .001 1.05 1.01 – 1.09 .022 1.10 1.06 – 1.14 < .001
Initial Beliefs ×
round number
1.06 1.02 – 1.10 .006 1.05 1.01 – 1.09 .017 1.08 1.04 – 1.12 < .001
Box A is Pro- cli-
mate
0.55 0.48 – 0.63 < .001 0.89 0.82 – 0.98 .012 0.93 0.85 – 1.02 .115
Position Sampling
Task
1.00 0.95 – 1.05 .931 1.01 0.96 – 1.07 .639
Random Effects
σ23.29 3.29 3.29
τ00 0.55participantID 0.07participantID 0.14participantID
ICC 0.14 0.02 0.04
N757participantID 674participantID 795participantID
Note. Regression models separate by country for effect of initial climate beliefs and round number on box choice. Counterbalancing
(whether Box A was the Pro- or Anti-climate box) and position/order of the sampling task within the larger panel study (not relevant for
China, where data was collected independently) were added as covariates.
Table 2
Results of a mixed effects regression model of the effect of box choice
on individual level climate change concern.
Roundwise Climate Concern
Predictors Estimates 95% CI p
(Intercept) 7.57 5.74 – 9.40 < .001
Box choice [Anti-climate] -1.29 -1.52 – -1.06 < .001
Initial Climate Concern 0.89 0.87 – 0.91 < .001
Country [Germany] 0.89 -0.47 – 2.25 .198
Country [USA] 0.33 -1.12 – 1.77 .658
Round number 0.00 -0.02 – 0.03 .833
Position Sampling Task 0.10 -0.61 – 0.80 .787
Box A is Pro-climate 1.23 0.26 – 2.21 .013
Random Effects
σ291.24
τ00participantID 131.57
τ00statementID 0.27
ICC 0.59
NparticipantID 2226
NstatementID 219
Note. Regression Model predicting variation in climate change concern af-
ter reading a Pro- or Anti- climate statement. The model controls for coun-
try, round number, counterbalancing (whether Box A was the Pro- or Anti-
climate box) and position/order of the sampling task within the larger panel
study (not relevant for China, where data was collected independently).

10 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
40%
45%
50%
55%
60%
65%
0 5 10 15
Round
Predicted Probability to choose Pro−Climate Box
Initial Beliefs −1 SD Mean +1 SD
Figure 3
The Effect of Initial beliefs on Box Choices Becomes More Pronounced With Increasing Round Number.
Note. The lines show how the predicted probability of choosing the Pro-climate box changes over the course of rounds,
depending on participants’ initial climate beliefs. Shaded areas show the 95%-confidence Intervals for predicted values of the
regression model.
Reinforcing Spiral Between Climate Concern and Infor-
mation Sampling
To examine preregistered Hypothesis 3, we investigated
whether the evolution of climate change concern varied as
a function of participants’ sampling patterns. Hypothesis 3
predicts a cumulative effect: multiple Pro-climate box choices
should lead to a steady increase in climate change concern.
Figure 5 displays trajectories of climate change concern for
participants grouped into tertiles based on the frequency of
Pro-climate box choices (many, medium and few). First, it
can be seen that the amount of box choices differentiates
between different levels of climate concern in the Western
countries, and, to a much lesser degree, in China. Second, it
also shows how climate change concern changes throughout
the experiment for participants with different box preferences.
For the U.S., visual inspection suggests a reinforcing feedback
loop, where participants’ initial climate change concern drifts
towards more extreme points. That is. the sampling task led to
an increase in the gap between participant expressing high and
low climate change concern. Statistical tests of this hypoth-
esis revealed that the interaction between round number and
number of Pro-climate box choices was significant and posi-
tive (β=0.14,95% CI [0.04; 0.24],p=.009). This indicates
that climate change concern increased over time when partici-
pants sampled more frequently from the Pro-climate box, and
decreased when participants sampled more frequently from
the Anti-climate box. The main effect of round number sug-
gested no overall uniform trend in climate concern evolution

SAMPLING CLIMATE (DIS)INFORMATION 11
China
Germany
U.S.
Anti−Climate Pro−Climate Anti−Climate Pro−Climate Anti−Climate Pro−Climate
−2
−1
0
1
2
Mean Change in Climate Concern
Figure 4
Mean Change in Within-person Centered Climate Change Concern After Reading a Pro- or Anti-climate Statement.
Note. The y-axis displays unstandardized deviations from a person’s mean climate concern (= zero) after reading a Pro- or
Anti-climate statement. The error-bars represent the 95% confidence interval of the mean.
(β=0.04,95% CI [−0.07; 0.14],p=.489). The main effect
of number of Pro-climate box choices on climate concern
was significant (β=0.62,95% CI [0.11; 1.12],p=.016).
Taken together, this result suggests that the effects of climate
concern on box choice and vice versa mutually influenced
each other leading to reinforcing spirals, where the relation-
ship between these two variables becomes stronger over time.
When adding the three-way interaction with country to test
for between-country differences in the reinforcing spirals, this
revealed that the effect was driven by the U.S. sample. Neither
for China (β=0.01,95% CI [−0.14; 0.15],p=.992), nor for
Germany (β=0.14,95% CI [−0.10; 0.38],p=.268), there
was a significant interaction between total Pro-climate box
choices and round number. Only the interaction with the U.S.
was significant, suggesting that participants from the U.S.
were driving the significant result in the overall regression
model that did not differentiate between countries. Specifi-
cally, the three-way interaction between the U.S., number of
Pro-climate box choices and round number was significant
(β=0.43,95% CI [0.16; 0.70],p=.002), suggesting that in
the U.S., different sampling patterns were predictive for the
change in climate concern over 15 rounds. Indeed, when com-
puting a regression model for U.S. participants only, again the
interaction between total Pro-climate box and round number
was significant (β=0.37,95% CI [0.16; 0.57],p=.001).
That is, in the U.S., climate change concern increased over
the 15 rounds for participants who frequently sampled from
the Pro-climate box and decreased for those who sampled
infrequently. As an additional, explorative test of Hypothesis
3, we fitted a latent growth curve model to examine the effect

12 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
of Pro-climate box choices on change in climate concern over
15 rounds, controlling for initial climate concern. The model
exhibited very good fit to the data (CFI = 99.2, TLI = 99.2,
RMSEA = 0.043 (90% CI = [0.04, 0.046]), SRMR = 0.008).
Initial climate concern was strongly and positively associated
with the intercept of climate concern (standardized coeffi-
cient =0.92,S E =0.01,p< .001), indicating that climate
change concern was largely stable across the 15 rounds of
the task. Cumulative Pro-climate box choices had a posi-
tive, and weak association with the slope of climate concern
(standardized c=0.06,S E <0.001,p=.044). This effect
was largely driven by the data from the U.S. sample. When
fitting the same model for the U.S. data only, the effect was
significant and the effect size was larger (standardized coef-
ficient =0.50,S E =0.004,p=.014), suggesting that in the
United states individual changes in climate concern could be
predicted by the number of Pro-climate box choices. When
fitting the same model for the Chinese or German sample sep-
arately, the effect of Pro-climate box choices on the individual
slopes of climate concern was insignificant (p=0.935 and
p=0.176, respectively). This finding corresponds with Fig-
ure 5, where visual inspection suggests that individual slopes
in climate concern are positive for participants who sample
many Pro-climate statements and negative for participants
with few Pro-climate box choices.
Discussion
This cross-cultural, pre-registered experiment used a novel
sequential information sampling paradigm to investigate in-
formation selection, belief updating and the interplay between
the two using real-world statements about climate change.
Our approach offers insights into how people process in-
formation and update their beliefs over time through select-
ing climate information and disinformation. The sequential
nature of the paradigm allowed us to examine aggregated
effects of statement sampling that may arise from selective
exposure to one-sided messages. Across all three countries,
China, Germany and the U.S., we observed confirmatory
box sampling, supporting our first hypotheses. Participants
were more likely to choose the box that was better aligned
with their initial views on climate change. Preferentially
consuming sources with one-sided and confirmatory content,
continually reinforces existing views. The selective exposure
can lead to increased certainty in one’s beliefs and reduce
openness to alternative perspectives. For example, consump-
tion of conservative media has been linked to decreased trust
in scientists and reduced acknowledgment of anthropogenic
climate change (Ophir et al., 2024). While climate disin-
formation can negatively affect everyone, some populations
segments may be more affected by it, as they actively seek
out sources featuring disinformation (McCright et al., 2016;
Roozenbeek et al., 2022).
Exploratively, we found that the preference for the more
belief-congruent box intensified over time, demonstrating that
the more participants learned about both boxes, the more se-
lective they became. This finding that confirmatory source
selection increased over time is consistent with reinforce-
ment learning and Bayesian belief updating where a belief-
confirming source can be perceived as increasingly credible
(Druckman & McGrath, 2019). One might also become
information fatigued which may create a need for messages
that are easier to process, which are more likely to be belief-
congruent. Belief-congruent messages cause less cognitive
dissonance and can be more easily integrated into one’s ex-
isting beliefs and be more emotionally rewarding (Jonas et
al., 2001). Emotion regulation may not only play a role in
confirmatory source selection, but may predict consumption
of climate disinformation more generally. Confrontation with
climate change can evoke unpleasant feelings of helplessness,
anxiety or guilt (Ágoston et al., 2022). People tend to avoid
unsettling information, which can include communication
about climate risks, and may prefer to maintain optimism or
comfort (Charpentier et al., 2018; Sharot & Garrett, 2016;
Sharot & Sunstein, 2020). Therefore, to maintain positive
emotions or to avoid potentially disturbing or worrisome news
about climate risks, one might prefer comforting disinforma-
tion that downplays these risks or promises easy solutions
(Golman et al., 2017; Yang & Kahlor, 2013).
Confirmatory information search can be exacerbated by
a confirmatory evaluation of climate-related messages. In
our study, participants agreed more with statements that con-
firmed their initial beliefs and rejected those that did not.
Again, there was a temporal dynamic such that the confirma-
tory evaluation became stronger over time. After participants
had formed an impression of a source, similar to a halo ef-
fect, this impression shaped how its messages were evaluated.
Comparing confirmatory evaluations between countries, we
found that it was stronger in the U.S. and in Germany than in
China. This might indicate that in the context of higher attitu-
dinal and political polarization of climate change, statements
are accepted or rejected more decisively than in the absence
of a polarized discourse (Hutchens et al., 2019).
Confirming our second hypothesis, climate change con-
cern varied systematically after box choices: it increased af-
ter reading a Pro-climate statement and decreased after Anti-
climate statements. We interpret the findings as evidence
that the real-world climate statements exerted persuasive ef-
fects among diverse audiences. We found that the statements
influenced climate-change concern across the entire sample,
which is in line with existing findings that counter-attitudinal
messages can have persuasive effects (Guess & Coppock,
2020; Haglin, 2017; Levitan & Visser, 2008). The effects
of the statements on climate change concern were small but
significant and consistent across all countries. Small effect
sizes are not surprising given that single statements are un-
likely to drastically alter climate change concern. Climate

SAMPLING CLIMATE (DIS)INFORMATION 13
China
Germany
U.S.
0 5 10 15 0 5 10 15 0 5 10 15
50
60
70
80
Round Number
Climate Change Concern
Total Pro Box Choices low middle upper tercile
Figure 5
The Evolution of Climate Change Concern (measured from 0 to 100) across the sampling task.
Note. The evolution of climate change concern across 15 rounds of choosing between Pro and Anti-climate statements is
shown for participants in the upper, middle or lower tercile for total Pro-climate box choices. The shaded areas show the
95%-confidence interval for the smoothed regression line based on a linear model.
change concern is likely influenced incrementally by numer-
ous messages over prolonged periods and can be reinforced
by the algorithmic curation of online media. Social media
platforms and personalized news feeds quickly identify the
content that resonates with their users. The algorithms pro-
vide their users with individually tailored feeds that are opti-
mized for engagement and dwell time rather than its factual
accuracy (Lazer et al., 2018; Rathje et al., 2024; Vosoughi
et al., 2018). On top of inherent preferences for attitude-
congruent information, recommendations from online media
can exacerbate the exposure to one-sided messages and con-
tribute to a drift towards more extreme opinions (Baumann
et al., 2020). However, other results question the influence
of content curation on the formation of extreme attitudes and
emphasize the importance of endogenous information search,
where users choose what to engage with themselves (Garrett,
2017; N. Liu et al., 2025).
The effects of climate messages on climate change concern
highlight the importance of considering the cumulative and
self-reinforcing impact of climate (dis)information sampling.
Albeit the limitations of a short-term online experimental
setting, the sequential sampling design allowed us to mimic
some aspects of real-world information consumption and pro-
cessing. Even without any influence of algorithmic curation,
we observed some evidence for an interplay between confir-
matory statement selection and cumulative effects on climate
concern in the U.S. sample. On an aggregate level, this
resulted in a positive feedback loop between preferring con-

14 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
firmatory statements and the statements’ persuasive effects:
while already starting at different levels, climate change con-
cern continued to diverge over the course of the experiment
between participants who predominantly chose the Pro- or
Anti-climate box. The U.S. ranks among the most polar-
ized countries with respect to climate-related issues (Egan
& Mullin, 2017; McCright & Dunlap, 2011). In our data,
we also find attitudinal polarization in climate change beliefs
and concern, with both measures correlating with political
attitudes. Taken together, our findings are in line with the
theoretical predictions of the reinforcing spirals model, which
posits that positive feedback loops are context-dependent and
more likely to occur in more polarized societies (Slater, 2015).
In contrast to the findings from the U.S., in the Chinese and
German samples, climate change concern remained stable
for most participants irrespective of their sampling pattern,
which underlines the importance to consider cultural speci-
ficities that contextualize empirical findings.
This raises the question of how the influence of disinfor-
mation on public opinion can be curbed. For a layperson, it
can be difficult to discern which messages are credible when
they navigate a jungle of contradictory content competing
for the one’s attention. In the online sphere, a person can
find reliable, evidence-backed information next to false and
unfounded claims, irrelevant remarks, noisy or inaccurate
statements, or emotional exaggerations. Identifying and fil-
tering relevant and trustworthy information sources can be a
complex task. Especially when resources and time are lim-
ited, a person may become vulnerable to emotionally charged
or easily digestible disinformation or disengage altogether
due to information overload (Hahnel et al., 2020; Heseltine
et al., 2024; Korteling et al., 2023; Momsen & Ohndorf,
2022). A person may rely on heuristics, such as accepting the
information that is shared by one’s favorite political party or
evaluating a messenger based on unrelated viewpoints.
Removing disinformation in the online sphere is a noto-
riously difficult task and the boundaries between free ex-
pressions of thoughts and disinformation campaigns can be
blurred. Fact-checking media articles and online posts is
resource-intensive, unscalable, and also contentious (Graves,
2018; Pennycook et al., 2020). As some social media plat-
forms turn away from fact checking and content moderation,
alternative strategies are needed (CBS News, 2025; Zannet-
tou, 2021). Our findings suggest that accurate scientific in-
formation can have protective effects, which underscores the
importance of maximizing encounters with credible climate
content. Practically, interspersing online news feeds with
officially verified information might be a promising inter-
vention that merits further investigation. Redirecting people
towards trustworthy science communicators that appeal to
diverse populations and universally agreed values can be a
key to reduce the impact of the climate countermovement.
Exposing the bias of some extreme influencers and offering
more promising alternatives can restore the trust in actors that
promote science-based content (Rathje et al., 2024).
Limitations
A potential limitation of our research is that the Pro- and
Anti-climate statements were derived from the platform X,
which is not available in China. While the authors of the state-
ments were selected to be diverse and come from 19 different
countries, none of the statements was from a Chinese author.
Therefore, the narratives reflected in these statements may
have been less familiar in the Chinese sample. We carefully
checked that the statements did not include specific references
to places or people or cultural references or required back-
ground knowledge to be understandable. Utilizing a database
of diverse and empirically pretested, real-world social media
posts, is an important step to increase the ecological validity
and reduce the likelihood for cultural biases (Westfall et al.,
2014) . Unlike artificial stimuli often used in experimental
research, real-world posts better resemble messages encoun-
tered in everyday life (Pennycook et al., 2021). Neverthe-
less, we cannot fully rule out that the statements may have
sounded less plausible to a Chinese compared to a Western
audience. Given that there were only two news sources par-
ticipants could select from, participants may have implicitly
assumed that they were expected to sample from both boxes
in a balanced way. However, it is also possible that people
have an inherent preference for a balanced consumption of
contradictory perspectives as this is sometimes perceived as
more neutral (Mont’Alverne et al., 2023). It is common in
experimental research to use binary choices to elicit prefer-
ences. Empirical evidence from process-tracing experiments
finds that attention is not split evenly between two competing
choice options (Motoki et al., 2021). Instead, findings sug-
gests that participants dwell more on choice options that are
instrumental to their goals (Sepulveda et al., 2020).
Beside climate concern, behavioral outcomes, policy sup-
port or voting behavior are a more tangible manifestation
of climate-related attitudes. Previous research found imper-
fect correlations between climate change concern, climate
beliefs and sustainable behavior (Spampatti et al., 2024)
(Bosshard et al., 2024; Spampatti, 2024). Since we did not
measure sustainable behavior, we do not know whether the
(dis)information sampling affected climate-relevant behavior.
However, the relationship with policy support, which is rele-
vant for large scale systemic implementation of climate mea-
surements, is usually highly correlated with climate concern
(Drews & Van Den Bergh, 2016; S. Wang et al., 2018)
Conclusion
Navigating the complex online information environment
surrounding climate change requires individuals to discern
credible sources from competing and counter-factual voices.
This study contributes to our understanding of the interplay

SAMPLING CLIMATE (DIS)INFORMATION 15
between selective exposure, belief updating, and attitude po-
larization in the context of climate change communication.
By demonstrating the cumulative impact of message exposure
and the context-dependent nature of reinforcing spirals, our
findings advance existing theories of belief formation and
contribute to a better understanding of cross-country differ-
ences in climate change discourses. We also demonstrate
that against diverse cultural backdrops, these processes un-
fold differently highlighting the role of contextual factors on
individual belief formation. Providing widespread access
to and fostering trust in credible scientific information on
climate change is key to countering the influence of climate
disinformation. Ultimately, effective science communication
should empower individuals to make better informed deci-
sions and can contribute to public discourses with a better
understanding of the scientific evidence.
Funding Statement. This research was funded as part
of the Eccellenza Grant (PCEFPI_203283) from the Swiss
National Science Foundation awarded to U.J.J.H.
Competing Interests. None.
Data Availability Statement. The supplementary mate-
rial, data and analysis scripts for this article can be found at
https://osf.io/e78da/. All implemented experimental condi-
tions and all measured variables and materials are disclosed.
Ethical Standards. The research meets all ethical guide-
lines, including adherence to the legal requirements of the
study country, and was approved by the Ethics Committee of
the University of Basel (003-23-1).
Author Contributions. Conceptualization: Z.R.A; T.S.;
S.G.; K-P.T.; U.J.J.H Methodology: Z.R.A; T.S.; S.G.; K-
P.T.; U.J.J.H Software: Z.R.A Investigation: Z.R.A. Data
curation: Z.R.A Formal analysis: Z.R.A Data visualization:
Z.R.A Writing – original draft: Z.R.A; Writing – review
& editing: Z.R.A; T.S.; S.G.; K-P.T.; U.J.J.H Supervision:
T.S.; S.G.; K-P.T.; U.J.J.H Funding acquisition: U.J.J.H All
authors approved the final submitted draft.
References
Ágoston, C., Csaba, B., Nagy, B., Kőváry, Z., Dúll, A.,
Rácz, J., & Demetrovics, Z. (2022). Identifying
Types of Eco-Anxiety, Eco-Guilt, Eco-Grief, and
Eco-Coping in a Climate-Sensitive Population: A
Qualitative Study. International Journal of Envi-
ronmental Research and Public Health,19(4), 2461.
https://doi.org/10.3390/ijerph19042461
Amasino, D. R., Pace, D. D., & van der Weele, J. J. (2024).
Fair Shares and Selective Attention. American Eco-
nomic Journal: Microeconomics,16(4), 259–290.
https://doi.org/10.1257/mic.20220275
Ballew, M. T., Thomas-Walters, L., Goldberg, M. H., Verner,
M., Lu, J., Marshall, J., Rosenthal, S. A., & Leis-
erowitz, A. (2025). Climate change messages can
promote support for climate action globally. Global
Environmental Change,90, 102951. https://doi.org/
10.1016/j.gloenvcha.2024.102951
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fit-
ting linear mixed-effects models using lme4. Jour-
nal of Statistical Software,67(1), 1–48. https://doi.
org/10.18637/jss.v067.i01
Baumann, F., Lorenz-Spreen, P., Sokolov, I. M., & Starnini,
M. (2020). Modeling Echo Chambers and Polariza-
tion Dynamics in Social Networks. Physical Review
Letters,124(4), 048301. https://doi.org/ 10 . 1103 /
PhysRevLett.124.048301
Berkebile-Weinberg, M., Goldwert, D., Doell, K. C., Van
Bavel, J. J., & Vlasceanu, M. (2024). The differen-
tial impact of climate interventions along the politi-
cal divide in 60 countries. Nature Communications,
15(1), 3885. https://doi.org/10.1038/s41467- 024-
48112-8
Bolsen, T., & Shapiro, M. A. (2018). The US News Media,
Polarization on Climate Change, and Pathways to
Effective Communication. Environmental Commu-
nication. Retrieved February 13, 2024, from https://
www.tandfonline.com/doi/full/10.1080/17524032.
2017.1397039
Boussalis, C., & Coan, T. G. (2016). Text-mining the sig-
nals of climate change doubt. Global Environmen-
tal Change,36, 89–100. https://doi.org/10.1016/j.
gloenvcha.2015.12.001
Brüggemann, M., & Engesser, S. (2017). Beyond false bal-
ance: How interpretive journalism shapes media
coverage of climate change. Global Environmental
Change,42, 58–67. https : / / doi . org / 10 . 1016 / j .
gloenvcha.2016.11.004
Brulle, R. J. (2014). Institutionalizing delay: Foundation
funding and the creation of U.S. climate change
counter-movement organizations. Climatic Change,
122(4), 681–694. https://doi.org/10.1007/s10584-
013-1018-7
Brulle, R. J. (2018). The climate lobby: A sectoral analysis
of lobbying spending on climate change in the USA,
2000 to 2016. Climatic Change,149(3), 289–303.
https://doi.org/10.1007/s10584-018-2241-z
Brulle, R. J., Aronczyk, M., & Carmichael, J. (2020). Corpo-
rate promotion and climate change: An analysis of
key variables affecting advertising spending by ma-
jor oil corporations, 1986–2015. Climatic Change,
159(1), 87–101. https://doi.org/ 10 .1007/s10584-
019-02582-8
Brulle, R. J., & Downie, C. (2022). Following the money:
Trade associations, political activity and climate

16 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
change. Climatic Change,175(3), 11. https : / / doi.
org/10.1007/s10584-022-03466-0
Carmichael, J. T., Brulle, R. J., & Huxster, J. K. (2017). The
great divide: Understanding the role of media and
other drivers of the partisan divide in public con-
cern over climate change in the USA, 2001–2014.
Climatic Change,141(4), 599–612. https://doi.org/
10.1007/s10584-017-1908-1
CBS News. (2025, January 7). Meta to end fact-checking,
replacing it with community-driven system akin to
Elon Musk’s X. Retrieved January 8, 2025, from
https:// www.cbsnews.com/news/meta- facebook-
instagram-fact-checking-mark-zuckerberg/
Charpentier, C. J., Bromberg-Martin, E. S., & Sharot, T.
(2018). Valuation of knowledge and ignorance in
mesolimbic reward circuitry. Proceedings of the
National Academy of Sciences,115(31), E7255–
E7264. https://doi.org/10.1073/pnas.1800547115
Chopra, F., Haaland, I., & Roth, C. (2024). The Demand
for News: Accuracy Concerns Versus Belief Confir-
mation Motives. The Economic Journal,134(661),
1806–1834. https://doi.org/10.1093/ej/ueae019
Chu, J., Zhu, Y., & Ji, J. (2023). Characterizing the se-
mantic features of climate change misinformation
on Chinese social media. Public Understanding of
Science,32(7), 845–859. https://doi.org/10.1177/
09636625231166542
Coan, T. G., Boussalis, C., Cook, J., & Nanko, M. O.
(2021). Computer-assisted classification of contrar-
ian claims about climate change. Scientific Reports,
11(1), 22320. https://doi.org/10.1038/s41598-021-
01714-4
Cook, J., Oreskes, N., Doran, P. T., Anderegg, W. R. L.,
Verheggen, B., Maibach, E. W., Carlton, J. S.,
Lewandowsky, S., Skuce, A. G., Green, S. A., Nuc-
citelli, D., Jacobs, P., Richardson, M., Winkler, B.,
Painting, R., & Rice, K. (2016). Consensus on
consensus: A synthesis of consensus estimates on
human-caused global warming. Environmental Re-
search Letters,11(4), 048002. https://doi.org/ 10.
1088/1748-9326/11/4/048002
Cook, J., Supran, G., Lewandowsky, S., Oreskes, N., &
Maibach, E. (2019). America Misled: How the fossil
fuel industry deliberately misled Americans about
climate change. Retrieved December 6, 2024, from
https : / / policycommons . net / artifacts / 2193396 /
america-misled/2949373/
Corner, A., Whitmarsh, L., & Xenias, D. (2012). Uncertainty,
scepticism and attitudes towards climate change: Bi-
ased assimilation and attitude polarisation. Climatic
Change,114(3–4), 463–478. https : / / doi . org/ 10 .
1007/s10584-012-0424-6
Doyle, J. (2011, September). Where has all the oil gone? BP
branding and the discursive elimination of climate
change risk. In N. Heffernan & D. A. Wragg (Eds.),
Culture, environment and ecopolitics (pp. 200–225).
Cambridge Scholars Publishing.
Drews, S., & Van Den Bergh, J. C. (2016). What explains
public support for climate policies? A review of
empirical and experimental studies. Climate Policy,
16(7), 855–876. https://doi.org/10.1080/14693062.
2015.1058240
Druckman, J. N., & McGrath, M. C. (2019). The evidence for
motivated reasoning in climate change preference
formation. Nature Climate Change,9(2), 111–119.
https://doi.org/10.1038/s41558-018-0360-1
Du, J., Liu, J. C.-E., & Mix, T. L. (2024). China deserves
its hamburger: The controversy over WildAid’s Shu
Shi campaign in China. Environmental Sociology,
10(3), 321–332. https://doi.org/10.1080/23251042.
2024.2319370
Dunlap, R. E., & McCright, A. M. (2010). Climate change
denial: Sources, actors and strategies. In Routledge
Handbook of Climate Change and Society. Rout-
ledge.
Dunlap, R. E., McCright, A. M., & Yarosh, J. H. (2016).
The Political Divide on Climate Change: Partisan
Polarization Widens in the U.S. Environment: Sci-
ence and Policy for Sustainable Development,58(5),
4–23. https :/ / doi . org / 10 . 1080 / 00139157 . 2016 .
1208995
Egan, P. J., & Mullin, M. (2017). Climate Change: US Public
Opinion. Annual Review of Political Science,20(1),
209–227. https://doi.org/10.1146/annurev-polisci-
051215-022857
Falkenberg, M., Galeazzi, A., Torricelli, M., Di Marco, N.,
Larosa, F., Sas, M., Mekacher, A., Pearce, W., Zollo,
F., Quattrociocchi, W., & Baronchelli, A. (2022).
Growing polarization around climate change on so-
cial media. Nature Climate Change,12(12), 1114–
1121. https://doi.org/10.1038/s41558-022-01527-x
Farrell, J. (2016). Corporate funding and ideological polar-
ization about climate change. Proceedings of the
National Academy of Sciences,113(1), 92–97. https:
//doi.org/10.1073/pnas.1509433112
Fielding, K. S., & Hornsey, M. J. (2016). A Social Identity
Analysis of Climate Change and Environmental At-
titudes and Behaviors: Insights and Opportunities.
Frontiers in Psychology,7. https://doi.org/10.3389/
fpsyg.2016.00121
Fischer, O., & Frey, R. (2023). The many flavors of polariza-
tion: A systematic comparison of different concep-
tualizations and contexts. https://doi.org/10.31234/
osf.io/bv496

SAMPLING CLIMATE (DIS)INFORMATION 17
Freelon, D., & Wells, C. (2020). Disinformation as Political
Communication. Political Communication,37(2),
145–156. https://doi.org/10.1080/10584609.2020.
1723755
Freeman, J. B., & Dale, R. (2013). Assessing bimodality
to detect the presence of a dual cognitive process.
Behavior Research Methods,45(1), 83–97. https :
//doi.org/10.3758/s13428-012-0225-x
Garrett, R. K. (2009). Echo chambers online?: Politically
motivated selective exposure among Internet news
users. Journal of Computer-Mediated Communica-
tion,14(2), 265–285. https:/ / doi.org/ 10 .1111/ j .
1083-6101.2009.01440.x
Garrett, R. K. (2017). The “Echo Chamber” Distraction: Dis-
information Campaigns are the Problem, Not Au-
dience Fragmentation. Journal of Applied Research
in Memory and Cognition,6(4), 370–376. https :
//doi.org/10.1016/j.jarmac.2017.09.011
Gluth, S., Deakin, J., & Rieskamp, J. (2024). A Theory of
Multi-Attribute Search and Choice. https://doi.org/
10.31234/osf.io/3qzak
Goldberg, M. H., & Gustafson, A. (2022). Communicating
the Scientific Consensus on Climate Change: Di-
verse Audiences and Effects Over Time. Environ-
ment and Behavior.
Golman, R., Hagmann, D., & Loewenstein, G. (2017). In-
formation Avoidance. Journal of Economic Litera-
ture,55(1), 96–135. https: / /doi.org/10.1257/jel .
20151245
Goodrich, B., Fenton, M., Penn, J., Bovay, J., & Mountain,
T. (2023). Battling bots: Experiences and strategies
to mitigate fraudulent responses in online surveys.
Applied Economic Perspectives and Policy,45(2),
762–784. https://doi.org/10.1002/aepp.13353
Graaf, J. D., Bal, M., Wit, J. de, & Stok, M. (2024). Climate
Change Doom Communication From a Fear Appeal
Perspective. European Journal of Health Communi-
cation,5(4), 26–50. https://doi.org/10.47368/ejhc.
2024.402
Graves, L. (2018). Boundaries Not Drawn: Mapping the in-
stitutional roots of the global fact-checking move-
ment. Journalism Studies,19(5), 613–631. https :
//doi.org/10.1080/1461670X.2016.1196602
Guenther, L., Brüggemann, M., & Elkobros, S. (2022). From
Global Doom to Sustainable Solutions: Interna-
tional News Magazines’ Multimodal Framing of our
Future with Climate Change. Journalism Studies,
23(1), 131–148. https://doi.org/10.1080/1461670X.
2021.2007162
Guess, A., & Coppock, A. (2020). Does Counter-Attitudinal
Information Cause Backlash? Results from Three
Large Survey Experiments. British Journal of Polit-
ical Science,50(4), 1497–1515. https://doi.org/10.
1017/S0007123418000327
Guo, J., Huang, X., & Fang, K. (2023). Authoritarian Envi-
ronmentalism as Reflected in the Journalistic Sourc-
ing of Climate Change Reporting in China. Envi-
ronmental Communication,17(5), 502–517. https:
//doi.org/10.1080/17524032.2023.2223774
Haglin, K. (2017). The limitations of the backfire effect. Re-
search & Politics,4(3), 2053168017716547. https:
//doi.org/10.1177/2053168017716547
Hahnel, U. J. J., Mumenthaler, C., & Brosch, T. (2020). Emo-
tional foundations of the public climate change di-
vide. Climatic Change,161(1), 9–19. https: / / doi.
org/10.1007/s10584-019-02552-0
Heseltine, M., Barnehl, H., & Wojcieszak, M. (2024). Par-
tisan temporal selective news avoidance: Evidence
from online trace data. American Journal of Politi-
cal Science,n/a(n/a), 1–18. https://doi.org/10.1111/
ajps.12944
Hornsey, M. J. (2021). The role of worldviews in shaping how
people appraise climate change. Current Opinion in
Behavioral Sciences,42, 36–41. https://doi.org/10.
1016/j.cobeha.2021.02.021
Hornsey, M. J., Harris, E. A., & Fielding, K. S. (2018).
Relationships among conspiratorial beliefs, conser-
vatism and climate scepticism across nations. Nature
Climate Change,8(7), 614–620. https://doi.org/10.
1038/s41558-018-0157-2
Hutchens, M. J., Hmielowski, J. D., & Beam, M. A. (2019).
Reinforcing spirals of political discussion and af-
fective polarization. Communication Monographs,
86(3), 357–376. https://doi.org/10.1080/03637751.
2019.1575255
IPCC (Ed.). (2023). Climate Change 2022 - Mitigation of
Climate Change: Working Group III Contribution to
the Sixth Assessment Report of the Intergovernmen-
tal Panel on Climate Change. Cambridge University
Press. https://doi.org/10.1017/9781009157926
Iyengar, S., & Hahn, K. S. (2009). Red Media, Blue Media:
Evidence of Ideological Selectivity in Media Use.
Journal of Communication,59(1), 19–39. https://
doi.org/10.1111/j.1460-2466.2008.01402.x
Jonas, E., Schulz-Hardt, S., Frey, D., & Thelen, N. (2001).
Confirmation bias in sequential information search
after preliminary decisions: An expansion of disso-
nance theoretical research on selective exposure to
information. Journal of Personality and Social Psy-
chology,80(4), 557–571. https://doi.org/10.1037/
0022-3514.80.4.557
Judge, M., Kashima, Y., Steg, L., & Dietz, T. (2023). Environ-
mental Decision-Making in Times of Polarization.
Annual Review of Environment and Resources,48,

18 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
477–503. https://doi.org/10.1146/annurev-environ-
112321-115339
Kahan, D. M., Jenkins-Smith, H., & Braman, D. (2011). Cul-
tural cognition of scientific consensus. Journal of
Risk Research,14(2), 147–174. https://doi.org/10.
1080/13669877.2010.511246
Kahan, D. M., Peters, E., Dawson, E. C., & Slovic, P.
(2017). Motivated numeracy and enlightened self-
government. Behavioural Public Policy,1(1), 54–
86. https://doi.org/10.1017/bpp.2016.2
Kerschreiter, R., Schulz-Hardt, S., Mojzisch, A., & Frey, D.
(2008). Biased Information Search in Homogeneous
Groups: Confidence as a Moderator for the Effect
of Anticipated Task Requirements. Personality and
Social Psychology Bulletin,34(5), 679–691. https:
//doi.org/10.1177/0146167207313934
Knobloch-Westerwick, S., & Meng, J. (2009). Looking
the Other Way: Selective Exposure to Attitude-
Consistent and Counterattitudinal Political Infor-
mation. Communication Research,36(3), 426–448.
https://doi.org/10.1177/0093650209333030
Korteling, J. E. (, Paradies, G. L., & Sassen-van Meer, J. P.
(2023). Cognitive bias and how to improve sustain-
able decision making. Frontiers in Psychology,14.
https://doi.org/10.3389/fpsyg.2023.1129835
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B.
(2017). lmerTest package: Tests in linear mixed ef-
fects models. Journal of Statistical Software,82(13),
1–26. https://doi.org/10.18637/jss.v082.i13
Lamb, W. F., Mattioli, G., Levi, S., Roberts, J. T., Capstick,
S., Creutzig, F., Minx, J. C., Müller-Hansen, F., Cul-
hane, T., & Steinberger, J. K. (2020). Discourses of
climate delay. Global Sustainability,3, e17. https:
//doi.org/10.1017/sus.2020.13
Lazer, D. M. J., Baum, M. A., Benkler, Y., Berinsky, A. J.,
Greenhill, K. M., Menczer, F., Metzger, M. J., Ny-
han, B., Pennycook, G., Rothschild, D., Schudson,
M., Sloman, S. A., Sunstein, C. R., Thorson, E. A.,
Watts, D. J., & Zittrain, J. L. (2018). The science of
fake news. Science,359(6380), 1094–1096. https :
//doi.org/10.1126/science.aao2998
Levitan, L. C., & Visser, P. S. (2008). The impact of the social
context on resistance to persuasion: Effortful versus
effortless responses to counter-attitudinal informa-
tion. Journal of Experimental Social Psychology,
44(3), 640–649. https: / /doi. org /10. 1016 /j . jesp.
2007.03.004
Lewandowsky, S. (2021). Climate Change Disinformation
and How to Combat It. Annual Review of Public
Health,42, 1–21. https://doi.org/10.1146/annurev-
publhealth-090419-102409
Lewandowsky, S., Ecker, U. K. H., Seifert, C. M., Schwarz,
N., & Cook, J. (2012). Misinformation and Its Cor-
rection: Continued Influence and Successful Debi-
asing. Psychological Science in the Public Inter-
est,13(3), 106–131. https : / / doi . org / 10 . 1177 /
1529100612451018
Lewandowsky, S., & van der Linden, S. (2021). Countering
Misinformation and Fake News Through Inocula-
tion and Prebunking. European Review of Social
Psychology,32(2), 348–384. https : / / doi . org/ 10 .
1080/10463283.2021.1876983
Lewis, G. B., Palm, R., & Feng, B. (2019). Cross-national
variation in determinants of climate change con-
cern. Environmental Politics,28(5), 793–821. https:
//doi.org/10.1080/09644016.2018.1512261
Liu, J. C.-E. (2015). Low carbon plot: Climate change skep-
ticism with Chinese characteristics. Environmental
Sociology,1(4), 280–292. https://doi.org/10.1080/
23251042.2015.1049811
Liu, J. C.-E. (2023). Public opinion on climate change in
China—evidence from two national surveys. PLoS
Climate,2(2), e0000065. Retrieved December 10,
2024, from https://journals.plos.org/climate/article?
id=10.1371/journal.pclm.0000065
Liu, N., Hu, X. E., Savas, Y., Baum, M. A., Berinsky,
A. J., Chaney, A. J. B., Lucas, C., Mariman, R.,
de Benedictis-Kessner, J., Guess, A. M., Knox,
D., & Stewart, B. M. (2025). Short-term exposure
to filter-bubble recommendation systems has lim-
ited polarization effects: Naturalistic experiments on
YouTube. Proceedings of the National Academy of
Sciences,122(8), e2318127122. https://doi.org/10.
1073/pnas.2318127122
Lynas, M., Houlton, B. Z., & Perry, S. (2021). Greater than
99% consensus on human caused climate change in
the peer-reviewed scientific literature. Environmen-
tal Research Letters,16(11), 114005. https : //doi.
org/10.1088/1748-9326/ac2966
McCright, A. M., Charters, M., Dentzman, K., & Dietz,
T. (2016). Examining the Effectiveness of Climate
Change Frames in the Face of a Climate Change
Denial Counter-Frame. Topics in Cognitive Science,
8(1), 76–97. https://doi.org/10.1111/tops.12171
McCright, A. M., & Dunlap, R. E. (2011). The Politicization
of Climate Change and Polarization in the Ameri-
can Public’s Views of Global Warming, 2001–2010.
The Sociological Quarterly,52(2), 155–194. https:
//doi.org/10.1111/j.1533-8525.2011.01198.x
Meng, K. C., & Rode, A. (2019). The social cost of lobbying
over climate policy. Nature Climate Change,9(6),
472–476. https: / /doi .org / 10. 1038/ s41558 - 019-
0489-6
Momsen, K., & Ohndorf, M. (2022). Information avoidance,
selective exposure, and fake (?) news: Theory and
experimental evidence on green consumption. Jour-

SAMPLING CLIMATE (DIS)INFORMATION 19
nal of Economic Psychology,88, 102457. https://
doi.org/10.1016/j.joep.2021.102457
Mont’Alverne, C., Badrinathan, S., Ross Arguedas, A., Toff,
B., Fletcher, R., & Nielsen, R. (2023). “Fair and
Balanced”: What News Audiences in Four Coun-
tries Mean When They Say They Prefer Impartial
News. Journalism Studies,24(9), 1131–1148. https:
//doi.org/10.1080/1461670X.2023.2201864
Motoki, K., Saito, T., & Onuma, T. (2021). Eye-tracking re-
search on sensory and consumer science: A review,
pitfalls and future directions. Food Research Inter-
national,145, 110389. https: //doi.org/ 10.1016/j.
foodres.2021.110389
Ophir, Y., Walter, D., Jamieson, P. E., & Jamieson, K. H.
(2024). The Politicization of Climate Science: Me-
dia Consumption, Perceptions of Science and Sci-
entists, and Support for Policy. Journal of Health
Communication,29, 18–27. https : / / doi . org / 10 .
1080/10810730.2024.2357571
Oreskes, N., & Conway, E. M. (2011). Merchants of doubt:
How a handful of scientists obscured the truth
on issues from tobacco smoke to global warming.
Bloomsbury Publishing USA.
Painter, J., Ettinger, J., Holmes, D., Loy, L., Pinto, J., Richard-
son, L., Thomas-Walters, L., Vowles, K., & Wetts, R.
(2023). Climate delay discourses present in global
mainstream television coverage of the IPCC’s 2021
report. Communications Earth & Environment,4(1),
118. https://doi.org/10.1038/s43247-023-00760-2
Pennycook, G., Bear, A., Collins, E. T., & Rand, D. G. (2020).
The Implied Truth Effect: Attaching Warnings to a
Subset of Fake News Headlines Increases Perceived
Accuracy of Headlines Without Warnings. Manage-
ment Science,66(11), 4944–4957. https://doi.org/
10.1287/mnsc.2019.3478
Pennycook, G., Binnendyk, J., Newton, C., & Rand, D. G.
(2021). A Practical Guide to Doing Behavioral Re-
search on Fake News and Misinformation. Collabra:
Psychology,7(1), 25293. https://doi.org/10.1525/
collabra.25293
Pennycook, G., & Rand, D. G. (2021). The Psychology of
Fake News. Trends in Cognitive Sciences,25(5),
388–402. https://doi.org/10.1016/j.tics.2021.02.007
Petersen, A. M., Vincent, E. M., & Westerling, A. L. (2019).
Discrepancy in scientific authority and media visi-
bility of climate change scientists and contrarians.
Nature Communications,10(1), 3502. https://doi.
org/10.1038/s41467-019-09959-4
Pringle, A., & Robbins, D. (2022). From denial to delay: Cli-
mate change discourses in Ireland. Administration,
70(3), 59–84. https://doi.org/10.2478/admin-2022-
0019
R Core Team. (2024). R: A Language and Environment for
Statistical Computing. R Foundation for Statistical
Computing. https://www.R-project.org/
Rathje, S., Pretus, C., He, J. K., Harjani, T., Roozenbeek, J.,
Gray, K., Linden, D. S. van der, & Bavel, J. J. V.
(2024). Unfollowing hyperpartisan social media in-
fluencers durably reduces out-party animosity. https:
//doi.org/10.31234/osf.io/acbwg
Rodenburg, E. (2024). Understanding the Dimensions of Cli-
mate Change Misinformation. Retrieved December
10, 2024, from https://dr.library.brocku.ca/handle/
10464/18218
Accepted: 2024-01-04T15:52:58Z.
Roozenbeek, J., Maertens, R., Herzog, S. M., Geers, M.,
Kurvers, R., Sultan, M., & Linden, S. van der.
(2022). Susceptibility to misinformation is consis-
tent across question framings and response modes
and better explained by myside bias and partisan-
ship than analytical thinking. Judgment and Deci-
sion Making,17(3), 547–573. https://doi.org/10.
1017/S1930297500003570
Roper, J., Ganesh, S., & Zorn, T. E. (2016). Doubt, Delay, and
Discourse: Skeptics’ Strategies to Politicize Climate
Change. Science Communication,38(6), 776–799.
https://doi.org/10.1177/1075547016677043
Rosseel, Y. (2012). Lavaan: An R Package for Structural
Equation Modeling. Journal of Statistical Software,
48, 1–36. https://doi.org/10.18637/jss.v048.i02
Sepulveda, P., Usher, M., Davies, N., Benson, A. A., Ortol-
eva, P., & De Martino, B. (2020). Visual attention
modulates the integration of goal-relevant evidence
and not value (C. Büchel, V. Wyar t, A. L. Filipowicz,
& S. Gluth, Eds.). eLife,9, e60705. https://doi.org/
10.7554/eLife.60705
Sharot, T., & Garrett, N. (2016). Forming Beliefs: Why Va-
lence Matters. Trends in Cognitive Sciences,20(1),
25–33. https://doi.org/10.1016/j.tics.2015.11.002
Sharot, T., & Sunstein, C. R. (2020). How people decide what
they want to know. Nature Human Behaviour,4(1),
14–19. https://doi.org/10.1038/s41562-019-0793-1
Slater, M. D. (2007). Reinforcing Spirals: The Mutual Influ-
ence of Media Selectivity and Media Effects and
Their Impact on Individual Behavior and Social
Identity. Communication Theory,17(3), 281–303.
https://doi.org/10.1111/j.1468-2885.2007.00296.x
Slater, M. D. (2015). Reinforcing Spirals Model: Concep-
tualizing the Relationship Between Media Content
Exposure and the Development and Maintenance of
Attitudes. Media Psychology,18(3), 370–395. https:
//doi.org/10.1080/15213269.2014.897236
Smith, E. K., & Mayer, A. (2019). Anomalous Anglophones?
Contours of free market ideology, political polar-
ization, and climate change attitudes in English-

20 RAHMANI ET AL. MANUSCRIPT UNDER REVIEW
speaking countries, Western European and post-
Communist states. Climatic Change,152(1), 17–34.
https://doi.org/10.1007/s10584-018-2332-x
Spampatti, T. (2024). Focusing on truth discernment may
blindside advances in fighting misinformation.
https://doi.org/10.31234/osf.io/kmzsj
Spampatti, T., Brosch, T., Mumenthaler, C., & Hahnel, U. J. J.
(2023). Blueprint of a smokescreen: Introducing the
validated Climate Disinformation Corpus for behav-
ioral research on combating climate disinformation.
https://doi.org/10.31234/osf.io/v7895
Spampatti, T., Hahnel, U. J. J., Trutnevyte, E., & Brosch, T.
(2024). Psychological inoculation strategies to fight
climate disinformation across 12 countries. Nature
Human Behaviour,8(2), 380–398. https://doi.org/
10.1038/s41562-023-01736-0
Spohr, D. (2017). Fake news and ideological polarization:
Filter bubbles and selective exposure on social me-
dia. Business Information Review,34(3), 150–160.
https://doi.org/10.1177/0266382117722446
Storozuk, A., Ashley, M., Delage, V., & Maloney, E. A.
(2020). Got Bots? Practical Recommendations to
Protect Online Survey Data from Bot Attacks. The
Quantitative Methods for Psychology,16(5), 472–
481. https://doi.org/10.20982/tqmp.16.5.p472
Sunstein, C. R., Bobadilla-Suarez, S., Lazzaro, S. C., &
Sharot, T. (2016–2017). How People Update Beliefs
about Climate Change: Good News and Bad News.
Cornell Law Review,102, 1431. https://heinonline.
org/HOL/Page?handle=hein.journals/clqv102&id=
1479&div=&collection=
Supran, G., Rahmstorf, S., & Oreskes, N. (2023). Assessing
ExxonMobil’s global warming projections. Science,
379(6628), eabk0063. https : / / doi . org / 10 . 1126 /
science.abk0063
Supran, G., & Oreskes, N. (2017). Assessing ExxonMobil’s
climate change communications (1977–2014). En-
vironmental Research Letters,12(8), 084019. https:
//doi.org/10.1088/1748-9326/aa815f
Supran, G., & Oreskes, N. (2021). Rhetoric and frame analysis
of ExxonMobil’s climate change communications.
One Earth,4(5), 696–719. https://doi.org/10.1016/
j.oneear.2021.04.014
Talluri, B. C., Urai, A. E., Tsetsos, K., Usher, M., & Don-
ner, T. H. (2018). Confirmation Bias through Selec-
tive Overweighting of Choice-Consistent Evidence.
Current Biology,28(19), 3128–3135.e8. https://doi.
org/10.1016/j.cub.2018.07.052
Tam, K.-P., & Chan, H.-W. (2023). Conspiracy theories and
climate change: A systematic review. Journal of En-
vironmental Psychology,91, 102129. https: / /doi .
org/10.1016/j.jenvp.2023.102129
Tam, K.-P., Leung, A. K.-y., & Clayton, S. (2021). Research
on climate change in social psychology publications:
A systematic review. Asian Journal of Social Psy-
chology,24(2), 117–143. https://doi.org/10.1111/
ajsp.12477
Treen, K. M. d’I. Williams, H. T. P., & O’Neill,
S. J. (2020). Online misinformation about climate
change. WIREs Climate Change,11(5), e665. https:
//doi.org/10.1002/wcc.665
van Eck, C. W. (2024). Opposing positions, dividing inter-
actions, and hostile affect: A systematic review and
conceptualization of “online climate change polar-
ization”. WIREs Climate Change,15(6), e906. https:
//doi.org/10.1002/wcc.906
van Valkengoed, A., Steg, L., & Perlaviciute, G. (2021). De-
velopment and validation of a climate change per-
ceptions scale. Journal of Environmental Psychol-
ogy,76, 101652. https://doi.org/10.1016/ j.jenvp.
2021.101652
van der Linden, S., Leiserowitz, A., Rosenthal, S., & Maibach,
E. (2017). Inoculating the Public against Misinfor-
mation about Climate Change. Global Challenges,
1(2), 1600008. https : / / doi . org / 10 . 1002 / gch2 .
201600008
Većkalov, B., Geiger, S. J., Bartoš, F., White, M. P., Rut-
jens, B. T., van Harreveld, F., Stablum, F., Akın, B.,
Aldoh, A., Bai, J., Berglund, F., Bratina Zimic, A.,
Broyles, M., Catania, A., Chen, A., Chorz˛epa, M.,
Farahat, E., Götz, J., Hoter-Ishay, B., . . . van der
Linden, S. (2024). A 27-country test of communi-
cating the scientific consensus on climate change.
Nature Human Behaviour,8(10), 1892–1905. https:
//doi.org/10.1038/s41562-024-01928-2
Vlasceanu, M., Doell, K. C., Bak-Coleman, J. B., Todorova,
B., Berkebile-Weinberg, M. M., Grayson, S. J., Patel,
Y., Goldwert, D., Pei, Y., Chakroff, A., Pronizius, E.,
van den Broek, K. L., Vlasceanu, D., Constantino,
S., Morais, M. J., Schumann, P., Rathje, S., Fang,
K., Aglioti, S. M., . . . Van Bavel, J. J. (2024). Ad-
dressing climate change with behavioral science: A
global intervention tournament in 63 countries. Sci-
ence Advances,10(6), eadj5778. https://doi.org/10.
1126/sciadv.adj5778
Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true
and false news online. Science,359(6380), 1146–
1151. https://doi.org/10.1126/science.aap9559
Wang, B., & Zhou, Q. (2020). Climate change in the Chinese
mind: An overview of public perceptions at macro
and micro levels. WIREs Climate Change,11(3),
e639. https://doi.org/10.1002/wcc.639
Wang, S., Leviston, Z., Hurlstone, M., Lawrence, C., &
Walker, I. (2018). Emotions predict policy sup-
port: Why it matters how people feel about climate

SAMPLING CLIMATE (DIS)INFORMATION 21
change. Global Environmental Change,50, 25–40.
https://doi.org/10.1016/j.gloenvcha.2018.03.002
West, E. A., & Iyengar, S. (2022). Partisanship as a Social
Identity: Implications for Polarization. Political Be-
havior,44(2), 807–838. https://doi.org/10.1007/
s11109-020-09637-y
Westfall, J., Kenny, D. A., & Judd, C. M. (2014). Statis-
tical power and optimal design in experiments in
which samples of participants respond to samples of
stimuli. Journal of Experimental Psychology: Gen-
eral,143(5), 2020–2045. https://doi.org/10.1037/
xge0000014
Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan,
L. D., François, R., Grolemund, G., Hayes, A.,
Henry, L., Hester, J., Kuhn, M., Pedersen, T. L.,
Miller, E., Bache, S. M., Müller, K., Ooms, J.,
Robinson, D., Seidel, D. P., Spinu, V., . . . Yutani, H.
(2019). Welcome to the tidyverse. Journal of Open
Source Software,4(43), 1686. https://doi.org/10.
21105/joss.01686
Winter, K., Hornsey, M. J., Pummerer, L., & Sassenberg,
K. (2024). Public agreement with misinformation
about wind farms. Nature Communications,15(1),
8888. https://doi.org/10.1038/s41467-024-53278-2
Wojcieszak, M., & Garrett, R. K. (2018). Social Identity, Se-
lective Exposure, and Affective Polarization: How
Priming National Identity Shapes Attitudes Toward
Immigrants Via News Selection. Human Communi-
cation Research,44(3), 247–273. https://doi.org/10.
1093/hcr/hqx010
Yang, Z. J., & Kahlor, L. (2013). What, Me Worry? The Role
of Affect in Information Seeking and Avoidance.
Science Communication,35(2), 189–212. https: / /
doi.org/10.1177/1075547012441873
Zannettou, S. (2021). "I Won the Election!": An Empirical
Analysis of Soft Moderation Interventions on Twit-
ter. Proceedings of the International AAAI Confer-
ence on Web and Social Media,15, 865–876. https:
//doi.org/10.1609/icwsm.v15i1.18110
Zeng, L. (2022). Chinese Public Perception of Climate
Change on Social Media: An Investigation Based
on Data Mining and Text Analysis. Journal of En-
vironmental and Public Health,2022(1), 6294436.
https://doi.org/10.1155/2022/6294436