University of Groningen

Climate change perceptions and their individual-level determinants

Poortinga, Wouter; Whitmarsh, Lorraine; Steg, Linda; Böhm, Gisela; Fisher, Stephen

Published in:
Global Environmental Change

DOI:
10.1016/j.gloenvcha.2019.01.007

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.

Document Version
Publisher's PDF, also known as Version of record

Publication date:
2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

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

Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.

Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.

Download date: 27-12-2022

 Global Environmental Change 55 (2019) 25–35

Contents lists available at ScienceDirect

Global Environmental Change

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

Climate change perceptions and their individual-level determinants: A cross- T

European analysis
Wouter Poortingaa,b, , Lorraine Whitmarshb, Linda Stegc, Gisela Böhmd,e, Stephen Fisherf
⁎

a
Welsh School of Architecture, Cardiff University, Cardiff, Wales, United Kingdom
b
School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom
c
Department of Psychology, Faculty of Behavioural and Social Sciences, University of Groningen, Groningen, the Netherlands
d
Department of Psychosocial Science, Faculty of Psychology, University of Bergen, Bergen, Norway
e
Department of Psychology, Inland Norway University of Applied Sciences, Lillehammer, Norway
f
Department of Sociology, University of Oxford, Oxford, United Kingdom

ARTICLE INFO ABSTRACT

Keywords: There is now an extensive literature on the question of how individual-level factors affect climate change per-
Climate change ceptions, showing that socio-political variables, notably values, worldviews and political orientation, are key
Perceptions factors alongside demographic variables. Yet little is known about cross-national differences in these effects, as
European social survey most studies have been conducted in a single or small number of countries and cross-study comparisons are
Cross-national analysis
difficult due to different conceptualisations of key climate change dimensions. Using data from the European
Social Survey Round 8 (n = 44,387), we examine how key socio-political and demographic factors are associated
with climate change perception across 22 European countries and Israel. We show that human values and po-
litical orientation are important predictors of climate change beliefs and concern, as are the demographics of
gender, age, and education. Certain associations with climate change perceptions, such as the ones for the self-
transcendence versus self-enhancement value dimension, political orientation, and education, are more con-
sistent across countries than for gender and age. However, even if the direction of the associations are to a large
extent consistent, the sizes of the effects are not. We demonstrate that the sizes of the effects are generally
smaller in Central and Eastern European countries, and that some demographic effects are larger in Northern
European as compared to Western European countries. This suggests that findings from one country do not
always generalize to other national contexts.

1. Introduction The now extensive literature on climate change perceptions has

contributed to a better insight into how different individuals perceive
1.1. Background and engage with climate change. Many studies on the topic have fo-
cused on individual-level factors in people’s beliefs and concerns about
Public perceptions of climate change have been extensively studied climate change. This research appears to show a consistent pattern
over the past two to three decades (Capstick et al., 2015). This research across different demographic groups. In particular, the research shows
is conducted on the understanding that climate change perceptions are that men, older age groups, and those with fewer years of formal
critical to public engagement and support for action on climate change education tend to be more doubtful about the reality and anthropogenic
(Bord et al., 2000; Corner et al., 2014). Ambitious targets, such as set nature of climate change, reflecting trend and attribution scepticism,
out in the Paris agreement (UNFCCC, 2017) and the European Com- respectively (Milfont et al., 2015; Poortinga et al., 2011), and that they
mission’s 2030 energy strategy (European Commission, 2014), require are less concerned about the impacts of climate change (Shi et al., 2016;
fundamental shifts in the way energy is used and produced to mitigate Whitmarsh, 2011). Explanations for this patterning include the ‘white
climate change. However, policymakers may be reluctant to take male effect’, showing that white men are generally more accepting of a
meaningful action, if their electorate do not think that climate change is range of environmental and technological risks – with a pattern that is
happening, anthropogenic, or a serious threat. distinct from almost any other demographic group (Finucane et al.,

⁎
Corresponding author at: Welsh School of Architecture, Cardiff University. Bute Building, King Edward VII Avenue, Cardiff, Wales, CF10 3NB, United Kingdom.
E-mail address: PoortingaW@cardiff.ac.uk (W. Poortinga).
@wouterpoortinga (W. Poortinga)

https://doi.org/10.1016/j.gloenvcha.2019.01.007
Received 9 May 2018; Received in revised form 14 December 2018; Accepted 23 January 2019
0959-3780/ © 2019 Elsevier Ltd. All rights reserved.
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

2000). This may reflect societal inequalities, as well as differences in 2005; Corner et al., 2014; Kahan et al., 2011; Poortinga et al., 2004).
the subjective experience of vulnerability in relation to these risks Values are understood as guiding principles in life, and are considered
(Satterfield et al., 2004). The effects may however be contingent on the an important part of what forms our beliefs and attitudes towards social
particular type of environmental risk concern (Hayes, 2001), and issues, including climate change (Milfont et al., 2015). This helps to
gender differences may not exist in relation to generic environmental shape the way we behave in relation to the environment (De Groot and
concern (Echavarren, 2017). Some scholars have pointed to the role of Steg, 2008; Poortinga et al., 2004; Stern, 2000). Values’ associations
conservative values (the ‘conservative male effect’) in combination with with climate-relevant attitudes and behaviour have mostly been studied
identity-protective cognition (Kahan et al., 2007; McCright and Dunlap, using either Schwartz’ theory of basic human values, which arranges
2013), whereby lower levels of risk perception, including those for ten distinct clusters across the two axes of conservation versus openness-
climate change, indicate a motivation to maintain prevailing social to-change and self-transcendence versus self-enhancement (Schwartz,
structures (Jylhä and Akrami, 2015; Jylhä et al., 2016). Climate scep- 1992), or the altruistic, egoistic, and biospheric trio of values derived from
ticism appears to be particularly common among politically con- the Value-Belief-Norm (VBN) model (Dietz et al., 2007; Stern, 2000).
servative men in a number of countries (McCright and Dunlap, 2011; Research has consistently shown that people who endorse self-trans-
Milfont et al., 2015; Whitmarsh, 2011); and there is evidence that cending (or: altruistic) values have higher levels of concern and are less
gender difference are only modest when key beliefs and values are likely to be sceptical about anthropogenic climate change (Brown and
taken into account (McCright, 2010). Kasser, 2005; Corner et al., 2014; De Groot and Steg, 2007; Poortinga
Age effects in climate change perceptions have been found con- et al., 2004), while the opposite is generally (if not always) true for self-
sistently across a large number of countries (e.g. Echavarren, 2017; enhancement (or: egoistic) values (Steg and De Groot, 2012). While
McCright, 2010; Milfont et al., 2015; Poortinga et al., 2011). Age ef- multiple studies have focused on the self-transcendence and self-en-
fects, just as gender effects, have been explained by differences in hancement value dimension, the role of the conservation and openness-
motivation to maintain prevailing social structures. Older people are to-change value dimensions in climate change perceptions has been
more integrated into existing social orders, and therefore may have explored less. There are indications that individuals who hold openness-
more to lose by changes that are needed to address environmental to-change values have stronger beliefs in the reality of climate change
challenges such as climate change. Age differences may also be ex- and its human cause (Milfont et al., 2015), and that those endorsing
plained by climate change having been a threat and/or part of the conservation values are less likely to be concerned about or willing to
formal education (Stevenson et al., 2014) when older age cohorts were make changes for the environment (Schultz and Zelezny, 1999; Stern
growing up. Furthermore, value orientations may change over the et al., 1998). Overall, the effects for openness-to-change and con-
lifecourse, which may have implications for how one feels about cli- servation values appear weaker than for self-transcendence and self-
mate change. There is evidence that people become more (politically) enhancement values, and a few studies found non-significant relation-
conservative as they age (Cornelis et al., 2009); and political values are ships with these dimensions (Milfont et al., 2015; Poortinga et al., 2004;
among the strongest socio-political determinants of climate change Steg and De Groot, 2012).
perceptions (see below) The question remains whether these findings can be generalised to
Education effects, i.e. of people with longer formal education ex- different countries or cultural contexts. Notably, most studies that have
pressing higher levels of concern about the environment in general and examined individual-level factors in relation to climate change per-
climate change in particular (Marquart-Pyatt, 2008; O’Connor et al., ceptions have been conducted in a single or a small number of coun-
1999), are interpreted in multiple ways. It is often implicitly assumed tries; and it is difficult to compare studies due to the use of different
that education is a proxy for knowledge or a better understanding of the measures and conceptualisations of key climate change dimensions. A
scientific underpinnings of climate change. However, climate change recent meta-analysis showed that individual-level effects were sig-
and/or scientific knowledge itself tends to be a poor predictor of cli- nificantly moderated by the type of measure used (Hornsey et al.,
mate change beliefs (Whitmarsh, 2011), with climate sceptics being 2016). There are indications that the importance of different demo-
generally as knowledgeable as non-sceptics (Hornsey et al., 2016). graphic and socio-political values in predicting climate change per-
Knowledge may even have diverging effects depending on people’s ceptions may vary cross-nationally independent of the type of measure.
political orientation (Kahan et al., 2012; Lee et al., 2015; Malka et al., For example, Shi et al. found that gender, age and education do not
2009), suggesting that a lack of belief or concern about climate change predict concern about climate change to the same extent in six different
may not be due to a deficit in knowledge (Snow and Dibners, 2016). countries (Shi et al., 2016). Similarly, political ideology has been shown
Socio-economic effects are further interpreted in reference to post- to predicts climate change beliefs in certain countries but not in others
materialism theory (Inglehart, 1990). Individuals who have satisfied (Capstick et al., 2015). Whereas McCright et al. (2016) found that po-
their basic material needs may shift their attention to more post- litical ideology is associated with public views on climate change across
materialist ones, such as freedom, quality of life and environmental multiple Western European countries, the effects are not as pronounced
protection (Fransson and Garling, 1999; Knight, 2016). While, overall, as in the US. Furthermore, non-significant effects were found for poli-
there are clear indications that the demographics of gender, age, and tical affiliation in former communist countries (McCright et al., 2016).
education are all important factors in climate change perceptions, it is There are suggestions that climate scepticism in the media is pre-
not known whether these effects are universal or that they vary across dominantly an Anglophone phenomenon (Painter and Ashe, 2012), and
countries. it can be expected that polarisation is the greatest in countries where
Strong associations of climate change perceptions have also been there is a political home for climate sceptical views through continued
found with a range of socio-political variables, such as political or- media attention and political representation (Dunlap and McCright,
ientation, human values and worldviews (Hornsey et al., 2016). It is 2011; Engels et al., 2013; Milfont et al., 2015; Poortinga et al., 2011;
well established that public views on climate change are divided along Tranter and Booth, 2015).
party-political lines in the US (Hoffman, 2011; McCright and Dunlap, Marquart-Pyatt (2008) concluded that the individual-level sources
2011; McCright et al., 2014); and there are indications that political for environmental concern, including demographics and knowledge,
orientation may be an important factor in other countries as well, in are largely consistent across nineteen industrialised countries; although
particular in Anglophone countries (Fielding et al., 2012; Milfont et al., there were some differences between them. In particular, a number of
2015; Poortinga et al., 2011). It is however not clear whether possible coefficients appeared different in former communist countries as com-
political divides in other countries are as pronounced as in the US. pared to advanced industrialised countries (Marquart-Pyatt, 2008). The
The role of human values and cultural worldviews in climate change study focused on general environmental concern, which may be less
perceptions has also attracted widespread attention (Brown and Kasser, politicised than attitudes to climate change. A recent meta-analysis

26
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

found high levels of variation in the strength of individual-level effects (European Social Survey (2015); Fitzgerald, 2015). Each country
across studies (Hornsey et al., 2016), and particularly revealed sig- needed to achieve a minimum effective sample size of 1500 (or 800 in
nificant differences in effect sizes between US and non-US samples. Yet, countries with populations smaller than 2 million), representative of
they did not explicitly examine cross-country variation. Moreover, the the resident population. Data collection involved strict random prob-
studies that were included in the meta-analysis were sourced from a ability sampling to obtain nationally-representative samples, and an
large number of countries, and used data from diverse representative extensive concept-based design process to ensure measurement
and non-representative samples that were collected at different time equivalence (Fitzgerald and Jowell, 2010). Interviews were conducted
periods, making it difficult to draw firm conclusions about possible face-to-face in respondents’ own homes with residents aged 15 years
country differences. Furthermore, the studies included a wide variety of and over. The sample sizes for the 23 countries are provided in Table 1.
measures reflecting different aspects of climate change perceptions. It is In total, 44,387 participants were available for the analyses. Data were
therefore possible that the reported variation in individual-level effects collected, usually within three-month, in the period from August 2016
is attributable to methodological (e.g. specific outcome measure or to December 2017. Post-stratification weight were used to take account
sampling strategy used) or contextual (e.g. country and period in which of unequal probabilities of selection, as well as of sampling and non-
study was conducted) differences between the different studies. There is response error. The full questionnaire and the complete European So-
thus a clear need for systematic international comparisons to better cial Survey Round 8 dataset can be downloaded from http://www.
understand the importance of individual factors for climate change europeansocialsurvey.org.
perceptions in different national contexts (Hopkins, 2015).
2.2. Measures
1.2. Aims of this paper
2.2.1. Dependent variables (climate change perceptions)
In this paper we make use of the European Social Survey Round 8 Climate change beliefs. Three questions were asked to assess
(European Social Survey, 2016) to examine how individual-level de- people’s beliefs regarding the existence, causes, and consequences of
mographic and socio-political factors are linked to climate change climate change, respectively. Trend scepticism was determined by asking
perceptions. In particular, we will examine levels of climate change respondents “You may have heard the idea that the world’s climate is
perceptions and their determinants in 22 European countries and Israel. changing due to increases in temperature over the past 100 years. What
The focus of the paper is on four dimensions of climate change percep- is your personal opinion on this? Do you think the world’s climate is
tions: public beliefs about the existence, causes, and consequences of changing?” Respondents could use the options: definitely not changing,
climate change, and climate change concern. We distinguish between probably not changing, probably changing, and definitely changing.
climate change beliefs, defined as propositional cognitions about the The 4-point response scale was dichotomised to 0 (probably/definitely
nature of climate change that may or may not correspond with reality changing) and 1 (probably/definitely not changing). Attribution scepti-
(i.e. beliefs regarding the reality, causes, and impacts of climate change, cism was assessed with the question “Do you think that climate change
which are often used to identify trend, attribution, and impact sceptical is caused by natural processes, human activity, or both?” The responses
views; Poortinga et al., 2011), and climate concern, defined as affective were coded as 1 (entirely/mainly by natural processes) and 0 (entirely/
evaluations of the seriousness of (the impacts of) climate change, in- mainly by human activity/about equally by natural processes and
dicated by personal feelings of worry about the issue (cf. Lo and Chow, human activity). The non-prompted option of “I don’t think climate
2015). We collectively refer to climate change beliefs and concern as cli- change is happening” was coded as missing to avoid overlap with trend
mate change perceptions. sceptical beliefs. Perceived impacts of climate change: respondents were
The paper has four aims. First, it will examine national differences in asked to indicate how good or bad they thought the impact of climate
climate change perceptions across the 23 countries. Second, it will ex- change would be on people across the world, on a scale from -5 (ex-
plore associations of different individual-level socio-political and de- tremely bad) and +5 (extremely good).
mographic predictors with climate change perceptions across all Climate concern was assessed by asking respondents “How worried
countries. Third, it will assess cross-national differences in the strength are you about climate change?” with the response options of 1 (not at
of the relationships between these individual-level predictors and cli- all worried), 2 (not very worried), 3 (somewhat worried), 4 (very
mate change perceptions, that is, whether the sizes of the regression worried), and 5 (extremely worried).
coefficients of individual-level socio-political and demographic pre-
dictors differ between countries. Fourth, the paper will explore whether 2.2.2. Independent variables (socio-political and demographic variables)
there are systematic differences in individual-level effects between Human values. A modified 21-item version of the Portrait Values
different European regions, namely Western, Central and Eastern, Questionnaire (PVQ) was used to measure peoples’ values (Schwartz,
Southern, and Northern European countries. By using high-quality, 2003). Each item consists of a short two-sentence, gender-matched
standardised measures of the key variables of interest, and coordinated description of a person. Respondents then indicate on a 6-point scale
data collection according to the highest methodological standards, the from 1 (very much like me) to 6 (not like me at all) how similar this
study is able to exclude methodological sources of variation. person is to themselves. The Schwartz (2015) approach was used to
transform the items into 10 values. Universalism, Benevolence,
2. Methods Achievement (reversed) and Power (reversed) values were subse-
quently combined into an internally consistent Self-transcendence vs.
2.1. The European social survey Self-enhancement dimension (α = 0.65), and Conformity, Security, Sti-
mulation (reversed) and Hedonism (reversed) values into an internally
The European Social Survey (ESS) is a biennial pan-European survey consistent Conservation vs. Openness-to-change dimension (α = 0.67).
that has been conducted since 2002. Each round contains two modules Higher positive values correspond to more self-transcendence and more
on key social themes. Round 8 of the ESS (European Social Survey, openness-to-change values, relative to self enhancement and openness-
2016) included a module on Climate and Energy, designed by the au- to-change respectively. The two value scales were standardised by
thors together with ESS headquarters and national coordinating teams. calculating the Z scores across all countries.
The ESS has a number of methodological standards regarding ques- Political orientation involved self-placement on a 10-point scale
tionnaire design, translation and data collection. The questionnaire was ranging from 0 (left) to 10 (right). The question read: “In politics people
developed in English through a two-year design process, which in- sometimes talk of ‘left’ and ‘right’. Using this card, where would you
cluded extensive testing, piloting and translation by national teams place yourself on this scale, where 0 means the left and 10 means the

27
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

Table 1
Mean scores and standard deviations for the four climate change perception variables.
(1)
Country Region Sample size Trend Scepticism Attribution scepticism(2) Perceived impacts of climate change(3) Concern about climate change(4)

N % % M (SD) M (SD)
Austria W 2,010 7.3 7.9 −1.75 (2.21) 3.07 (0.90)
Belgium W 1,766 3.6 5.9 −1.64 (2.34) 3.17 (0.86)
Czech Republic CE 2,269 10.7 9.9 −1.57 (1.98) 2.77 (1.05)
Estonia CE 2,019 8.6 10.8 −1.40 (2.07) 2.65 (0.95)
Finland N 1,925 5.9 6.0 −1.49 (2.00) 3.05 (0.82)
France W 2,070 3.7 6.1 −1.99 (2.21) 3.21 (0.93)
Germany W 2,852 4.5 5.1 −2.04 (1.93) 3.36 (0.85)
Hungary CE 1,614 8.2 6.9 −2.26 (2.06) 3.05 (0.85)
Iceland N 880 2.3 5.3 −2.28 (1.96) 3.13 (0.92)
Ireland W 2,757 3.8 8.4 −1.46 (2.37) 2.83 (0.92)
Israel – 2,557 12.0 11.7 −1.07 (2.65) 2.64 (1.05)
Italy S 2,626 5.0 6.1 −1.74 (2.34) 3.21 (0.84)
Lithuania CE 2,122 11.0 15.4 −1.69 (2.04) 2.82 (0.91)
Netherlands W 1,681 3.7 8.0 −1.13 (2.07) 3.01 (0.86)
Norway N 1,545 7.1 12.0 −1.64 (1.97) 3.00 (0.83)
Poland CE 1,694 7.1 9.7 −1.68 (2.06) 2.75 (0.86)
Portugal S 1,270 3.0 6.2 −2.55 (2.32) 3.48 (0.92)
Russian Federation CE 2,430 16.5 12.7 −1.42 (2.16) 2.75 (0.97)
Slovenia CE 1,307 3.5 6.9 −1.69 (2.30) 3.17 (0.86)
Spain S 1,958 4.1 4.0 −2.90 (2.00) 3.42 (0.88)
Sweden N 1,551 3.2 7.5 −2.11 (1.93) 2.86 (0.87)
Switzerland W 1,525 3.5 5.4 −1.80 (2.10) 3.12 (0.85)
United Kingdom W 1,959 6.4 8.8 −1.46 (2.24) 2.96 (0.94)

Note: Post-stratification weights have been applied for country-level analysis; W = Western Europe; N = Northern Europe; CE = Central and Eastern Europe;
S = Southern Europe (1) Coding: 0 probably/definitely changing, 1 probably/definitely not changing; (2) coding: 0 entirely/mainly by human activity/about equally
by natural processes and human activity, 1 entirely/mainly by natural processes; (3) coding: scale from -5 extremely bad to +5 extremely good; (4) coding: scale from
1 not at all worried to 5 extremely worried.

right?” The political orientation variable was standardised by calcu- were used to estimate the overall associations of the independent
lating Z scores across all countries. variables with the four climate change perception dimensions across
Demographics. Gender was indicated as 0 (female) and 1 (male). the 23 countries, while allowing the countries to vary with respect to
The age variable was centred on its grand mean of 47.64, and expressed their average level on the dependent variable in the respective regres-
in 10 year deviations from that mean. Level of education was indicated sion model. Second, a series of random intercept, random slope models
by the ESS version of the International Standard Classification of were constructed for the four dependent climate change belief and
Education (ISCED). The level of education variable was centred on its concern variables. This means that Model 1 was extended by allowing
grand mean of 4.14. the slopes of the independent variables to vary across countries (Model
European regions. The European Regions included Western Europe 2). Separate regression analyses were conducted for each of the six
(Austria, Belgium, Switzerland, Germany, France, United Kingdom, independent variables in their associations with the four climate change
Ireland, and the Netherlands), Central and Eastern Europe (Czech perception dimensions. That is, all independent variables were in-
Republic, Estonia, Hungary, Lithuania, Poland, the Russian Federation, cluded, but only one slope was allowed to vary in each regression
and Slovenia), Southern Europe (Spain, Italy, and Portugal), and model. This approach was chosen, as the number of countries involved
Northern Europe (Finland, Iceland, Norway, and Sweden). This dis- is insufficient to reliably estimate all parameters simultaneously. Only
tinction was made, as previous research has predominantly been con- the random effects of the multilevel regression analyses are reported.
ducted in Western and Northern European countries, and there are The random effects indicate the cross-country variation in the strength
indications that differences in climate change perceptions across dif- of the association between the individual-level socio-political and de-
ferent socio-political and demographic groups are smaller in former mographic variables on the one hand and the climate change belief and
communist Central and Eastern European countries (Marquart-Pyatt, concern variables on the other. Third, a series of analyses was con-
2012; McCright et al., 2016). Israel was excluded from the region ducted to examine whether there are any systematic differences be-
comparisons as the only non-European country involved in the ESS tween countries from different European regions. This was done by
(Israel was included in all the other analyses). adding the regions as dummy variables (Model 3a), and subsequently
their interactions with the socio-political and demographic variables
2.3. Data analysis (Model 3b). Separate regression analyses were conducted for each of
the six independent variables in their associations with the four climate
Multilevel modelling was used to analyse the data, utilising the change perception dimensions. This means that each multilevel re-
MLwiN 2.36 software package, with individuals (level 1) nested within gression model had three dummies indicating Central and Eastern
countries (level 2). Logistic models were constructed for trend and at- European, Southern European, and Northern European countries, re-
tribution scepticism, and linear models for the perceived impacts of spectively (using Western European countries as the reference cate-
climate change and concern about climate change. Three sets of ana- gory), and three interaction terms of Central and Eastern European,
lyses were conducted, in addition to the descriptive statistics to assess Southern, and Northern European countries with the respective socio-
national differences in climate change perceptions. First, we constructed political and demographic factors. Only the interaction effects are re-
a series of random intercept multilevel regression models that included ported. The interaction effects indicate the extent to which the in-
human values, political orientation, and demographics of gender, age, dividual-level effects in those regions differ from the ones found in the
and level of education as predictors (Model 1). Only the fixed effects of Western European countries.
the multilevel regression models are reported. Hence, these models

28
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

3. Results The results demonstrate that men were more likely to have trend
and attribution sceptical beliefs across the 23 countries, and generally
3.1. National differences in climate change perceptions had lower levels of concern about climate change than women. In
contrast to these findings, men perceived the impacts of climate change
In line with several other studies (see e.g. Capstick et al., 2015 for to be more negative than women did. Furthermore, older respondents
an overview), we find that levels of trend and attribution scepticism are were more likely to have trend or attribution sceptical views, perceived
low in most countries (see Table 1). This means that an overwhelming the impacts of climate change to be less negative, and had lower levels
majority of the European population thinks that climate change is of concern about climate change than younger respondents. Finally, the
happening and is at least partly caused by human activity. However, results show that level of education was negatively associated with
there are substantial differences across the participating countries. trend and attribution sceptical beliefs. Respondents with higher levels
Trend scepticism ranged from 2.3% in Iceland to 16.5% in the Russian of education also perceived the impacts of climate change to be more
Federation, and attribution scepticism from 4.0% in Spain to 15.4% in negative, and had higher levels of concern about climate change.
Lithuania. Attribution scepticism in Norway (12.0%) was surprisingly
high, given that it has a mid-sized level of trend scepticism (7.1%) and a
just-above average level of concern about climate change (see below). 3.3. Cross-national differences in the strength of effects
On average, the perceived impacts of climate change were seen to be
negative in all participating countries, and ranged from -1.07 in Israel In order to investigate cross-national variation in the strength of
to -2.55 in Portugal, suggesting that most people think that the impacts individual-level effects, we conducted a series of random intercept,
of climate change around Europe (and Israel) will only be slightly ne- random slope multilevel regression analyses, in which not only the in-
gative. Average levels of concern ranged from 2.64 in Israel and 2.65 in tercepts but also slopes of the regression coefficients were allowed to
Estonia to 3.42 in Spain and 3.48 in Portugal. This means that in all vary across countries (Model 2). Table 3 presents the cross-country
countries concern hovered around the scale midpoint of 3, which variation (σ²) in the strength of the associations between the individual-
equates to “somewhat worried”. These differing results show the im- level socio-political and demographic variables on the one hand and the
portance of distinguishing between different types of climate change climate change perception dimensions on the other. It shows that the
beliefs and concern. cross-country variation in the associations of the self-transcendence vs
self-enhancement value dimension with trend scepticism, attribution
scepticism and the perceived impacts of climate change were sig-
3.2. Individual-level effects of climate change perceptions nificant, but not with concern about climate change. All of the asso-
ciations of the conservation vs. openness-to-change value dimension
We subsequently explored the associations of the individual-level varied significantly across the 23 countries. For political orientation,
socio-political and demographic factors with the four climate change the cross-country variation in the associations with attribution scepti-
perception dimensions across the 23 countries. This was done with a cism, perceived impacts of climate change and concern about climate
series of random intercept multilevel models, in which the individual- change were significant, but not the one with trend scepticism. The
level factors were included as independent variables (Model 1). The associations of gender with attribution scepticism concern about cli-
models assumed the regression coefficients to be constant but allowed mate change varied significantly across the 23 countries, as did the
the intercepts to vary across the participating countries. This type of association of gender with concern about climate change. The asso-
analysis allows us to explore the overall associations, while taking into ciations of age with trend scepticism, attribution scepticism and the
consideration that the countries differ with respect to their means. As perceived impacts of climate change were significant, but not the one
can be seen in Table 2, the six socio-political and demographic variables with concern about climate change. Similarly, the associations of edu-
were significant predictors of a number or all climate perception di- cation with trend scepticism, attribution scepticism and the perceived
mensions. Individuals who prioritise self-transcendence over self-enhan- impacts of climate change were significant, but not the one with con-
cing values were less likely to have trend or attribution sceptical views cern about climate change.
(as is indicated by odds ratios below 1), perceived the impacts of cli- Figs. 1 and 2 show the country-level regression lines for the socio-
mate change to be more negative, and had higher levels of concern. The political and demographic variables, respectively. The figures visualise
conservation vs openness-to-change value dimension was non-sig- the degree to which the associations vary across the 23 countries. The
nificantly associated with trend and attribution scepticism, but in- raw regression coefficients and their confidence intervals are provided
dividuals prioritising conservation over openness-to-change values in the Supplementary Information document (see Tables A to F) to show
tended to perceive the impacts of climate change as slightly less ne- the strength of the associations in the individual countries. The figures
gative and to have slightly lower levels of concern. Individuals who appear to show that certain associations are more consistent across
placed themselves on the right hand side of the political spectrum were countries than others, and that the cross-national variations in the
more likely to have trend or attribution sceptical views, perceived the strength of individual-level effects sometimes but not always lead to
impacts of climate change to be less negative, and had lower levels of different conclusions regarding their importance. Table A in the sup-
concern. porting information shows that the self-transcendence vs. self-

Table 2
Associations of socio-political and demographic factors with the four climate change perception variables (Model 1).
Trend Scepticism Attribution scepticism Perceived impacts of climate change Concern about climate change

OR (95% CI) OR (95% CI) B (95% CI) B (95% CI)

Self-transcendence (vs. self-enhancement) 0.78 (0.74 to 0.82) *** 0.79 (0.75 to 0.83)*** −0.25 (-0.27 to -0.22)*** 0.11 (0.10 to 0.12)***
Conservation (vs. openness-to-change) 0.97 (0.92 to 1.01)n.s. 0.99 (0.95 to 1.03)n.s. 0.05 (0.02 to 0.07)*** −0.03 (-0.04 to -0.02)***
Political orientation: right (vs. left) 1.11 (1.06 to 1.16)*** 1.18 (1.14 to 1.23)*** 0.22 (0.19 to 0.24)*** −0.08 (-0.09 to -0.07)***
Gender: male (vs female) 1.37 (1.26 to 1.49)*** 1.31 (1.22 to 1.42)*** −0.09 (-0.13 to -0.040)*** −0.09 (-0.10 to -0.07)***
Age 1.06 (1.03 to 1.08)*** 1.12 (1.10 to 1.15)*** 0.11 (0.09 to 0.12)*** −0.03 (-0.04 to -0.02)***
Level of education 0.91 (0.89 to 0.94)*** 0.91 (0.89 to 0.93)*** −0.12 (-0.13 to -0.11)*** 0.04 (0.04 to 0.05)***

n.s.
Note: *** p < 0.001; non-significant ; OR = odds ratio; CI = confidence interval.

29
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

Table 3
Cross-country variation in the associations of the socio-political and demographic factors with the four climate change perception variables (Model 2).
Trend scepticism Attribution scepticism Perceived impacts of climate change Concern about climate change

σ² (SE) σ² (SE) σ² (SE) σ² (SE)

Self-transcendence (vs. self-enhancement) 0.040 (0.020)* 0.016 (0.008)* 0.014 (0.006)* 0.001 (0.001) n.s.
Conservation (vs. openness-to-change) 0.024 (0.012)* 0.013 (0.006) * 0.024 (0.009)** 0.002 (0.001)*
Political orientation: right (vs. left) 0.007 (0.004) n.s. 0.017 (0.008)* 0.011 (0.005)* 0.003 (0.001)**
Gender: male (vs female) 0.021 (0.015) n.s. 0.060 (0.029)* 0.010 (0.006) n.s. 0.008 (0.003)**
Age 0.007 (0.003) * 0.002 (0.001)*· 0.007 (0.003)* 0.001 (0.001) n.s.
Level of education 0.004 (0.002)*· 0.004 (0.002)* 0.005 (0.002)* 0.000 (0.000) n.s.

n.s.
Note: * p < 0.05; ** p < 0.01; non-significant.

enhancement value dimension is consistently and positively associated to be significantly negative, the associations were non-significant in a
with concern about climate change. It is also consistently associated majority of the individual countries, with only a few exceptions.
with the perceived negative impacts of climate change, with only a few The associations of political orientation with the four climate per-
exceptions. While the self-transcendence vs. self-enhancement value ception dimensions were consistent across the 23 countries (Table C).
dimension was generally negatively associated with trend and attribu- Individuals with a right-leaning political orientation were less likely to
tion scepticism, the associations were non-significant in twelve and five perceive negative impacts and to be concerned about climate change in
countries, respectively (Table A). a majority of countries. While a right-leaning political orientation was
With regard to the conservation vs. openness-to-change value dimen- generally positively associated with trend and attribution scepticism,
sion, this factor was non-significantly associated with trend and attri- the associations were non-significant in seven and nine countries, re-
bution scepticism in the overwhelming majority of countries (Table B). spectively.
It was only significantly associated with trend scepticism in the Czech The association of gender with trend scepticism was generally con-
Republic and with attribution scepticism in Israel. However, while the sistent across the 23 countries (Table D). The association was non-sig-
overall association of the factor with the perceived impacts of climate nificant in five countries. The associations of gender with the other
change was non-significant (Table 2), there were a number of countries three climate perception dimensions were less consistent. While gender
in which the association was significantly negative and a number of was positively associated with attribution scepticism overall (Table 2),
countries where the association was significantly positive. Furthermore, the association was non-significant in ten out of the 23 countries. Si-
whereas the overall association of the conservation vs. openness-to- milarly, while gender was negatively associated with concern about
change value dimension with concern about climate change was found climate change overall, the association was non-significant in twelve

Fig. 1. Associations of (1) self-transcendence versus self-enhancement, (2) conservation versus openness-to-change, and (3) political orientation, with (a) trend
scepticism, (b) attribution scepticism, (c) perceived impacts of climate change, and (d) concern about climate change in 23 countries.

30
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

Fig. 2. Associations of (1) gender, (2) age, and (3) level of education, with (a) trend scepticism, (b) attribution scepticism, (c) perceived impacts of climate change,
and (d) concern about climate change in 23 countries.

out of the 23 countries. While men were found to perceive less negative 1.06–2.29) scepticism are more common in Central and Eastern Europe
impacts overall, the associations of gender with the perceived impacts as compared to Western Europe; that the perceived impacts are more
of climate change were non-significant in all but two countries. negative and that concern about climate change is higher in Southern
Age was consistently associated with attribution scepticism: in vir- Europe; and that there are no significant differences between Northern
tually all countries older respondents were more likely to have doubts and Western Europe in any of the four climate perception dimensions.
about the anthropogenic nature of climate change (Table E). However, The interaction effects (Model 3b) indicate the extent to which the
its association with the other three climate perceptions dimensions was individual-level effects in Central and Eastern, Southern and Northern
more variable. In a majority of countries, older respondents were more European countries differ from the ones found in Western European
likely to hold trend sceptical views, to perceive less negative impacts, countries. The interaction effects need to be compared to the regression
and to be less concerned about climate change; but the associations coefficients of the different factors (see Table 4), which reflect their
were non-significant in ten countries for each of the three dimensions. association with the respective climate perception dimensions in Wes-
The association between age and concern was even significantly posi- tern European countries. That is, where the overall regression coeffi-
tive in Lithuania. cient is positive, a negative interaction term generally indicates a
Respondents with a higher level of education were generally less weaker effect and a positive interaction term a stronger effect for that
likely to hold trend and attribution sceptical beliefs, perceived more factor in the region of interest. Reversely, where the overall regression
negative impacts, and were more concerned about climate change coefficient is negative, a negative interaction term generally indicates a
(Table F). These effects were consistent, in particular for attribution stronger effect and a positive interaction term a weaker effect. Where
scepticism and concern about climate change. The associations were the overall regression coefficient is close to zero (e.g. for conservation
non-significant in four and three countries respectively. The results for vs openness-to-change), a negative interaction term may indicate a
trend scepticism and the perceived impacts of climate change were negative effect and a positive interaction term a positive effect for that
somewhat more variable. The associations were non-significant in nine factor in the region of interest.
and eight countries, respectively. Table 4 shows that the effects of the self-transcendence vs. self-en-
hancement value dimension were generally weaker in Central and
3.4. Differences in the strength of effects between European regions Eastern European countries than in Western European countries, as
indicated by the positive interaction terms for attribution scepticism
We subsequently conducted a series of analyses to examine whether and perceived impacts of climate change, and the negative interaction
there are any systematic differences between countries in different term for concern about climate change. The effects of self-transcen-
European regions. Table 4 presents the main (Model 3a) the interaction dence vs. self-enhancement values were also weaker in Northern Eur-
effects (Model 3b) of Central and Eastern, Southern, and Northern opean countries as compared to Western European countries, although
European countries that took part in the ESS. It shows that trend the interaction effects for attribution scepticism and concern about
(OR = 2.05, 95%CI 1.26–3.25) and attribution (OR = 1.56, 95%CI climate change were non-significant. No significant differences were

31
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

Table 4
Interactions of European regions with the socio-political and demographic factors for the four climate change perception variables (Model 3a and 3b).
Trend scepticism Attribution scepticism Perceived impacts of climate Concern about climate
change change

B (95% CI) B (95% CI) B (95% CI) B (95% CI)

Model 3a (main effects)
Region
Central and Eastern Europe 0.717 (0.235 to 1.179)** 0.446 (0.062 to 0.830)* −0.126 (-0.444 to 0.192) n.s. −0.165 (-0.357 to 0.027) n.s.

Southern Europe −0.282 (-0.933 to 0.369) n.s. −0.341 (-0.890 to 0.208) n.s. −0.870 (-1.301 to -0.439)*** 0.348 (0.117 to 0.579)**
Northern Europe 0.067 (-0.527 to 0.661) n.s. 0.143 (-0.384 to 0.670) n.s. −0.115 (-0.515 to 0.285) n.s. −0.152 (-0.338 to 0.034) n.s.

Model 3b (interactions)
Self-transcendence (vs. self-enhancement) −0.284 (-0.372 to −0.281 (-0.354 to -0.208)*** −0.301 (-0.338 to -0.264)*** 0.114 (0.098 to 0.130)***
-0.196)***
Central and Eastern Europe 0.063 (-0.055 to 0.181) n.s. 0.097 (0.009 to 0.185)* 0.098 (0.035 to 0.161)** −0.025 (-0.049 to -0.001)*
Southern Europe −0.135 (-0.341 to 0.071) n.s. 0.041 (-0.131 to 0.213) n.s. −0.059 (-0.141 to 0.023) n.s. 0.029 (0.004 to 0.062) n.s.
Northern Europe 0.183 (0.020 to 0.346)* 0.101 (-0.030 to 0.232) n.s. 0.194 (0.121 to 0.267)*** −0.027 (-0.056 to 0.002) n.s.
Conservation (vs. openness-to-change) −0.034 (-0.017 to 0.085) n.s −0.037 (-0.082 to 0.008) n.s −0.011 (-0.046 to 0.024) n.s −0.031 (-0.05 to -0.017)***
Central and Eastern Europe −0.089 (-0.193 to 0.015) n.s. 0.036 (-0.058 to 0.130) n.s. −0.167 (-0.224 to -0.110)*** 0.023 (-0.001 to 0.047) n.s.
Southern Europe −0.131 (-0.309 to 0.047) n.s. 0.197 (0.044 to 0.350)* −0.153 (-0.224 to -0.082)*** −0.012 (-0.041 to 0.017) n.s.
Northern Europe 0.210 (0.059 to 0.361)** 0.193 (0.070 to 0.316)** 0.146 (0.079 to 0.213)*** −0.003 (-0.030 to 0.024) n.s.
Political orientation: right (vs. left) 0.204 (0.120 to 0.288)*** 0.183 (0.110 to 0.256)*** 0.266 (0.229 to 0.303)*** −0.099 (-0.11 to -0.08)***
Central and Eastern Europe −0.191 (-0.297 to -0.085)** −0.125 (-0.223 to -0.027)* −0.132 (-0.189 to -0.075)*** 0.067 (0.043 to 0.091)***
Southern Europe −0.068 (-0.237 to 0.101) n.s. −0.040 (-0.183 to 0.103) n.s. −0.150 (-0.221 to -0.079)*** 0.052 (0.025 to 0.079)***
Northern Europe −0.085 (-0.242 to 0.072) n.s. 0.152 (0.023 to 0.281)* 0.029 (-0.038 to 0.096) n.s. −0.024 (-0.051 to 0.003) n.s.
Gender: male (vs female) 0.360 (0.201 to 0.519)*** 0.304 (0.173 to 0.435)*** −0.126 (-0.193 to -0.059)*** −0.030 (-0.06 to 0.00) n.s
Central and Eastern Europe −0.083 (-0.287 to 0.121) n.s. −0.031 (-0.209 to 0.147) n.s. 0.099 (0.001 to 0.197)* −0.104 (-0.149 to -0.059)***
Southern Europe −0.246 (-0.611 to 0.119) n.s. −0.182 (-0.472 to 0.108) n.s. −0.024 (-0.165 to 0.117) n.s. 0.028 (-0.031 to 0.087) n.s.
Northern Europe 0.136 (-0.168 to 0.440) n.s. 0.297 (0.054 to 0.540)* 0.052 (-0.073 to 0.177) n.s. −0.183 (-0.238 to -0.128)***
Age 0.097 (0.056 to 0.138)*** 0.149 (0.112 to 0.186)*** 0.112 (0.092 to 0.132)*** −0.031 (-0.04 to -0.02)***
Central and Eastern Europe −0.076 (-0.131 to -0.021)** −0.078 (-0.129 to -0.027)** −0.070 (-0.099 to -0.041)*** 0.017 (0.003 to 0.031)*
Southern Europe −0.027 (-0.121 to 0.067) n.s 0.025 (-0.059 to 0.109) n.s. −0.046 (-0.085 to -0.007)* 0.008 (-0.008 to 0.024) n.s.
Northern Europe 0.008 (-0.068 to 0.084) n.s 0.014 (-0.049 to 0.077) n.s. 0.129 (0.094 to 0.164)*** −0.021 (-0.035 to -0.007)**
Level of education −0.102 (-0.145 to −0.127 (-0.162 to -0.092)*** −0.150 (-0.168 to -0.132)*** 0.052 (0.040 to 0.060)
-0.059)***
Central and Eastern Europe 0.055 (-0.004 to 0.114) n.s 0.090 (0.037 to 0.143)*** 0.098 (0.067 to 0.129)*** −0.013 (-0.015 to 0.012) n.s.

Southern Europe 0.077 (-0.013 to 0.167) n.s −0.066 (-0.152 to 0.020) n.s. 0.096 (0.061 to 0.131)*** −0.008 (-0.024 to 0.008) n.s.

Northern Europe −0.100 (-0.180 to -0.020)* 0.006 (-0.061 to 0.073) n.s. −0.056 (-0.091 to -0.021)** −0.002 (-0.018 to 0.014) n.s.

Note: Western Europe is the reference region;*** p < 0.001; n.s. non-significant ; CI = confidence interval; The odds ratios for trend and attribution scepticism can be
calculated by EXP (B).

found between Southern and Western Europe. relatively uniform across the different regions of Europe. In Central and
Individuals living in Southern European countries, who prioritise Eastern and Northern European countries men were generally less
conservation over openness-to-change values, were more likely to hold concerned about climate change as compared to women in these re-
attribution sceptical views and to perceive less negative climate change gions, while the association of gender and climate concern was non-
impacts, as indicated by positive interaction terms. Individuals living in significant in Western European countries. The gender effects in rela-
Northern European countries, who prioritise conservation over open- tion to the perceived impacts of climate change were weaker in Central
ness-to-change values, were more likely to hold trend and attribution and Eastern European countries as compared to Western European
sceptical views and to perceive less negative climate change impacts, as countries. Gender effects were stronger in terms of attribution scepti-
indicated by positive interaction terms. This is in contrast to Western cism In Northern European countries as compared to Western European
Europe where the associations were non-significant. The only sig- countries.The other interaction effects for gender were non-significant.
nificant interaction of Central and Eastern Europe with the conservation Age effects were generally weaker in Eastern European countries
vs. openness-to-change value dimension was found for the perceived than in Western European countries, as indicated by negative interac-
impacts of climate change. The negative interaction term indicates that, tion terms for trend scepticism, attribution scepticism and perceived
on average, conservation vs. openness-to-change values are associated impacts of climate change, and a positive interaction term for concern
with more negative perceived climate change impacts in Eastern about climate change. Age effects for the perceived impacts and con-
European countries. cern about climate change were generally stronger in Northern
The political orientation effects were generally weaker in Central and European countries, as indicated by a positive and a negative interac-
Eastern European countries as compared to Western European coun- tion term, respectively. The only significant interaction for Southern
tries, as indicated by the negative interaction terms for trend scepticism Europe was found for the perceived impacts of climate change, in-
and the perceived impacts of climate change, and by the negative in- dicating that, on average, the age effects for the perceived impacts of
teraction term for concern about climate change. In Southern Europe, climate change were weaker in Southern European countries as com-
political orientation effects were weaker for the perceived impacts and pared to Western European countries.
concern about climate change, as indicated by a positive and a negative The relationships between level of education on the one hand and the
interaction term, respectively. The only significant interaction effect of perceived impacts of climate change on the other appear stronger in
Northern Europe with political orientation was for attribution scepti- Northern European countries, but weaker in Central and Eastern and
cism. This suggests that there is a bigger political divide with regard to Southern European countries (as indicated by negative and positive
attribution scepticism in Northern European countries as compared to interactions, respectively). The link between education and attribution
Western European countries. scepticism appears weaker in Eastern European countries as compared
Table 4 further shows that there were only a small number of sig- to Western European countries (as indicated by a positive interaction),
nificant interaction effects for gender, suggesting that effects are while the link between education and trend scepticism appears stronger

32
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

in Northern European as compared to Western European countries. The divide in environmental attitudes. These mostly focus on the legacy left
other interaction effects were non-significant. by decades of communist rule as well as the profound impact of its
collapse in the late 1980s. For example, it has been argued that the
4. Discussion political and economic uncertainty following the collapse of communist
regimes may have prioritised economic survival over environmental
This paper examined the associations of climate change perceptions protection (cf. Inglehart, 1990). Others point to the speed of economic
with a range of individual-level factors, and how the importance of and social change, and a possibility of a cultural lag in environmental
these factors may vary cross-nationally. Building upon previous re- attitudes (Brinkman and Brinkman, 1997; Balžekiene and Telešiene,
search on the individual-level determinants of climate change percep- 2017). Environmental attitudes may be ‘sticky’, in particular when they
tions, we show that both socio-political and demographic factors are emerge from fundamental orientations and beliefs (Chaisty and
significant predictors of climate change beliefs and concern across 22 Whitefield, 2015). This means that they may have to play catch-up in a
European countries and Israel (Capstick et al., 2015; Poortinga et al., fast changing world.
2011; Shi et al., 2016; Steentjes et al., 2017). This demonstrates that General explanations for cross-national differences include differ-
both socio-political and demographic factors are needed to understand ences in experiences with extreme weather events (e.g. flooding and
public perceptions of climate change. droughts) and vulnerability to the impacts of climate change (Brody
We show that some of the associations are remarkably consistent et al., 2008; Demski et al., 2017; Deryugina, 2013; Donner and
across the 23 countries that participated in the European Social Survey McDaniels, 2013; Spence et al., 2011), elite cues and media coverage
(cf. Marquart-Pyatt, 2008). In particular political orientation, level of (Carmichael and Brulle, 2017; Feldman et al., 2017), and current CO2
education and self-transcendence vs. self-enhancement values were con- emissions and dependence on fossil fuels (Lee et al., 2015). While there
sistently linked to the four different dimensions of climate change are indications that all these factors are important for climate-relevant
perceptions. That is, people who place themselves on the right-hand environmental attitudes, they are less appropriate for explaining dif-
side of the political spectrum, have a lower level of education, and ferences in the size of individual-level effects. Here again we may need
prioritise self-enhancement over self-transcendence values are more to look at possible historical and political explanations. According to
likely to hold climate sceptical views, perceive fewer negative impacts, Rohrschneider and Miles (2015), one reason as to why left-right divi-
and are less likely to be concerned about climate change in all or a great sions are smaller in Central and Eastern Europe is because the en-
majority of countries. The other individual-level effects were more vironment is less of an issue for party competition in these countries. If
variable. For example, gender and age were significantly associated political parties do not compete for the green vote, they are less likely to
with climate change perceptions in some but non-significantly so in polarise the public in return (Rohrschneider and Miles, 2015). Fur-
other countries; and whereas conservation vs openness-to-change va- thermore, if there is no platform for climate sceptical views, either as
lues were non-significantly associated with the four climate perception part of party politics and/or the media, there is less opportunity for the
dimensions in most countries, there were a number of countries in public to become polarised through elite cues (cf. Brulle et al., 2012;
which the association was significantly negative and a number of Carmichael and Brulle, 2017)
countries in which the association was significantly positive. Another It is less clear as to why some of the demographic and value effects
interesting case is the association of gender with the perceived impacts vary across countries and regions; and there is no coherent literature
of climate change. While the association is significant overall, it is only available to draw upon. Gender differences in environmental risk per-
so in a small minority of individual countries (associations are non- ception are often explained by social inequalities, and it could be ar-
significant in the other countries). gued that such effects are therefore less likely to emerge in more gender
Even if the direction of the associations were to a large extent equal societies (Norgaard and York, 2005). This view is however not
consistent, the sizes of the effects were not. This shows the importance supported by the results of the current study. Gender effects appeared
of cross-cultural research, and the need to validate results in multiple stronger in Northern European countries that tend to have higher levels
countries and cultural contexts before assuming certain effects are of gender equality. Further research is needed to see wat may explain
universal. For example, while climate change perceptions are fairly the effects.
consistently linked to political orientation, they are not equally po- There is a need to be cautious when interpreting the reported
larised in all countries. This not only applies to the socio-political fac- findings. The study involved a relatively small number of countries
tors but also to the demographic ones. Evidence was found that the (n = 23). This means that the models only have the statistical power to
effects for the demographic (e.g. age) and socio-political (e.g. political detect large national-level differences (Button et al., 2013). One criti-
orientation) factors are generally weaker in Eastern as compared to cism of current climate perception research is that the vast majority of
Western European countries. Some of the demographic effects (e.g. empirical focus has been on a small number of mainly affluent Western
gender and age) appeared stronger in Northern European countries. countries (Hopkins, 2015). A strength of our study is that there was a
Demographics can reflect important socio-cultural categories, as illu- range of countries, including a number of Eastern European countries
strated by the ‘conservative male effect’ (cf. Jylhä and Akrami, 2015). with smaller and mainly national literatures on climate change per-
The phenomenon that a specific demographic subgroup holds very ceptions (Gwiazda and Kolbowska, 2009; Balžekienė et al., 2008;
distinct attitudes to a range of risk issues, from climate change and gun Vladyka, 2007; Soasepp, 2016). The inclusion of these, as well as other
control to financial markets, emerged from and has mainly been found countries across the continent, allowed for systematic comparisons
in the US (Dunlap and McCright, 2011; Finucane et al., 2000); and these between countries from different European regions.
socio-cultural categories may mean different things in different coun- Most research on cross-national differences has predominantly been
tries. That is, a conservative male in the US is most likely socio-cultu- conducted using data from the International Social Survey Programme
rally not the same as a conservative male in a different country. (ISSP) and the World Values Survey WVS), which contain question
The question of course is where the cross-national differences come modules on more generic environmental attitudes and preferences for
from, and how they can be explained. The results of our study appear to environmental protection. While others used Eurobarometer or inter-
confirm previous research showing systematic differences in individual- national opinion poll data, Round 8 of the ESS was the first interna-
level effects between Central and Eastern European on the one hand tional survey with a dedicated and theory-driven module on climate
and other European countries (McCright et al., 2016) or advanced in- change perceptions, allowing this study to explore associations with
dustrialised countries (Marquart-Pyatt, 2008) on the other. Specifically, different aspects of climate change perceptions in a systematic way.
there appears to be less of a political divide in former communist Climate change perceptions can be understood to have different di-
countries. Several explanations have been offered for an East-West mensions, and thus may be influenced by different sets of determinants.

33
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

Results relating to cross-national variation may also depend on the 0013916506297831.

specific combination of countries included in the analysis (Lo and Demski, C., Capstick, S., Pidgeon, N., Sposato, R.G., Spence, A., 2017. Experience of ex-
treme weather affects climate change mitigation and adaptation responses. Clim.
Chow, 2015). The countries included in the ESS are a relatively co- Change 140 (2), 149–164. https://doi.org/10.1007/s10584-016-1837-4.
herent and affluent set of countries at the world stage. It can be as- Deryugina, T., 2013. How do people update? The effects of local weather fluctuations on
sumed that differences in effects could be even greater when a more beliefs about global warming. Clim. Change 118 (2), 397–416. https://doi.org/10.
1007/s10584-012-0615-1.
geographically, economically and culturally diverse set of countries is Dietz, T., Dan, A., Shwom, R., 2007. Support for climate change policy : social psycho-
sampled and compared (Lee et al., 2015). Future research should at- logical and social structural influences. Rural Sociol. 72 (2), 185–214. https://doi.
tempt to expand the number of countries to improve the estimates of org/10.1526/003601107781170026.
Donner, S.D., McDaniels, J., 2013. The influence of national temperature fluctuations on
cross-national effects, and to explore the contextual factors that shape opinions about climate change in the U.S. Since 1990. Clim. Change 118 (3),
the differences in individual-level effects. This will be the focus of fu- 537–550. https://doi.org/10.1007/s10584-012-0690-3.
ture analyses. The time and resources needed to conduct high-quality Dunlap, R.E., McCright, A.M., 2011. Cool dudes: the denial of climate change among
conservative white males in the United States. Glob. Environ. Change 21,
cross-national social research with valid, culturally equivalent measures
1163–11172.
should not be underestimated, and can only be delivered through Echavarren, J.M., 2017. From objective environmental problems to subjective environ-
substantial research infrastructures, such as the European Social Survey mental concern: a multilevel analysis using 30 indicators of environmental quality.
(ESS) or the International Social Survey Programme (ISSP) (Fitzgerald Soc. Nat. Resour. 30 (2), 145–159 http://doi.org/.
Engels, A., Hüther, O., Schäfer, M., Held, H., 2013. Public climate-change skepticism,
and Jowell, 2010; Haller et al., 2009). These international collabora- energy preferences and political participation. Glob. Environ. Change 23 (5),
tions and investments help to improve our understanding of the cultural 1018–1027. https://doi.org/10.1016/j.gloenvcha.2013.05.008.
dependency of how climate change is perceived, as well as which po- European Commission, 2014. 2030 Framework for Climate and Energy Policies.
Retrieved from. European Commission, Brussels. http://ec.europa.eu/energy/2030_
licies and sustainable behaviours are considered acceptable. en.htm.
European Social Survey, 2016. European Social Survey Round 8 Data. File Edition 2.0.
Acknowledgments Bergen, Norway: Norwegian Centre for Research Data. Retrieved from. http://
www.europeansocialsurvey.org/data.
European Social Survey, 2015. Round 8 Survey Specification for ESS ERIC Member,
The European Social Survey (ESS) is a European Research Observer and Guest Countries. Retrieved from. ESS ERIC Headquarters, City
Infrastructure Consortium (ERIC). Participating countries contribute to University, London. https://www.europeansocialsurvey.org/docs/round8/methods/
ESS8_project_specification.pdf.
the central coordination costs of the ESS ERIC as well as covering the Feldman, L., Hart, P.S., Leiserowitz, A., Maibach, E., Roser-Renouf, C., 2017. Do hostile
costs of their own fieldwork and national coordination. media perceptions lead to action? The role of hostile media perceptions, political
efficacy, and ideology in predicting climate change activism. Communic. Res. 44 (8),
1099–1124. https://doi.org/10.1177/0093650214565914.
References
Fielding, K.S., Head, B.W., Laffan, W., Western, M., Hoegh-Guldberg, O., 2012. Australian
politicians’ beliefs about climate change: political partisanship and political ideology.
Balžekiene, A., Telešiene, A., 2017. Vulnerable and insecure? Environmental and tech- Environ. Polit. 21 (5), 712–733. https://doi.org/10.1080/09644016.2012.698887.
nological risk perception in Europe. In: Telešiene, A., Gross, M. (Eds.), Green Finucane, M.L., Slovic, P., Mertz, C.K., Flynn, J., Satterfield, T.A., 2000. Gender, race, and
European. Environmental Behaviour and Attitudes in Europe in a Historical and perceived risk: the “white male” effect. Health Risk Soc. 2 (2), 159–172.
Cross-Cultural Comparative Perspective. Routledge, Abingdon, pp. 31–55. Fitzgerald, R., 2015. Sailing in Unchartered Waters: Structuring and Documenting Cross-
Balžekienė, A., Telešienė, A., Rinkevičius, L., 2008. Klimato kaita: socialinio rizikos su- National Questionnaire Design (GESIS papers). GESIS., Mannheim.
vokimo ir žiniasklaidos diskurso Lietuvoje konfigūracijos. Sociologija: Mintis ir Fitzgerald, R., Jowell, R., 2010. Measurement Equivalence in Comparative Surveys: The
Veiksmas 2 (22), 5–19. European Social Survey (ESS)—From Design to Implementation and Beyond. In:
Bord, R.J., O’Connor, R.E., Fisher, A., 2000. In what sense does the public need to un- Harkness, J.A., Braun, M., Edwards, B., Johnson, T.P., Lyberg, L., Mohler, P.P., Smith,
derstand global climate change? Public Underst. Sci. 9 (3), 205–218. https://doi.org/ T.W. (Eds.), Survey Methods in Multinational, Multiregional, and Multicultural
10.1088/0963-6625/9/3/301. Contexts. John Wiley & Sons, Inc, Hoboken, NJ, USA p. Fitzgerald, R. and Jowell, R.
Brinkman, R.L., Brinkman, J.E., 1997. Cultural lag: conception and theory. Int. J. Soc. (2010) Measurement E.
Econ. 24 (6), 609–627. Fransson, N., Garling, T., 1999. Environmental concern: concepual definitions, mea-
Brody, S.D., Zahran, S., Vedlitz, A., Grover, H., 2008. Examining the relationship between surement methods, and research findings. J. Environ. Psychol. 19 (4), 369–382.
physical vulnerability and public perceptions of global climate change in the United https://doi.org/10.1006/jevp.1999.0141.
States. Environ. Behav. 40 (1), 72–95. https://doi.org/10.1177/0013916506298800. Gwiazda, M., Kolbowska, A., 2009. Polacy o zmianach klimatu. Centrum Badania Opinii
Brown, K.W., Kasser, T., 2005. Are psychological and ecological well-being compatible? SpoĹ., Warsaw.
The role of values, mindfulness, and lifestyle. Soc. Indic. Res. 74 (2), 349–368. Haller, M., Jowell, R., Smith, T.W., 2009. The International Social Survey Programme
https://doi.org/10.1007/s11205-004-8207-8. 1984-2009: Charting the Globe. Routledge, Abingdon.
Brulle, R.J., Carmichael, J., Jenkins, J.C., 2012. Shifting public opinion on climate Hayes, B.C., 2001. Gender, scientific knowledge, and attitudes toward the environment: a
change: an empirical assessment of factors influencing concern over climate change crossnational analysis. Polit. Res. Q. 54, 657–671. https://doi.org/10.1177/
in the US 2002–2010. Clim. Change 114 (2), 169. https://doi.org/10.1007/s10584- 106591290105400309.
012-0403-y. Hoffman, A.J., 2011. The growing climate divide. Nat. Clim. Change 1, 195. https://doi.
Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., org/10.1038/nclimate1144. Retrieved from.
Munafò, M.R., 2013. Power failure: why small sample size undermines the reliability Hopkins, D., 2015. Country comparisons. Nat. Clim. Change 5, 975. https://doi.org/10.
of neuroscience. Nat. Rev. Neurosci. 14, 365. https://doi.org/10.1038/nrn3475. 1038/nclimate2730. Retrieved from.
Retrieved from. Hornsey, M.J., Harris, E.A., Bain, P.G., Fielding, K.S., 2016. Meta-analyses of the de-
Capstick, S., Whitmarsh, L., Poortinga, W., Pidgeon, N., Upham, P., 2015. International terminants and outcomes of belief in climate change. Nat. Clim. Change 6 (6),
trends in public perceptions of climate change over the past quarter century. Wiley 622–626. https://doi.org/10.1038/nclimate2943.
Interdiscip. Rev. Clim. Change 6 (1), 35–61. https://doi.org/10.1002/wcc.321. Inglehart, R., 1990. Culture Shift in Advanced Industrial Society. Princeton University
Carmichael, J.T., Brulle, R.J., 2017. Elite cues, media coverage, and public concern: an Press, Princeton, New Jersey.
integrated path analysis of public opinion on climate change, 2001–2013. Environ. Jylhä, K.M., Akrami, N., 2015. Social dominance orientation and climate change denial:
Polit. 26 (2), 232–252. https://doi.org/10.1080/09644016.2016.1263433. the role of dominance and system justification. Pers. Individ. Dif. 86, 108–111.
Chaisty, P., Whitefield, S., 2015. Attitudes towards the environment: are postcommunis https://doi.org/10.1016/j.paid.2015.05.041.
societies (still) different? Environ. Polit. 24 (4), 598–616. https://doi.org/10.1080/ Jylhä, K.M., Cantal, C., Akrami, N., Milfont, T.L., 2016. Denial of anthropogenic climate
09644016.2015.1023575. change: social dominance orientation helps explain the conservative male effect in
Cornelis, I., Van Hiel, A., Roets, A., Kossowska, M., 2009. Age differences in conservatism: Brazil and Sweden. Pers. Individ. Dif. 98, 184–187. https://doi.org/10.1016/j.paid.
evidence on the mediating effects of personality and cognitive style. J. Pers. 77 (1), 2016.04.020.
51–88. https://doi.org/10.1111/j.1467-6494.2008.00538.x. Kahan, D.M., Braman, D., Gastil, J., Slovic, P., Mertz, C.K., 2007. Culture and identity-
Corner, A., Markowitz, E., Pidgeon, N., 2014. Public engagement with climate change: the protective cognition: explaining the white-male effect in risk perception. J. Empir.
role of human values. Wiley Interdiscip. Rev. Clim. Change 5 (3), 411–422. https:// Leg. Stud. 4 (3), 465–505. https://doi.org/10.1111/j.1740-1461.2007.00097.x.
doi.org/10.1002/wcc.269. Kahan, D.M., Jenkins-Smith, H., Braman, D., 2011. Cultural cognition of scientific con-
De Groot, J.I.M., Steg, L., 2007. Value orientations and environmental beliefs in five sensus. J. Risk Res. 14 (2), 147–174. https://doi.org/10.1080/13669877.2010.
countries: validity of an instrument to measure egoistic, altruistic and biospheric 511246.
value orientations. J. Cross. Psychol. 38 (3), 318–332. https://doi.org/10.1177/ Kahan, D.M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L.L., Braman, D., Mandel, G.,
0022022107300278. 2012. The polarizing impact of science literacy and numeracy on perceived climate
De Groot, J.I.M., Steg, L., 2008. Value orientations to explain beliefs related to en- change risks. Nat. Clim. Change 2 (10), 732–735. https://doi.org/10.1038/
vironmental significant behavior: how to measure egoistic, altruistic, and biospheric nclimate1547.
value orientations. Environ. Behav. 40 (3), 330–354. https://doi.org/10.1177/ Knight, K.W., 2016. Public awareness and perception of climate change: a quantitative

34
W. Poortinga et al. Global Environmental Change 55 (2019) 25–35

cross-national study. Environ. Sociol. 2 (1), 101–113. https://doi.org/10.1080/ Satterfield, T.A., Mertz, C.K., Slovic, P., 2004. Discrimination, vulnerability, and justice in
23251042.2015.1128055. the face of risk. Risk Anal. 24 (1), 115–129. https://doi.org/10.1111/j.0272-4332.
Lee, T.M., Markowitz, E.M., Howe, P.D., Ko, C.-Y., Leiserowitz, A.A., 2015. Predictors of 2004.00416.x.
public climate change awareness and risk perception around the world. Nat. Clim. Schultz, P.W., Zelezny, L., 1999. Values as predictors of environmental attitudes: evidence
Change 5 (11), 1014–1020. https://doi.org/10.1038/nclimate2728. Retrieved from. for consistency across 14 countries. J. Environ. Psychol. 19 (3), 255–265. https://doi.
Lo, A.Y., Chow, A.T., 2015. The relationship between climate change concern and na- org/10.1006/jevp.1999.0129.
tional wealth. Clim. Change 131 (2), 335–348. https://doi.org/10.1007/s10584-015- Schwartz, S.H., 1992. Universals in the content and structure of values: theoretical ad-
1378-2. vances and empirical tests in 20 countries. Adv. Exp. Soc. Psychol. 25, 1–65.
Malka, A., Krosnick, J.A., Langer, G., 2009. The association of knowledge with concern Schwartz, S.H., 2003. A Proposal for Measuring Value Orientations Across Nations.
about global warming: trusted information sources shape public thinking. Risk Anal. Retrieved from. European Social Survey, London. http://www.
29 (5), 633–647. https://doi.org/10.1111/j.1539-6924.2009.01220.x. europeansocialsurvey.org/docs/methodology/core%7B_%7Dess%7B_
Marquart-Pyatt, S.T., 2008. Are there similar sources of environmental concern? %7Dquestionnaire/ESS%7B_%7Dcore%7B_%7Dquestionnaire%7B_%7Dhuman%7B_
Comparing industrialized countries*. Soc. Sci. Q. 89 (5), 1312–1335. https://doi.org/ %7Dvalues.pdf.
10.1111/j.1540-6237.2008.00567.x. Shi, J., Visschers, V.H.M., Siegrist, M., Arvai, J., 2016. Knowledge as a driver of public
McCright, A.M., 2010. The effects of gender on climate change knowledge and concern in perceptions about climate change reassessed. Nat. Clim. Change 6, 759. https://doi.
the American public. Popul. Environ. 32 (10), 66–87. https://doi.org/10.1007/ org/10.1038/nclimate2997. Retrieved from.
s11111-010-0113-1McCright. Snow, C.E., Dibners, K.E., 2016. Science Literacy: Concepts, Contexts, and Consequences.
McCright, A., Dunlap, R.E., 2011. The politization of climate change and polarization in National Academies Press, Washington, DC.
the American public’s views of global warming, 2001–2010. Sociological Quarterly Soasepp, S., 2016. Eesti elanike hoiakud CO2-põhise automaksu suhtes ja sellega seotud
52 (2), 155–194. https://doi.org/10.1111/j.1533-8525.2011.01198.x. tegurid. Doctoral dissertation. Tartu: Tartu Ülikool.
McCright, A.M., Dunlap, R.E., 2013. Bringing ideology in: the conservative white male Spence, A., Poortinga, W., Butler, C., Pidgeon, N.F.N., 2011. Perceptions of climate
effect on worry about environmental problems in the USA. J. Risk Res. 16 (2), change and willingness to save energy related to flood experience. Nat. Clim. Change
211–226. https://doi.org/10.1080/13669877.2012.726242. 1 (1), 46–49. https://doi.org/10.1038/NCLIMATE1059.
McCright, A.M., Dunlap, R.E., Xiao, C., 2014. The impacts of temperature anomalies and Steentjes, K., Pidgeon, N.F., Poortinga, W., Arnold, A., Böhm, G., Mays, C., Tvinnereim,
political orientation on perceived winter warming. Nat. Clim. Change 4, 1077. E., 2017. European Perceptions of Climate Change (EPCC): Topline Findings of a
https://doi.org/10.1038/nclimate2443. Retrieved from. Survey Conducted in Four European Countries in 2016. Retrieved from. Cardiff
McCright, A.M., Dunlap, R.E., Marquart-Pyatt, S.T., 2016. Political ideology and views University, Cardiff. http://orca.cf.ac.uk/98660/7/EPCC.pdf.
about climate change in the European Union. Environ. Polit. 25 (2), 338–358. Steg, L., De Groot, J.I.M., 2012. Environmental values. In: Clayton, S. (Ed.), The Oxford
https://doi.org/10.1080/09644016.2015.1090371. Handbook of Environmental and Conservation Psychology. Oxford University Press,
Milfont, T.L., Milojev, P., Greaves, L.M., Sibley, C.G., 2015. Socio-structural and psy- Oxford, pp. 81–92. Retrieved from. http://www.oxfordhandbooks.com/view/10.
chological foundations of climate change beliefs. N. Z. J. Psychol. 44 (1), 17–30. 1093/oxfordhb/9780199733026.001.0001/oxfordhb-9780199733026-e-5.
Norgaard, K., York, R., 2005. Gender equality and state environmentalism. Gend. Soc. 19 Stern, P.C., 2000. Toward a coherent theory of environmentally significant behavior. J.
(4), 506–522. https://doi.org/10.1177/0891243204273612. Soc. Issues 56 (3), 407–424. https://doi.org/10.1111/0022-4537.00175.
O’Connor, R.E., Bord, R.J., Fisher, A., 1999. Risk perceptions, general environmental Stern, P.C., Dietz, T., Guagnano, G.A., 1998. A brief inventory of values. Educ. Psychol.
beliefs, and willingness to address climate change. Risk Anal. 19 (3), 461–471. Meas. 58 (6), 984–1001. https://doi.org/10.1177/0013164498058006008.
https://doi.org/10.1023/A:1007004813446. Stevenson, K.T., Peterson, M.N., Bondell, H.D., Moore, S.E., Carrier, S.J., 2014.
Painter, J., Ashe, T., 2012. Cross-national comparison of the presence of climate scepti- Overcoming skepticism with education: interacting influences of worldview and
cism in the print media in six countries, 2007–10. Environ. Res. Lett. 7 (4), 44005. climate change knowledge on perceived climate change risk among adolescents.
Retrieved from. http://stacks.iop.org/1748-9326/7/i=4/a=044005. Clim. Change 126 (3–4), 293–304. https://doi.org/10.1007/. s10584-014-1228-7.
Poortinga, W., Steg, L., Vlek, C., 2004. Values, environmental concern, and environ- Tranter, B., Booth, K., 2015. Scepticism in a changing climate: a cross-national study.
mental behavior: a study into household energy use. Environ. Behav. 36 (1), 70–93. Glob. Environ. Change 33, 154–164. https://doi.org/10.1016/j.gloenvcha.2015.05.
https://doi.org/10.1177/0013916503251466. 003.
Poortinga, W., Spence, A., Whitmarsh, L., Capstick, S., Pidgeon, N.F., 2011. Uncertain UNFCCC, 2017. The Paris Agreement. United Nations Framework Convention on Climate
climate: an investigation into public scepticism about anthropogenic climate change. Change.
Glob. Environ. Change 21 (3, SI), 1015–1024. https://doi.org/10.1016/j.gloenvcha. Vladyka, M., 2007. Globální změna klimatu jako sociální konstrukt. Doctoral disserta-
2011.03.001. tion. Masarykova Univerzita, Fakulta sociálních studií, Brno.
Rohrschneider, R., Miles, M.R., 2015. Representation through parties? Environmental Whitmarsh, L., 2011. Scepticism and uncertainty about climate change: dimensions, de-
attitudes and party stances in Europe in 2013. Environ. Polit. 24 (4), 617–640. terminants and change over time. Glob. Environ. Change 21 (2), 690–700. https://
https://doi.org/10.1080/09644016.2015.1023579. doi.org/10.1016/j.gloenvcha.2011.01.016.

35