Journal of Environmental Psychology 93 (2024) 102210

Contents lists available at ScienceDirect

Journal of Environmental Psychology

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

Effects of communicating the rise of climate migration on public

perceptions of climate change and migration
Kaitlin T. Raimi a, *, Melanie A. Sarge b, Nathaniel Geiger b, Ash Gillis c, Julia Lee Cunningham a
a
University of Michigan, USA
b
University of Indiana-Bloomington, USA
c
Vanderbilt University, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Climate change-induced migration is a rapidly increasing phenomenon estimated to affect millions of people in
Climate change migration the coming decades. With increasing media coverage of climate migration, including within the United States, it
Climate change mitigation is critical to understand how to effectively communicate about this issue. In two high-powered experiments of U.
Climate change policy support
S. Democrats and Republicans (& leaners; Study 1: N = 1452; Study 2: N = 1873), we test the effects of fictional
Framing climate change
Climate change communication
news coverage about domestic and international climate-induced migration on climate change and migration risk
Episodic framing perceptions, warmth toward migrants, and policy support. We also tested whether including a personal story
about a specific migrant alongside numerical trends could increase support for climate change mitigation and
aiding potential migrants. Impacts on climate outcomes were minimal: in Study 1, those exposed to different
message frames reported similar climate policy support and risk perception in Study 1, and, in Study 2, domestic
(but not international) climate migration coverage slightly increased climate risk perceptions and marginally
increased mitigation policy support compared to a no-message baseline. In both studies, international migration
was perceived as riskier to Americans than domestic migration, and coverage of international migration heighted
this perceived risk. Coverage of domestic climate migration, on the other hand, did not boost risk perceptions of
that type of migration. Furthermore, coverage of international climate migration led to greater support for
adaptation infrastructure to help potential migrants stay home but did not affect support for policies to help
migration. Contrary to expectations, the inclusion of personal stories did not affect warmth toward migrants in
either study. Political affiliation dominated associations with all climate change and migration outcomes, likely
obscuring potential framing effects. These results suggest that coverage about climate-induced migration may
face challenges in shifting entrenched public opinion and policy support. Communicators should be aware of the
limitations of using this type of coverage to promote constructive actions to address climate change and
migration.

1. Introduction S. (The White House, 2021; United States Government Accountability

Office, 2020). The drivers of migration are clear: looking only at the
Climate change is causing growing numbers of people to migrate financial costs (which are easier to quantify though pale in comparison
away from increasingly vulnerable home regions and communities. to human suffering), U.S. federal obligations for disaster assistance have
Reports estimate that over 20 million people per year have been dis­ already totaled at least $460 billion since 2005 (United States Govern­
placed due to climate change since 2008 (Bilak et al., 2016; IPCC, 2022). ment Accountability Office, 2020) in addition to the nearly $10.5 billion
This trend is expected to accelerate in coming years, with hundreds of per year spent on humanitarian assistance and protection provided to
millions potentially displaced between 2050 and 2100 (IPCC, 2022). those born elsewhere does (The White House, 2021). These and other
Governments are beginning to incorporate climate migration in their related costs are expected to grow as climate change worsens. In
planning, including the U.S., which has started publishing noteworthy response, climate migration has received increasing media attention in
reports on climate migrants both from other countries and within the U. the United States and elsewhere (Sakellari, 2022; Schäfer, Scheffran, &

* Corresponding author. Gerald R. Ford School of Public Policy, University of Michigan, 735 S. State St, Ann Arbor, MI, 48103, USA.
E-mail address: [email protected] (K.T. Raimi).

https://doi.org/10.1016/j.jenvp.2023.102210
Received 31 March 2023; Received in revised form 1 December 2023; Accepted 4 December 2023
Available online 10 December 2023
0272-4944/© 2023 Elsevier Ltd. All rights reserved.
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Penniket, 2016), including a series published in New York Times migration is domestic, with people moving from one region to another
Magazine and ProPublica reporting on geopolitical consequences of within their home country, yet some climate migration is expected to
international migration and the expected prevalence of migration within cross borders (Bilak et al., 2016; IPCC, 2022; Rigaud et al., 2018, pp.
the U.S. (Lustgarten, 2020a; 2020b; 2020c). Popular press coverage 1–256). In addition to differences in policies and relevant government
often reflects government and non-government reports on climate levels for addressing domestic vs. international climate migration, these
migration that explicitly call for policy actions to address the underlying different types of migration may inspire different reactions from
cause (i.e., reducing greenhouse gas emissions) and to manage the receiving communities. Discussing the topic of climate migration as an
consequences, including aiding climate migrants and funding infra­ international issue may engender an outgroup evaluation of climate
structure for climate adaptation efforts (Blake, Clark-Ginsberg, & Bala­ migrants given their spatial and social distance, whereas emphasizing
gna, 2021; Podesta, 2019; Pörtner et al., 2022; United Nations OHCHR, the domestic aspect of climate migration could provoke an ingroup
2018). mentality (Trope & Liberman, 2012).
Yet, it remains unclear whether this reporting changes public per­ Decades of research finds that people often prefer members of their
ceptions of climate change, climate migration, and related policies ingroups over outgroup members (Everett, Faber, & Crockett, 2015;
(Geiger, Raimi, Roy, & Cunningham, 2023). Discussing the migration Tajfel, 1974; Tajfel & Turner, 1986) and feel more threatened by out­
implications of climate change may highlight this issue among Repub­ group members (O’Donnell, Neumann, Duffy, & Paolini, 2019; Riek,
lican audiences, given that immigration is a topic Republicans tend to be Mania, & Gaertner, 2006). This may lead to a lack of warmth toward
more concerned about than Democrats (Jones, 2021). Perhaps international climate migrants who are often considered outgroup
describing how climate change will impede conservative goals and members and an increased perception of risk regarding the likelihood
priorities could capture attention and relate to the concerns of Re­ and threat of negative consequences as a result of international climate
publicans in the U.S., who continue to be more dismissive of climate migration (i.e., outgroup members migrating to an ingroup’s territory).
change and policies to address it than Democrats (Leiserowitz et al., Further, while help is not usually offered to an outgroup (Bernhard,
2022). For example, framing climate change as an issue that shares Fischbacher, & Fehr, 2006; Tajfel & Turner, 2004), in this case, one
conservative values improved Republican’s understanding of climate remedy to ameliorate risk perceptions about migration is to provide
change (Goldberg, Gustafson, Rosenthal, & Leiserowitz, 2021). Further, support to help people stay in their home communities and prevent the
using a public health frame generated more hopeful feelings regarding need for migration. Coverage of international migration may therefore
potential benefits of adopting climate change prevention and mitigation be more effective in inspiring support for adaptation infrastructure
solutions (Myers, Nisbet, Maibach, & Leiserowitz, 2012) and likening development rather than policies that encourage migration.
climate change to a medical disease altered decision-relevant beliefs If domestic climate migrants are considered ingroup members, peo­
about climate change (Raimi, Stern, & Maki, 2017) across the political ple should feel more warmth toward them and their migration should be
spectrum. However, whether highlighting climate migration increases viewed as less threatening (i.e., lower risk perceptions). Supporting this
support for climate change policies depends on whether the message conjecture, research on perceptions of domestic migrants in Bangladesh
resonates with an audience (Luong, Garrett, & Slater, 2019). Climate found that both geographic proximity and social proximity (similarity in
change frames that connect the issue with politicized topics, such as wealth and education) predicted more welcoming attitudes toward mi­
discussing the national security impacts of climate change, can poten­ grants (Lujala, Bezu, Kolstad, Mahmud, & Wiig, 2020). In the U.S.
tially heighten the salience of climate change among political conser­ context, a majority of Americans believed that weather or
vatives (Gainous & Merry, 2022) but could also trigger motivated climate-related issues may influence them to move at some point in the
reasoning in some recipients, potentially resulting in message rejection future (Kim, Kay, & Schuldt, 2021), suggesting that many respondents
or unintended backlash (Hart & Nisbet, 2012; Myers et al., 2012; Singh may see domestic climate migrants as facing similar circumstances as
& Swanson, 2017). Indeed, some studies suggest that presenting brief themselves. Additionally, people tend to act more pro-socially toward
information about climate migration may not engender climate action ingroup members than outgroup members (e.g., Stürmer, Snyder, &
(Arias & Blair, 2022), and could instead result in more negative attitudes Omoto, 2005). Indeed, recent research suggests that Americans may be
toward migrants (Gillis, Geiger, Raimi, Cunningham, & Sarge, 2023). more supportive of policies to help domestic climate migrants than in­
The literature is mixed on whether climate migrants are seen as more ternational climate migrants (Stanley et al., 2022). Thus, we expect that
similar to economic migrants or to refugees fleeing political persecution Americans may be more inspired to help ingroup domestic migrants
(who tend to receive more empathy) (Arias & Blair, 2022; Castellano, than outgroup international migrants, both in terms of policies that
Dolšak, & Prakash, 2021; Hedegaard, 2022; Helbling, 2020; Spilker, provide adaptation support or help relocate people and to address the
Nguyen, Koubi, & Böhmelt, 2020; Stanley, Ng Tseung-Wong, & Levis­ underlying cause of climate change that threatens them.
ton, 2022). While some of this discrepancy may be due to pre-existing As shown in Table 1, we therefore predict 1:
differences in perceptions of various migrants, some may also be due
Hypothesis 1. (H1): Overall, Americans prefer domestic climate mi­
to how these migration patterns and people are described. Prior research
grants over international climate migrants, specifically (H1a) lesser
has yet to reach a consensus on what aspects of climate migration may
perceived risk toward domestic (vs. international) migration, (H1b)
affect these and other outcomes, especially given the politicized nature
more warmth toward domestic (vs. international) migrants as compared
of this issue. In sum, while the growing body of literature noted above
to native-born residents, and (H1c) greater policy support toward
has examined how climate migrants are perceived compared to other
helping domestic (vs. international) migrants.
groups, there is not much evidence for how people react to different
framings of climate migration. Thus, there is a considerable gap in un­
derstanding which aspects of this issue should be emphasized to
encourage acceptance of the message and avoid rejection. Accordingly,
the present research experimentally examines four message frames that 1
may alter the way people evaluate climate change and perceive climate All hypotheses were pre-registered for Study 2 (see https://osf.io/s2bnt/?
view_only=3bf64fc539bf4b78a04150aa34a7dcc0). The order and in some
migrants.
cases the precise wording of the hypotheses differed in the pre-registration, but
the content is the same. One notable difference is that all interactions in these
1.1. Domestic vs. international migration analyses were probed using adjusted p-values with Sidak corrections to mini­
mize Type 1 error – this was not included in the original preregistration. The
One aspect of climate migration that may affect public perception is preregistration also included other research questions that are beyond the scope
the origin of migrants: domestic or international. Most climate of this paper and analyzed in a different project.

2
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Table 1
Overview of hypotheses.
Hypothesis or Research Question Supported

Study 1 Study 2

Hypothesis 1 (H1): Overall, Americans prefer domestic climate migrants over Supported Supported
international climate migrants, specifically (H1a) lesser perceived risk toward
domestic (vs. international) migration, (H1b) more warmth toward domestic (vs.
international) migrants as compared to native-born residents, and H1c) greater
policy support toward helping domestic (vs. international) migrants.
Hypothesis 2 (H2): Exposure to messages about either domestic or international climate Partial Support (H2a supported, H2c Partial Support (H2a supported, others not
migration will heighten the perceived risk of that form of migration and increase partially supported, others not supported)
support for policies to slow it. Specifically, (H2a) exposure to a message about supported),
international (vs. domestic) climate migration will increase perceived risk of
international migration; (H2b) exposure to a message about domestic (vs.
international) climate migration will increase perceived risk of domestic migration.
(H2c) Exposure to a message about international (vs. domestic) climate migration
will increase policy support for helping potential immigrants stay in their home
countries and (H2d) decrease policy support to help potential immigrants move to
the U.S.; (H2e) exposure to a message about domestic (vs. international) climate
migration will increase policy support to help potential domestic migrants stay in
their home regions and (H2f) decrease policy support for helping potential migrants
move within the U.S.
Hypothesis 3 (H3): Exposure to a message about any impacts of climate change (as n/a Partial Support (H3a supported for domestic
compared to a baseline with no message exposure) will (H3a) increase perceived conditions, H3b for domestic combined
risks of climate change and (H3b) increase support for policies to reduce greenhouse condition)
gas emissions.
Hypothesis 4 (H4): Exposure to information about climate migration that includes both Not Supported Not Supported
episodic and thematic information formats (instead of just thematic information) will
reduce preferences for native-born residents over migrants and increase support for
policies to help migrants.
RQ1: Will exposure to a message about climate migration that includes both thematic No significant effect No significant effect
and episodic information (vs. thematic information only) elicit greater perceived risk
of climate change (RQ1a) and more support for climate policies (RQ1b)?
Hypothesis 5 (H5): Overall, Republicans (vs. Democrats) will (H5a) perceive climate Supported Supported
change as less risky and (H5b) report lesser climate policy support. In contrast,
Republicans (vs. Democrats) will (H5c) perceive migration as riskier, (H5d) show
greater preferences for native-born residents over migrants, and be (H5e) less
supportive of policies that provide aid to potential migrants.
Hypothesis 6 (H6): Any effects of climate-migration messages will be moderated by Not Supported Not Supported
political affiliation.

1.2. Migration and climate change risk salience and (H2f) decrease policy support for helping potential migrants move
within the U.S.
Work on the risk salience hypothesis suggests that for many in the
Similarly, regardless of which type of migration is discussed, the
Global North, including the U.S., climate change is not perceived as an
mention of climate change as the cause is likely to make audiences
immediate risk (Whitmarsh, 2008). However, drawing attention to the
prioritize the issue more than they otherwise would. In line with theories
need for drastic action on climate change can make its risks feel more
of risk salience and previous work on support for climate action (e.g.,
salient and personally threatening in ways that also might boost efforts
Carrico et al., 2015), we predict that messages about climate migration
to address it (Jebari et al., 2021). For example, highlighting the need for
will not only make migration (and its attendant risks) more salient, but
climate adaptation can not only increase worry about climate change
they would also raise concerns about climate change as a whole and
(Urban, Vačkářová, & Badura, 2021) but also increase support for both
elicit greater support for policies to address it. However, we do note that
personal mitigation behaviors (Evans, Milfont, & Lawrence, 2014;
contrary to our prediction, a preliminary study found that
Urban et al., 2021) and climate mitigation policies (Carrico, Truelove,
climate-driven immigration (vs. a message about immigration from
Vandenbergh, & Dana, 2015). Similarly, making a specific form of
unspecified causes) did not increase climate concern (Gillis et al., 2023).
climate-induced migration salient may increase perceived risk for that
Nonetheless, we hypothesize the following.
form of migration (i.e., domestic vs. international) and thus boost sup­
port for policies to address it. This leads to following hypotheses. Hypothesis 3. (H3): Exposure to a message about any impacts of
climate change (as compared to a baseline with no message exposure)
Hypothesis 2. (H2): Exposure to messages about either domestic or
will (H3a) increase perceived risks of climate change and (H3b) increase
international climate migration will heighten the perceived risk of that
support for policies to reduce greenhouse gas emissions.
form of migration and increase support for policies to slow it. Specif­
ically, (H2a) exposure to a message about international (vs. domestic)
1.3. Thematic and episodic information formats
climate migration will increase perceived risk of international migra­
tion; (H2b) exposure to a message about domestic (vs. international)
Information formats may also affect public reactions to climate
climate migration will increase perceived risk of domestic migration.
migration. Media discussions of climate change use a variety of infor­
(H2c) Exposure to a message about international (vs. domestic) climate
mation formats, most commonly, general news reports or human-
migration will increase policy support for helping potential immigrants
interest stories (Haynes, Merolla, & Ramakrishnan, 2016). General
stay in their home countries and (H2d) decrease policy support to help
news reports are characterized by thematic information involving
potential immigrants move to the U.S.; (H2e) exposure to a message
numbers, trends and abstract aspects of the issue context or process
about domestic (vs. international) climate migration will increase policy
(Iyengar, 1991). Human-interest stories contain episodic information,
support to help potential domestic migrants stay in their home regions
which entails detailed accounts of a specific event, case study, or

3
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Table 2 Hypothesis 4. (H4): Exposure to information about climate migration

Participant demographics. that includes both episodic and thematic information formats (instead of
Variable Study 1 Study 2 just thematic information) will reduce preferences for native-born resi­
dents over migrants and increase support for policies to help migrants.
N 1768 2400
Age (Mean, SD) 36.55 (12.99) 46.63 (17.56) We also investigate whether the addition of episodic information
Gender (%)
might heighten concern regarding the cause of the exemplar’s plight
Male 47.1 46.5
Female 50.3 53.5 (climate change) and increase support for mitigation policies. Thus, we
Non-binary/non-conforming 2.2 –a propose the following non-directional research question.
Prefer not to answer 0.3 –
Education (% with bachelor’s degree) 54.7 33.8 RQ1. Will exposure to a message about climate migration that includes
Household income per year (Median) b $50,000- $40,000- both thematic and episodic information (vs. thematic information only)
$59,999 $49,000 elicit greater perceived risk of climate change (RQ1a) and more support
Race/Ethnicityc
for climate policies (RQ1b)?
Caucasian/White % 78.2 64.8
African-American/Black % 11.0 11.5
Hispanic, Latino, or Spanish % 8.3 14.6 1.4. Political affiliation, climate change, and immigration
Asian % 11.6 2.9
Native Hawaiian/Pacific Islander % 0.6 –
As with all aspects of climate change in the U.S., political affiliation is
American Indian/Alaska Native % 2.2 0.8
Middle Eastern – 0.3
likely to play an outsized role in perceptions of climate migration. Re­
Two or more races – 2.7 publicans are consistently less concerned about climate change and its
Other Race/Ethnicity % 0.0 2.4 consequences than Democrats, and less supportive of policies to address
Political Affiliation it (Leiserowitz et al., 2022). In fact, a meta-analysis suggests that po­
Democrat or Democrat-leaning % 49.4 44.8
litical partisanship is one of the strongest predictors of acknowledging
Republican or Republican-leaning % 32.7 33.3
Independents/Something Else (not 17.9 22.0 the existence of climate change in the first place (Hornsey, Harris, Bain,
leaning) % & Fielding, 2016). Climate migration may be further politically polar­
a ized in the U.S. due to its ties with another divisive issue—immigration:
– indicates that this option was not presented in this study.
b
Study 1 Income level options were Less than $10,000, $10,000 to $19,999, Republicans tend to be much more wary of immigrants and
$20,000 to $29,999, $30,000 to $39,999, $40,000 to $49,999, $50,000 to pro-immigration policies than Democrats (Younis, 2020).
$59,999, $60,000 to $69,999, $70,000 to $79,999, $80,000 to $89,999, Given these well-documented associations, we expect large partisan
$90,000 to $99,999, $100,000 to $149,999, and $150,000 or more. Study 2 differences in risk perceptions, preferences for native-born residents,
options were Less than $10,000, $10,000 to $19,999, $20,000 to $29,999, and policy support, especially when people are considering interna­
$30,000 to $39,999, $40,000 to $49,999, $50,000 to $59,999, $60,000 to tional migration and among Republicans.
$69,999, $70,000 to $79,999, $80,000 to $89,999, $90,000 to $99,999,
$100,000 to $119,999, $120,000 to $149,999, $150,000 to $199,999, $200,000 Hypothesis 5. (H5): Overall, Republicans (vs. Democrats) will (H5a)
to $249,999, $250,000 to $349,999, $350,000 to $499,999, and $500,000 or perceive climate change as less risky and (H5b) report lesser climate
more. policy support. In contrast, Republicans (vs. Democrats) will (H5c)
c
Racial/ethnic categories were not mutually exclusive in Study 1 but were in perceive migration as riskier, (H5d) show greater preferences for native-
Study 2. born residents over migrants, and be (H5e) less supportive of policies
that provide aid to potential migrants.
personal narrative that serves as an exemplar (Iyengar, 1991).
However, issue perceptions are not the only outcome associated with
The effects of thematic or episodic information formats in news
political ideologies. People often interpret messages about politicized
coverage have been studied extensively. Both framings can be effective
topics such as climate change or migration through the lens of their
in garnering policy support, though they work through different chan­
political affiliation or ideology. A couple different theoretical frame­
nels (e.g., Aarøe, 2011; Gross, 2008; Hart, 2011). Thematic information
works, motivated reasoning (Kunda, 1990) and the
provides a broader, comprehensive picture of an issue that can stimulate
Orientations-Stimulus-Orientations-Response (O–S–O-R) model (Mar­
support for considerable societal or governmental action (Iyengar, 1990,
kus & Zajonc, 1985), propose explanations that support this claim. Ac­
1991). Episodic information, on the other hand, serves as heuristic for
cording to motivated reasoning, predispositions (e.g., political
determining how quantifiable, representative, or available aspects of an
affiliation) can lead people to biasedly interpret information out of an
issue are, which often causes people to overestimate risks and triggers
unconscious motivation to reinforce beliefs associated with these pre­
strong emotions capable of influencing beliefs and behaviors (Zillmann,
dispositions (Kunda, 1990). The O–S–O-R model would discuss political
1999, 2006). In the context of immigration, episodic information about
ideology as the “Orientation” variable that dictates which aspects of a
an identifiable victim can increase positive public sentiment as opposed
frame resonate with people and thus influence interpretation of the
to purely thematic statistical information (Heizler & Israeli, 2021).
message and its subsequent effects.
Given the importance of both information formats, a combination of
These theories have guided previous research using communication
the two may be most effective (Allen et al., 2000; Haynes et al., 2016).
framings and thus, provide rationale for political affiliation affecting
This also reflects the external validity of much climate migration
reactions to the messages in the present studies. For instance, Hart and
coverage; rather than presenting purely episodic information, personal
Nisbet (2012) presented motivated reasoning (Kunda, 1990) as a
stories in news articles are usually paired with thematic information
framework to predict that the impact of social distance cues (similar to
about the big picture of migration (e.g., Lustgarten, 2020a; 2020b;
our domestic and international migration conditions) on identification
2020c). Indeed, frames that included both thematic and episodic ele­
with potential victims of climate change and support for mitigation
ments were more effective in garnering support for legalization of im­
policies would be contingent on party affiliation. Messages that cued
migrants than episodic information formats alone (Haynes et al., 2016).
high social distance were associated with lower levels of identification
Therefore, incorporating episodic information depicting a climate mi­
with climate change victims and decreased support for climate mitiga­
grant’s personal story into a thematic message may make individuals
tion among Republicans; while Democrats reported the same level of
feel more warmth toward climate migrants (and thus reduce preference
identification with the victims regardless of the message frame but
for native-born residents over migrants) and increase their support for
increased policy support after exposure to the high social distance
any policy aimed at helping migrants.
message. Additionally, Boukes (2022) predicted differential effects of

4
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

thematic and episodic framing on attributions of responsibly due to 2.1. Methods

political ideology based on the O–S–O-R model (Markus & Zajonc,
1985). While Boukes (2022) did not measure policy support as a mes­ 2.1.1. Participants
sage effect, he did examine responsibility attributions for political ac­ A priori power analyses using the pwr2 package (v1.0; Lu, Liu, &
tors, which has been positively associated with policy support (Weiner, Koestler, 2017) in R (v4.0.5; R Core Team, 2021) indicated that 1616
2006). He found that episodic framing lowered responsibility attribu­ participants would provide 90% power for our most restrictive analyses
tions for political actors among both ideologies but to a greater extent assuming an effect size of Cohen’s d = 0.20 (considered a “small” effect;
for right-wing participants and that thematic framing increased such Cohen, 1988). We recruited 1800 U.S. adults through Prolific in June
attributions of responsibility only among right-wing participants. Based 2021. Prolific pools tend to overrepresent Democrats relative to the U.S.
on this literature, we expect that political affiliation will interact with public, so to recruit a political mix closer to the U.S. public we recruited
our messages, though there is not enough prior work testing our fram­ 800 participants without restrictions on political affiliation, and the
ings in this context to predict specific interactions. Therefore, we pro­ remaining 1000 participants excluding self-identified Democrats. One
pose a general hypothesis in line with the theoretical frameworks participant did not complete the study, 12 indicated that we should not
mentioned above. use their data, and 21 failed the open-ended attention check (final N =
1768). Participant demographics are shown in Table 2 and our
Hypothesis 6. (H6): Any effects of climate-migration messages will be
pre-registered criteria for excluding participants is shown in the sup­
moderated by political affiliation.
plementary materials.
We do not have predictions for true (non-leaning) Independents, who
tend to be a much more heterogeneous and politically disengaged group 2.1.2. Design
than those with party affiliations and leanings (Doherty, Kiley, & Participants first read a fictional news article reporting that climate
Johnson, 2019). We therefore focus on those who self-identified as change was causing people to leave their home communities and that
Democrats or Republicans (or leaning toward those parties) in the pre­ this trend was likely to increase. In a 2x2 design, participants were
sent set of studies. randomly assigned to receive an article that described either (1) do­
mestic or international migration, presented with either (2) just the­
1.5. Present research matic information or combined thematic and episodic information. The
international migration articles discussed people moving from other
We conducted two online experiments that were both approved as countries to the U.S. due to climate change, while the domestic migra­
exempt by the University of Michigan IRB to test hypotheses listed tion articles discussed people moving within the U.S. from one region to
above. Study 1 tested all hypotheses except H3 via recruiting a conve­ another due to climate change. The thematic-only framed articles pre­
nience sample of US adults and randomly assigning them to read sented statistics and graphics about the number of people migrating
different framings of fictional news articles about climate-driven (based on real data from published reports), but no information about
migration. Study 2 (pre-registered) tested all hypotheses by recruiting any particular person. The combined articles presented thematic ele­
a nationally representative sample of US adults and randomly assigning ments with the addition of a personal story about one climate migrant.
them to listen to one of four different variants of a fictional podcast As such, participants were randomly assigned to either the Domestic
about climate-driven migration or to a control condition in which they Thematic, Domestic Combined, International Thematic, or International
completed a questionnaire without podcast exposure. Combined condition (see Supplementary Materials).

2. Study 1 2.1.3. Measures

After reading the article, participants completed a survey with the
In Study 1, a convenience sample of U.S. adults were randomly following (see Supplementary Table 1 for item wording, descriptive, and
assigned to read one of four fictional newspaper articles about climate reliability statistics). All multi-items scales for both studies were inter­
migration. The articles varied in the type of migration discussed (do­ nally consistent (α > 0.80).
mestic vs. international) and information format used (thematic-only,
presenting issue trends using statistics and graphs vs. combined the­ 2.1.3.1. Climate change risk perceptions. Participants indicated climate
matic and episodic, presenting issue trends using statistics and graphs change risk perceptions along two dimensions: likelihood (0 not at all
with the inclusion of a personal story about one climate migrant). We likely to 4 strongly likely), and magnitude of threat (0 none at all to 4 a
examined how these message frames affected perceptions of climate great deal). The threat measure was a composite of three items asking
change, migration, and policy support for participants across the polit­ about perceived threat to themselves, their friends and family, and the
ical spectrum. Datasets, preregistration for power analyses, exclusion U.S. The likelihood item and threat composite were strongly correlated
criteria and data collection plan are available at https://osf.io/s2bnt/? so we averaged them together for a single climate risk composite
view_only=3bf64fc539bf4b78a04150aa34a7dcc0.2 measure.

2.1.3.2. Climate change policy support. Participants indicated support

for three climate change policies on a scale of − 3 (strongly oppose) to
+3 (strongly support) (e.g., “Requiring electric utilities to produce all of
their electricity from clean (carbon-neutral) energy sources by the year 2035,
even if this raises prices for consumers.”).
2
Note that we also preregistered preliminary hypotheses (see above link) but
realized when running analyses that we were missing some key predictions (for
2.1.3.3. Migration risk perceptions. Participants indicated the perceived
example, about the non-experimental effects of political affiliation and type of risks of migration using measures similar to the climate change risk
migration outcome) and specificity. Thus, we do not consider this study to be perceptions composite, but with “climate change” replaced with “immi­
fully pre-registered (in contrast to Study 2), though these pre-registered hy­ gration to the U.S.” and “migration from one region to another within the
potheses guided our analysis. Note that Hypothesis 3 (listed above in the U.S.”
introduction of our paper) was not planned or tested for this study as we were
testing the relative effects of different message framings in relation to one 2.1.3.4. Feelings toward migrants. Participants indicated their feelings
another and thus did not include a control condition.

5
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Fig. 1. Study 1 Perception of migration risk and migrants.

Fig. 2. Study 1 Support for migration policy outcomes.

6
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Table 3
Repeated-measure ANOVAs testing effects of experimental conditions and political affiliation on migration outcomes.
Predictor Study 1 Study 2

Migration risk Preference for native-born Migration risk Preference for native-born

F P η2p F p η2p F p η2p F p η2p

Between-subjects
Mig. 0.04 .843 .000 0.11 .741 .000 2.53 .112 .002 0.160 .689 .000
Format 0.031 .861 .000 2.71 .100 .002 0.42 .518 .000 1.00 .318 .001
Pol. 321.91 <.001 .182 390.77 <.001 .213 269.23 <.001 .153 419.72 <.001 .220
Mig. x Format 0.04 .839 .000 0.02 .904 .000 0.03 .864 .000 0.77 .379 .001
Mig. x Pol 0.34 .561 .000 0.00 .967 .000 0.00 .962 .000 0.97 .324 .001
Format x Pol. 3.95 .047 .003 0.24 .625 .000 0.10 .752 .000 0.49 .486 .000
Mig. x Format x Pol. 5.56 .019 .004 2.13 .145 .001 2.31 .129 .002 0.19 .660 .000
Within-subjects
Outcome 226.84 <.001 .136 56.41 <.001 .038 258.79 <.001 .148 85.12 <.001 .054
Outcome x Mig. 10.45 .001 .007 0.02 .876 .000 7.15 .008 .005 0.35 .551 .000
Outcome x Format 0.02 .883 .000 0.18 .670 .000 2.68 .102 .002 0.30 .585 .000
Outcome x Pol. 258.05 <.001 .152 155.99 <.001 .098 321.04 <.001 .177 195.93 <.001 .117
Outcome x Mig. x Format 0.10 .750 .000 0.159 .690 .000 0.62 .432 .000 0.12 .731 .000
Outcome x Mig x Pol. 1.95 .163 .001 1.66 .198 .001 3.22 .073 .002 1.61 .204 .001
Outcome x Format x Pol. 0.48 .487 .000 0.13 .722 .000 0.01 .904 .000 1.51 .220 .001
Outcome x Mig. x Format x Pol. 0.54 .463 .000 0.03 .865 .000 2.04 .153 .001 0.24 .625 .000

Note: Study 1 df = (1, 1444), Study 2 df for risk effects = (1, 1492), df for preference effects = (1, 1485).
Mig. = Type of migration described in article (international = 1); Format = Information format condition (thematic = 1); Outcome = Type of migration being rated for
this outcome measure (international = 1); Pol. = Political affiliation (Republican = 1).

of warmth toward (a) recent immigrants, (b) native-born Americans, (c) migration participants read about (domestic vs. international), infor­
Americans who recently moved to their state, and (d) people who were mation format (thematic-only vs. combined) and political affiliation
born in their state on a feeling thermometer with a 0–100 slider scale. (Republican vs. Democrat), as well as the within-subject effects of the
Two feeling differentials (for international and domestic migrants) were type of migration outcome they were asked about (domestic vs. inter­
created by subtracting warmth toward migrants from warmth toward national). Means for these analyses are shown in Figs. 1 and 2; ANOVA
native-born residents in each case, such that positive ratings indicated results are shown in Table 3 and Table 4. Significant simple effects tests
preference for native-born residents and negative ratings indicated probing interaction effects are shown in Table 6.
preference for migrants. Migration risk perceptions. There were large main effects of political
affiliation and migration outcome type on migration risk perceptions.
2.1.3.5. Climate-induced migration policy support. Participants indicated Compared to Democrats, Republicans consistently rated migration as
their support for six policies aimed to help either international or do­ riskier (supporting H5), and people rated international migration as
mestic migrants 1) move permanently, 2) move temporarily, or 3) help riskier than domestic migration (supporting H1a). These main effects
their home communities adapt to climate change on a -3 (strongly were qualified by a significant interaction of political affiliation and
oppose) to +3 (strongly support) scale. These items were analyzed outcome type. Qualifying H1a, simple effects tests showed that Re­
separately because we expected that people may treat policies to help publicans rated international migration as riskier than domestic
migrants stay home differently than policies to help them move. migration. No such effects emerged for Democrats.
There was also a significant interaction between the type of climate
2.1.3.6. Demographics and attention checks. Participants were asked migration participants read about and the type of migration outcome
“Generally speaking, do you usually think of yourself as a Republican, they were rating. Supporting H2a, simple effects tests showed that the
Democrat, Independent, or something else?” Those who selected “Inde­ perception that international migration was riskier than domestic
pendent” were then asked, “Do you think of yourself as closer to the migration was heightened among those who had also read about inter­
Republican Party or to the Democratic Party?” or “neither”. Those who national climate migration, compared to those who read about domestic
indicated leaning toward one party were assigned to that party in ana­ climate migration.3
lyses, all others were treated as “Independent” (n = 316) and dropped Feelings toward migrants. We next analyzed preference for native-born
from analyses with political affiliation. This left a sample of 1452 in the residents over migrants. Supporting H5d, Republicans showed greater
analyses that follow (Domestic Thematic: n = 352; Domestic Combined: perceived warmth for native-born residents over migrants than did
n = 369; International Thematic: n = 363; International Combined: n = Democrats. Supporting H1b, the preference for native-born residents
368). was stronger when respondents considered international migration than
Finally, participants reported demographics indicated in Table 2 and when rating domestic migration.
completed data quality checks. As with the migration risk perceptions, there was also a significant
interaction of political affiliation and migration outcome. Simple effects
tests found that for both migration outcomes, Republicans showed
2.2. Results greater perceived warmth for native-born residents over migrants than
Democrats, however this effect was magnified when considering
Analyses for both studies were conducted using SPSS Version 29.

2.2.1. Migration outcomes

First, we tested the effects of experimental condition on migration- 3
Two other interactions emerged. One was two-way interaction between
related outcomes (risk perceptions, feelings toward migrants, and information format and political affiliation and one was a three-way interaction
migration policies). Because we had parallel questions related to do­ between these variables and the type of migration participants read about. As
mestic and international migration outcomes, we ran repeated-measures we had no hypotheses about these interactions and because these effects did not
ANOVAs testing the between-subject effects of the type of climate replicate in Study 2, we discuss them in Supplementary Materials.

7
 K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

international migration and more muted when considering domestic

Mig. = Type of migration described in article (international = 1); Form = Information format condition (thematic = 1); Outcome = Type of migration being rated for this outcome measure (international = 1); Pol. =
migration. In addition, Republicans preferred native-born residents

.000
.001
.361
.003
.000
.000
.001

.311
.000
.000
.041
.000
.000
.001
.000
Support for helping stay particularly when considering international migration; this effect was

η2p
reversed for Democrats (see Fig. 1). Democrats only preferred native-

<.001

<.001

<.001
born residents when considering domestic migration, and even

.461
.358

.030
.742
.931
.205

.889
.571

.877
.647
.298
.402
showed a slight preference for migrants over native-born residents when
p

considering international migration. In contrast to H4, the information

842.63

672.70

63.39
format (thematic-only vs. combined) had no effect on preferences for
home

0.54
0.84

4.71
0.11
0.01
1.61

0.02
0.32

0.02
0.21
1.08
0.70
native-born residents.
F

Migration policies. We next tested support for policies to help poten­

tial climate migrants either move permanently, move temporarily, or
fund infrastructure to allow them to stay in their home communities (see
Support for helping move

.000
.000
.320
.002
.000
.001
.000

.301
.001
.000
.152
.000
.000
.000
.000
η2p

Fig. 2 and Table 4).

Democrats consistently supported all three types of policies more
<.001

<.001

<.001
than Republicans (supporting H5e). Participants were also more sup­
.714
.440

.133
.675
.325
.749

.337
.710

.907
.678
.756
.747
p

portive of policies to help Americans (vs. international migrants) move

temporarily

temporarily or stay home (supporting H1c). No such effect emerged for

701.55

641.43

267.02
0.923
0.14
0.60

2.26
0.18
0.97
0.10

0.14

0.01
0.17
0.10
0.10
policies to help people move permanently.
F

Across all three types of policies, these main effects were qualified by
significant interactions between political affiliation and the type of
migration outcome being considered. Simple effects test showed that
Democrats supported these policies more than Republicans, but that this
Support for helping move

.000
.000
.400
.000
.000
.000
.001

.001
.000
.000
.008
.000
.001
.001
.001
η2p

political polarization was stronger when considering international

migration policies than domestic ones.
<.001

<.001
.786
.826

.436
.431
.483
.380

.248
.533
.767

.414
.162
.345
.214

Another set of simple effects tests found that for helping people move
permanently

temporarily or stay home, both Democrats and Republicans supported

Study 2

domestic migration policies more than international ones, but this effect
994.72

11.62
0.07
0.05

0.61
0.62
0.49
0.77

1.33
0.39
0.09

0.67
1.95
0.89
1.55

was stronger among Republicans than Democrats. Indeed, as seen in

F

Fig. 2, Republican sentiment went below the midpoint to slight oppo­

Repeated-measure ANOVAs testing effects of experimental conditions and political affiliation on policy support outcomes.

sition when considering support for potential international migrants.

.004
.002
.179
.000
.000
.000
.000

.362
.002
.000
.058
.001
.000
.001
.000

Simple effects tests (and examination of Fig. 2) found a different

Support for helping stay

η2p

pattern when rating support for policies to promote permanent moves.

<.001

<.001

<.001

Here, Republicans continued to support domestic migration more than

.011
.066

.999
.568
.941
.673

.071
.713

.161
.802
.325
.579

international migration; notably however, the size of this effect was

p

much smaller than with Republican ratings of other policies due to slight
315.13

818.77

opposition to domestic migrants moving permanently (as opposed to

88.71
home

6.45
3.37

0.00
0.33
0.01
0.18

3.27
0.14

1.97
0.06
0.97
0.31

moderate support for the other forms of policies). Democrats also treated
F

permanent moves differently than the other policies; they actually

reversed their preferences by supporting permanent international
migration more than permanent domestic migration.
.001
.000
.182
.000
.000
.001
.000

.426
.002
.000
.164
.000
.000
.001
.000
Support for helping move

η2p

Our manipulations had only one effect: those who read about in­
ternational climate migration (vs. domestic climate migration) were
<.001

<.001

<.001
.346
.712

.648
.795
.272
.667

.069
.933

.448
.552
.250
.980

more supportive of policies to build infrastructure to help potential

p

migrants to stay home. This was somewhat in line with H2c, although
temporarily

we found this effect regardless of whether the policies were meant to

1070.13
320.58

282.59
0.577
0.354

help potential migrants from international or domestic locales. In

0.89
0.14

0.21
0.07
1.21
0.19

3.31
0.01

1.32
0.00

contrast to H4, support for these policies was not affected by the infor­
F

mation format of the articles.

2.2.2. Climate outcomes

Support for helping move

.000
.000
.263
.000
.000
.000
.000

.001
.000
.000
.054
.000
.001
.000
.000

Note: Study 1 df = (1, 1444), Study 2 df = (1, 1492).

η2p

While we did not have a control condition in Study 1 that would

allow formal tests of H3, we were curious whether the message frames
<.001

<.001

influenced climate change risk perceptions or policy support and

.929
.420

.941
.864
.561
.875

.164
.510
.559

.532
.186
.735
.442
permanently

whether this differed by political affiliation. We tested this question with

two 2 (domestic vs. international) x 2 (thematic-only vs. combined) x 2
Study 1

515.91

81.69

Political affiliation (Republican = 1).

0.01
0.65

0.01
0.03
0.34
0.03

1.94
0.44
0.34

0.39
1.75
0.11
0.59

(Republican vs. Democrat) factorial ANOVAs. Results are shown in

F

Table 5 and Fig. 5. Only political affiliation significantly predicted

climate change risk perceptions or policy support, with Democrats
Outcome x Mig. x Form x Pol.

showing higher risk perceptions and more support for mitigation pol­
Outcome x Mig. x Form

icies than Republicans. The experimental conditions had no significant

Outcome x Form x Pol.
Outcome x Form x Pol.

effects on these outcomes.

Mig. x Form x Pol.

Outcome x Form

Thus, we found no evidence of experimental effects on climate out­

Outcome x Mig.
Between-subjects

Outcome x Pol.
Within-subjects

comes. Answering RQ1, we found no evidence that the combined in­

Mig. x Form

Form x Pol.
Mig. x Pol
Predictor

Outcome

formation format differentially affected climate change risk perceptions

Table 4

Form

or policy support compared to the thematic-only format.

Mig.

Pol.

8
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Table 5
ANOVAs testing effects of experimental condition and political affiliation on climate change outcomes.
Predictor Study 1

Climate change risk Climate policy support

F p η2p F p η2p
Migration article condition 1.67 .197 .001 0.66 .418 .000
Information format condition 1.73 .189 .001 0.12 .734 .000
Political affiliation 594.01 <.001 .291 725.01 <.001 .334
Migration article x Format 0.09 .760 .000 0.23 .632 .000
Migration article x Political affil. 0.07 .796 .000 0.02 .896 .000
Format x Political affil. 0.74 .389 .001 0.03 .871 .000
Migration article x Format x Political affil. 0.31 .578 .000 0.52 .471 .000

Predictor Study 2

Climate change risk Climate policy support

F p η2p F p η2p
Condition 3.32 .010 .004 2.30 .057 .003
Political affiliation 1631.77 <.001 .465 1779.90 <.001 .487
Condition x Political affil. 0.15 .964 .000 0.46 .766 .001

Note: Study 1 N = 1452, df = (1, 1444); Study 2 N = 1873, df = (1, 1863). Cohen’s d showing differences between experimental conditions for the climate outcomes is
shown in Supplementary Table 4.

Table 6
Study 1: Simple effects tests for significant interactions.
Interaction probed Significant Simple Effects Migration risk Preference for native-born

Marginal means F η2p Marginal means F η2p

Outcome x Pol. Outcome type within Republicans Dom = 1.28; 402.16*** .218 Dom = 15.25; 165.72*** .103
Int = 2.09 Int = 32.02
Outcome type within Democrats Dom = 0.89; 0.63 ns (p = .427) .000 Dom = 3.03; 15.63*** .011
Int = 0.86 Int = − 1.13
Political party within International outcome Rep = 32.02; 441.91*** .235
Dem = − 1.13
Political party within Domestic outcome Rep = 15.25; 95.99*** .062
Dem = 3.04
Outcome x Migration article condition Outcome type within International condition Dom = 1.04; 169.75*** .105
Int = 1.51
Outcome type within Domestic condition Dom = 1.13; 68.97*** .046
Int = 1.44

Interaction Significant Simple Effects Move permanently Move temporarily Stay home
probed
Marginal means F η2p Marginal means F η2p Marginal means F η2p
Outcome x Outcome type within Dom = − 0.20; 45.16*** .030 Dom = 1.88; 1018.13*** . Dom = 1.18; 600.48*** .294
Pol. Republicans Int = − 0.69 Int = − 0.28 414 Int = − 0.43
Outcome type within Dom = 1.00; 36.74*** .025 Dom = 2.21; 158.93*** .099 Dom = 2.01; 231.58*** .138
Democrats Int = 1.37 Int = 1.52 Int = 1.21
Political party within Rep = − 0.69; 601.08*** 294 Rep = − 0.28; 432.76*** .231 Rep = − 0.43; 329.07*** .186
International outcome Dem = 1.37 Dem = 1.52 Dem = 1.21
Political party within Rep = − 0.20; 186.42*** .114 Rep = 1.88; 32.20*** .022 Rep = 1.18; 136.93*** .087
Domestic outcome Dem = 1.00 Dem = 2.21 Dem = 2.01

df for all effects = (1, 1444), ***p values for all tests <0.001.

2.3. Discussion 3. Study 2

The largest effects in Study 1 were not tied to how climate migration To ensure the small (and null) effects from Study 1 were not due to a
was discussed, but rather by the non-experimental predictors of political lack of attention to the stimuli, in Study 2 we employed podcast clip
affiliation and the type of migration participants evaluated. In contrast manipulations (rather than newspaper articles) to make the experi­
to hypotheses, including episodic information in messages with thematic mental manipulations more vivid and interactive. Participants listened
information did not decrease preferences for native-born residents over to one of four short podcast clips with the same conditions as Study 1.
migrants or increase support for policies to help migrants. However, the We also added a control group who did not hear about climate-induced
type of migration discussed affected risk perceptions: reading about migration. We surveyed a nationally representative sample of U.S.
international climate migration heightened its perceived risks. adults, allowing us to use this control group to test baseline U.S. public
Study 1 also found that how climate-induced migration was dis­ opinion on these issues. All hypotheses, exclusion criteria, and power
cussed made no difference for climate change outcomes. Yet, it could not analyses were pre-registered prior to data collection (see https://osf.
answer how these various communication techniques compare to a io/s2bnt/?view_only=3bf64fc539bf4b78a04150aa34a7dcc0).4
baseline of not mentioning climate-induced migration at all. For
example, it is possible that the act of reporting on climate-induced
migration does inspire climate change risk perceptions and action, but 4
Note that the OSF document was modified in January 2023 to create an
how it is framed makes no difference. anonymous link – the content of the preregistration was not changed.

9
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Fig. 3. Study 2 Perception of migration risk and migrants.

Fig. 4. Study 2 Migration policy support.

10
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Fig. 5. Climate change outcomes.

3.1. Methods 3.1.2. Materials and measures

Participants were randomly assigned to one of five conditions: four
3.1.1. Participants involved listening to a podcast matching Study 1 conditions (interna­
We enlisted YouGov to recruit 2400 U.S. adults to participate in an tional vs. domestic migration X thematic-only vs. combined), and those
online experiment in February 2022. This sample size was determined in the fifth condition (control) did not listen to a podcast and directly
through a priori power analyses. YouGov recruited all participants from completed the dependent measures. Transcriptions of the clips are in
its opt-in panel members, with a matching and weighting algorithm to Supplementary Materials and audio is available at https://osf.io/s2bnt
approximate a nationally representative sample of U.S. population. This /?view_only=3bf64fc539bf4b78a04150aa34a7dcc0.
procedure involves active recruiting of a diverse pool of respondents, a All measures were identical to Study 1 except as noted below (see
sample matching procedure to ensure that a representative pool of Supplementary Materials for descriptive statistics).
participants is selected from that larger panel to best match the U.S.
population based on factors such as demographics and political ideol­ 3.1.2.1. Risk perceptions. In Study 2, risk perceptions (of climate
ogy, followed by the provision of weights to account for nonresponse. change, international migration, and domestic migration) were assessed
While this is not a true probability sample, estimates show that this with two questions asking about the likelihood and severity of negative
procedure performs as well as (if not better than) traditional probability consequences from each source. Likelihood was measured as in Study 1.
designs in creating nationally representative samples (Cohn, 2019; For severity, participants responded on a scale of 0 (not a problem) to 4
Kennedy et al., 2016). (very serious) to the following question: “How serious do you think the
We used sample weights provided by YouGov to calculate descriptive negative consequences of [climate change/immigration to the U.S./migration
statistics for the full sample and separately for the control condition to from one region to another within the U.S.] will be.” We used the mean of
test baseline U.S. public sentiment (Supplementary Tables 1 and 2). these two risk components as a composite measure of risk perceptions.
Otherwise we did not use weights, following findings that using weights
in YouGov survey experiments does not add accuracy and instead de­ 3.1.2.2. Political affiliation. Political affiliation was assessed similarly
creases the precision and thus power of estimates (Miratrix et al., 2018). to Study 1 with some slight changes (see Supplementary Materials). As
All analyses reported in the main text therefore used unweighted scores. in Study 1, true Independents (n = 527) were excluded from analyses
Data quality checks and any exclusion of participants were handled by with political affiliation. This left a sample of 1873 in the analyses that
YouGov prior to sharing the dataset with researchers (see Supplemen­ follow (Domestic Thematic: n = 370; Domestic Combined: n = 377;
tary Materials for pre-registered exclusion criteria). Participant de­ International Thematic: n = 378; International Combined: n = 375;
mographics are shown in Table 2. Control: n = 373).

11
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Table 7
Study 2: Simple effects tests for significant interactions.
Interaction probed Significant Simple Effects Migration risk Preference for native-born

Marginal means F η2p Marginal means F η2p

Outcome x Pol. Outcome type within Republicans Dom = 1.87; Int = 502.92*** .252 Dom = 19.13; Int = 233.74*** .136
2.86 39.18
Outcome type within Democrats Dom = 1.57; Int = 1.97 ns (p = .001 Dom = 0.64; Int = 4.76 13.45*** .009
1.52 .161)
Political party within International Rep = 39.18; Dem = 507.11*** .255
outcome 0.64
Political party within Domestic Rep = 19.13; Dem = 108.93*** .068
outcome 4.76
Outcome x Migration article Outcome type within International Dom = 1.72; Int = 177.52*** .106
condition condition 2.27
Outcome type within Domestic Dom = 1.72; Int = 89.19*** .056
condition 2.11

Interaction probed Significant Simple Move permanently Move temporarily Stay home
Effects
Marginal F η2p Marginal F η2p Marginal F η2p
means means means

Outcome x Pol. Outcome type within Dom = − 1.14; 2.21ns (p = .001 Dom = 0.82; 755.12*** .336 Dom = 0.07; 499.78*** .251
Republicans Int = − 1.23 .137) Int = − 0.97 Int = − 1.38
Outcome type within Dom = 0.97; 12.25*** .008 Dom = 1.87; 47.47*** .031 Dom = 1.81; 189.96*** .113
Democrats Int = 1.16 Int = 1.48 Int = 1.04
Political party within Rep = − 1.23; 874.53*** .370 Rep = − 0.97; 899.37*** .376 Rep = − 1.38; 814.44*** .
International outcome Dem = 1.16 Dem = 1.48 Dem = 1.04 353
Political party within Rep = − 1.14; 615.32*** .292 Rep = 0.82; 191.88*** .114 Rep = 0.07; 450.41*** .232
Domestic outcome Dem = 0.97 Dem = 1.87 Dem = 1.81
Information format x Migration article within Dom = 0.25; 4.22* (p = .003
Migration article Thematic condition Int = 0.46 .040)
condition Migration article within Dom = 0.47; 1.03ns (p = .001
Combined condition Int = 0.37 .310)
Information format Them = 0.46; 0.79 ns (p = .001
within International Comb = 0.37 .374)
condition
Information format Them = 0.25; 4.73* (p = .003
within Domestic Comb = 0.47 .030)
condition

df for risk and policy effects = (1, 1492), df for preference effects = (1, 1485), ***unless otherwise noted p values for all tests <0.001.

3.2. Results Feelings toward migrants. As in Study 1, political affiliation, migration

outcome type, and their interaction predicted preference for native-born
3.2.1. Migration outcomes residents over migrants. Replicating Study 1, simple effects test found
Repeated-measures ANOVAs tested the between-subject effects of that Republicans preferred native-born residents over migrants more
the type of climate migration participants heard about (domestic vs. than Democrats, which was magnified when considering international
international), information format (thematic-only vs. combined) and migration rather than domestic migration. As in Study 1, among Re­
political affiliation (Republican vs. Democrat), and the within-subject publicans, warmer feelings toward native-born residents was higher
effects of the type of migration outcome being evaluated (domestic vs. when considering international migration. This effect was reversed for
international) on migration outcomes. To make easy comparisons to Democrats; see Fig. 3. No effects emerged for either the type of climate
Study 1, we did not include the control condition in these analyses (N = migration participants heard about nor the information format (in
1500). Means are shown in Figs. 3 and 4; other values are shown in contrast to H4).
Tables 3 and 4. Significant simple effects tests probing interactions are Migration policies. As in Study 1, political affiliation interacted with
shown in Table 7. the migration outcome being considered for all three types of policies.
Migration risk perceptions. Replicating Study 1, Republicans rated all Again replicating Study 1, simple effects tests showed that Democrats
migration as risker than Democrats (supporting H5c) and participants supported these policies more than Republicans, especially when
rated international (vs. domestic) migration as riskier (supporting H1a). considering international migration rather than domestic ones.
As before, simple effects tests of the significant interaction between Replicating Study 1, simple effects tests showed that for policies
these two non-experimental variables found that Republicans (but not supporting temporary moves or helping migrants stay home, Re­
Democrats) rated international migration as riskier than domestic. publicans supported domestic migration policies more than interna­
Also replicating Study 1, there was a significant interaction between tional ones and that Democrats showed the same preference to a lesser
the type of climate migration participants heard about and the type of degree. As before, Republican responses reflected opposition when
migration outcome they were rating. Regardless of what form of climate considering support for international migrants. Meanwhile, when rating
migration they heard about, people rated international migration as support for policies to promote permanent moves, Republicans did not
riskier than domestic. However, as in Study 1, simple effects showed that show a preference for domestic over international migration; in Study 1
the heightened perceived risk of international (vs. domestic) migration there was a slight preference. As in Study 1, Democrats reversed their
was stronger among those who had just heard about international preferences by supporting permanent international migration more than
climate migration instead of domestic climate migration (supporting permanent domestic migration.
H2a). In contrast, the perceived risk of domestic migration was not Further replicating Study 1, the only effect of experimental condition
higher among those who had heard about domestic climate migration was on policies to help migrants to stay home, although here there was a
(rejecting H2b). significant interaction between information format and the type of

12
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

climate migration people heard about (rather than in Study 1, where it 4. General discussion
was just a main effect of the type of migration). Simple effects found that
within the thematic-only condition, people who heard about interna­ Our two studies tested effects both of whether climate-induced
tional (vs. domestic) climate migration were more supportive of helping migration is discussed (Study 2) and how it is discussed (both studies).
migrants stay home. There was no such effect within the combined We assume that those communicating about climate migration, such as
condition. Among those who heard about domestic climate migration, journalists, researchers, and activists, often aim to increase action on
those in the combined condition supported helping migrants stay home climate mitigation and support for those affected by climate change.
more than those in the thematic-only condition. There was no such effect However, our studies suggest that such efforts may not always be
among those who heard about international climate migration. effective and may even have unintended consequences.
A second set of (non-preregistered) exploratory analyses tested how One barrier to the effect of any communication efforts in this arena,
the experimental conditions compared to the baseline control for each of at least in the U.S., is the dominant influence of political ideology on
the migration outcomes. As shown in Supplementary Table 3 and Sup­ attitudes toward both climate change and international migration.
plementary Fig. 1, the conditions only affected the perceived risks of Indeed, the strongest predictors of risk perceptions, preferences for
international and domestic migration (in each case, qualified by a sig­ native-born residents over migrants, and policy support were political
nificant interaction of condition and political affiliation). For both forms affiliation and the type of migration being considered (international vs.
of migration, there was no effect of condition among Republicans. domestic). Supporting H5, Republicans (vs. Democrats) consistently
Among Democrats, those in the control condition reported less rated climate change as less risky and migration as more risky and
perceived risk of domestic migration than in any of the experimental supported policies in line with those beliefs. In addition, international
conditions and less risk of international migration than in any other migration was generally seen as more risky than domestic migration,
conditions except the domestic thematic-only condition. and policy support for the former was more muted (supporting H1). The
effect sizes of these relationships were much larger than the effects of the
3.2.2. Climate outcomes message frames tested in the present study, indicating that strong
We ran two 5 (Control, Domestic Thematic, Domestic Combined, partisanship (e.g., Hornsey et al., 2016) and in-group favoritism (Everett
International Thematic, International Combined) x 2 (Republican vs. et al., 2015) may make it difficult to change opinions through message
Democrat) factorial ANOVAs to test H3. For these analyses, we excluded framing alone. Thus, any one story about climate-induced migration
527 independents who did not lean toward a party, leaving 1873 par­ faces a considerable uphill battle to overcome these entrenched starting
ticipants in analyses. ANOVA test statistics are detailed in Table 5 and positions if it is to move the needle on public opinion and policy support.
patterns are visually illustrated in Fig. 5. Despite these barriers, we find some limited effects of how climate
Climate change risk perceptions. As in Study 1, Democrats perceived migration is discussed.
significantly more risk from climate change than Republicans. Unlike in
Study 1, there was also a main effect of condition. Post-hoc multiple 4.1. Impacts on migration risk and preference for native-born residents
comparison tests with Sidak corrections found that both of the domestic over migrants
climate migration conditions (Thematic-only: M = 2.51, SE = .06, p =
.021; Combined: M = 2.57, SE = .06, p = .003) increased climate change Our results suggest that communicators should proceed with caution
risk above the baseline Control condition (M = 2.31, SE = .06). No other when discussing climate-induced migration. Study 2 found that
significant differences emerged between conditions, and there was not a regardless of how it was presented, this type of messaging increased the
significant interaction between condition and political affiliation. perceived risk of migration among Democrats. Republicans did not show
Climate change policy support. As in Study 1, only political affiliation this pattern, possibly because they already view migration as risky.
predicted climate change mitigation policy support, with Democrats There were no significant differences between the baseline control and
more supportive than Republicans. The effect for condition was mar­ other migration outcomes, however, suggesting that Democrats’
ginal (p = .057), and post-hoc comparisons with Sidak correction found increased risk perceptions after hearing about climate migration may
that like climate change risk perceptions, this was due to higher support not translate into widespread changes to policy support.
for mitigation policies in the Domestic Combined condition (M = 0.30, Supporting H2a, participants rated international migration as riskier
SE = .08) than the Control condition (M = − 0.02, SE = .08, p = .010), after consuming news about it, which would have made that form of
though here the control did not differ from the Domestic Thematic-only migration and its potential impacts more salient. In contrast to H2b,
condition. There was again no significant interaction. domestic migration was not rated as riskier when people read or heard
Thus, H3 was only partially supported: some, but not all, conditions about that form of climate migration. Neither the type of migration
increased climate risk and policy support compared to baseline. discussed nor the information format affected feelings of warmth toward
migrants compared to native-born residents. Fortunately, and unlike in
3.3. Discussion other recent studies (Gillis et al., 2023), this means that exposure to
information about international migrants and the thematic information
As in Study 1, the most robust set of predictors in Study 2 were po­ formats employed in the present studies did not lead to a dehumanizing
litical affiliation and the type of migration being evaluated. Replicating lack of warmth toward migrants. However, it is worth noting that the
Study 1, we found that hearing about international climate migration warmth measures in the present studies asked about feelings of warmth
made participants rate it as riskier. We also replicated Study 1 in that toward migrants in general (not climate migrants in particular). It is
episodic information about one climate migrant did not affect prefer­ possible that having a more specific set of warmth questions would have
ences for native-born residents over migrants. Unlike in Study 1, picked up on more coldness to the particular type of migrants described
episodic information slightly increased support for helping potential in these materials. These results suggests that coverage of people coming
migrants stay home, but only when hearing about domestic migrants. from other countries to the U.S. due to climate change may be particu­
The inclusion of a baseline control in Study 2 allowed us to test larly likely to inspire fear and backlash (though not reduced warmth),
whether discussing climate migration at all makes a difference. We while coverage about U.S. migrants may be less inflammatory.
found that it did, with mixed effects: discussing domestic climate
migration increased fears of climate change and, if paired with episodic 4.2. Impacts on support for migration policies
information, increased support for policies to reduce emissions. How­
ever, discussing any form of climate migration also raised Democrats’ As expected, participants responded differently to the three types of
perceived risks of migration. migration policies: helping people move permanently, helping them

13
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

move temporarily, and building adaptation infrastructure to help them place.

stay in their home communities. This is notable, as previous work that
has examined Americans’ support for climate migration policies found 4.3. Impacts on climate change outcomes
no effect of whether those policies were in aid of permanent (vs. tem­
porary) moves (Stanley et al., 2022). Yet, the current studies found that For climate change outcomes, the results suggest that discussing
policies to help people move permanently inspired a more partisan climate-induced migration may—at best—shift beliefs slightly. The
response than those to help them move temporarily or stay home. difference between experimental conditions were quite small: neither
Whereas Republicans showed at least modest support for helping do­ study found significant differences between experimental conditions
mestic migrants move temporarily or stay home, their average opinion (domestic vs. international; thematic-only vs. combined) on climate
turned to opposition when it came to helping domestic migrants move change risk perceptions or policy support. Furthermore, while both
permanently. Meanwhile Democrats supported domestic migrants more studies found that political affiliation strongly predicted climate out­
than international ones when it came to temporary moves or infra­ comes, it did not moderate the effects of condition, disconfirming H6.
structure to help them stay home. Yet their preferences switched to Yet, compared to baseline, there may be slight benefits in reporting
support international migrants more than domestic ones when it came to on Americans who may have to migrate: partially supporting H3, Study
supporting permanent moves. Thus, policies promoting permanent 2 found that discussing domestic climate migration increased climate
moves are likely to experience the biggest partisan divisions, and even change risk perceptions above the baseline, and that discussing domestic
domestic migrants may face resistance when seeking support to move climate migration with a personal story increased climate mitigation
permanently, as has been seen following storms in the U.S. (Shelton & policy support. Thus, in answer to RQ1, including episodic information
Coleman, 2009). may inspire climate action, but only for domestic migration. Given that
In both studies, the only effects of experimental condition on policy both ingroup members (Stürmer et al., 2005) and episodic information
support emerged for policies to help migrants to stay home. Supporting formats (Gross, 2008) tend to invoke sympathy and prosocial action, it
H2c, learning about international (vs. domestic) climate migration makes sense that this combination was most effective in promoting
increased support for funding infrastructure to help people stay in their climate change mitigation, even if it didn’t inspire more warmth.
home communities (though this effect was restricted to those receiving Many studies have found that discussing various outcomes and
the thematic-only framing in Study 2). However, the same effect was consequences of climate change can increase risk salience (Rode et al.,
found whether this infrastructure was to be built in the U.S. or else­ 2021). Therefore, the null result for the international migration out­
where, meaning that learning about domestic climate migration comes and small effects for domestic migration outcomes may imply
decreased (not increased) support for policies to help domestic migrants that discussing climate migration is less effective in motivating action
stay in their home communities. compared to other climate change consequences. Future research might
In Study 2, among respondents who heard about domestic climate explore this by comparing climate migration to other consequences of
migration, those who were exposed to the combined information format climate change, such as health impacts or environmental degradation.
supported helping migrants stay home more than those exposed to the Although our message frames made a politically controversial
thematic-only format (supporting H4). Together these results reflect connection between climate change and immigration, none of them
that, particularly in Study 2, the Domestic Thematic-only condition reduced climate change risk perceptions or engagement outcomes below
inspired the least support when it came to helping people stay home. baseline. Other message frames (e.g., national security) have had mixed
This may be because Americans are less familiar with the idea of effects, with some studies finding them to be useful tools for increasing
climate-induced domestic migration, as they tend to think climate concerns about climate change, at least when delivered by a relevant
change is more of a threat to people in other countries (Gifford et al., messenger (i.e., a military source) (Bolsen, Palm, & Kingsland, 2019;
2009). Without a story to make sense of domestic migration, they might Motta, Ralston, & Spindel, 2021) whereas others find that they lead
therefore not be moved to address infrastructure changes to deal with it. those on the political right to feel more anger (Myers et al., 2012) and
Alternatively, when picturing domestic adaptation infrastructure, they view climate change policy as less important (Hart & Nisbet, 2012), even
may picture wealthy coastal dwellers rather than people in need. Future below baseline levels. The current findings do not suggest that climate
research on lay theories of domestic climate migrants would help un­ communicators should avoid discussing climate migration entirely, but
pack these findings. rather that if they hope to use migration outcomes to motivate climate
Thus, where we do find communication effects on policy support, action, they should focus on migration within the U.S. and include
they point to opinions on infrastructure-building policies as the most migrant narratives as exemplars.
malleable. This may reflect the fact that these infrastructure-building Given that where changes to climate change risk perceptions did
policies are not cleanly pro-immigration policies: while they provide occur, they were even more muted for policy support, future research
aid to potential migrants, the intended outcome of such aid is reduced may explore which additional steps may be needed to translate changes
migration. Support for these items could encompass both anti- in climate change risk perceptions due to migration into support for
immigration sentiment and pro-migrant sentiment. Furthermore, these mitigation policies. These efforts may require actions to reduce partisan
infrastructure-building policies are climate adaptation efforts, and as entrenchment on climate polices or efforts to help weigh the costs of
such, should reflect beliefs and policy support about climate change. climate change policies against the costs of inaction.
Thus, just as communications that combine climate change and migra­
tion may lead to unexpected results by combining different controversial 4.4. Future directions and limitations
topics, support for policies that address multiple of these issues may also
not behave similarly as support for just one topic or the other. Additional As with any vignette studies, the messages tested here only represent
research should examine how these policies are interpreted by various one example of a climate migration story. Different protagonists, such as
publics and to what extent we could or should be altering their framing. those of different races or from different geographic areas, may inspire
No experimental effects emerged for policies to support migrants mov­ stronger responses. For example, if the more positive response to do­
ing, perhaps because these policies are more tied to traditional party mestic (vs. international) migration stories was due to participants’
stances and rhetoric on immigration (Younis, 2020) and thus less sus­ heightened feelings of similarity to other Americans (vs. immigrants),
ceptible to novel frames. These results suggest that while coverage about this gap could be mitigated by using personalized audience segmenta­
international climate migration might worryingly increase the perceived tion approaches to provide immigrant exemplars with similar back­
risk of migration, it may also inspire Americans to help migrants, at least grounds to a particular message recipient. Conversely, there may be
if that help also serves to discourage people from migrating in the first further unintended consequences of these frames that were not

14
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

measured here, such as creating new stereotypes of climate migrants Declaration of interest
based on these stories.
These studies also didn’t explore the role of the source of the infor­ None.
mation message. An immigration-focused message may only persuade
when delivered by messengers who are perceived to be relevant and Funding
trusted by the target audience.
This work is also limited in its focus on the U.S. and for the specific This work was funded by a grant from the University of Michigan
policies asked about in these studies. While the U.S. has an outsized role Carbon Neutrality Acceleration Program.
when it comes to climate emissions, it is not the only contributor and is
hardly the only country that is facing domestic or international migra­ Acknowledgments
tion due to climate change. In particular, our results tied to U.S. political
parties are unlikely to neatly generalize to other countries and the The authors thank Ann Lin and Daniel Raimi for their help in crafting
support for the specific types of policies that we asked about may not and evaluating survey materials.
capture support for other policies in this realm. The extreme partisan
divide on climate change found in the U.S. is an outlier, rather than the Appendix A. Supplementary data
norm, and so the role of party affiliation in these decisions may play a
smaller role in other countries (Smith & Mayer, 2019). Yet, together Supplementary data to this article can be found online at https://doi.
with work focused on other countries (Stanley, Ng Tseung-Wong, Lev­ org/10.1016/j.jenvp.2023.102210.
iston, & Walker, 2021; Stanley & Williamson, 2021), these experiments
can help to start painting a picture of what public reaction is likely to be References
to this new source of migration across the globe.
Finally, as with all self-report studies on controversial topics, there is Aarøe, L. (2011). Investigating frame strength: The case of episodic and thematic frames.
Political Communication, 28(2), 207–226. https://doi.org/10.1080/
also the possibility that participants engaged in socially desirable 10584609.2011.568041
responding rather than responding as they really would if these policies Allen, M., Bruflat, R., Fucilla, R., Kramer, M., McKellips, S., Ryan, D. J., et al. (2000).
were on a ballot. It will be important for future research to compare Testing the persuasiveness of evidence: Combining narrative and statistical forms.
Communication Research Reports, 17(4), 331–336. https://doi.org/10.1080/
these types of attitudes and self-reported support studies to real-world 08824090009388781
policy actions as climate-induced migration becomes more of a focus Arias, S. B., & Blair, C. W. (2022). Changing tides: Public attitudes on climate migration.
of political attention. The Journal of Politics, 84(1), 560–567. https://doi.org/10.1086/715163
Bernhard, H., Fischbacher, U., & Fehr, E. (2006). Parochial altruism in humans. Nature,
442(7105), 912–915. https://doi.org/10.1038/nature04981
4.5. Conclusions Bilak, A., Cardona-Fox, G., Ginnetti, J., Rushing, E. J., Scherer, I., Swain, M., et al.
(2016). Global report on internal displacement. Internal Displacement Monitoring
Centre (IDMC).
Amid the growing rate of climate-induced migration and expecta­
Blake, J. S., Clark-Ginsberg, A., & Balagna, J. (2021). Addressing climate migration: A
tions that climate change will cause vast increases in migration in future review of national policy approaches. RAND Corporation. https://www.rand.org/pub
years, journalists, academics, and advocates have started to discuss this s/perspectives/PEA1085-1.html.
consequence of climate change. However, the impact of this coverage is Bolsen, T., Palm, R., & Kingsland, J. T. (2019). The impact of message source on the
effectiveness of communications about climate change. Science Communication, 41
less certain. Other work on framing in climate communication has found (4), 464–487. https://doi.org/10.1177/1075547019863154
presenting climate change alongside an issue that resonates with audi­ Boukes, M. (2022). Episodic and thematic framing effects on the attribution of
ences’ values or public health concerns can enhance understanding or responsibility: The effects of personalized and contextualized news on perceptions of
individual and political responsibility for causing the economic crisis. The
engender hope for actions to mitigate it (Goldberg et al., 2021; Myers International Journal of Press/Politics, 27(2), 374–395. https://doi.org/10.1177/
et al., 2012). Therefore, it is reasonable that climate migration could be 1940161220985241
yet another avenue to spur more awareness about this topic. Carrico, A. R., Truelove, H. B., Vandenbergh, M. P., & Dana, D. (2015). Does learning
about climate change adaptation change support for mitigation? Journal of
Yet these studies should give climate communicators pause when Environmental Psychology, 41(1), 19–29. https://doi.org/10.1016/j.
communicating with those who identify as a Republican or Democrat (or jenvp.2014.10.009
learn toward either of these parties), as they may have deep-rooted Castellano, R., Dolšak, N., & Prakash, A. (2021). Willingness to help climate migrants: A
survey experiment in the korail slum of dhaka, Bangladesh. PLoS One, 16(4), Article
biases associated with their political orientation. These experiments e0249315. https://doi.org/10.1371/journal.pone.0249315
show that, at best, discussing climate-induced migration has small ef­ Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Erlbaum.
fects on climate change concern and policy support that is limited to Cohn, N. (2019). No one picks up the phone, but which online polls are the answer? The New
York Times. https://www.nytimes.com/2019/07/02/upshot/online-polls-analyzi
discussion of domestic migration within the U.S. Instead, discussions of
ng-reliability.html.
climate migration seem to increase Americans’ (especially Democrats’) Doherty, C., Kiley, J., & Johnson, B. (2019). Political independents: Who they are are, what
perception that migration is risky compared to not discussing this topic. they think. Pew Research Center.
While it is important to continue public discussion about climate Evans, L., Milfont, T. L., & Lawrence, J. (2014). Considering local adaptation increases
willingness to mitigate. Global Environmental Change, 25, 69–75. https://doi.org/
migration, as adapting to changing migration drivers will require it, the 10.1016/j.gloenvcha.2013.12.013
current studies suggest that, where possible, public-facing discussions Everett, J. A. C., Faber, N. S., & Crockett, M. (2015). Preferences and beliefs in ingroup
should highlight the impact of climate change within the U.S. rather favoritism. Frontiers in Behavioral Neuroscience, 9. https://doi.org/10.3389/
fnbeh.2015.00015
than from other countries to the U.S. and may benefit from the inclusion Gainous, J., & Merry, M. K. (2022). Climate change as a national security issue:
of individual migrant narratives. However, if the goal to increase sup­ Examining framing effects across party. American Politics Research, 50(2), 199–212.
port for policies to build climate adaptation infrastructure, communi­ https://doi.org/10.1177/1532673X211053213
Geiger, N., Raimi, K. T., Roy, S., & Cunningham, J. L. (2023). Public opinion and
cators may wish to consider highlighting international immigration to communication on climate-driven migration. Under review.
the U.S., with the caveat that this may undermine other efforts to help Gifford, R., Scannell, L., Kormos, C., Smolova, L., Biel, A., Boncu, S., et al. (2009).
migrants move. Future research should attempt to replicate findings and Temporal pessimism and spatial optimism in environmental assessments: An 18-
nation study. Journal of Environmental Psychology, 29(1), 1–12. https://doi.org/
extend them by not only testing with multiple messages, but also 10.1016/j.jenvp.2008.06.001
determining effects of climate migration coverage on less politically Gillis, A., Geiger, N., Raimi, K., Cunningham, J. L., & Sarge, M. A. (2023). Climate
biased audiences. change–induced immigration to the United States has mixed influences on public
support for climate change and migrants. Climatic Change, 176(5), 48. https://doi.
org/10.1007/s10584-023-03519-y

15
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Goldberg, M. H., Gustafson, A., Rosenthal, S. A., & Leiserowitz, A. (2021). Shifting O’Donnell, A. W., Neumann, D. L., Duffy, A. L., & Paolini, S. (2019). Learning to fear
Republican views on climate change through targeted advertising. Nature Climate outgroups: An associative learning explanation for the development and reduction of
Change, 11(7), 573–577. https://doi.org/10.1038/s41558-021-01070-1 intergroup anxiety. Social and Personality Psychology Compass, 13(3). https://doi.org/
Gross, K. (2008). Framing persuasive appeals: Episodic and thematic framing, emotional 10.1111/spc3.12442
response, and policy opinion. Political Psychology, 29(2), 169–192. https://doi.org/ Podesta, J. (2019). The climate crisis, migration, and refugees. Brookings Blum
10.1111/j.1467-9221.2008.00622.x Roundtable, 1–6.
Hart, P. S. (2011). One or many? The influence of episodic and thematic climate change Pörtner, H.-O., Roberts, D. C., Tignor, M. M. B., Poloczanska, E. S., Mintenbeck, K.,
frames on policy preferences and individual behavior change. Science Alegría, A., et al. (Eds.). (2022). Climate change 2022: Impacts, adaptation and
Communication, 33(1), 28–51. https://doi.org/10.1177/1075547010366400 vulnerability. Contribution of working group II to the sixth assessment report of the
Hart, P. S., & Nisbet, E. C. (2012). Boomerang effects in science communication: How intergovernmental panel on climate change.
motivated reasoning and identity cues amplify opinion polarization about climate Raimi, K. T., Stern, P. C., & Maki, A. (2017). The promise and limitations of using
mitigation policies. Communication Research, 39(6), 701–723. https://doi.org/ analogies to improve decision-relevant understanding of climate change. PLoS One,
10.1177/0093650211416646 12(1), 1–20. https://doi.org/10.1371/journal.pone.0171130
Haynes, C., Merolla, J. L., & Ramakrishnan, S. K. (2016). Framing immigrants: News Riek, B. M., Mania, E. W., & Gaertner, S. L. (2006). Intergroup threat and outgroup
coverage, public opinion, and policy. Russell Sage Foundation. http://muse.jhu.edu/b attitudes: A meta-analytic review. Personality and Social Psychology Review, 10(4),
ook/46876. 336–353. https://doi.org/10.1207/s15327957pspr1004_4
Hedegaard, T. F. (2022). Attitudes to climate migrants: Results from a conjoint survey Rigaud, K. K., de Sherbinin, A., Jones, B., Bergmann, J., Clement, V., Ober, K., et al.
experiment in Denmark. Scandinavian Political Studies, 45(1), 25–45. https://doi.org/ (2018). Groundswell: Preparing for internal climate migration. International Bank for
10.1111/1467-9477.12213 Reconstruction and Development/The World Bank.
Heizler, O., & Israeli, O. (2021). The identifiable victim effect and public opinion toward Rode, J. B., Dent, A. L., Benedict, C. N., Brosnahan, D. B., Martinez, R. L., & Ditto, P. H.
immigration; a natural experiment study. Journal of Behavioral and Experimental (2021). Influencing climate change attitudes in the United States: A systematic
Economics, 93, Article 101713. https://doi.org/10.1016/j.socec.2021.101713 review and meta-analysis. Journal of Environmental Psychology, 76, Article 101623.
Helbling, M. (2020). Attitudes towards climate change migrants. Climatic Change, 160(1), https://doi.org/10.1016/j.jenvp.2021.101623
89–102. https://doi.org/10.1007/s10584-020-02697-3 Sakellari, M. (2022). Media coverage of climate change induced migration: Implications
Hornsey, M. J., Harris, E. A., Bain, P. G., & Fielding, K. S. (2016). Meta-analyses of the for meaningful media discourse. Global Media and Communication, 18(1), 67–89.
determinants and outcomes of belief in climate change. Nature Climate Change, 6(6), https://doi.org/10.1177/17427665211064974
622. Schäfer, M. S., Scheffran, J., & Penniket, L. (2016). Securitization of media reporting on
IPCC. (2022). Technical summary. Climate change 2022: Impacts, adaptation and climate change? A cross-national analysis in nine countries. Security Dialogue, 47(1),
vulnerability. In Contribution of working group II to the sixth assessment report of the 76–96. https://doi.org/10.1177/0967010615600915
intergovernmental panel on climate change (pp. 37–118). Cambridge University Press. Shelton, J. E., & Coleman, M. N. (2009). After the storm: How race, class, and
https://doi.org/10.1017/9781009325844.002. immigration concerns influenced beliefs about the Katrina evacuees. Social Science
Iyengar, S. (1990). Framing responsibility for political issues: The case of poverty. Quarterly, 90(3), 480–496. https://doi.org/10.1111/j.1540-6237.2009.00627.x
Political Behavior, 12(1), 19–40. https://doi.org/10.1007/BF00992330 Singh, S. P., & Swanson, M. (2017). How issue frames shape beliefs about the importance
Iyengar, S. (1991). Is anyone responsible?. How television frames political issues. University of climate change policy across ideological and partisan groups. PLoS One, 12(7),
of Chicago Press. Article e0181401. https://doi.org/10.1371/journal.pone.0181401
Jebari, J., Andrews, T. M., Aquila, V., Beckage, B., Belaia, M., Clifford, M., et al. (2021). Smith, E. K., & Mayer, A. (2019). Anomalous anglophones? Contours of free market
From moral hazard to risk-response feedback. Climate Risk Management, 33, Article ideology, political polarization, and climate change attitudes in English-speaking
100324. https://doi.org/10.1016/j.crm.2021.100324 countries, western European and post-communist states. Climatic Change, 152(1),
Jones, J. M. (2021). Americans remain divided on preferred immigration levels. Gallup. https 17–34. https://doi.org/10.1007/s10584-018-2332-x
://news.gallup.com/poll/352664/americans-remain-divided-preferred-immigrati Spilker, G., Nguyen, Q., Koubi, V., & Böhmelt, T. (2020). Attitudes of urban residents
on-levels.aspx towards environmental migration in Kenya and Vietnam. Nature Climate Change, 10
Kennedy, C., Mercer, A., Keeter, S., Hatley, N., McGeeney, K., & Gimenez, A. (2016). (7), 622–627. https://doi.org/10.1038/s41558-020-0805-1
Evaluating online nonprobability surveys. https://www.pewresearch.org/methods/201 Stanley, S. K., Ng Tseung-Wong, C., & Leviston, Z. (2022). Welcoming climate refugees to
6/05/02/evaluating-online-nonprobability-surveys/. the United States: Do attitudes depend on refugee origins, numbers, or permanence?
Kim, B., Kay, D. L., & Schuldt, J. P. (2021). Will I have to move because of climate Journal of Environmental Psychology, 83, Article 101874. https://doi.org/10.1016/j.
change? Perceived likelihood of weather- or climate-related relocation among the US jenvp.2022.101874
public. Climatic Change, 165, 9. https://doi.org/10.1007/s10584-021-03026-y Stanley, S. K., Ng Tseung-Wong, C., Leviston, Z., & Walker, I. (2021). Acceptance of
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), climate change and climate refugee policy in Australia and New Zealand: The case
480–498. https://doi.org/10.1037/0033-2909.108.3.480 against political polarisation. Climatic Change, 169(3–4), 26. https://doi.org/
Leiserowitz, A. A., Maibach, E. W., Rosenthal, S. A., Kotcher, J. E., Carman, J. P., 10.1007/s10584-021-03278-8
Neyers, L., et al. (2022). Politics & global warming. Yale Program on Climate Change Stanley, S. K., & Williamson, J. (2021). Attitudes towards climate change aid and climate
Communication. April 2022. refugees in New Zealand: An exploration of policy support and ideological barriers.
Lu, P., Liu, J., & Koestler, D. (2017). Power and Sample Size Analysis for One-way and Two- Environmental Politics, 30(7), 1259–1280. https://doi.org/10.1080/
way ANOVA Models (1.0) [Computer software]. 09644016.2021.1892982
Lujala, P., Bezu, S., Kolstad, I., Mahmud, M., & Wiig, A. (2020). How do host–migrant Stürmer, S., Snyder, M., & Omoto, A. M. (2005). Prosocial emotions and helping: The
proximities shape attitudes toward internal climate migrants? Global Environmental moderating role of group membership. Journal of Personality and Social Psychology,
Change, 65, Article 102156. https://doi.org/10.1016/j.gloenvcha.2020.102156 88(3), 532–546. https://doi.org/10.1037/0022-3514.88.3.532
Luong, K. T., Garrett, R. K., & Slater, M. D. (2019). Promoting persuasion with Tajfel, H. (1974). Social identity and intergroup behaviour. Social Science Information, 13
ideologically tailored science messages: A novel approach to research on emphasis (2), 65–93. https://doi.org/10.1177/053901847401300204
framing. Science Communication, 41(4), 488–515. https://doi.org/10.1177/ Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In
1075547019862559 S. Worchel, & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24).
Lustgarten, A. (2020a). The great climate migration has begun. The New York Times. https Nelson Hall.
://www.nytimes.com/interactive/2020/07/23/magazine/climate-migration.html. Tajfel, H., & Turner, J. C. (2004). The social identity theory of intergroup behavior. In
Lustgarten, A. (2020). How climate migration will reshape America. The New York Times. Political psychology. Psychology Press.
https://www.nytimes.com/interactive/2020/09/15/magazine/climate-crisis-mig Trope, Y., & Liberman, N. (2012). Construal level theory. In P. A. M. Van Lange,
ration-america.html. A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology
Lustgarten, A. (2020c). December 16. How Russia wins the climate crisis. The New York (Vol. 1). SAGE Publications Ltd. https://doi.org/10.4135/9781446249215.
Times https://www.nytimes.com/interactive/2020/12/16/magazine/russia-climat United Nations, O. H. C. H. R. (2018). Addressing human rights protection gaps in the context
e-migration-crisis.html. of migration and displacement of persons across international borders resulting from the
Markus, H. R., & Zajonc, R. B. (1985). In E. Aronson, & G. Lindzey (Eds.) (3rd ed., Vol. 1, adverse effects of climate change and supporting the adaptation and mitigation plans of
The handbook of social psychology (pp. 137–230). Random House. developing countries to bridge the protection gaps. Office of the United Nations High
Miratrix, L. W., Campos, L. F., Theodoridis, A. G., Sekhon, J. S., Theodoridis, A. G., & Commissioner for Human Rights (OHCHR).
Campos, L. F. (2018). Worth weighting? How to think about and use weights in United States Government Accountability Office. (2020). Climate change: A climate
survey experiments. Political Analysis, 26(3), 275–291. https://doi.org/10.1017/ migration pilot program could enhance the nation’s resilience and reduce federal fiscal
pan.2018.1 exposure (GAO-20-488). United States Government Accountability Office.
Motta, M., Ralston, R., & Spindel, J. (2021). A call to arms for climate change? How Urban, J., Vačkářová, D., & Badura, T. (2021). Climate adaptation and climate mitigation
military service member concern about climate change can inform effective climate do not undermine each other: A cross-cultural test in four countries. Journal of
communication. Environmental Communication, 15(1), 85–98. https://doi.org/ Environmental Psychology, 77, 101658. https://doi.org/10.1016/j.
10.1080/17524032.2020.1799836 jenvp.2021.101658
Myers, T. A., Nisbet, M. C., Maibach, E. W., & Leiserowitz, A. A. (2012). A public health Weiner, B. (2006). Social motivation, justice, and the moral emotions. Psychology Press.
frame arouses hopeful emotions about climate change. Climatic Change, 113(3–4), https://doi.org/10.4324/9781410615749
1105–1112. https://doi.org/10.1007/s10584-012-0513-6 White House, T. (2021). Report on the impact of climate change on migration. The White
House.

16
K.T. Raimi et al. Journal of Environmental Psychology 93 (2024) 102210

Whitmarsh, L. (2008). Are flood victims more concerned about climate change than Zillmann, D. (1999). Exemplification theory: Judging the whole by some of its parts.
other people? The role of direct experience in risk perception and behavioural Media Psychology, 1(1), 69–94. https://doi.org/10.1207/s1532785xmep0101_5
response. Journal of Risk Research, 11(3), 351–374. https://doi.org/10.1080/ Zillmann, D. (2006). Exemplification effects in the promotion of safety and health.
13669870701552235 Journal of Communication, 56, S221–S237. https://doi.org/10.1111/j.1460-
Younis, M. (2020). Americans want more, not less, immigration for first time. Gallup. https 2466.2006.00291.x
://news.gallup.com/poll/313106/americans-not-less-immigration-first-time.aspx

17