Journal of Environmental Psychology 76 (2021) 101652

Contents lists available at ScienceDirect

Journal of Environmental Psychology

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

Development and validation of a climate change perceptions scale

A.M. van Valkengoed *, L. Steg, G. Perlaviciute
Faculty of Social and Behavioural Sciences, University of Groningen, Grote Kruisstraat 2/1, 9712TS, Groningen, the Netherlands

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

Handling Editor: Sander van der Linden People’s perceptions of climate change have been of great interest for understanding how people respond to
climate change. Yet, studies differ greatly in how they assess climate change perceptions, which makes it difficult
Keywords: to compare and integrate findings. The aim of this paper is to develop and validate a scale to reliably measure
Climate change perceptions people’s climate change perceptions. We define climate change perceptions as people’s perceptions of the reality
Scale validity
and causes of climate change, and the perceived valence, spatial distance, and temporal distance of consequences
Scale reliability
of climate change. To measure these perceptions, an item pool was developed based on a literature search and
Climate change beliefs
expert review. The resulting scale was tested in three empirical studies conducted in the US and the Netherlands,
which supported the factor structure; reliability; convergent, predictive, and discriminant validity; and mea­
surement invariance of the scale. Implications for future research are discussed.

1. Introduction
While most people agree that climate change is real (Steg, 2018),
people may hold different beliefs about the extent to which climate
change is caused by humans and what consequences it will have, where,
and when. Such perceptions of climate change play a role in whether
people support climate policies, and take action to mitigate and/or
adapt to climate change (Brügger, Morton, & Dessai, 2015; Lorenzoni,
Nicholson-Cole, & Whitmarsh, 2007). Many studies have examined
people’ climate change perceptions (e.g., Clayton et al., 2015; Hornsey,
Harris, Bain, & Fielding, 2016; Weber, 2010, 2016). Yet, these studies
vary greatly in which types of climate change perceptions they assessed
and how they assessed them (Motta, Chapman, Stecula, & Haglin, 2019).
This hinders the comparison and integration of results to develop of a
coherent, cumulative understanding of climate change perceptions. To
address this gap in the literature, we aim to develop and validate a scale
to assess people’s perceptions of climate change.
1.1. Theories on climate change perceptions
The first critical step of scale development is to define the constructs
to be assessed (Boateng, Neilands, Frongillo, Melgar-Quiñonez, &
Young, 2018). It has been theorised that three types of climate change
perceptions can be distinguished: beliefs about the reality of climate
change (i.e., the extent to which people believe that climate change is
happening), causes of climate change (i.e., the extent to which people
ascribe human versus natural causes to climate change), and conse­
quences of climate change (i.e., the extent to which people perceive the
consequences of climate change as negative or positive) (Bostrom et al.,
2012; Guy, Kashima, Walker, & O’Neill, 2014; Heath & Gifford, 2006;
Poortinga, Whitmarsh, Steg, Böhm, & Fisher, 2019). This tripartite
framework is based on theorizing in the domain of climate change
denial/scepticism. Based on analyses of climate denial tactics over time,
it has been proposed that the arguments used by interest groups that
deny climate change have developed over time. While denial strategies
would at first focus on outright denial of the reality of climate change,
this position became untenable due to mounting scientific evidence.
Denial strategies then moved towards denying that humans caused
climate change and to downplaying the extent to which climate change
would bring about negative consequences (Björnberg, Karlsson, Gilek, &
Hansson, 2017; Mann, 2014; McCright, 2016; Rahmstorf, 2004; Van
Rensburg, 2015). This theoretical distinction between perceptions
related to reality, causes, and consequences of climate change has been
validated in the literature, with studies showing that the extent to which
extent people perceive climate change as real, human-caused, and as
having negative consequences can be distinguished from each other and
that they contribute unique variance in explaining outcomes such as
mitigation behaviours and support for climate policy (Bostrom et al.,
2012; Ding, Maibach, Zhao, Roser-Renouf, & Leiserowitz, 2011; Guy
et al., 2014; Heath & Gifford, 2006; Milfont, Milojev, Greaves, & Sibley,
2015; Poortinga, Spence, Whitmarsh, Capstick, & Pidgeon, 2011;
Zawadzki, Bouman, Steg, Bojarskich, & Druen, 2020).

* Corresponding author.
E-mail address: a.m.van.valkengoed@rug.nl (A.M. van Valkengoed).

https://doi.org/10.1016/j.jenvp.2021.101652
Received 31 March 2020; Received in revised form 8 April 2021; Accepted 11 July 2021
Available online 12 July 2021
0272-4944/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A.M. van Valkengoed et al.
It has been further theorised that besides the perceived valence of the
consequences of climate change (i.e., positive versus negative), the
perceived proximity of these consequences is another important
dimension of people’s beliefs about climate change (McDonald, Chai, &
Newell, 2015; Spence, Poortinga, & Pidgeon, 2012). Specifically, on the
basis of Construal Level Theory (Trope & Liberman, 2010), it has been
theorised that two key dimensions of psychological distance are relevant
for climate change perceptions: whether people perceive consequences
of climate change as occurring nearby or far away (i.e., perceived spatial
distance), and as occurring soon or in the distant future (i.e., temporal
distance) (McDonald et al., 2015). Theoretically, it has been argued that
these dimensions are distinct from the perceived negative consequences
of climate change, as people who think that climate change is real,
human caused, and that it will have negative consequences, may still not
engage in climate change mitigation and adaptation if they believe that
these consequences will occur far away in time and/or space (e.g., Jones,
Hine, & Marks, 2017; van der Linden, Maibach, & Leiserowitz, 2015).
Studies have confirmed this theoretical framework, showing that per­
ceptions of the spatial and temporal distance of consequences of climate
change can be distinguished from other climate change perceptions and
that they add unique variance in explaining relevant outcomes of in­
terest (Bloodhart, Maibach, Myers, & Zhao, 2015; Hine et al., 2013;
Rosentrater et al., 2013; Singh, Zwickle, Bruskotter, & Wilson, 2017;
Spence et al., 2012). In line with this theorizing, we propose that the
perceived consequences of climate change encompasses the perceived
valence, spatial distance, and temporal distance. Importantly, a full
model comparing all five climate change perceptions as distinct factors
has not yet been tested in the literature, and represents a unique
contribution of this paper to the literature.
1.2. Developing and validating the climate change perceptions scale
To develop and validate a scale to measure these five types of per­
ceptions of climate change (namely reality, causes, valence of conse­
quences, spatial distance of consequences, and temporal distance of
consequences), we first establish an item pool based on previous studies
that reflect our definition of the five types of climate change perceptions.
Next, we ask experts on climate change perceptions to evaluate the
relevance and quality of the items for measuring the climate change
perceptions, in order to select the most appropriate items. Finally, we
test and validate the climate change perceptions scale in three studies.
Specifically, we assess the factor structure of the scale using confirma­
tory factor analysis, as we have a clear proposed factor structure based in
theoretical understanding of the construct of climate change percep­
tions. We assess the reliability of the subscales using Cronbach’s alpha.
We also assess the convergent, predictive, and discriminant validity of
the scale as well as measurement invariance.
Convergent validity is established when a scale correlates with theo­
retically related constructs (Walton & Jones, 2018). We will test the
correlations between climate change perceptions and three theoretically
related constructs. First, the New Ecological Paradigm (NEP) reflects
people’s views about human-environment relationships, including the
perceived fragility of nature and vulnerability to human influences
(Dunlap, Van Liere, Mertig, & Jones, 2000). The more strongly people
endorse the NEP, the more they think that nature is negatively influ­
enced by humans now and on a global scale, and that humans do not
have the right to rule over nature (Dunlap et al., 2000). We expect that
the more strongly people endorse the NEP, the more they think that
climate change is real, human caused, and has negative consequences
that occur already now and close by (Brenkert-Smith, Meldrum, &
Champ, 2015; O’Connor, Bard, & Fisher, 1999).
Second, we expect that climate change perceptions are related to
negative affect towards climate change, such as worry, concern or fear
(van der Linden, 2017). We expect negative affect to be stronger the
more people perceive climate change as real, human caused, and having
negative consequences that occur already now and close by (Berry &
2
Journal of Environmental Psychology 76 (2021) 101652
Peel, 2015; Kittipongvises & Mino, 2013).
Third, we expect that climate change perceptions are associated with
climate change risk perception, namely the extent to which people think
that climate change threatens valued entities, for example close others,
nature, or the national economy (cf. Kellstedt, Zahran, & Vedlitz, 2008;
Paek & Hove, 2017). Please note that climate change risk perception
differs from the perceived valence of consequences of climate change as
it pertains to impacts on a specific valued entity. We expect that people
will perceive more risks from climate change, the more they perceive
climate change as real and having negative consequences that occur
already now and close by (Li, Juhász-Horváth, Harrison, Pintér, &
Rounsevell, 2017; Safi, Smith, & Liu, 2012). Similarly, we expect that
perceived risks from climate change will be higher when people perceive
climate change as caused by humans, as people seem to perceive more
risks from a hazard (e.g., an oil spill in the sea) when it is caused by
humans rather than by nature (Siegrist & Sütterlin, 2014).
The predictive validity of a scale is established by examining whether
the scale predicts outcomes of interest (Rosenthal & Rosnow, 2007). We
expect climate change perceptions to predict climate change mitigation
and adaptation behaviours. Mitigation behaviours are actions aimed at
reducing climate change (e.g., saving electricity, reducing car use),
while adaptation behaviours are actions aimed at protecting oneself
against the negative impacts of climate change (e.g., purchasing flood
insurance, preparing an emergency kit; Brügger et al., 2015). Besides
individual behaviours, policy measures are important to effectively
respond to climate change (Dietz, Dan, & Shwom, 2007; Shwom, Bid­
well, Dan, & Dietz, 2010). We expect that the more people perceive that
climate change is real, human caused, and has negative consequences
that occur already now and close by, the more they will be likely to
undertake mitigation and adaptation behaviours and to support miti­
gation and adaptation policies. Importantly, it has been theorised that
particularly perceptions of temporal and spatial distance could be rele­
vant predictors of climate action: As previously mentioned, even if
people think that climate change is real, human caused, and that it will
have negative consequences, they may not engage in climate change
mitigation and adaptation if they believe that these consequences will
occur far away in time and/or space (e.g., Jones et al., 2017; van der
Linden et al., 2015). We will therefore examine whether perceptions of
temporal and spatial distance correlate more strongly with climate ac­
tion and policy support compared to perceptions of the reality, causes, or
valence of consequences of climate change.
Discriminant validity is established by examining whether scale scores
differ across groups for which differences may be expected based on
theory (Hattie & Cooksey, 1984). Women are more likely to believe in
the reality and anthropogenic causes of climate change than men
(McCright, 2010). A possible explanation is that being concerned and
caring towards the environment is more closely aligned with the societal
gender role for women than for men (McCright, 2010). Therefore, we
expect that women will think more than men that climate change is real
and human caused, and perceive the consequences of climate change as
more negative and proximate in time and space.
Furthermore, Democratic voters in the US are more likely to endorse
the reality and anthropogenic causes of climate change than Republican
voters (McCright & Dunlap, 2011). One proposed reason for why Re­
publicans believe less in anthropocentric climate change is because they
disagree with mitigation measures, such as increasing taxes on gasoline
or regulating industries’ CO2 emissions, and are therefore motivated to
downplay climate change altogether (McCright & Dunlap, 2011).
Following this reasoning, we expect that Republicans will less strongly
perceive climate change as real and human caused, and perceive the
consequences of climate change as less negative and farther away in
time and space, compared to Democrats.
To establish measurement invariance, we will examine whether the
factor structure, factor loadings, intercepts, and residuals of the scale are
similar between men and women, and between Republicans and Dem­
ocrats (Fischer & Karl, 2019). While the mean scores on these scales may
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

differ between men and women and Republicans and Democrats, Table 1
respectively, we expect that the factor structure, factor loadings, in­ Items selected in the expert review and results of the oblique multiple group
tercepts, and residuals of the scale will be similar across these groups. method in Study 1.
Moreover, we will assess measurement invariance between respondents Reality Causes Consequences Spatial Temporal
from different cultures, specifically the US and the Netherlands, again Reality
expecting no differences in factor structure, factor loadings, intercepts, 1. I believe that .935 .699 .708 -.695 -.670
or residuals between these groups. climate change is
real.
2. Climate change is .942 .686 .750 -.698 -.633
1.3. Item selection NOT occurring. (r)
3. The world’s .916 .668 .701 -.692 -.619
A literature search was conducted on the terms “climate change climate is
changing.
perceptions” and “climate change beliefs” in PsycINFO, which resulted
4. I do NOT believe .930 .693 .737 -.673 -.648
in a database of 234 items from 54 papers. Of these, we selected 67 items that climate
by removing highly similar items and items that did not match our change is real. (r)
definitions of the perceived reality, causes, and consequences (i.e., 5. Climate change is .929 .690 .699 -.707 -.670
valence, spatial distance, temporal distance) of climate change.1 happening.

Following best practices in scale development (Boateng et al., 2018),

Causes
experts (n = 14) on the topic of climate change perceptions were asked 1. Human activities .699 .892 .736 -.660 -.647
to rate the items based on the quality and the relevance to the type of are a major cause
climate change perception that these items were supposed to assess of climate change.
(definitions of each type of climate change perceptions were provided), 2. Climate change is .591 .821 .713 -.628 -.644
mainly due to
on a scale from 1 (‘Terrible’) to 5 (‘Excellent‘). Based on the experts’
natural causes. (r)
ratings, 5 items for each type of climate change perception were 3. Climate change is .669 .894 .730 -.643 -.664
selected, resulting in 25 items in total (see Table 1). For further details mostly caused by
on the literature search, item pool generation, and the expert review, see human activity.
4. The main causes .693 .909 .737 -.640 -.661
Supplemental Material. We tested the reliability and validity of the
of climate change
resulting scale in three studies: two in the US, and one in the are human
Netherlands. Institutional ethical approval was obtained for all studies, activities.
as well as for the expert review. 5. Climate change is .611 .739 .698 -.614 -.667
caused entirely by
natural processes.
2. Study 1 (r)

2.1. Method Valence of consequences

1. Overall, climate .690 .678 .772 -.620 -.577
2.1.1. Participants and procedure change will bring
more negative
In this study (and Studies 2 and 3), we aimed for a sample size of at than positive
least 300 participants, based on sample size recommendations for ana­ consequences to
lysing factor structure (Tabachnick & Fidell, 2013). Using Amazon the world.
MTurk, 401 US participants were recruited. After a data quality check 2. Climate change .478 .600 .689 -.540 -.523
will mostly have
(see below), 353 responses remained. The study was advertised on
positive
Amazon MTurk as a short questionnaire on current affairs. Those willing consequences. (r)
to participate were provided with a link to the online questionnaire 3. Climate change .702 .731 .784 -.702 -.655
programme Qualtrics, where the study was completed. The participants will bring about
first read the description of the aims of the research and provided their serious negative
consequences.
informed consent. Next, they evaluated the climate change perceptions
4. The consequences .764 .774 .786 -.724 -.682
items. Participants then completed the scales included to validate the of climate change
climate change perceptions scale in a random order. will be very
Compared to US census data (United States Census Bureau, 2018), serious.
5. There will be .515 .564 .701 -.556 -.556
our sample consisted of more men and white respondents, and relatively
mostly positive
fewer African American and Latino/a respondents (see Table 2). The consequences of
general education level of our sample was somewhat higher (more climate change.
participants had a high school degree or higher), while the median in­ (r)
come level was somewhat lower than the US census.
Spatial distance
1. My local area will -.683 -.638 -.667 .836 .690
2.1.2. Materials be influenced by
Climate change perceptions. Participants responded to the 25 climate climate change.
change perceptions items (see Table 1) on a seven-point Likert-scale (r)
from ‘strongly disagree’ (1) to ‘strongly agree’ (7). The perceived reality 2. Climate change -.485 -.529 -.575 .664 .626
only influences
of climate change items were presented first (in a random order),
locations far away
from me.
3. The region where -.696 -.631 -.647 .794 .658
1 I live will
For example, questions pertaining to perceived climate beliefs among sci­
entists (e.g., ‘Most scientists agree that humans are causing climate change’) or experience the
consequences of
opinions on the media’s reporting on climate change (e.g., ‘The media is
overstating the evidence about global warming’) were excluded as these do not (continued on next page)
fit our definitions of any of the five types of climate change perceptions.

3
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 1 (continued ) exists would not be able to answer the questions about the different
Reality Causes Consequences Spatial Temporal types of climate change perceptions.2
New Ecological Paradigm (NEP). We used the 6-item version of the
climate change.
(r)
NEP (Whitmarsh, 2008, 2009; see Table 3). Responses were given on a
4. The consequences -.496 -.491 -.589 .648 .609 7-point Likert scale (1 = ‘strongly disagree’, 7 = ‘strongly agree’). We
of climate change averaged the scores on these six items (α = .83, M = 5.24, SD = 1.15); a
will only take higher score indicates that people agreed more that nature is fragile and
place in distant
negatively influenced by humans.
locations.
5. Climate change -.675 -.659 -.671 .789 .679 Negative affect related to climate change was measured with the
will also influence following three items: ‘I worry about climate change’, ‘I am concerned
the place where I about climate change’ and ‘When I think about climate change, I get
live. (r) distressed’ (adapted from van der Linden, 2017). The items were
Temporal distance
answered using a 7-point Likert scale (1 = ‘strongly disagree’, 7 =
1. The consequences -.763 -.734 -.729 .725 .675 ‘strongly agree’). We averaged the scores on these three items (α = .93,
of climate change M = 4.95, SD = 1.72); a higher score indicates more negative affect
are visible now. related to climate change.
(r)
Climate change risk perception. Participants indicated to what extent
2. It will be a long -.585 -.626 -.653 .687 .773
time before the they think that climate change poses risks for them and their family, the
consequences of national economy, the environment, and public health, respectively
climate change (adapted from Kellstedt et al., 2008). Responses were given on a 7-point
are felt. Likert scale (1 = ‘strongly disagree’, 7 = ‘strongly agree’). We averaged
3. Only future -.232 -.299 -.266 .340 .403
generations will
the scores on these four items (α = .96, M = 5.45, SD = 1.55); a higher
experience the score indicates higher perceived risks from climate change.
consequences of Climate change mitigation behaviour. We included 8 items from the
climate change. General Ecological Behaviour scale (Arnold, Kibbe, Hartig, & Kaiser,
4. The consequences -.533 -.591 -.602 .662 .774
2018; Kaiser & Wilson, 2004) reflecting both environmentally friendly
of climate change
will only be and harmful behaviours. Two items were added to represent behaviour
experienced in the with a relatively large climate impact, namely eating meat and
far future. refraining from flying. Items were answered using a 7-point Likert scale
5. The effects of -.679 -.706 -.661 .732 .674 (1 = ‘never’, 7 = ‘always’). We averaged the scores on the items (after
climate change
reverse-coding environmentally harmful behaviour; α = .63, M = 4.61,
will be felt very
soon. (r) SD = 0.78); a higher score indicates that people engage more frequently
in mitigation behaviours.
Coefficients in bold represent the highest overall total-item correlations. Bold
Climate change adaptation behaviour. A set of 9 items was compiled to
face indicates that the assignment of the item was in line with theory, underline
assess adaptive behaviour that may protect people from different types
indicates that the item correlated more strongly with a different subscale than
the subscale to which it was assigned based on theory. of climate-related hazards (see Table 3). Three items were adapted from
Consequences = valence of consequences, Spatial = spatial distance, Temporal the Impact Oriented Adaptive Responses Scale (Elrick-Barr, Smith,
= temporal distance. Preston, Thomsen, & Baum, 2016), whereas the other items were
developed based on an overview of household adaptive behaviours
(Wamsler & Brink, 2014). One additional item assessed whether people
Table 2 check weather forecasts. Responses were given on an ordinal scale (0 =
Comparison of the characteristics of the Study 1 and Study 2 sample to the US ‘No’, 1 = ‘No, but I am planning to do this’, 2 = ‘Yes’). An ordinal
public. response format was selected because most adaptive behaviours are
Study 1 (n = Study 2 (n = US Census (n = performed only once rather than on a daily basis. The items were
353) 345) 327,167,434) combined into one reliable scale (α = .87) by adding up the scores on
Median age 35 36 37.9a each item (M = 8.90, SD = 5.46); a higher score indicates that people
% Women 45.0 53.3 50.8 have implemented more adaptation measures, and/or have the intention
% White 74.4 71.3 60.4 to do so.
% Black/African 9.4 12.5 13.4
American
Climate change policy support. Policy support was assessed with 10
% Hispanic 4.3 5.8 18.3 items, five for mitigation policies and five for adaptation policies (see
% American Indian 1.1 0.03 1.3 Table 3). All policy support items were answered using a 7-point Likert
% Asian or Pacific 5.7 5.2 6.1 scale (1 = ‘strongly oppose’, 7 = ‘strongly support’). Two mitigation
Islander
policy items were adapted from the rotating module on public attitudes
% Multi-ethnic 4.3 3.8 2.7
% High school 99.4% 99.7% 87% to climate change of the eighth round of the European Social Survey
graduate or (2016). Two items were adapted from the Six America’s survey (Lei­
higher serowitz, Maibach, & Roser-Renouf, 2009). The other mitigation policy
Median household $35,000–49,999 $35,000–49,999 $55,332 item was developed for this study. Two adaption policies items were
income
adapted from previous studies (Hagen, Middel, & Pijawka, 2016; Singh
a
This number was obtained from https://www.statista.com/statistics/ et al., 2017), while the other three additional adaptation policy items
241494/median-age-of-the-us-population/. All other data were obtained were developed for this study. We averaged the scores of the evaluations
through the United States Census Bureau (2018). of the five mitigation policies (α = .91, M = 5.50, SD = 1.39) and
adaptation policies (α = .91, M = 5.17, SD = 1.42); a higher score
followed by all other climate change perceptions items (in a random
order). For all items except those measuring the perceived reality of
climate change, the option ‘I do not believe climate change exists’ (0) 2
In case the participants chose ‘0’, the response was coded as ‘missing’, and
was added because participants who do not believe that climate change subsequently excluded from the analyses.

4
A.M. van Valkengoed et al.
Table 3
Measures in Study 1 and Study 2.
New Ecological Paradigm (From Whitmarsh, 2008)
1. Humans have the right to modify the natural environment to suit their needs *
2. Humans are severely abusing the environment.
3. Plants and animals have as much right as humans to exist.
4. The balance of nature is strong enough to cope with the impact of modern industrial
nations. *
5. Humans were meant to rule over the rest of nature. *
6. The balance of nature is very delicate and easily upset.
Climate change mitigation behaviour
1. Riding a bicycle or taking public transportation rather than taking the car. (from
Arnold et al., 2018)
2. Recycling paper. (from Arnold et al., 2018)
3. Buying seasonal produce. (from Arnold et al., 2018)
4. Waiting until I have a full load before doing laundry. (from Arnold et al., 2018)
5. Showering for longer than 5 minutes * (from Arnold et al., 2018)
6. Taking the train or car rather than a plane when possible. (This Study)
7. Eating meat with my dinner. * (This Study)
8. Leaving my cell phone charger plugged in when not in use. * (from Arnold et al.,
2018)
9. Turning the lights off when I am the last person to leave a room. (from Arnold et al.,
2018)
10. Purchasing an energy efficient device when I purchase a new electronic appliance.
(from Arnold et al., 2018)
Climate change adaptation behaviour
1. Preparing a household emergency kit, containing for example a flashlight, a radio,
emergency blankets, first aid kit. (from Elrick-Barr et al., 2016)
2. Adjusting my home to better withstand natural hazards, for example installing wind
shutters against hurricanes or painting my house in a lighter color to reduce heating.
(from Elrick-Barr et al., 2016)
3. Preparing an evacuation plan in the event of a natural hazard. (from Elrick-Barr
et al., 2016)
4. Storing bottled water and canned food in case a natural hazard occurs. (based on
Wamsler & Brink, 2014)
5. Purchasing insurance against losses from natural hazards. (based on Wamsler &
Brink, 2014)
6. Looking up information about whether my house is at risk of natural hazards. (based
on Wamsler & Brink, 2014)
7. Looking up information about what I can do to prepare for natural hazards. (based
on Wamsler & Brink, 2014)
8. Making sure that my home is maintained well to avoid damage from natural
hazards, for example cleaning gutters and repairing the roof. (based on Wamsler &
Brink, 2014)
9. Actively checking and following weather forecasts. (This Study)
Policy support (mitigation)
1. Using public money to subsidise renewable energy such as wind and solar power. (
European Social Survey, 2016)
2. Increasing taxes on fossil fuels, such as oil, gas, and coal. (European Social Survey,
2016)
3. Providing tax rebates for energy efficient vehicles and solar panels. (Leiserowitz
et al., 2009)
4. Requiring energy companies to produce at least 35% of their electricity from wind,
solar, or other renewable energy sources. (Leiserowitz et al., 2009)
5. Setting national targets to reduce carbon emissions. (This Study)
Policy support (adaptation)
1. Implementing heat warning systems in every city so that people can better protect
themselves against heatwaves caused by climate change. (This Study)
2. Investing public money to make vital infrastructure (for example, energy utilities,
power lines, cell towers) more resistant to climate change risks. (This Study)
3. Launching an awareness campaign that provides people with information on how to
prepare for climate change risks. (This Study)
4. Requiring cities to invest in protection against increases in flooding due to climate
change. (from Hagen et al., 2016)
5. Increasing funding for farmers to help them prepare for increases in flooding or
drought. (from Singh et al., 2017)
* = reverse-coded item.
indicates more support for mitigation/adaptation policies.
Demographic information. Participants indicated their gender (‘male’,
‘female’, ‘other’), voting behaviour (‘Democrat’, ‘Independent’,
‘Republican’, ‘Other’, ‘No party/not interested in politics’), age,
5
Journal of Environmental Psychology 76 (2021) 101652
ethnicity, income, household composition, and education level.
2.1.3. Data quality check
To secure data quality, the data were cleaned based on the following
a priori established criteria. First, as it would be unrealistic to complete
the full questionnaire within 3 minutes, we removed respondents that
completed the questionnaire within 3 minutes (n = 11). Second, we
removed responses of participants who inconsistently identified as a
climate change denier (i.e., selecting ‘I do not believe climate change
exists’ on some items but responding to other climate change perception
items using the ‘agree-disagree’ scale) (n = 19). Third, participants were
removed that ‘straight-lined’ the reality items (n = 14). Straight-lining
was defined as providing responses that were consistently above or
below the scale midpoint (i.e., only providing response ‘5’, ‘6’, ‘7’ or ‘1’,
‘2’, ‘3’) on at least four (out of 5) of the items measuring the perceived
reality of climate change. Because the reality items were mutually
exclusive (e.g., ‘I believe that climate change is real’ and ‘I do NOT
believe that climate change is real’), straight-lining these items is evi­
dence of content non-responsivity (Meade & Craig, 2012). Fourth, only
the first responses from an IP address were included and subsequent
responses from the same IP address were excluded (n = 4). One pair of
responses from the same IP address was kept because the starting and
completion times indicated that the questionnaires were completed
simultaneously, which makes it unlikely that they were completed by
the same person. Overall, 48 responses (11.97% of the data) were
excluded.
2.1.4. Data analyses
First, we conduct confirmatory factor analysis to examine whether
the theoretically proposed factor structure is supported by the data. We
use the oblique multiple group method, which is an established
approach to confirmatory factor analysis, to assess whether individual
items correlate strongest with the subscale to which they are assigned
based on theory (Stuive, Kiers, & Timmerman, 2009)3,4.We then
assessed the reliability of the final subscales by calculating Cronbach’s
alpha. The convergent and predictive validity of the final scales was
determined by correlating the climate change perceptions with the
relevant scales described above. The discriminant validity of the scale
was examined by investigating whether men and women and Democrats
and Republicans, respectively, differ in their scores on the climate
change perceptions. Finally, measurement invariance was assessed by
testing whether the factor structure, factor loadings, intercepts, and
residuals of the scale are the same for men and women, and Democrats
and Republicans, respectively.
2.2. Results
2.2.1. Constructing the final scale
Table 1 presents the results of the oblique multiple group analysis, in
which each item in a scale is correlated with the subscale it is assigned to
and all other subscales, corrected for self-correlation (see Bouman, Steg,
& Kiers, 2018). The factor structure of the scale is confirmed if the items
correlate more strongly with the subscale to which they are assigned
based on theory. The items measuring the perceptions of the reality and
the causes of climate change, and the valence and spatial distance of
3
We use this method here as its results are more easily interpretable for a
general audience compared to the structural equation approach to factor
analysis (Stuive et al., 2009). We however employ the structural equation
approach later on to assess the measurement invariance of the scale, as this
approach offers a formal test to examine whether the factor structure, factor
loadings, intercepts, and residuals differ across groups.
4
Upon a reviewer’s request, we also performed exploratory factor analysis on
these data. The results from this analysis largely aligned with the confirmatory
factor analysis; the results are presented in the Supplemental Material.
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

consequences of climate change all correlated most strongly with the Table 5
subscales to which they were assigned on the basis of our theoretical Pearson’s correlation coefficients of the climate change perceptions with vari­
framework, confirming the factor structure of the scale. To form the final ables to test convergent and predictive validity in Study 1.
subscales, three items with the highest item-total correlations for each Reality Causes Consequences Spatial Temporal
type of climate change perceptions were selected (see Table 3). Two of Reality .70 .79 -.71 -.59
the items measuring the perceived temporal distance of consequences of Cause .79 -.68 -.56
climate change (i.e. ‘The consequences of climate change are visible Consequence -.76 -.61
now’ and ‘The effects of climate change will be felt very soon’) corre­ Spatial .62
lated more strongly with the items measuring the perceptions of the
Convergent validity
reality and the spatial distance of consequences of climate change,
NEP .54 .66 .71 -.57 -.63
respectively. These items were therefore excluded from the final scale. Negative affect .71 .72 .75 -.66 -.56
Another item measuring the temporal distance of consequences of Risk Perception .81 .81 .87 -.78 -.63
climate change (i.e. ‘Only future generations will experience the con­
sequences of climate change’) showed a low item-total correlation with Predictive validity
Mitigation .27 .29 .34 -.32 -.33
the subscale to which it was assigned (r = .40) and was also excluded
behaviour
from the final scale. Consequently, two items were included in the final Adaptation .19 .16 .22 -.23 -.16
subscale to measure perceived temporal distance of consequences of behaviour
climate change. The final climate change perceptions scale and the Mitigation policy .66 .73 .73 -.63 -.55
descriptive statistics for each item are presented in Table 4. support
Adaptation policy .58 .58 .61 -.51 -.41
The reliabilities of the subscales measuring each type of climate support
change perception were excellent as they were all higher than .80 (see
Table 4) (Panayides, 2013). The different types of climate change per­ Note: all correlations are significant at p < .01. NEP = New Ecological Paradigm,
Consequences = valence of consequences, Spatial = spatial distance, Temporal
ceptions are all significantly positively correlated (see Table 5). There
= temporal distance.
were especially strong correlations between perceiving climate change

as real and having negative consequences (r = .79), and between

Table 4
perceiving climate change as human caused and having negative con­
Items of the final scale and descriptive statistics in Study 1.
sequences (r = .79). Still, the five-factors model fit the data significantly
Mean SD Min Max N α better than models that collapsed the strongly correlated factors into one
Reality .98 factor, or a unidimensional model (see Supplemental Material).
1. I believe that climate change is 5.98 1.42 1 7 353
real.
2.2.2. Convergent validity
2. Climate change is NOT occurring. 5.99 1.43 1 7 353
(r) Table 5 shows that the correlations between the different types of
3. I do NOT believe that climate 5.97 1.51 1 7 353 climate change perceptions and the three theoretically related con­
change is real. (r) structs (i.e., NEP, negative affect related to climate change, and climate
change risk perception) were all significant and in the expected direc­
Causes .97
tion, ranging from .54 to .87 and − .56 to − .78, corresponding to a
1. Human activities are a major 5.59 1.65 1 7 341
cause of climate change. medium to large effect size (Ferguson, 2009). The more people
2. Climate change is mostly caused 5.37 1.71 1 7 341 perceived climate change as real, human caused, and having negative
by human activity. consequence that occur already now and close by, the more they
3. The main causes of climate 5.44 1.67 1 7 341 endorsed the NEP, experienced negative affect related to climate
change are human activities.
change, and perceived risks from climate change. We therefore conclude
Valence of consequences .90
that the convergent validity of the scale was good. The correlations
1. Overall, climate change will 5.78 1.46 1 7 341 between the climate change perceptions and climate change risk
bring more negative than positive perception were very strong, especially for the perceived valence of
consequences to the world. consequences of climate change (r = .87), suggesting that these con­
2. Climate change will bring about 5.74 1.42 1 7 341
structs may be difficult to distinguish empirically.
serious negative consequences.
3. The consequences of climate 5.73 1.48 1 7 341
change will be very serious. 2.2.3. Predictive validity
Correlations between the climate change perceptions and mitigation
Spatial distance .93 behaviour, adaptation behaviour, and mitigation and adaptation policy
1. My local area will be influenced 2.51 1.44 1 7 341
support were all significant and in the expected direction, ranging from
by climate change. (r)
2. The region where I live will 2.47 1.45 1 7 341
.16 to .73 and − .16 to − .63 (see Table 5). The more people perceived
experience the consequences of climate change as real, human caused, and having negative conse­
climate change. (r) quences that occur already now and close by, the more they engaged in
3. Climate change will also 2.53 1.43 1 7 341 mitigation and adaptation behaviours, and the more they supported
influence the place where I live.
mitigation and adaptation policies. The predictive validity of the scale
(r)
ranged from a small-to-medium effect size to a strong effect, and most
Temporal distance .89 correlations were around or larger than (− ).20, which is a recommended
1. It will be a long time before the 2.63 1.67 1 7 341 guideline for practical significance (Ferguson, 2009). Especially corre­
consequences of climate change lations between climate change perceptions and mitigation and adap­
are felt.
tation policy support were strong. Hence, the predictive validity of the
2. The consequences of climate 2.52 1.59 1 7 341
change will only be experienced scale was excellent for policy support, and good for adaptation and
in the far future. mitigation behaviours.
Note: N differs between Reality and other subscales as climate change deniers
are excluded from other subscales.

6
A.M. van Valkengoed et al.
2.2.4. Discriminant validity
We compared climate change perceptions between men and women
and Democrats and Republicans using non-parametric MANOVA with
the R package ‘npmv’ (Ellis, Burchett, Harrar, & Bathke, 2017), because
the data were not normally distributed. We found that there was no
significant difference between men (n = 193) and women (n = 159)5 in
any of the climate change perceptions (F(1.55, 518.38) = 1.48, p = .23,
partial η2 = 0.004). Climate change perceptions of participants who
indicated to vote Republican (n = 77) and Democrat (n = 171)6 did
significantly differ (F(1.77, 351.31) = 59.55, p < .01, partial η2 = 0.23).
Post hoc Mann-Whitney U comparisons (one-sided) with
Bonferroni-corrected significance levels indicate that, as expected, Re­
publicans believed less strongly in the reality of climate change,
perceived climate change as less caused by humans, and perceived the
consequences of climate change as less negative and less close in time
and space than Democrats (see Table 6). Therefore, the scale distin­
guished between Democrats and Republicans, but not between men and
women.
2.2.5. Measurement invariance
We assessed whether the factor structure, factor loadings, intercepts,
and residuals of the scale are similar across the proposed subgroups
(Fischer & Karl, 2019). A CFA using the confirmatory common factor
method was conducted in R (version 4.0.3., R Core Team, 2020) using
the Lavaan package (version 0.6–7, Rosseel, 2012). The fit of the factor
structure of the scale was excellent if men and women were considered
separately (omnibus Satorra-Bentler scaled χ 2(134) = 167.84, χ 2/df =
1.25, p = .03, scaled CFI = 0.987, scaled RMSEA = 0.039, 90% CI
[0.021, 0.053], SRMR = 0.025).7 There was no evidence of measure­
ment non-invariance across factor loadings (χ 2(9) = 5.18, p = .82), in­
tercepts (χ 2(9) = 5.41, p = .80), or residuals (χ 2(14) = 16.47, p = .29)
between men and women.
Similarly, the fit of the factor structure of the scale was excellent if
Democrats and Republicans were considered separately (omnibus
Satorra-Bentler scaled χ 2 (134) = 160.21, χ 2/df = 1.20, p = .06, scaled
CFI = 0.978, scaled RMSEA = 0.040, 90% CI [0.013, 0.058], SRMR =
0.036). There was some evidence of measurement non-invariance of the
factor loadings across Democrats and Republicans (χ 2(9) = 17.50, p =
.04). Yet, the change in model fit was marginal (ΔCFI = − 0.007
ΔRMSEA = 0.005). The model constraining factor loadings also no
longer differed significantly from an unconstrained model after the
constraints on the factor loadings for one item (“The consequences of
climate change will only be experienced in the far future”) were
removed (χ 2(8) = 14.13, p = .08), showing that the scale was partially
invariant. There was no further evidence of measurement non-
invariance across the intercepts (χ 2(9) = 14.97, p = .09) or residuals
(χ 2(14) = 12.37, p = .58) of the scale between Republicans and
Democrats.8
2.3. Discussion
In Study 1, we conducted confirmatory factor analysis using the
5
The responses of one participant who selected ‘other’ for the gender ques­
tion were excluded for this analysis.
6
Participants that reported to vote for Independent’ or ‘Other’, or who
selected ‘No party/not interested in politics’ were not considered in these
analyses.
7
Generally accepted guidelines for good model fit are as follows: ratio χ2/df
≤ 3, Comparative Fit Index (CFI) ≥ 0.95, Root Mean Square Error of Approx­
imation (RMSEA) ≤ 0.06, Standardized Root Mean Square Residual (SRMR) ≤
0.08 (Hu & Bentler, 1999).
8
Upon a reviewer’s request, we also tested whether the scale could be
considered measurement invariant across age groups. There was no evidence of
measurement non-invariance across age groups in Study 1 (see Supplemental
Material).
7
Journal of Environmental Psychology 76 (2021) 101652
oblique multiple group method and selected three items per each type of
climate change perceptions (2 items for the perceived temporal dis­
tance) for the final scale. The final scale had good reliability, with the
reliabilities of the subscales measuring the perceptions of the reality and
causes of climate change even approaching unity, which indicates that
these dimensions may be over identified and that some items may be
redundant (Streiner, 2003). The scale correlated with the theoretically
related variables in the expected direction, confirming its convergent
validity. The scale was also associated in the expected direction with
adaptation behaviour, mitigation behaviour, and support for mitigation
and adaptation policies, and differentiated between Democratic and
Republican voters (but not men and women), confirming its predictive
and discriminant validity, respectively. Contrary to expectation, the
perceptions of the spatial and temporal distance of consequences of
climate change did not correlate notably stronger with any of the out­
comes compared to three core perceptions of reality, causes, and valence
of consequences. Lastly, the factor structure, factor loadings, intercepts,
and residuals of the scale did not differ across men and women, and
showed only slightly higher factor loadings for Republicans compared to
Democrats.
It is noteworthy that for measuring the perceptions of the causes of
climate change and the valence and spatial distance of its consequences,
items got selected that measure a single end of each dimension, namely
perceived human causes, negative consequences, and spatially close
consequences, respectively (see Table 4). The items measuring the
opposite ends of these dimensions, namely perceived natural causes,
positive consequences, and spatially distant consequences of climate
change had lower factor loadings, and were therefore not selected for
the final scale. However, the fact that these items correlated less strongly
with the respective subscales may not necessarily mean they do not
measure these perceptions well, but rather that they measure a different
dimension of these perceptions. Notably, this may signal that percep­
tions of the causes, valence of consequences, and spatial distance of
consequences of climate change may be two-dimensional constructs. For
example, people may at the same time think that climate change has
positive and negative consequences, natural and human causes, and
spatially close and distant consequences. This does not apply to the
perceived reality of climate change, as agreeing that climate change is
real excludes agreeing that climate change is not real, and vice versa. For
the perceived temporal distance of consequences, only one dimension
seems relevant, namely at which point in time people think that the
consequences of climate change will begin. Agreeing that the conse­
quences will begin later excludes thinking that the consequences begin
already now, and vice versa. Study 2 aims to test whether measuring the
two-dimensional nature of the perceived causes, valence of conse­
quences and spatial distance of consequences of climate change may be a
more accurate representation of people’s climate change perceptions,
and could contribute to the predictive and discriminant validity of the
scale.
3. Study 2
In Study 2, we aim to replicate and extend the findings from Study 1.
We use the items selected in Study 1 to measure the perceived reality,
human causes, negative consequences, and spatial proximity and tem­
poral distance of consequences of climate change. Additionally, we
examine whether conceptualizing the perceptions of the causes, valence
of consequences, and spatial distance of consequences of climate change
as encompassing two dimensions rather than one can add to the pre­
dictive and discriminant validity of the climate perceptions scale.
3.1. Method
3.1.1. Participants and procedure
Using Amazon MTurk, 402 American participants were recruited.
The procedure of the study was identical to Study 1. After data quality
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 6
Post-hoc Mann Whitney U tests to test discriminant validity in Study 1, Study 2, and Study 3.
Men Women Democrats Republicans

Mean SD Mean SD U p 95% CI Mean SD Mean SD U p 95% CI

Study 1
Reality 5.99 1.33 5.96 1.53 14,858 .30 [-0.27, 0.34] 6.56 0.78 4.94 1.75 10,616 <.01 [1.21, 2.03]
Causes 5.41 1.66 5.53 1.58 13,720 .27 [-0.48, 0.22] 6.09 1.02 4.21 1.99 9521 <.01 [1.38, 2.37]
Consequences 5.71 1.29 5.79 1.37 13,508 .20 [-0.36, 0.21] 6.23 0.96 4.71 1.54 9727 <.01 [1.13, 1.91]
Spatial distance 2.59 1.37 2.41 1.32 15,435 .09 [-0.11, 0.47] 2.04 1.05 3.48 1.64 2834 <.01 [-1.86, − 1.02]
Temporal distance 2.68 1.51 2.46 1.59 15,898 .03 [-0.11, 0.55] 2.11 1.28 3.61 1.76 2995 <.01 [-1.97, − 1.05]

Study 2
Reality 6.23 1.16 6.09 1.38 15,088 .67 [-0.13, 0.41] 6.51 0.88 5.14 1.78 9661 <.0045 [0.93, 1.82]
Human Causes 5.80 1.37 5.76 1.30 14,232 .73 [-0.24, 0.34] 6.17 0.97 4.75 1.82 8230 <.0045 [0.93, 1.90]
Natural Causes 3.67 1.75 3.47 1.60 14,516 .18 [-0.16, 0.57] 3.02 1.54 4.76 1.65 2438 <.0045 [-2.21, − 1.26]
Neg.Consequences 6.05 1.14 6.03 1.09 14,080 .67 [-0.23, 0.25] 6.33 1.00 5.31 1.32 8217 <.0045 [0.65.1.38]
Pos.Consequences 2.68 1.58 2.32 1.33 15,308 .03 [0.03, 0.67] 2.23 1.43 3.24 1.53 3280 <.0045 [-1.45, − 0.57]
Spatial proximity 5.80 1.10 5.90 1.16 12,888 .17 [-0.34, 0.15] 6.10 0.99 5.12 1.36 7914 <.0045 [0.61, 1.35]
Spatial distance 6.00 1.08 6.05 1.01 13,305 .31 [-0.28, 0.17] 6.21 0.99 5.44 1.13 7766 <.0045 [0.45, 1.09]
Temporal distance 3.00 1.64 2.60 1.52 15,668 .01 [0.05, 0.74] 2.48 1.54 3.66 1.60 3186 <.0045 [-1.64, − 0.71]
Bipolar cause 5.24 1.45 5.21 1.31 14,166 .71 [-0.28, 0.32] 5.69 0.98 3.95 1.71 8747 <.0045 [1.28, 2.20]
Bipolar consequence 5.85 1.11 5.91 0.99 13,456 .38 [-0.29, 0.17] 6.10 0.90 5.25 1.29 7752 <.0045 [0.50, 1.20]
Bipolar spatial 3.82 0.77 3.87 0.51 13,260 .26 [-.20, 0.09] 3.87 0.62 3.74 0.80 6083 .06 [-0.10, 0.35]

Study 3
Reality 6.29 0.92 6.29 0.98 113,841 .39 [-0.12, 0.12]
Causes 5.38 1.25 5.55 1.18 123,179 .02 [0.02, 0.33]
Consequences 5.68 1.09 5.78 1.00 119,014 .17 [-0.04, 0.22]

Note: Corrected alpha level for tests is 0.05/5 = 0.01 in Study 1, 0.05/11 = 0.0045 in Study 2, and 0.05/3 = 0.017 in Study 3. Consequences = valence of consequences,
Neg.Consequences = negative consequences, Pos.Consequences = positive consequences.

check (see below) 345 responses remained. Compared to US census data
(United States Census Bureau, 2018), our sample consisted of more
women and white respondents, and fewer Latino/a respondents. The
average education level of our sample was somewhat higher (more
participants had a high school degree or higher), while the median in­
come level was somewhat lower than the US census.
3.1.2. Materials
Climate change perceptions. We used the items selected in Study 1 to
measure climate change perceptions. Since only two items were selected
to measure perceptions of the temporal distance of consequences of
climate change in Study 1, we added one additional item, namely
‘Climate change will only take place in the long term’.
To assess the two-dimensional nature of the perceived causes,
valence of consequences, and spatial distance of consequences of climate
change, we first measured both dimensions independently. That is,
people indicated for each dimension separately the extent to which they
agree or disagree with it, for example to what extent they think that
climate change is caused by humans and to what extent they think that
climate change has natural causes. Next to the items from Study 1 that
measure the perceived human causes, negative consequences, and
spatially proximate consequences of climate change, we added items to
measure the perceived natural causes, positive consequences, and
spatially distant consequences of climate change (each containing three
items; see Table 7). All items were answered on a seven-point Likert-
scale, from ‘strongly disagree’ (1) to ‘strongly agree’ (7). In addition, we
revised two items from Study 1 to avoid overlap with the newly added
items.9
9 with ‘straight-lined’ reality items were removed. For two pairs of re­
Specifically, we replaced ‘Overall, climate change will bring more negative
than positive consequences to the world’ with ‘Climate change will bring
negative consequences to the world’ and replaced ‘Climate change will also
influence the place where I live’ with ‘Climate change will influence the place
where I live’. Also, we added the word ‘negative’ to one item measuring
perceived negative consequences, to distinguish it from the new items
measuring perceived positive consequence (i.e. ‘The negative consequences of
climate change will be very serious’).
Second, we measure endorsement of each dimension relative to the
other dimension using a bipolar scale. Specifically, participants evalu­
ated the perceived causes of climate change on a scale from ‘only by
natural process’ (1) to ‘equally natural processes and human activities’
(4) to ‘only human activities’ (7), the perceived valence of consequences
of climate change on a scale from ‘completely negative’ (1) to ‘equally
negative and positive’ (4) to ‘completely positive’, and the perceived
spatial distance of consequences of climate change on a scale from ‘only
locations far away’ (1), ‘equally locations far away and my local region’
(4) to ‘only my local region’ (7). We included two respective items for
perceived causes, valence of consequences, and spatial distance of
consequences of climate change (see Table 7).
In Study 2, participants who did not believe climate change exists
automatically skipped the questions about the causes and consequences
of climate change. These were 12 (3.5%) participants who selected
‘Strongly disagree’ or ‘Disagree’ for the item ‘I believe that climate
change is real’, and ‘Strongly agree’ or ‘Agree’ for the items ‘Climate
change is NOT occurring’ and ‘I do NOT believe that climate change is
real’.
Other scales. We measured in the same way as in Study 1 the NEP (M
= 5.22, SD = 1.19, α = .84), negative affect related to climate change (M
= 5.31, SD = 1.55, α = .92), climate change risk perception (M = 5.74,
SD = 1.39, α = .94), mitigation behaviour (M = 4.67, SD = 0.77, α =
.64), mitigation policy support (M = 5.62, SD = 1.33, α = .91), adap­
tation behaviour (M = 9.43, SD = 5.46, α = .87), adaptation policy
support (M = 5.46, SD = 1.28, α = .90), and demographic information.
3.1.3. Data quality check
The data were checked according to similar criteria as in Study 1.
Nine responses were completed in less than 3 minutes and 46 responses
sponses that came the same IP addresses, we only included the first
response. Another three pairs of responses from the same IP address
were kept because recorded starting and completion times indicated that
the questionnaires were completed simultaneously, which makes it un­
likely that they were completed by the same person.

8
A.M. van Valkengoed et al.
Table 7
Items included in Study 2 to measure the two-dimensional nature of the
perceived causes, valence of consequences, and spatial distance of consequences
of climate change.
Item
Perceived natural causes
1. Natural processes are a major cause
of climate change.
2. Climate change is mostly caused by
natural processes.
3. The main causes of climate change
are natural processes.
Perceived positive consequences
1. Climate change will bring positive
consequences to the world.
2. Climate change will bring about
important positive consequences.
3. The positive consequences of climate
change will be substantial.
Perceived spatial distance
1. Regions far away from me will be
influenced by climate change.
2. Regions far away from me will
experience the consequences of
climate change
3. Climate change will influence
regions far away from me.
Bipolar: natural-human causes
1. Climate change is caused …
2. The causes of climate change are …
Bipolar: valence of consequences
1. The consequences of climate change
will be …
2. Climate change will bring about …
Bipolar: spatial distance
1. Climate change will influence …
2. The consequence of climate change
will be experienced …
Answering scale
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
1 (completely disagree) – 7 (completely
agree)
only by natural processes (1), mostly by
natural processes (2), slightly more by
natural processes than by human activities
(3), equally by natural processes and
human activities (4), slightly more by
human activities than by natural processes
(5), mostly by human activities (6), only
by human activities (7)
only natural processes (1), mostly natural
processes (2), slightly more natural
processes than human activities (3),
equally natural processes and human
activities (4), slightly more human
activities than natural processes (5),
mostly human activities (6), only by
human activities (7)
only positive (1), mostly positive (2),
slightly more positive than negative (3)
equally positive and negative (4), slightly
more negative than positive (5), mostly
negative (6), only negative (7)
only positive consequences (1), mostly
positive consequences (2), slightly more
positive than negative consequences (3),
equally positive and negative
consequences (4), slightly more negative
than positive consequences (5), mostly
negative consequences (6) only negative
consequences (7)
only regions far away (1), mostly regions
far away (2), slightly more regions far
away than my local area (3), equally
regions far away and my local area (4),
slightly more my local area than regions
far away (5), mostly my local area (6),
only my local area (7)
Only by regions far away (1), mostly by
regions far away (2), slightly more by
regions far away than by my local area (3),
equally by regions far away and by my
local area (4), slightly more by my local
9
Journal of Environmental Psychology 76 (2021) 101652
Table 7 (continued )
Item Answering scale
area than by regions far away (5), mostly
by my local area (6), only by my local area
(7)
3.2. Results
3.2.1. Evaluating factor structure
The oblique multiple group method was used to determine whether
the items correlated most strongly with the type of climate change
perception to which they were assigned based on theory. The analysis
confirmed the assignment of all items to their designated subscales (see
Table 8). As regards the two-dimensional constructs, perceived natural
causes, positive consequences, and spatially distant consequences of
climate change could be empirically distinguished from the perceived
human causes, negative consequences and spatially proximate conse­
quences of climate change, respectively. The theory-based structure of
the scale fit the data significantly better than a model that collapsed the
two dimensions of the perceived causes, valence of consequences, and
spatial distance of consequences of climate change, or other more
parsimonious models (see Supplemental Material). This further supports
the finding that the two dimensions of perceived causes, valence of
consequences and spatial distance can be empirically distinguished.
Table 9 presents the descriptive statistics for all climate change
perception subscales. The subscales had good reliabilities, ranging from
.83 to .98 (see Table 9). The correlations between the climate change
perceptions were strong, except for the perceived spatial distance of
consequences of climate change measured with a bipolar scale, which
correlated weakly or not significantly with the other climate change
perceptions (See Table 10). This is likely because of the low variance in
this measure: 75% of the respondents (n = 250) selected the middle
option ‘Climate change will influence equally regions far away and my
local area’.
3.2.2. Convergent validity
Table 10 shows that all climate change perceptions (except the bi­
polar scale of perceived spatial distance of consequences of climate
change) correlated significantly and strongly with the theoretically
related constructs. The more people perceived climate change as real,
human caused, and having consequences that are negative, happening
both nearby and faraway and happening already now, the stronger their
endorsement of the NEP, the more they experienced negative affect
related to climate change, and the more risks they perceived from
climate change. Conversely, the more people perceived climate change
as naturally caused, and having positive consequences that are far away
in time, the less they endorsed the NEP, experienced negative affect
related to climate change, and perceived less risks from climate change.
The results from Study 1 were therefore replicated. While all measures of
climate change perceptions correlated significantly with theoretically
related constructs, these correlations were stronger for the original di­
mensions measured in Study 1 (i.e. perceived human causes, negative
consequences, spatial proximity of consequences) than for the newly
added dimensions (perceived natural causes, positive consequences,
spatial distance of consequences) and the bipolar measures added in
Study 2 (see Table 10).
3.2.3. Predictive validity
The correlations between the climate change perceptions and miti­
gation behaviour and support for mitigation and adaptation policy were
all significant (except the bipolar scale of perceived spatial distance of
consequences of climate change, see Table 10). The more people
perceived climate change as real, human caused, and having conse­
quences that are negative, and both spatially proximate and spatially
distant, the more they engaged in mitigation behaviours and supported
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 8
Results of the Multiple Group Method for the items included in Study 2.
Reality Human Natural Negative Positive Spatial Spatial Temporal
causes causes consequences consequences proximity distance distance

Reality
1. I believe that climate change is real. .897 .610 -.458 .659 -.459 .702 .621 -.562
2. Climate change is NOT occurring. (r) .938 .576 -.486 .661 -.540 .667 .621 -.608
3. I do NOT believe that climate change is .941 .566 -.461 .647 -.514 .677 .636 -.579
real. (r)

Human causes
1. Human activities are a major cause of .623 .820 -.604 .683 -.493 .638 .560 -.509
climate change.
2. Climate change is mostly caused by human .531 .854 -.579 .624 -.440 .566 .471 -.409
activity.
3. The main causes of climate change are .536 .861 -.590 .632 -.460 .557 .511 -.437
human activities.

Natural causes
1. Natural processes are a major cause of -.418 -.531 .801 -.492 .522 -.420 -.357 .524
climate change.
2. Climate change is mostly caused by natural -.468 -.609 .853 -.590 .598 -.459 -.465 .592
processes
3. The main causes of climate change are -.458 -.618 .835 -.588 .594 -.456 -.440 .567
natural processes.

Negative consequences
1. Climate change will bring negative .611 .617 -.542 .799 -.643 .675 .571 -.604
consequences to the world.
2. Climate change will bring about serious .634 .637 -.554 .817 -.605 .676 .585 -.589
negative consequences.
3. The negative consequences of climate .613 .645 -.557 .799 -.599 .638 .603 -.579
change will be very serious.

Positive consequences
1. Climate change will bring positive -.495 -.498 .599 -.656 .764 -.528 -.518 .674
consequences to the world.
2. Climate change will bring about important -.516 -.497 .583 -.650 .765 -.518 -.477 .621
positive consequences.
3. The positive consequences of climate -.380 -.335 .460 -.484 .633 -.382 -.417 .592
change will be substantial.

Spatial proximity
1. My local area will be influenced by climate .639 .583 -.469 .727 -.492 .827 .620 -.603
change.
2. The region where I live will experience the .655 .561 -.411 .626 -.486 .838 .646 -.590
consequences of climate change.
3. Climate change will influence the place .666 .603 -.455 .659 -.512 .831 .618 -.615
where I live.

Spatial distance
1. Regions far away from me will be .596 .509 -.368 .544 -.494 .586 .763 -.521
influenced by climate change.
2. Regions far away from me will experience .602 .495 -.425 .569 -.457 .629 .776 -.496
the consequences of climate change.
3. Climate change will influence regions far .549 .482 -.439 .620 -.487 .615 .775 -.565
away from me.

Temporal distance
1. It will be a long time before the -.600 -.481 .597 -.617 .691 -.619 -.554 .838
consequences of climate change are felt.
2. The consequences of climate change will -.559 -.447 .542 -.609 .656 -.621 -.538 .827
only be experienced in the far future.
3. Climate change will only take place in the -.513 -.411 .541 -.562 .632 -.563 -.532 .814
long term.

mitigation and adaptation policies. Conversely, the more people
perceived climate change as naturally caused, and having positive
consequences that are far away in time, the less they engaged in miti­
gation behaviours, and the less they supported mitigation and adapta­
tion policies. Surprisingly, none of the climate change perceptions were
significantly correlated with adaptation behaviour, unlike in Study 1.
The results of Study 1 were therefore partly replicated. The original
dimensions measured in Study 1 (perceived human causes, negative
consequences, and spatial proximity of consequences) performed equally
well or better compared to the newly added dimensions (perceived
natural causes, positive consequences, and spatial distance of conse­
quences) and the bipolar scales in predicting mitigation behaviour and
support for mitigation and adaptation policies (See Table 10).
3.2.4. Discriminant validity
Using the R package ‘npmv’ (Ellis et al., 2017), a non-parametric
MANOVA was conducted to compare the climate change perceptions
of men and women, and Republicans and Democrats, respectively. No

10
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 9 Table 9 (continued )

Items of the final scale and descriptive statistics in Study 2. Mean SD Min Max N α
Mean SD Min Max N α
1. Climate change is caused … only 5.23 1.40 1 7 333
Reality .98 by natural processes (1) – only by
1. I believe that climate change is 6.17 1.27 1 7 345 human activities (7)
real. 2. The causes of climate change are 5.23 1.40 1 7 333
2. Climate change is NOT occurring. 6.13 1.32 1 7 345 … only natural processes (1) –
(r) only human activities (7)
3. I do NOT believe that climate 6.16 1.33 1 7 345
change is real. (r) Bipolar – Valence of consequences .92
1. The consequences of climate 5.89 1.09 1 7 333
Human causes .92 change will be … only positive
1. Human activities are a major 5.91 1.35 1 7 333 (1) – only negative (7)
cause of climate change. 2. Climate change will bring about 5.87 1.09 1 7 333
2. Climate change is mostly caused 5.71 1.44 1 7 333 … only positive consequences (1)
by human activity. – only negative consequences (7)
3. The main causes of climate 5.71 1.49 1 7
change are human activities. Bipolar - Spatial distance .83
1. Climate change will influence … 3.87 0.70 1 7 333
only regions far away (1) – only
Natural causes .92
my local area (7)
1. Natural processes are a major 3.95 1.76 1 7 333
2. The consequence of climate 3.82 0.69 1 6 333
cause of climate change.
change will be experienced …
2. Climate change is mostly caused 3.41 1.82 1 7 333
only by regions far away (1) –
by natural processes.
only by my local area (7)
3. The main causes of climate 3.32 1.84 1 7 333
change are natural processes. Note: N differs between Reality and other subscales as climate change deniers
are excluded from other subscales.
Negative consequences .90
1. Climate change will bring 6.09 1.16 1 7 333
negative consequences to the significant difference was found between men (n = 160) and women (n
world. = 184) (F(2.45, 804.16) = 1.15, p = .32, partial η2 = 0.003, see
2. Climate change will bring about 5.99 1.27 1 7 333 Table 6)10. There was a significant difference in climate change per­
serious negative consequences.
ceptions between Democrats (n = 178) and Republicans (n = 71) (F
3. The negative consequences of 6.04 1.21 1 7 333
climate change will be very (2.74, 495.67) = 35.13, p < .01, partial η2 = 0.16).11 Post hoc
serious. Mann-Whitney U comparisons (one-sided) with Bonferroni-corrected
significance levels indicate that Republicans were less likely to
Positive consequences .85 perceived climate change as real, caused by humans (and more likely to
1. Climate change will bring 2.31 1.56 1 7 333
be naturally caused), having negative consequences (and more likely to
positive consequences to the
world.
have positive consequences), having consequences either close or far
2. Climate change will bring about 2.40 1.57 1 7 333 away, and having consequences that are close in time than Democrats.
important positive consequences. The results of Study 1 were therefore replicated. Additionally, the bi­
3. The positive consequences of 2.75 1.86 1 7 333 polar scales measuring the perceived causes and the valence of conse­
climate change will be
quences of climate change also distinguished between Republicans and
substantial.
Democrats in the expected direction, only the bipolar scale measuring
Spatial proximity .92 the spatial distance of consequences of climate change did not.
1. My local area will be influenced 5.83 1.25 1 7 333
by climate change. 3.2.5. Measurement invariance
2. The region where I live will 5.88 1.20 1 7 333
The fit of the factor structure of the scale was excellent if men and
experience the consequences of
climate change. women were considered separately (omnibus Satorra-Bentler scaled χ 2
3. Climate change will influence the 5.86 1.21 2 7 333 (448) = 567.75, χ 2/df = 1.27, p < .01, scaled CFI = 0.978, scaled
place where I live. RMSEA = 0.040, 90% CI [0.032, 0.048], SRMR = 0.037). There was no
evidence of measurement non-invariance across factor loadings (χ 2(16)
Spatial distance .88 = 14.56, p = .56), intercepts (χ 2(16) = 17.31, p = .37), or residuals
1. Regions far away from me will be 6.02 1.19 1 7 333
influenced by climate change.
(χ 2(24) = 31.01, p = .15) between men and women.
2. Regions far away from me will 6.05 1.13 1 7 333 Similarly, the fit of the factor structure of the scale was excellent if
experience the consequences of Democrats and Republicans were considered separately (Satorra-Bentler
climate change. scaled χ 2 (448) = 573.84, χ 2/df = 1.28, p < .01, scaled CFI = 0.970,
3. Climate change will influence 6.02 1.13 1 7 333
scaled RMSEA = 0.048, 90% CI [0.038, 0.058], SRMR = 0.042). There
regions far away from me.
was some evidence of measurement non-invariance of the intercepts
Temporal distance .91 across Democrats and Republicans (χ 2(16) = 37.10, p = .002). Yet, the
1. It will be a long time before the 2.81 1.74 1 7 333 reduction in model fit was marginal (ΔCFI = − 0.004, ΔRMSEA =
consequences of climate change 0.002). Moreover, the difference between the model constraining in­
are felt.
tercepts and factor loadings and the model constraining only factor
2. The consequences of climate 2.76 1.70 1 7 333
change will only be experienced
in the far future.
3. Climate change will only take 2.79 1.71 1 7 333 10
place in the long term.
The responses of one participant who selected ‘other’ for the gender ques­
tion were excluded for this analysis.
11
Bipolar - Causes .96 Participants that reported to vote for Independent’ or ‘Other’, or who
selected ‘No party/not interested in politics’ were not considered in these
analyses.

11
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 10
Pearson’s correlation coefficients of the climate change perceptions with variables to test convergent and predictive validity in Study 2.
Reality Human Natural Negative Positive Spatial Spatial Temporal Causes Consequences Spatial
causes causes consequences consequences proximity distance distance bipolar bipolar bipolar

Reality .60 -.48 .68 -.52 .71 .65 -.60 .53 .55 .09
Human causes -.63 .69 -.50 .63 .55 -.48 .78 .56 .10
Natural causes -.60 .62 -.48 -.46 .61 -.76 -.56 .00
Negative -.67 .73 .64 -.65 .68 .68 .13
consequences
Positive -.54 -.53 .71 -.53 -.70 .07
consequences
Spatial .68 -.65 .52 .55 .21
proximity
Spatial distance -.59 .46 .54 -.08
Temporal -.50 -.60 .02
distance
Causes bipolar .65 .19
Consequences .11
bipolar

Convergent validity
NEP .54 .61 -.58 .64 -.61 .57 .60 -.65 .57 .58 .02
Negative affect .73 .72 -.53 .68 -.46 .67 .53 -.52 .62 .52 .17
Risk perception .82 .77 -.53 .76 -.55 .77 .65 -.59 .63 .59 .11

Predictive validity
Mitigation .29 .28 -.30 .30 -.35 .31 .24 -.32 .26 .33 .00
behaviour
Adaptation .06 .03 .05 .03 .05 .08 .01 .01 -.04 .00 .04
behaviour
Mitigation .67 .72 -.49 .66 -.47 .66 .55 -.52 .59 .51 .17
policy
Adaptation .66 .57 -.36 .57 -.39 .57 .48 -.43 .46 .45 .14
policy

Note: Non-significant correlations marked in cursive.

loadings was marginal after the constraints on the intercepts for one
item (“It will be a long time before the consequences of climate change
are felt”) were removed (χ 2(15) = 25.42, p = .045), showing that the
scale was partially invariant. There was no further evidence of mea­
surement non-invariance across the factor loadings (χ 2(16) = 24.37, p =
.08) or residuals (χ 2(24) = 19.21, p = .74) of the scale between Re­
publicans and Democrats.12
3.3. Discussion
In Study 2, we replicated the findings from Study 1 showing that the
different climate change perceptions could be empirically distinguished,
and that the items had good internal reliability. Again, the reliability of
the subscale measuring the perceptions of the reality of climate change
approached unity and may therefore be over identified. The perceptions
of the reality, human causes, negative consequences, spatial proximity
and temporal distance of consequences of climate change correlated
significantly with the NEP, negative affect related to climate change,
climate change risk perception, mitigation behaviour, and support for
mitigation and adaptation policies. Perceptions of the spatial and tem­
poral distance/proximity of consequences of climate change again did
not correlate noticeably stronger with any of the outcome variables
compared to the three core perceptions of reality, causes, and valence of
consequences of climate change. Democrats more than Republicans
perceived climate change as real, human-caused, and as having conse­
quences that are negative, both spatially proximate and spatially distant,
and temporally proximate, while there was no difference in climate
change perceptions between men and women. Moreover, the structure
of the scale, factor loadings, intercepts, and residuals did not vary
12
We also tested whether the scale was invariant across age groups. There
were some minor deviancies from invariance, but they were not systematic and
the scale proved to be at least partially invariant, see Supplemental Material.
between men and women, and there was only a slight deviation in the
intercepts between Republicans and Democrats. Unlike in Study 1,
however, climate change perceptions were not related to adaptation
behaviours.
While the two dimensions of the perceived causes, valence of con­
sequences, and the spatial distance of consequences of climate change
could be empirically distinguished, this did not seem to have added
value to the validity of the scale. Notably, the originally measured
human causes, negative consequences, and spatial proximity of conse­
quences of climate change correlated more strongly with the NEP,
negative affect related to climate change, climate change risk percep­
tion, mitigation behaviours, and support for mitigation and adaptation
policies, compared to the newly added perceived natural causes, positive
consequences, and spatial distance of consequences of climate change.
Also, measuring the two-dimensional nature of these climate change
perceptions with bipolar scales did not improve the predictive or
convergent validity of the scale. The original items measuring the
climate change perceptions in Study 1 are therefore preferred to
examine how climate change perceptions relate to outcomes of interest.
4. Study 3
Study 1 and 2 tested the scale with a convenience sample (MTurk)
and in a single cultural context (the US). The aim of Study 3 is to
examine the generalizability of the findings with a different sample in a
different country: members of the general public in the Netherlands. For
this purpose, we conducted a large door-to-door survey on public per­
ceptions of the sustainable energy transition. As the previous two studies
showed that the perceptions of the temporal and spatial distance of
consequences of climate change did not correlate noticeably stronger
with any of the outcome variables than the core perceptions of climate
change, we focused on perceptions of the reality, causes, and valence of
consequences in this study.
We will conduct a conceptual replication of the convergent and

12
A.M. van Valkengoed et al.
predictive validity of the scale, which is of critical importance to assure
that the scale can be used with a wide variety of different outcomes and
in different contexts (Crandall & Sherman, 2016). We therefore examine
how the scale relates to a novel set of theoretically relevant variables,
specifically focused on sustainable energy use and production. For
convergent validity, we assess to what extent the climate change per­
ceptions correlate with biospheric values, that is, the extent to which
people value and care for the environment (de Groot & Steg, 2008). We
expect that, similar to the NEP, people with stronger biospheric values
care more about the state of the environment and possible threats to the
environment (Steg, De Groot, Dreijerink, Abrahamse, & Siero, 2011),
and are therefore likely to perceive climate change as more real,
human-caused, and having negative consequences.
To assess predictive validity, we examine whether the scale predicts
whether people have the aspiration to reduce their energy use, and to
use energy from renewable resources. Both reflect people’s more
conscious efforts to mitigate climate change. Because energy use is a key
factor influencing climate change, we expect that the more people
perceive climate change perceptions as real, human-caused, and having
negative consequences, the stronger their aspiration to reduce energy
use and to use more renewable energy. We further examine whether the
scale predicts the extent to which people find renewable energy projects
in their local area acceptable. In Study 1 and 2, we focused on support
for national level policies, and we therefore focus now on whether
climate change perceptions are also related to the perceived accept­
ability of local developments to mitigate climate change. Again, since
renewable energy is a key strategy to mitigate climate change, we expect
that the more people perceive climate change perceptions as real,
human-caused, and having negative consequences, the more they
perceive renewable energy developments in their local area as accept­
able. We assess discriminant validity and measurement invariance by
comparing the scale scores, factor structure, factor loadings, intercepts,
and residuals between men and women.13 To assess measurement
invariance across countries, we compare the factor structure factor
loadings, intercepts, and residuals of the scale between the Dutch par­
ticipants of Study 3 to the US participants from Study 1.14
4.1. Methods
4.1.1. Participants and procedure
Participants were residents of a Northern province of the
Netherlands. Participants were approached at their home, via door-to-
door methods and invited to fill in a questionnaire, resulting in a more
diverse participant pool compared to using online panels (Harms &
DeSimone, 2015). The questionnaire was picked up at a later time point.
Overall, 1125 useable questionnaires were returned, and after data
quality check (see below), 1060 responses were included in the study.
The final sample consisted of slightly more men (55%) than women
(45%). The average age was 51. Compared to the general Dutch popu­
lation, the current sample had on average a higher education level, but
lower income. For a full overview of the methods, see Perlaviciute and
Squintani (2020).
4.1.2. Materials
Climate change perceptions. Perceptions of the reality, causes, and
valence of consequences of climate change were assessed using the items
selected in Study 1. In Study 1 and 2, the reliability of the reality sub­
scale was extremely high, indicating that the use of three items may be
redundant. To reduce the length of this subscale, only two items were
13
We do not consider political identity as a relevant factor as climate change
is far less politicized in the Netherlands compared to the US.
14
We compare the data from Study 3 with the data from Study 1, as these
studies had the exact same wording of the items. In Study 2, some of the items
had slightly different wordings to avoid overlaps with the bipolar scales.
13
Journal of Environmental Psychology 76 (2021) 101652
therefore used in the current study (‘I believe that climate change is
real’, ‘Climate change is NOT occurring’).15 For the perceived causes
and valence of consequences of climate change, the three items selected
in Study 1 were used. The items were translated into Dutch by the first
author. A back translation into English by a native speaker confirmed
the accuracy of the translation.
Biospheric values. Biospheric values were measured with an estab­
lished scale to measure hedonic, egoistic, altruistic and biospheric
values (Steg, Perlaviciute, van der Werff, & Lurvink, 2014). Four items
measured biospheric values: (‘Respecting the earth’, ‘Being one with
nature’, ‘Protecting the environment’, ‘Preventing environmental
pollution’). Participants indicated for each value the extent to which
they endorse the value (ranging from − 1 = ‘opposed to my values’ to 0
= ‘not important’ to 7 = ‘of supreme importance’). A mean score was
calculated across the four items (M = 5.37, SD = 1.11 α = .83).
Aspiration to use less energy was assessed with one item: ‘I want to
reduce my energy use’. Responses were given on a 7-point Likert scale
(1 = ‘strongly disagree’, 7 = ‘strongly agree’; M = 5.14, SD = 1.48).
Aspiration to use renewable energy was assessed with one item: ‘I want
to use more energy from renewable resources’. Responses were given on
a 7-point Likert scale (1 = ‘strongly disagree’, 7 = ‘strongly agree’; M =
5.26, SD = 1.35).
Acceptability of wind energy projects. Participants were asked to what
extent they found the development of both offshore and onshore wind
parks in their local area acceptable and necessary, respectively. Re­
sponses were given on a 7-point Likert scale (1 = ‘strongly disagree’, 7 =
‘strongly agree’). A mean score was calculated across the four items (M
= 5.07, SD = 1.25, α = .81).
Acceptability of solar energy projects. Participants were asked whether
they found the development of both solar parks and residential solar
panels in their local area acceptable and necessary, respectively. Re­
sponses were given on a 7-point Likert scale (1 = ‘strongly disagree’, 7 =
‘strongly agree’). A mean score was calculated across the four items (M
= 5.72, SD = 1.04, α = .76).
4.1.3. Data quality check
To secure the quality of the data, participants were removed that
‘straight-lined’ the reality items (n = 65), which was defined as
providing responses that were consistently above or below the scale
midpoint (i.e., only providing response ‘5’, ‘6’, ‘7’ or ‘1’, ‘2’, ‘3’) on both
items measuring the perceived reality of climate change.
4.2. Results
4.2.1. Evaluating factor structure
The oblique multiple group method showed that all items correlated
most strongly with their designated subscale, confirming the factor
structure of the scale (see Table 11). The reliabilities of the three sub­
scales were good to excellent (αreality = .79, αcauses = .91, αconsequences =
.86).
4.2.2. Convergent and predictive validity
As depicted in Table 12, the more people perceived climate change as
real, human caused, and having negative consequences, the stronger
they endorsed biospheric values, confirming the convergent validity of
the scale. The more people perceived climate change as real, human
caused, and having negative consequences, the stronger aspiration to
reduce their energy use and to use energy from renewable resources, and
the more acceptable they find wind and solar energy developments in
their local area, confirming the predictive validity of the scale.
15
We chose to drop the item ‘I do NOT believe that climate change is real’ as it
was very similar to ‘I believe climate change is real’.
A.M. van Valkengoed et al. Journal of Environmental Psychology 76 (2021) 101652

Table 11 0.021), implying that the factor loadings differ significantly between the
Oblique multiple group method in Study 3. two groups.16 The decrease in model fit was however much smaller, and
Reality Causes Consequences approached the threshold for non-invariance, if the constraints on one
item (“Human activities are a major cause of climate change”) were
Reality
I believe that climate change is real. .666 .415 .510 released (ΔCFI = − 0.015, ΔRMSEA = 0.010), and virtually absent if
Climate change is NOT occurring. (r) .666 .209 .281 constraints on one more item were released (“Climate change is NOT
Causes occurring“) (ΔCFI = − 0.004, ΔRMSEA = 0.001), suggesting that the
Human activities are a major cause of climate .422 .755 .611 scale is partially invariant. Further examination also does not offer
change.
Climate change is mostly caused by human .273 .870 .571
convincing evidence of measurement non-invariance for intercepts
activity. (ΔCFI = − 0.012, ΔRMSEA = 0.006), or residuals (ΔCFI = 0.002,
The main causes of climate change are human .280 .882 .627 ΔRMSEA = − 0.008).17
activities.
Consequences
Overall, climate change will bring more .331 .573 .706 4.3. Discussion
negative than positive consequences to the
world.
Climate change will bring about serious .402 .566 .771
In Study 3, we replicated and extended the key findings from Studies
negative consequences. 1 and 2. We replicated the factor structure of the scale and the re­
The consequences of climate change will be .423 .582 .720 liabilities of the subscales, further confirming the theoretical structure of
very serious. the scale. Moreover, the reality subscale also functioned well with two
Coefficients in bold represent the highest overall total-item correlations. items, confirming that this subscale can be shortened without reducing
the validity of the scale. We found that the three types of climate change
perceptions correlated in the expected direction with biospheric values,
Table 12 aspirations to use less energy and to use energy from renewable sources,
Pearson’s correlation coefficients of the climate change perceptions with vari­ and acceptability of wind and solar energy projects, conceptually
ables to test convergent and predictive validity in Study 1. replicating the convergent and predictive validity of the scale. We found
Reality Causes Consequences no differences between men and women on the climate change per­
Causes .35 ceptions, and the factor structure, factor loadings, intercepts, and re­
Consequences .44 .65 siduals of the scale also did not differ across gender, suggesting that the
scale performs equally well with both men and women. While the factor
Convergent validity structure of the scale was similar for Dutch and American respondents,
Biospheric values .17 .26 .32
the factor loadings were somewhat stronger for American respondents
on some items. The scale was invariant between Dutch and American
Predictive validity
Aspiration reduce energy .22 .25 .31 respondents with regard to the intercepts and residuals.
Aspiration renewable energy .28 .35 .41
Acceptability wind energy .15 .29 .25 5. Exploratory analyses: single-items scale
Acceptability solar energy .27 .32 .32

Note: all correlations significant at p < .05.
4.2.3. Discriminant validity and measurement invariance
Using the R package ‘npmv’ (Ellis et al., 2017), a non-parametric
MANOVA was conducted to compare the climate change perceptions
of men (n = 600) and women (n = 491). We found no significant dif­
ference between men and women on any of the climate change per­
ceptions (F(1.90, 1683.31) = 1.97, p = .14, partial η2 = 0.002; see
Table 6). Next, we assessed whether the factor structure, factor loadings,
intercepts, and residuals of the scale are similar for men and women
(Fischer & Karl, 2019). The fit of the factor structure is adequate if men
and women considered separately (omnibus Satorra-Bentler scaled
χ 2(34) = 125.152, χ2/df = 3.68, p < .001, scaled CFI = 0.967, scaled
RMSEA = 0.077, 90% CI [0.066, 0.089], SRMR = 0.045). There was no
evidence of measurement non-invariance across factor loadings (χ 2(5)
= 6.83, p = .23), intercepts (χ 2(5) = 5.29, p = .38), or residuals (χ 2(8) =
3.80, p = .87) between men and women.
To investigate whether factor structure, factor loadings, intercepts,
and residuals of the scale are similar for Dutch and American re­
spondents, we compared the data collected in the Netherlands (Study 3)
with the data collected in Study 1 in the US, using the items that are
available in both studies. The fit of the scale is adequate if both countries
are considered separately, confirming that the factor structure of the
scale is similar in both countries (omnibus Satorra-Bentler scaled (34) =
159.697, χ 2/df = 4.70, p < .01, scaled CFI = 0.967, scaled RMSEA =
0.076, 90% CI [0.066, 0.086], SRMR = 0.036). A model that assumes
that the factor loadings for both the US and Dutch sample are equal
decreased the CFI and RMSEA model fit indices by at least 0.010 and
0.015, respectively, from an unconstrained model, which we considered
evidence of measurement non-invariance (ΔCFI = − 0.029, ΔRMSEA =
In all our studies, the reliabilities of the sets of items measuring the
same climate change perceptions were relatively high. This might
indicate that a concept is over-identified by the use of multiple items,
and that it may be possible to assess the construct with fewer items
(Postmes, Haslam, & Jans, 2013). We therefore explored whether a scale
consisting of single items to measure each type of climate change per­
ceptions performs similarly to the full scale where each type of climate
change perceptions is measured using three items. Based on the factor
loadings in Study 1 and the face validity of the items, we selected the
following five items to represent the five types of climate change per­
ceptions: ‘I believe that climate change is real’ (reality), ‘The main
causes of climate change are human activities’ (causes), ‘Climate change
will bring about serious negative consequences’ (valence of conse­
quences), ‘My local area will be influenced by climate change’ (spatial
distance of consequences),’ It will be a long time before the conse­
quences of climate change are felt’ (temporal distance of consequences).
We reanalysed the data from Studies 1, 2 and 3 to examine the val­
idity of the scale with single items. Across all three studies, the single
items correlated significantly and in the expected direction with the
theoretically related items, mitigation behaviour, and policy support,
16
Notably, we do not rely on the χ2 test to compare nested models here,
because the large sample size of this study makes this test overly sensitive to
small deviations (Putnick & Bornstein, 2016). Alternatively, we examine the
changes in model fit between nested models. We consider a change in 0.010 in
CFI and 0.015 in RMSEA as evidence of measurement non-invariance (Cheung
& Rensvold, 2002).
17
We also tested whether the scale was invariant across age groups. There was
no evidence of measurement non-invariance across age groups in Study 3 (see
Supplemental Material).

14
A.M. van Valkengoed et al.
supporting the convergent and predictive validity of the single-items
scale (see Table S6). The items also correlated with adaptation behav­
iour in Study 1, but not Study 2, recreating the pattern of results ob­
tained with the full scale. While the correlations were overall a bit
stronger for the multiple-items scale, the difference is in all cases smaller
than 0.10 and therefore of limited practical relevance.
The results of the discriminant validity analyses are also almost
identical to those of the multi-item scale, showing that there were no
differences in climate change perceptions between men and women, but
significant differences between Democrats and Republicans (see
Table S7). One exception was a significant difference between men and
women in the perceived causes of climate change in Study 3 (women
endorsed the human causes of climate change more strongly than men,
U = 125,842, p = .01); this difference was not significant when
perceived causes of climate change were measured with multiple items.
The analyses for measurement invariance for the multi-item scale
already showed that none of the items selected for the single-items scale
were identified as non-invariant. The only exception is the item ‘It will
be a long time before the consequences of climate change are felt’, which
was identified as non-invariant on the dimension of scalar invariance, in
Study 2. However, this item was not flagged as non-invariant in Study 1,
which suggests that the non-invariance in Study 2 may have occurred
due to sampling fluctuations.
In conclusion, overall, the single-items scale performed comparably
well to the multi-item scale, although it might be more susceptive to
influence of measurement error and sampling fluctuations. The single-
items scale has the key practical advantage of being easily includable
in research studies that aim to minimize the length of the questionnaire
without compromising the validity of the scale too much. Yet, we still
recommend the use of the multi-item scale if climate change perceptions
form a central construct of interest and optimal measurement accuracy
is desired.
6. General discussion
People’s perceptions of the reality, causes, and consequences of
climate change can have important implications for their behaviour and
support for policies to mitigate and adapt to climate change. Multiple
studies have examined which factors affect climate change perceptions,
and how perceptions of climate change in turn affect behaviour and
policy support (e.g., Clayton et al., 2015; Hornsey et al., 2016; Weber,
2010, 2016). Yet, these studies vary in which type of climate change
perceptions are assessed and how, making it difficult to compare and
integrate findings across studies. We addressed this gap in the literature
by developing and validating a theory-based scale to measure different
types of climate change perceptions: the perceived reality, causes, and
consequences of climate change, with the latter encompassing the
perceived valence, spatial distance, and temporal distance of conse­
quences of climate change.
We first defined 5 types of climate change perceptions on the basis of
current theorizing, and selected items reflecting these five types of
climate change perceptions based on a literature review and experts’
assessment. Next, we conducted three studies to test the scale, following
key steps of the scale development process, namely confirmatory factor
analysis, internal reliability, convergent, predictive, and discriminant
validity, measurement invariance, and conceptual replication. The
result is a reliable and valid scale to measure the different types of
climate change perceptions. Specifically, confirmatory factor analyses
showed that the different climate change perceptions could be empiri­
cally distinguished, supporting the theoretical structure of the scale.
High Cronbach’s alphas for each subscale supported the internal reli­
ability of the scale. The climate change perceptions were associated with
theoretically related constructs (i.e. the New Ecological Paradigm,
negative affect related to climate change, climate change risk percep­
tion, and biospheric values) in the expected direction, confirming the
convergent validity of the scale. Furthermore, the climate change
15
Journal of Environmental Psychology 76 (2021) 101652
perceptions were associated with (aspired) mitigation behaviour, sup­
port for mitigation and adaptation policies, including local renewable
energy projects, in the expected direction, confirming the predictive
validity of the scale. Republican and Democratic voters differed in the
scores on the climate perception scale in the expected direction, con­
firming the discriminant validity of the scale. In contrast to earlier
studies (McCright, 2010), we found no gender differences in the climate
change perceptions. Such gender effects tend to be rather small (around
r = 0.03) and much larger sample sizes may be needed to detect these
effects (Davidson & Haan, 2012; Hamilton, Hartter, Lemcke-Stampone,
Moore, & Safford, 2015; McCright, 2010).
In all three studies, the scale proved to be fully invariant for men and
women. For Republicans and Democrats, the scale was found to be non-
invariant in Study 1 and Study 2. However, in both cases, the non-
invariance could be traced to single items, suggesting that the scale is
partially invariant. Moreover, the non-invariance affected the model fit
only marginally, and was not consistent (i.e. different items and
different types of non-invariance were identified in Study 1 and Study
2), suggesting that the non-invariance is likely not of practical concern
(Putnick & Bornstein, 2016). Lastly, the factor loadings appeared to
differ between Dutch and American respondents, but this could again be
traced to two items only, suggesting the scale is also partially invariant
across the samples from different countries. Future research is needed to
determine whether this effect is robust, or occurred due to sampling
fluctuations.
The three studies showed that the climate change perceptions could
be empirically distinguished, offering support for the theoretical
distinction between perceptions of the reality, causes, and consequences
(including valence, spatial distance, and temporal distance) of climate
change. Study 2 further showed that two dimensions could be distin­
guished for some climate change perceptions: people can perceive
climate change as having both natural and human causes, positive and
negative consequences, and having consequences that occur both far
away and close by. Moreover, the two dimensions of the causes and
valence of consequences could also be measured using a bipolar scale.
However, the original dimensions measured in Study 1, namely per­
ceptions of the human causes, negative consequences, and spatial
proximity of consequences correlated most strongly with mitigation
behaviour and support for climate policy. The newly added dimensions
of perceived natural causes, positive consequences, and spatial distance
therefore did not improve the predictive validity of the scale. This sug­
gests that the original items measuring the climate change perceptions
in Study 1 are preferred to examine how climate change perceptions
relate to outcomes of interest.
While the different climate change perceptions could be empirically
distinguished, the correlations between the climate change perceptions
were strong, particularly in the US. This indicates that it may not be
necessary to include all climate change perceptions in a given research
project. This may be particularly relevant when researchers look to
reduce the length of the complete scale. As perceptions of the negative
consequences of climate change were overall the strongest predictor of
relevant outcome variables such as pro-environmental behaviour and
policy support, this may be the most relevant subscale for studies that
aim to explain climate-related behaviour or policy support. This choice
can also be theoretically substantiated. Even if people perceive climate
change as real, human-caused, and occurring close by in time and space,
they may not be motivated to act unless they perceive the valence of
consequences of climate change to be negative. Indeed, even though it
has been argued that the perceived psychological distance of climate
change should be particularly important to motivate climate-relevant
behaviour (e.g., Jones et al., 2017; van der Linden et al., 2015), we
found that perceptions of the spatial and temporal distance of conse­
quences of climate change were not stronger predictors of
climate-related behaviour and policy support than perceptions of the
negative consequences of climate change. However, the other type
climate change perceptions included in the scale may still be useful to
A.M. van Valkengoed et al.
answer different types of research questions. For example, the reality
and causes subscales may be particularly relevant for studying the per­
ceptions of more climate sceptic groups, such as political conservatives
(McCright & Dunlap, 2011), the far-right (Forchtner, 2019), farmers
(Hyland, Jones, Parkhill, Barnes, & Williams, 2016; Woods, Ørsted
Nielsen, Branth Pedersen, & Kristofersson, 2017), or those inclined to
conspiratorial ideation (Hornsey, Harris, & Fielding, 2018). Similarly,
the perceived spatial or temporal distance of consequences of climate
change may be particularly relevant when focussing on regions that are
least affected by climate change.
The correlations between the multiple items measuring different
types of climate change perceptions were rather high. Therefore, we
examined whether single items could be used to measure the different
types of climate change perceptions, further facilitating the practical
application of the scale. Across the three studies, the single-items scale
performed comparably well to the multi-items scale in convergent val­
idity, predictive validity, and discriminant validity. Yet, the correlations
between climate change perceptions measured with single items and
constructs of interests were somewhat weaker compared to the multi-
item scale. Moreover, the single-items scale revealed that women
perceived climate change more as human-caused than men in Study 3,
which was not found with the multi-item scale. Overall, the findings
provide initial evidence that the single-item scale could be adequate to
assess climate change perceptions. Yet, future studies are needed to
further assess the reliability of this version of the scale, for example by
establishing low measurement error and test-re-test reliability.
While we found that the climate change perceptions, and particularly
the perceived negative consequences, consistently predicted mitigation
behaviour and policy support, the climate change perceptions were
however inconsistent predictors of adaptation behaviours across Studies
1 and 2. Our measure of adaptation behaviours consisted of behaviours
that were more directly aimed at coping with climate change risks (such
as adjusting the home to withstand natural hazards), as well as behav­
iours that could also be performed for other reasons rather than to
protect oneself against climate change risks (e.g. people may prepare a
household emergency kit because they want to be prepared for any ac­
cidents or risk). The relationship between climate change perceptions
and adaptation behaviour may be more pronounced for behaviours that
are more directly aimed at coping with climate change risks, compared
to behaviours that people also engage in for other reasons. Future
studies may disentangle to what extent and under which conditions
climate change perceptions predict different types of adaptation
behaviour.
Overall, we found that most respondents strongly believed that
climate change is real, caused by humans, and has negative conse­
quences. Only 3.3% (Study 1), 3.5% (Study 2), and 1.3% (Study 3) of the
respondents denied the reality of climate change. Most respondents also
thought that the consequences of climate change would happen soon
and close by, in contrast to previous studies that suggest that many
people perceive climate change as psychologically distant (e.g., Jones
et al., 2017). We found that even most Republican voters agreed that
climate change is real, human-caused, and will have negative conse­
quences nearby in time and space. This is an important finding, as the
United States is often considered a ‘hotbed of climate change denial’
(Milman & Harvey, 2019). These findings align with research in Europe
with representative samples, which suggests that climate change denial
and scepticism may be less widespread than often assumed (Steg, 2018).
Yet, our results should be interpreted with caution, as the samples of
Study 1 and 2 were not fully representative of the United States, which
might have influenced our findings. Overall, these findings indicate that
policy may not need to focus on communicating to people about the
reality, human causes, or negative consequences of climate change, as
most people already acknowledge this. Rather, it may be more impor­
tant to encourage and empower people to translate their climate change
perceptions into climate actions and policy support (Steg & Vlek, 2009).
Future research could examine to what extent the scale is applicable
16
Journal of Environmental Psychology 76 (2021) 101652
and useable for groups that may hold different climate change percep­
tions than the general public, such as farmers, who are usually more
sceptical about climate change (van Valkengoed & Steg, 2019), or
indigenous communities, who may interpret climatic changes on the
basis of indigenous knowledge (Boillat & Berkes, 2013). Future studies
could also assess test-retest reliability, referring to the extent to which
participants’ responses on the same scale are similar over time (McArdle
& Woodcock, 1997). Establishing test-retest reliability is especially
important for establishing whether climate change perceptions are a
stable construct or susceptible to change, for example due to increasing
frequency and severity of climate-related hazards, and with climate
change becoming increasingly important in the media and politics
(Deeg, Lyon, Leiserowitz, Maibach, & Marlon, 2019). To establish
test-retest reliability, a longitudinal assessment of climate change per­
ceptions over multiple time points will be required.
In sum, we developed and validated a scale to measure people’s
climate change perceptions, including perceived reality, causes, and
consequences of climate change. We hope this scale will prove useful to
researchers that aim to study people’s climate change perceptions and
their relationship to other constructs of interest.
Authors’ contributions
AV and LS came up with the research idea and designed the first
study. AV, LS, and GP came up with the idea and design for the second
study. AV collected the data for Study 1 and Study 2. GP collected the
data for Study 3. AV analysed the data for all studies. All authors
contributed to writing the manuscript.
Data availability
The data and R scripts to reproduce the analyses reported in this
paper are available via: https://osf.io/v65q7/.
Declaration of competing interest
The authors have no conflicts of interest to disclose.
Acknowledgements
We thank dr. Lorenzo Squintani (Faculty of Law, University of Gro­
ningen) for collaboration in carrying out Study 3, and the following
research assistants for their contribution to collecting the data for Study
3: Rasmus Geertsema, Maud Groot Beumer, Yorick Juffer, Gido Metz,
Mauritz Mulder, Lies Rademaker, and Julia van der Struijf. Study 3 was
funded by the Province of Groningen, the Netherlands.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.jenvp.2021.101652.
References
Arnold, O., Kibbe, A., Hartig, T., & Kaiser, F. G. (2018). Capturing the environmental
impact of individual lifestyles: Evidence of the criterion validity of the general
ecological behavior scale. Environment and Behavior, 50(3), 350–372.
Berry, H. L., & Peel, D. (2015). Worrying about climate change: Is it responsible to
promote public debate? BJPsych International, 12(2), 31–32.
Björnberg, K. E., Karlsson, M., Gilek, M., & Hansson, S. O. (2017). Climate and
environmental science denial: A review of the scientific literature published in
1990–2015. Journal of Cleaner Production, 167, 229–241. https://doi.org/10.1016/j.
jclepro.2017.08.066
Bloodhart, B., Maibach, E., Myers, T., & Zhao, X. (2015). Local climate experts: The
influence of local TV weather information on climate change perceptions. PloS One,
10(11), Article e0141526. https://doi.org/10.1371/journal.pone.0141526
Boateng, G. O., Neilands, T. B., Frongillo, E. A., Melgar-Quiñonez, H. R., & Young, S. L.
(2018). Best practices for developing and validating scales for health, social, and
behavioral research: A primer. Frontiers in Public Health, 6, 149.
A.M. van Valkengoed et al.
Boillat, S., & Berkes, F. (2013). Perception and interpretation of climate change among
Quechua farmers of Bolivia: Indigenous knowledge as a resource for adaptive
capacity. Ecology and Society, 18(4), art21.
Bostrom, A., O’Connor, R. E., Böhm, G., Hanss, D., Bodi, O., Ekström, F., et al. (2012).
Causal thinking and support for climate change policies: International survey
findings. Global Environmental Change, 22(1), 210–222. https://doi.org/10.1016/j.
gloenvcha.2011.09.012
Bouman, T., Steg, L., & Kiers, H. A. L. (2018). Measuring values in environmental
research: A test of an environmental portrait value questionnaire. Frontiers in
Psychology, 9, 564. https://doi.org/10.3389/fpsyg.2018.00564
Brenkert-Smith, H., Meldrum, J. R., & Champ, P. A. (2015). Climate change beliefs and
hazard mitigation behaviors: Homeowners and wildfire risk. Environmental Hazards,
14(4), 341–360. https://doi.org/10.1080/17477891.2015.1080656
Brügger, A., Morton, T. A., & Dessai, S. (2015). Hand in hand: Public endorsement of
climate change mitigation and adaptation. PloS One, 10(4), Article e0124843.
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing
measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9
(2), 233–255. https://doi.org/10.1207/S15328007SEM0902_5
Clayton, S., Devine-Wright, P., Stern, P. C., Whitmarsh, L., Carrico, A., Steg, L., et al.
(2015). Psychological research and global climate change. Nature Climate Change, 5
(7), 640–646. https://doi.org/10.1038/nclimate2622
Crandall, C. S., & Sherman, J. W. (2016). On the scientific superiority of conceptual
replications for scientific progress. Journal of Experimental Social Psychology, 66,
93–99.
Davidson, D. J., & Haan, M. (2012). Gender, political ideology, and climate change
beliefs in an extractive industry community. Population and Environment, 34(2),
217–234.
Deeg, K., Lyon, E., Leiserowitz, A., Maibach, E., & Marlon, J. (2019). Who is changing their
mind about global warming and why? Yale Program on Climate Change Communication.
https://climatecommunication.yale.edu/publications/who-is-changing-their-mind-
about-global-warming-and-why/.
Dietz, T., Dan, A., & Shwom, R. (2007). Support for climate change policy: Social
psychological and social structural influences. Rural Sociology, 72(2), 185–214.
Ding, D., Maibach, E. W., Zhao, X., Roser-Renouf, C., & Leiserowitz, A. (2011). Support
for climate policy and societal action are linked to perceptions about scientific
agreement. Nature Climate Change, 1(9), 462–466. https://doi.org/10.1038/
nclimate1295
Dunlap, R. E., Van Liere, K. D., Mertig, A. G., & Jones, R. E. (2000). Measuring
endorsement of the new ecological Paradigm: A revised NEP scale. Journal of Social
Issues, 56(3), 425–442.
Ellis, A. R., Burchett, W. W., Harrar, S. W., & Bathke, A. C. (2017). Nonparametric
inference for multivariate data: The R package npmv. Journal of Statistical Software,
76(4), 1–18. https://doi.org/10.18637/jss.v076.i04
Elrick-Barr, C. E., Smith, T. F., Preston, B. L., Thomsen, D. C., & Baum, S. (2016). How are
coastal households responding to climate change? Environmental Science & Policy, 63,
177–186.
European Social Survey. (2016). ESS round 8 source questionnaire. ESS ERIC c/o City
University London.
Ferguson, C. J. (2009). An effect size primer: A guide for clinicians and researchers.
Professional Psychology: Research and Practice, 40(5), 532–538.
Fischer, R., & Karl, J. A. (2019). A primer to (Cross-Cultural) multi-group invariance
testing possibilities in R. Frontiers in Psychology, 10, 1507. https://doi.org/10.3389/
fpsyg.2019.01507
Forchtner, B. (2019). Climate change and the far right. Wiley Interdisciplinary Reviews:
Climate Change, 10(5), e604. https://doi.org/10.1002/wcc.604
de Groot, J. I. M., & Steg, L. (2008). Value orientations to explain beliefs related to
environmental significant behavior: How to measure egoistic, altruistic, and
biospheric value orientations. Environment and Behavior, 40(3), 330–354. https://
doi.org/10.1177/0013916506297831
Guy, S., Kashima, Y., Walker, I., & O’Neill, S. (2014). Investigating the effects of
knowledge and ideology on climate change beliefs: Knowledge, ideology, and
climate change beliefs. European Journal of Social Psychology, 44(5), 421–429.
Hagen, B., Middel, A., & Pijawka, D. (2016). European Climate Change Perceptions:
Public support for mitigation and adaptation policies. Environmental Policy and
Governance, 26(3), 170–183.
Hamilton, L. C., Hartter, J., Lemcke-Stampone, M., Moore, D. W., & Safford, T. G. (2015).
Tracking public beliefs about anthropogenic climate change. PloS One, 10(9), Article
e0138208. https://doi.org/10.1371/journal.pone.0138208
Harms, P. D., & DeSimone, J. A. (2015). Caution! MTurk workers ahead—fines doubled.
Industrial and Organizational Psychology, 8(2), 183–190. https://doi.org/10.1017/
iop.2015.23
Hattie, J., & Cooksey, R. W. (1984). Procedures for assessing the validities of tests using
the ‘known-groups’ method. Applied Psychological Measurement, 8(3), 295–305.
Heath, Y., & Gifford, R. (2006). Free-market ideology and environmental degradation:
The case of belief in global climate change. Environment and Behavior, 38(1), 48–71.
Hine, D. W., Reser, J. P., Phillips, W. J., Cooksey, R., Marks, A. D. G., Nunn, P., et al.
(2013). Identifying climate change interpretive communities in a large Australian
sample. Journal of Environmental Psychology, 36, 229–239. https://doi.org/10.1016/
j.jenvp.2013.08.006
Hornsey, M. J., Harris, E. A., Bain, P. G., & Fielding, K. S. (2016). Meta-analyses of the
determinants and outcomes of belief in climate change. Nature Climate Change, 6(6),
622–626.
Hornsey, M. J., Harris, E. A., & Fielding, K. S. (2018). Relationships among conspiratorial
beliefs, conservatism and climate scepticism across nations. Nature Climate Change, 8
(7), 614–620. https://doi.org/10.1038/s41558-018-0157-2
17
Journal of Environmental Psychology 76 (2021) 101652
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation Modeling:
A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/
10705519909540118
Hyland, J. J., Jones, D. L., Parkhill, K. A., Barnes, A. P., & Williams, A. P. (2016).
Farmers’ perceptions of climate change: Identifying types. Agriculture and Human
Values, 33(2), 323–339. https://doi.org/10.1007/s10460-015-9608-9
Jones, C., Hine, D. W., & Marks, A. D. G. (2017). The future is now: Reducing
psychological distance to increase public engagement with climate change. Risk
Analysis, 37(2), 331–341.
Kaiser, F. G., & Wilson, M. (2004). Goal-directed conservation behavior: The specific
composition of a general performance. Personality and Individual Differences, 36(7),
1531–1544. https://doi.org/10.1016/j.paid.2003.06.003
Kellstedt, P. M., Zahran, S., & Vedlitz, A. (2008). Personal efficacy, the information
environment, and attitudes toward global warming and climate change in the United
States. Risk Analysis, 28(1), 113–126.
Kittipongvises, S., & Mino, T. (2013). The influence of psychological factors on global
climate change perceptions held by the rural citizens of Thailand. Ecopsychology, 5
(2), 126–135. https://doi.org/10.1089/eco.2012.0080
Leiserowitz, A., Maibach, E., & Roser-Renouf, C. (2009). Global warming’s six America’s.
Yale Program on Climate Change Communication.
Li, S., Juhász-Horváth, L., Harrison, P. A., Pintér, L., & Rounsevell, M. D. A. (2017).
Relating farmer’s perceptions of climate change risk to adaptation behaviour in
Hungary. Journal of Environmental Management, 185, 21–30.
van der Linden, S. (2017). Determinants and measurement of climate change risk
perception, worry, and concern. In S. van der Linden (Ed.), Oxford research
encyclopedia of climate science. Oxford University Press. https://doi.org/10.1093/
acrefore/9780190228620.013.318.
van der Linden, S., Maibach, E., & Leiserowitz, A. (2015). Improving public engagement
with climate change: Five “best practice” insights from psychological science.
Perspectives on Psychological Science, 10(6), 758–763. https://doi.org/10.1177/
1745691615598516
Lorenzoni, I., Nicholson-Cole, S., & Whitmarsh, L. (2007). Barriers perceived to engaging
with climate change among the UK public and their policy implications. Global
Environmental Change, 17(3–4), 445–459. https://doi.org/10.1016/j.
gloenvcha.2007.01.004
Mann, M. E. (2014). The hockey stick and the climate wars: Dispatches from the front lines
(Paperback edition). Columbia University Press.
McArdle, J. J., & Woodcock, R. W. (1997). Expanding test-retest designs to include
developmental time-lag components. Psychological Methods, 2(4), 403–435.
McCright, A. M. (2010). The effects of gender on climate change knowledge and concern
in the American public. Population and Environment, 32(1), 66–87.
McCright, A. M. (2016). Anti-reflexivity and climate change skepticism in the US general
public. Human Ecology Review, 22(2). https://doi.org/10.22459/HER.22.02.2016.04
McCright, A. M., & Dunlap, R. E. (2011). The politicization of climate change and
polarization in the American public’s views of global warming, 2001–2010. The
Sociological Quarterly, 52(2), 155–194.
McDonald, R. I., Chai, H. Y., & Newell, B. R. (2015). Personal experience and the
‘psychological distance’ of climate change: An integrative review. Journal of
Environmental Psychology, 44, 109–118.
Meade, A. W., & Craig, S. B. (2012). Identifying careless responses in survey data.
Psychological Methods, 17(3), 437–455.
Milfont, T. L., Milojev, P., Greaves, L. M., & Sibley, C. G. (2015). Socio-structural and
psychological foundations of climate change beliefs. New Zealand Journal of
Psychology, 44(1), 17–30.
Milman, O., & Harvey, F. (2019). US is hotbed of climate change denial, major global survey
finds. The Guardian. https://www.theguardian.com/environment/2019/may/07/u
s-hotbed-climate-change-denial-international-poll.
Motta, M., Chapman, D., Stecula, D., & Haglin, K. (2019). An experimental examination
of measurement disparities in public climate change beliefs. Climatic Change, 154
(1–2), 37–47.
O’Connor, R. E., Bard, R. J., & Fisher, A. (1999). Risk perceptions, general environmental
beliefs, and willingness to address climate change. Risk Analysis, 19(3), 461–471.
Paek, H.-J., & Hove, T. (2017). Risk perceptions and risk characteristics. In H.-J. Paek, &
T. Hove (Eds.), Oxford research encyclopedia of communication. Oxford University
Press. https://doi.org/10.1093/acrefore/9780190228613.013.283.
Panayides, P. (2013). Coefficient alpha: Interpret with caution. Europe’s Journal of
Psychology, 9(4), 687–696.
Perlaviciute, G., & Squintani, L. (2020). Public participation in climate policy making:
Toward reconciling public preferences and legal frameworks. One Earth, 2(4),
341–348. https://doi.org/10.1016/j.oneear.2020.03.009
Poortinga, W., Spence, A., Whitmarsh, L., Capstick, S., & Pidgeon, N. F. (2011). Uncertain
climate: An investigation into public scepticism about anthropogenic climate
change. Global Environmental Change, 21(3), 1015–1024.
Poortinga, W., Whitmarsh, L., Steg, L., Böhm, G., & Fisher, S. (2019). Climate change
perceptions and their individual-level determinants: A cross-European analysis.
Global Environmental Change, 55, 25–35. https://doi.org/10.1016/j.
gloenvcha.2019.01.007
Postmes, T., Haslam, S. A., & Jans, L. (2013). A single-item measure of social
identification: Reliability, validity, and utility. British Journal of Social Psychology, 52
(4), 597–617. https://doi.org/10.1111/bjso.12006
Putnick, D. L., & Bornstein, M. H. (2016). Measurement invariance conventions and
reporting: The state of the art and future directions for psychological research.
Developmental Review, 41, 71–90. https://doi.org/10.1016/j.dr.2016.06.004
R Core Team. (2020). R: A language and environment for statistical computing (4.0.3.)
[Computer software]. R Foundation for Statistical Computing.
A.M. van Valkengoed et al.
Rahmstorf, S. (2004). The climate sceptics. Potsdam Institute for Climate Impact Research.
www.pik-potsdam.de/stefan/Publications/Other/rahmstorf_climate_sceptics_2004.
pdf.
Rosenthal, R., & Rosnow, R. (2007). Essentials of behavioural research: Methods and data
analysis. McGraw-Hill.
Rosentrater, L. D., Sælensminde, I., Ekström, F., Böhm, G., Bostrom, A., Hanss, D., et al.
(2013). Efficacy trade-offs in individuals’ support for climate change policies.
Environment and Behavior, 45(8), 935–970. https://doi.org/10.1177/
0013916512450510
Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of
Statistical Software, 48(2). https://doi.org/10.18637/jss.v048.i02
Safi, A. S., Smith, W. J., & Liu, Z. (2012). Rural Nevada and climate change:
Vulnerability, beliefs, and risk perception: Vulnerability and climate change. Risk
Analysis, 32(6), 1041–1059. https://doi.org/10.1111/j.1539-6924.2012.01836.x
Shwom, R., Bidwell, D., Dan, A., & Dietz, T. (2010). Understanding U.S. public support
for domestic climate change policies. Global Environmental Change, 20(3), 472–482.
https://doi.org/10.1016/j.gloenvcha.2010.02.003
Siegrist, M., & Sütterlin, B. (2014). Human and nature-caused hazards: The affect
heuristic causes biased decisions: Human and nature-caused hazards. Risk Analysis,
34(8), 1482–1494. https://doi.org/10.1111/risa.12179
Singh, A. S., Zwickle, A., Bruskotter, J. T., & Wilson, R. (2017). The perceived
psychological distance of climate change impacts and its influence on support for
adaptation policy. Environmental Science & Policy, 73, 93–99. https://doi.org/
10.1016/j.envsci.2017.04.011
Spence, A., Poortinga, W., & Pidgeon, N. (2012). The psychological distance of climate
change. Risk Analysis, 32(6), 957–972. https://doi.org/10.1111/j.1539-
6924.2011.01695.x
Steg, L. (2018). Limiting climate change requires research on climate action. Nature
Climate Change, 8(9), 759–761. https://doi.org/10.1038/s41558-018-0269-8
Steg, L., De Groot, J. I. M., Dreijerink, L., Abrahamse, W., & Siero, F. (2011). General
antecedents of personal norms, policy acceptability, and intentions: The role of
values, worldviews, and environmental concern. Society & Natural Resources, 24,
349–367.
Steg, L., Perlaviciute, G., van der Werff, E., & Lurvink, J. (2014). The significance of
hedonic values for environmentally relevant attitudes, preferences, and actions.
Environment and Behavior, 46(2), 163–192. https://doi.org/10.1177/
0013916512454730
Steg, L., & Vlek, C. (2009). Encouraging pro-environmental behaviour: An integrative
review and research agenda. Journal of Environmental Psychology, 29(3), 309–317.
https://doi.org/10.1016/j.jenvp.2008.10.004
18
Journal of Environmental Psychology 76 (2021) 101652
Streiner, D. L. (2003). Starting at the beginning: An introduction to coefficient alpha and
internal consistency. Journal of Personality Assessment, 80(1), 99–103. https://doi.
org/10.1207/S15327752JPA8001_18
Stuive, I., Kiers, H. A. L., & Timmerman, M. E. (2009). Comparison of methods for
adjusting incorrect assignments of items to subtests: Oblique multiple group method
versus confirmatory common factor method. Educational and Psychological
Measurement, 69(6), 948–965. https://doi.org/10.1177/0013164409332226
Tabachnick, B. G., & Fidell, L. S. (2013). Using Multivariate Statistics. Pearson.
Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance.
Psychological Review, 117(2), 440–463. https://doi.org/10.1037/a0018963
United States Census Bureau. (2018). QuickFacts: United States. https://www.census.
gov/quickfacts/fact/table/US/PST045217.
van Valkengoed, A. M., & Steg, L. (2019). The psychology of climate change adaptation (1st
ed.). Cambridge University Press. https://doi.org/10.1017/9781108595438
Van Rensburg, W. (2015). Climate change scepticism: A conceptual Re-evaluation. SAGE
Open, 5(2). https://doi.org/10.1177/2158244015579723, 215824401557972.
Walton, T. N., & Jones, R. E. (2018). Ecological identity: The development and
assessment of a measurement scale. Environment and Behavior, 50(6), 657–689.
https://doi.org/10.1177/0013916517710310
Wamsler, C., & Brink, E. (2014). Interfacing citizens’ and institutions’ practice and
responsibilities for climate change adaptation. Urban Climate, 7, 64–91. https://doi.
org/10.1016/j.uclim.2013.10.009
Weber, E. U. (2010). What shapes perceptions of climate change?: What shapes
perceptions of climate change? Wiley Interdisciplinary Reviews: Climate Change, 1(3),
332–342. https://doi.org/10.1002/wcc.41
Weber, E. U. (2016). What shapes perceptions of climate change? New research since
2010: What shapes perceptions of climate change? Wiley Interdisciplinary Reviews:
Climate Change, 7(1), 125–134. https://doi.org/10.1002/wcc.377
Whitmarsh, L. (2008). Are flood victims more concerned about climate change than
other people? The role of direct experience in risk perception and behavioural
response. Journal of Risk Research, 11(3), 351–374. https://doi.org/10.1080/
13669870701552235
Whitmarsh, L. (2009). Behavioural responses to climate change: Asymmetry of intentions
and impacts. Journal of Environmental Psychology, 29(1), 13–23. https://doi.org/
10.1016/j.jenvp.2008.05.003
Woods, B. A., Ørsted Nielsen, H., Branth Pedersen, A., & Kristofersson, D. (2017).
Farmers’ perceptions of climate change and their likely responses in Danish
agriculture. Land Use Policy, 65, 109–120.
Zawadzki, S. J., Bouman, T., Steg, L., Bojarskich, V., & Druen, P. B. (2020). Translating
climate beliefs into action in a changing political landscape. Climatic Change, 161(1),
21–42. https://doi.org/10.1007/s10584-020-02739-w