Introduction

The awareness of the climate change issues is an essential precondition for individuals to behave responsibly in their everyday lives and to demand and support political measures to mitigate climate change. Since environmentally friendly behavior often comes at some cost to individuals, at least in terms of convenience and comfort, it is reasonable to expect that individuals who acknowledge climate change and have a high level of climate change awareness will be more likely to act in accordance with environmental protection. Indeed, studies have shown that environment-related attitudes can significantly influence personal commitment to climate protection, as well as the acceptance of climate policies (Engels et al., 2013). Therefore, examining climate change awareness and denial is crucial for understanding how individuals, nation-states, and the international community respond to the threat of climate change.

In this paper, we employ survey data from the European Social Survey (ESS) to investigate the following questions: To what extent do climate-related attitudes differ among and within populations across European countries? Which individual-level and country-specific factors could account for these differences? Does climate change denial and various levels of environmental consciousness represent a gradual continuum within a one-dimensional concept of climate change awareness, or do they capture distinct phenomena with proponents differing in their values and socio-demographic profiles?

Numerous studies have investigated individual-level attitudes towards climate change using survey data. Until recently, most cross-national studies were conducted in only a few countries (Poortinga et al., 2019); yet, a growing body of literature has documented cross-national differences in climate change attitudes. These differences manifest in terms of the intensity of concern, support for national and international action, and willingness to change individual environmental behavior. One set of studies has focused on the perception of the seriousness of climate change, either by rating the severity of the climate change challenge (Morales-Giner & Gedik, 2023) or by ranking climate change in relation to other important global challenges (Baiardi & Morana, 2021; Echavarren et al., 2019), or both (Diekmann & Franzen, 1999). Also, views on global warming as an important aspect of climate change have been studied from a cross-national perspective (Kvaløy et al., 2012; Mostafa, 2016).

Numerous studies have directed attention to climate change beliefs, risk perceptions, concerns, or worries (Arıkan & Günay, 2021; Czarnek et al., 2021; Echavarren et al., 2019; Franzen & Meyer, 2010; Lee et al., 2015; Poortinga et al., 2019; Zafar & Ammara, 2024), as well as knowledge about climate change and behavior (Shi et al., 2015; Smith & Mayer, 2019; Zafar & Ammara, 2024). Related issues include public support for environmental protection, policy preferences, and behavioral intentions (Hao et al., 2020; Inglehart, 1995; Kim & Wolinsky-Nahmias, 2014; Kulin & Johansson Sevä, 2020; Marquart-Pyatt et al., 2019).

Another significant body of research has focused on beliefs regarding the existence of climate change, including the denial of global warming, the denial of its anthropogenic sources and the denial of its seriousness (Dunlap & McCright, 2010). Although these studies often refer to denial, they actually examine various specifications of climate change skepticism. Furthermore, the concept of climate change skepticism is often used casually, without consideration of the complexity it represents. Scholars interpret the concept in varied ways, with different meanings and scopes. “These shades of the phenomenon are largely lost in both the public and scholarly discourse where the blanket labels climate change scepticism/denial and climate change sceptic/denier still dominate”, as critically remarked by Van Rensburg (2015).

Some researchers have used Rahmstorf’s (2004) classification to examine the dimensions of trend, attribution and impact skepticism in Germany (Engels et al., 2013) and the UK (Poortinga et al., 2011), using specific national-level survey data. Such specialized surveys allow the creation of a skepticism scale based on a battery of attitude statements related to climate change (Poortinga et al., 2011; Whitmarsh, 2011). In a cross-national comparison, such detailed information is generally not available. Using the ESS data, Poortinga et al. (2019) included in their measure of skepticism respondents who claim that climate change is definitely or probably not changing, which conflates the concepts of denial and uncertainty. Similarly, when examining European regions, Lübke (2022) defined climate change denial and climate change uncertainty by combining information on climate change acknowledgement and the recognition of anthropogenic sources of climate change. In her specification, climate change denial refers to individuals who believe that the climate is definitely or probably not changing or that climate change is mainly or entirely due to natural processes. Uncertainty refers to individuals stating that the climate is probably or definitely changing, but that climate change is about equally caused by human activity and natural processes.

Our approach to measuring climate changes awareness and climate change denial departs from existing studies. The literature on individual indicators related to climate change views, such as worries, concern, or knowledge about the origins and consequences of climate change, has provided important insights into the roles certain determinants play in various aspects of climate change attitudes. However, people may have diffuse ideas and feelings about climate change and going beyond the individual indicators might provide additional value in understanding the general public’s perception of climate change, particularly in cross-national comparison. We endorse the view that a composite score is theoretically a more appropriate measure of climate change awareness than a single item.

We understand awareness as a multidimensional construct and a psychological concept that incorporates opinions, concerns, and various viewpoints regarding the issue at hand (for the complexity of measuring climate change risk perception, worry, and concern see van der Linden, 2017). This calls for more comprehensive measures beyond one-item scales. For example, in a tailored survey administered to teenagers in Germany and Austria, Kuthe et al. (2019) decomposed climate change awareness into knowledge, climate-friendly behavior, multiplicative actions, attitude, and personal concern. They demonstrated that participants differed in their cognitive, affective and conative aspects of climate change awareness.

Since we are limited to existing measures in cross-national surveys, we utilize the ESS battery of questions on climate change and combine them into a novel measurement. Climate change awareness in our study is a composite index that goes beyond climate risk perception and worry. It encompasses knowledge and beliefs regarding the existence of climate change, its causes, the severity of its impact, concerns about climate issues, and views on personal responsibility. Studies that use such composite measures are scarce, with Franzen and Meyer (2010) being a notable exception. They utilized data from the International Social Survey Programme (ISSP) to examine public concern for the state of the natural environment.

The second outcome variable, climate change denial, is measured in our study as a categorical assertion that climate change is definitely not occurring. This measurement differs from commonly used concepts in the literature, such as denial of the predominant anthropogenic sources of climate change (Lübke, 2022) or the combination of climate change denial and skepticism (Poortinga et al., 2019), which does not entail absolute rejection of climate-related findings but rather reflects distrust of these findings (Radtke et al., 2020; Whitmarsh, 2011). Our measurement of denial does not refer to skepticism and uncertainty about climate change. Instead, our aim is to better understand the small minority that does not express uncertainty but confidently claims that climate change is definitely not occurring. In our study, uncertainty and skepticism about the existence and/or causes of climate change are understood as expressions of the level of climate change awareness. These indicators are included in the composite index of climate change awareness, contributing to a low value on this index.

In addition to analyzing the determinants of both the level of climate change awareness and the denial of climate change using novel measurements, our second key contribution lies in testing a comprehensive set of explanations for variations in climate change awareness and denial. We categorize these explanations under individuals' human and socio-political value orientations, socio-demographic characteristics, and societal national-level characteristics. While these potential explanatory factors are not novel and have been considered previously, the breadth of factors examined simultaneously in our study exceeds that of existing literature.

Third, we hypothesize that contextual factors play an important yet under-researched role in understanding climate change attitudes. Therefore, we employ a multilevel framework, considering socio-economic, natural, and cultural contexts at the nation-state level alongside individual characteristics.

Climate change awareness and denial: Empirical findings, theoretical frameworks, and hypotheses

Climate change, as assessed in the European Social Survey, refers to long-term changes in the state of the climate system observed over extended periods at local, regional, and global scales (ESS, 2016). Beliefs regarding the existence, concern, causes, impact, and personal responsibility concerning climate change exhibit significant variation among individuals and countries. We assess the range of these beliefs through a composite indicator termed “climate change awareness.”

The second phenomenon under examination in this paper is climate change denial. Denial is understood as the refusal to acknowledge commonly known facts (Brüggemann, 2017), in this instance, the existence of climate change. For both phenomena, we present an overview of previous findings and theoretical explanations, consolidating literature that typically concentrates on a limited range of explanatory factors.

Individual value orientations, political, and ideological attitudes

While much research and many reports focus on sociodemographic correlates as intuitive explanations for climate change beliefs, a meta-analysis by Hornsey et al. (2016) found that values, ideologies, worldviews, and political orientation have greater predictive power.

Political conservatism has been identified as a significant predictor of low climate change awareness and climate change denial. People who identify as politically conservative often exhibit a tendency to justify and maintain the existing social order. This motivated perception, known as system justification theory, may lead conservatives to downplay the threat of climate change or deny its existence (Feygina et al., 2010; van der Linden, 2017). Additionally, a right-wing political orientation is associated with lower belief in climate change and a reduced perception of its severity (McCright et al., 2016). The polarization of viewpoints between political elites and selective media consumption has contributed to significant divides on environmental issues, especially in the United States (McCright & Dunlap, 2016).

Social groups that support right-wing populist parties are more likely to reject mainstream environmental concerns for ideological reasons, leading to lower climate change awareness (Althoff, 2022). Since science, political parties, and the media are often perceived as representatives of the mainstream, a decline in trust toward these institutions is frequently associated with opposition to environmental causes (Inglehart & Norris, 2016). Conversely, individuals who have greater trust in the government and its capabilities are more likely to support climate-friendly policies (Konisky et al., 2008; Kulin & Johansson Sevä, 2020) and exhibit higher climate change awareness (Droste & Wendt, 2021). Trust in political parties reflects the level of confidence individuals have in political organizations to represent their interests and act in the public's best interest. Media consumption also plays a crucial role in shaping climate change awareness, with low media consumption potentially leading to reduced awareness of climate change (Lee et al., 2015).

In addition to political attitudes, individuals’ values play a significant role. Schwartz's theory of basic human values classifies value orientations into two dimensions: "openness to change vs. conservation" and "self-transcendence vs. self-enhancement." Individuals with high self-transcending values (e.g., universalism and benevolence) tend to exhibit greater climate change awareness, while those with self-enhancing values (e.g., power and achievement) are more likely to show lower climate change awareness and recognition (Corner et al., 2014; Milfont et al., 2015). Regarding the "openness to change vs. conservation" dimension, individuals with a high degree of openness, self-direction, and stimulation values may display higher climate change awareness due to their readiness for new experiences and independent action. Conversely, those with high values in security, conformity, and tradition (conserving values) may exhibit lower climate change awareness due to their emphasis on self-restraint and resistance to change (Milfont et al., 2015; Poortinga et al., 2011). This value dimension closely parallels other value concepts, such as Inglehart’s materialism/post-materialism attitudes, which are not directly measured by items in the European Social Survey (Schwartz, 2003).

  • Hypotheses 1a and 1b: Left-wing self-positioning on the political left–right spectrum, high trust in political parties, high media consumption, high self-transcending and low self-enhancing values, high openness to change and low conservation values are associated with higher climate change awareness (H1a), and lower likelihood of climate change denial (H1b).

Sociodemographic predictors

Socio-demographic characteristics, such as age, education, gender, ethnicity, parenthood, and religiosity, have been found to be associated with various aspects of climate awareness and environmental attitudes. We consider these sociodemographic characteristics alongside values and attitudes not only as control variables but also to explore how they fare when multiple and diverse explanatory factors are examined simultaneously.

For example, several studies have demonstrated that age is a significant predictor of climate change awareness (Grunenberg & Kuckartz, 2003; Hornsey et al., 2016; Milfont et al., 2015; Poortinga et al., 2019). This association may reflect an independent effect, as younger individuals, with longer life expectancies, have more at stake in avoiding climate risks and stand to lose more from the exacerbation of the climate crisis. However, younger cohorts may have received more climate change education than older ones (Stevenson et al., 2014), potentially confounding age with varying levels of educational exposure. Additionally, older individuals are often more integrated into the existing social order and may have more at stake in societal changes prompted by environmental challenges. This aligns with theoretical assumptions regarding the relationship between conservative value attitudes and climate change awareness. Without considering these value orientations and ideological positions, age may merely serve as a proxy for individuals' values and attitudes. Only by jointly examining demographic and value-oriented explanatory factors can we disentangle these observed associations.

Previous studies have consistently found a positive effect of higher education levels and more years of schooling on individuals' climate change awareness (Droste & Wendt, 2021; Kuckartz & Rheingans-Heintze, 2006). Higher education is associated with greater knowledge and a better understanding of climate change and its consequences, leading to increased environmental consciousness (Ranney & Velautham, 2021; van der Linden, 2017). However, an important interaction between education and political orientation has also been found. In the United States, a higher level of education correlates with higher risk perception of climate change among liberals and Democrats, but with lower risk perception among conservatives and Republicans (McCright & Dunlap, 2016). However, in a large study across 64 countries, involving diverse questions and datasets, Czarnek et al. (2021) found that the positive effects of education on pro-climate beliefs for the political left and the negative effects for the political right were observed in just a small number of countries: the US and other highly developed, mainly English-speaking countries. This aligns with the findings of Smith and Mayer (2019). However, at high levels of a country’s development, such as in the European context, the positive effects of education become attenuated by right-wing ideology (Czarnek et al., 2021). We will reexamine this relationship in our study.

Women tend to have higher climate change awareness and are less likely to deny climate change compared to men (Knight, 2019; Weber, 2015; Zafar & Ammara, 2024). McCright (2010) argues that these gender differences could be attributed to different socialization experiences, which can lead to varying levels of trust in science and climate change awareness. In terms of ethnic and racial minorities, Stevenson et al. (2014) found that non-white respondents in the United States exhibit higher climate change awareness than white respondents. Additionally, when considering the joint impact of gender and race on risk perception, a "white male effect" was identified (Finucane et al., 2000). This refers to the phenomenon where white (conservative) men are generally less concerned about risks such as climate change and approach them with a higher degree of skepticism compared to women and non-white individuals (McCright & Dunlap, 2011). Furthermore, climate-related attitudes among various population groups can arise from differences in subjective experiences and vulnerability to environmental risks (Satterfield et al., 2004). For example, ethnic minorities often tend to live in areas more affected by negative climate impacts (Stevenson et al., 2014). Additionally, women and members of ethnic minorities tend to be more vulnerable, have less control, and benefit less from the status quo (Finucane et al., 2000).

Findings concerning individuals' income and climate-related attitudes are not consistent. Mildenberger and Leiserowitz (2017) found no evidence in the USA that changes in either individual economic well-being or state and local economic conditions explain the public belief that climate change is happening. However, several studies have corroborated the importance of economic factors. On the one hand, low-income groups tend to be more vulnerable to climate change impacts than those with more financial resources. This vulnerability may lead to a more pronounced awareness among low-income individuals and a higher risk assessment towards climate change (Scruggs & Benegal, 2012; Tschakert, 2007). On the other hand, previous studies have often concluded that individual income and higher socio-economic status have a positive effect on climate change awareness and environmental attitudes (Bronfman et al., 2015; Hornsey et al., 2016; Kemmelmeier et al., 2002; Kuckartz & Rheingans-Heintze, 2006; Lee et al., 2015; Shen & Saijo, 2008), which we aim to test in this study. This is also in line with the proposition of a "finite pool of worry" (Weber, 2015), which implies that people can only pay attention to and perceive a limited number of threats at the same time. Individuals who worry about their own or national financial security have less capacity to perceive climate change as a problem (Whitmarsh, 2011). This explanation is thus consistent with the assumed positive income effects on individuals' climate change awareness. However, since the theory is specifically based on the perception of risks and worries, we supplement the income-related research hypotheses with ones that include the perception of financial hardship in the household and satisfaction with the state of the economy in one's own country.

Having children can contribute to climate change awareness, as anxiety about the well-being of one's children and their future often involves factors outside of direct parental control, such as the threat climate change poses to future generations. A survey in Sweden found that parenthood increased concern about the effects of climate change on the next generation, particularly for fathers (Ekholm, 2020). Finally, when examining religiosity as a predictor of climate-related attitudes, most studies outside the USA and other Anglophone countries (Lewis et al., 2019) have found a positive effect of religiosity on environmental concerns (Kvaløy et al., 2012; Mostafa, 2016). However, some studies have not found a significant relationship (Arikan & Günay, 2021).

Based on theoretical propositions and previous findings, we generate two hypotheses linking sociodemographic profiles of individuals to the level of climate change awareness and climate change denial:

  • Hypotheses 2a and 2b: Women, younger age, higher educational level, belonging to an ethnic minority, high household income, low perceived financial hardship, high satisfaction with the state of the economy of one's country, parenthood, and religiosity are associated with higher climate change awareness (H2a), and with a lower likelihood of climate change denial (H2b).

Country-specific predictors of climate-related attitudes

National gross domestic product (GDP) has been widely studied as a contextual factor affecting climate change attitudes. Studies have found mixed results, with some indicating a positive effect of wealth on individual climate change awareness, acknowledgement, and risk perception (Arikan & Günay, 2021; Franzen & Meyer, 2010; Knight, 2018; Shum, 2012), while others suggest a negative or no effect (Lee et al., 2015; Mildenberger & Leiserowitz, 2017; Mostafa, 2016). The prosperity hypothesis suggests that in affluent societies, where the demand for a higher quality environment rises, climate-related concerns become more important (Diekmann & Franzen, 1999). However, being directly affected by the negative consequences of climate change may also lead to higher climate change awareness in poorer countries, highlighting the importance of national risk factors.

Experiences with local warming and extreme weather incidents have been shown to impact individual climate-related attitudes. Inhabitants of countries with a higher prevalence of extreme weather events may be more aware of climate change risks and its consequences (Echavarren et al., 2019; McCright et al., 2014; Shum, 2012; Zaval et al., 2014). However, this effect has not been found in the USA (Marquart-Pyatt et al., 2014).

National carbon footprint is another country-level contextual factor. From a cost–benefit standpoint, individuals in countries with high national per capita CO2 emissions and environmentally damaging economic systems may fear increased costs during the transition to a more environmentally-friendly economy. Indeed, studies have shown that greater reliance on fossil fuel production is linked to lower public awareness of environmental issues (Knight, 2018) and fewer behavioral intentions aimed at reducing climate change impact (Hao et al., 2020). Perceived threats to socioeconomic security may drive a desire to protect one's economic interests and maintain the status quo (Lübke, 2022), potentially resulting in lower climate change awareness or increased climate change denial.

Hofstede’s (1980) pioneering study on culture across modern nations identified a cultural dimension known as "individualism vs. collectivism," which appears to be associated with climate change awareness (Stevenson et al., 2014; Xiang et al., 2019). Hofstede categorized societies as either individualistic or collectivistic based on the value orientations of individuals within those societies. However, it is important to note the distinctions between our measurement of values at the individual and societal levels. For individual-level value orientation, we utilize Schwartz’s theory of basic human values, which are essentially intercultural values rooted in universal human motivations. At the country level, we apply Hofstede’s cultural dimension of individualism vs. collectivism, which serves to distinguish between cultures and societies. Collectivism emphasizes group norms and collective harmony, while individualism prioritizes personal goals over group goals. This cultural orientation is particularly pertinent in the context of climate change, where the individualistic perspective can prove counterproductive when dealing with global issues necessitating collective action and the prioritization of collective goals over individual ones.

Finally, the post-communism hypothesis suggests that there is a widening gap in environmental policy orientations and individual climate-related attitudes between Western and post-socialist Eastern European countries. Several studies have confirmed differences in climate change attitudes (Baiardi & Morana, 2021; Marquart-Pyatt et al., 2019; Scruggs & Benegal, 2012; Zafar & Ammara, 2024). In Eastern European countries, environmental issues tend to be less integrated into the political culture, and there is a weak connection between environmental issues and other, more salient, economic and political questions (Chaisty & Whitefield, 2015).

  • Hypotheses 3a and 3b: High GDP per capita, high national risk factors (temperature changes, extreme weather events), low national CO2 emissions, low values on cultural individualism, and not living in a post-socialist country are associated with higher climate change awareness of country inhabitants (H3a), and lower propensity to deny climate change (H3b).

Data and methods

We obtained individual-level data for this study from Round 8 of the European Social Survey, European Research Infrastructure (ESS ERIC, 2020). The European Social Survey (ESS) has been conducted bi-annually since 2002 across more than 30 countries. The survey measures the attitudes, beliefs, and behavioral patterns of diverse populations in participating nations. ESS adheres to rigorous sampling requirements. Countries’ samples must be representative of all individuals aged 15 and above, encompassing residents in private households in each country, irrespective of nationality, citizenship, or language. Various sampling frames, including individuals, households, and addresses, may be employed, with individuals selected through random probability methods at each stage.

The ESS Round 8 was collected in 2016/2017 and includes a module on climate and energy. In total, the dataset includes 44,387 respondents from 22 European countries and Israel. For the analysis of the composite index of climate change awareness, we can utilize 39,439 cases, and for the analysis of climate change denial, 42,877 cases. Country-specific context data were derived from various other sources, as specified below, and assigned to the respective countries in the ESS data set.

Operationalization of dependent variables

Climate change awareness: To quantify the level of climate change awareness among survey respondents, we created a composite variable using several measures. The first relevant item assessed people's belief in climate change. Respondents were asked, "You may have heard the idea that the world's climate is changing due to increases in temperature over the past 100 years. What is your personal opinion on this? Do you think the world's climate is changing?" The response options were: 'definitely changing,' 'probably changing,' 'probably not changing,' and 'definitely not changing.' Respondents who claimed that climate change is definitely not happening were excluded, and the scale was reversed so that higher values denote higher awareness.

Further measures included information on climate concern ("How worried are you about climate change?"; a five-point Likert scale ranging from ‘not at all’ to ‘extremely worried’), causes ("Do you think that climate change is caused by natural processes, human activity, or both?"; a five-point scale ranging from ‘entirely by natural processes’ to ‘entirely by human activity’), climate change impacts (“How good or bad do you think the impact of climate change will be on people across the world?”; an 11-point scale from ‘extremely bad’ to ‘extremely good’, reversed), and responsibility ("To what extent do you feel a personal responsibility to try to reduce climate change?"; an 11-point scale from ‘not at all’ to ‘a great deal’).

After conducting factor analysis on these five original variables, which confirmed that all items load onto a single factor, we computed Bartlett scores. These scores were used in subsequent analyses to determine an individual's placement on the composite variable, Climate change awareness, for respondents who do not outright deny the existence of climate change.

For respondents who claimed that climate change is 'definitely not changing,' none of the subsequent questions on climate change beliefs were applicable, as these questions presuppose acknowledgment of climate change. To capture this outright denial, we created a second, binary outcome variable: Climate change denial. This allows us to assess whether the factors associated with climate change denial are the same as those associated with the level of climate change awareness, thus determining if they reflect different positions on a single climate change-related spectrum.

Independent individual-level variables

Individuals’ value orientation was measured in ESS with a modified 21-item version of the Portrait Values Questionnaire (PVQ) to measure human values according to Schwartz (2003), see also Bilsky et al. (2011). Each item describes a person, and respondents indicate on a six-point scale how similar this person is to themselves, ranging from 1 ‘very much like me’ to 6 ‘not like me at all’. Values are organized along two bipolar dimensions: openness to change vs. conservation as well as self-transcendence vs. self-enhancement. The scales were constructed from the corresponding items in the ESS. The (reversed) values on twelve items in ESS were used to create an index Openness vs. conservation, with higher values indicating that respondents emphasize independence of thought, action, and feelings, as well as readiness for change, while lower scores indicate a preference for conservation and resistance to change. Similarly, the Self-transcendence vs. self-enhancement dimension was measured by an index where higher values reflect a greater orientation toward self-transcendence, i.e. motivational value orientation emphasizing concern for the welfare and interests of others.

Left–right: To measure political ideology, respondents were asked to position themselves on a scale ranging from 0 ‘left’ to 10 ‘right’. We also created an interaction between self-placement on this left–right scale and educational attainment. Trust in political parties was assessed on a scale from 0 ‘no trust at all’ to 10 ‘complete trust’. Media consumption was measured by the number of hours respondents typically spend per day watching, reading, or listening to news about politics and current affairs.

Age was calculated from the respondent’s birth year. Gender was recoded into a binary variable Woman, with women coded as 1 and men as the reference category. Education was categorized into eight levels according to the International Standard Classification of Education (ISCED11): less than primary, lower secondary, upper secondary, post-secondary non-tertiary, short-cycle tertiary, bachelor or equivalent, master or equivalent, and doctoral or equivalent. Household income was measured as total net income and categorized into country-specific deciles to account for cross-national differences in income levels and distribution. Financial hardship was assessed by how respondents felt about their household’s income, on a scale from 1 ‘living comfortably on present income’ to 4 ‘finding it very difficult on present income’. Satisfaction with the economy was measured by respondents' general satisfaction with the state of their country’s economy, on a scale from 0 ‘extremely dissatisfied’ to 10 ‘extremely satisfied’.

Ethnic minority is a binary variable indicating whether the respondent belongs to an ethnic minority within their country. Parental status is not directly asked in the ESS; therefore, it was derived from the household grid, which details current household members, and from the question on whether the respondent has ever had children. A dummy variable, Parent, was created to indicate whether a respondent has or had children, including adopted, step, or foster child(ren). Religiosity was measured on a scale from 0 ‘not at all religious’ to 10 ‘very religious’.

Independent country-level variables

To measure national prosperity, the logarithm of GDP per capita was included in the analysis. GDP, measured in 2016 U.S. dollars, was obtained from the World Bank Group's World Development Indicators dataset (World Bank Group, 2022). The range of GDP per capita spans from $8,705 in Russia to $83,073 in Switzerland.

National risk factors include year-to-year temperature changes, with data sourced from the Food and Agriculture Organization of the United Nations (FAOSTAT, 2022). These temperature changes are measured relative to a baseline climatology for the period 1951–1980. In 2016, the minimum temperature change was 1.0 °C in Ireland, and the maximum was 2.5 °C in Russia. The impact of climate change consequences was assessed using the Climate-Risk Index (CRI), developed by Germanwatch (Eckstein et al., 2018). The CRI measures the quantified impacts of extreme weather events between 1998 and 2017, including both fatalities and economic losses. France experienced the highest impact, with the lowest CRI score of 38.67, while Iceland was least affected, with the highest CRI score of 117. Carbon footprint was measured by country-specific CO2 emissions for 2016, obtained from the World Bank Group's World Development Indicators dataset (World Bank Group, 2022). Sweden recorded the lowest annual CO2 emissions at 3.9 tons per capita, and Estonia had the highest at 11.9 tons.

The Individualism vs. collectivism score is based on Hofstede's dataset of cultural dimensions at the national level (Hofstede, 2015), with a range from 1, indicating the lowest level of individualism, to 120, indicating the highest. Portugal and Slovenia are the least individualistic countries with the score of 27, while Great Britain is at the other end of the spectrum, with a score of 89. A dummy variable Postsocialism is coded as 1 for post-socialist countries in ESS (Czech Republic, Estonia, Hungary, Lithuania, Poland, Russia, and Slovenia) and 0 for all other countries.

Methods

In addition to conducting descriptive statistics and factor analysis to develop the measure for environmental awareness, we employed multilevel modeling with random intercepts at the country level for the core data analysis, accounting for the hierarchical structure of the data (Snijders & Bosker, 2012). Multilevel modeling allows us to analyze individuals (level-1 units of analysis) nested within countries (level-2 units). The random intercept captures variations in attitudes towards climate change across countries and helps determine whether these cross-country differences are influenced by country-level variables. We constructed linear models to examine the determinants of climate change awareness and logistic models to analyze climate change denial. The analyses were performed separately for each dependent variable.

In the initial null model, we included only the variable distinguishing countries to assess the proportion of variance explained by differences between countries. This step determines whether multilevel regression is appropriate for our data analysis. Next, we built a series of random intercept multilevel regression models. Model 1 included individual value orientations, political, and ideological attitudes; Model 2 incorporated only sociodemographic predictors; and Model 3 combined all individual-level predictors. Except for dummy variables, all independent variables were centered at the overall mean (grand mean centering). Subsequently, we introduced country-level variables into the models (Models 4–10), adding each level-2 variable to Model 3 sequentially, one at a time. This approach was necessary due to the limited number of countries, which constrained our ability to reliably estimate all level-2 parameters simultaneously.

Regarding missing data, some individual-level ESS variables had incomplete data. Although most missing values were minimal, the self-placement on the left–right political scale had up to 13% missing cases. To avoid data loss, we opted against case deletion. Instead, we imputed missing values with the country-specific arithmetic mean of the respective variable (excluding dummy variables) and created indicator variables for missingness. These indicators were used to assess whether missing cases significantly differed from average individuals with complete data.

Results

Existence of climate change and awareness index

Over 80% of respondents in European countries believe that climate change is either definitely or probably occurring, with Portugal reporting the highest proportion (77.9%) who affirm it definitely exists (Fig. 1). Weighted data reveal that only 2.3% of Europeans deny the existence of climate change, regardless of its cause. However, there are notable differences between countries. Russia has the highest percentage of climate change deniers at 9.7%, followed by Israel at 6.9% and Lithuania at 6.4%.

Fig. 1
figure 1

Attitudes about climate change in Europe. Responses to the question "You may have heard the idea that the world's climate is changing due to increases in temperature over the past 100 years. What is your personal opinion on this? Do you think the world's climate is changing?". Source: European Social Survey, Wave 8

Full size image

The climate change awareness index was calculated for ESS respondents who acknowledged the possibility of climate change. This index, our primary dependent variable, is a composite score derived from an extracted factor score based on measures of beliefs about the existence of climate change, its causes, environmental concern, its impact, and personal responsibility (see section "Operationalization of dependent variables"). The Kaiser–Meyer–Olkin criterion (KMO = 0.71) confirmed the data’s suitability for factor analysis, indicating moderate adequacy and suggesting that factor analysis would produce interpretable factors. A KMO value above 0.7 indicates that factor analysis is appropriate, with values above 0.8 being very good and values below 0.5 generally unacceptable. All variables loaded onto a single factor, labeled as climate change awareness. Factor scores were then computed and used in subsequent analyses using Bartlett's method. Bartlett scores, generated through maximum likelihood estimates, provide the most likely representation of the "true" factor scores (DiStefano et al., 2009). The resulting coefficients have a mean of 0 and are unbiased, reflecting the cases' positions on the latent continuum in the population (ibid.). Individual values ranged from -4.33 to 2.99, with higher values indicating greater environmental awareness.

Figure 2 displays country-specific mean values of climate change awareness, revealing significant variation across European countries. Spain has the highest average climate change awareness score of 0.66 in Europe, while Estonia has the lowest average score of -0.62. Notably, the countries with the lowest climate change awareness scores—Estonia, the Czech Republic, Russia, Poland, and Lithuania—are post-socialist countries. This finding supports Hypothesis 3a, which postulates a post-socialism effect on climate change awareness.

Fig. 2
figure 2

Variability in climate change awareness score in Europe by country. The climate change awareness score is mean centered. Source: Authors’ estimates from the European Social Survey, Wave 8

Full size image

Determinants of climate change awareness

This section explores individual sociodemographic characteristics, attitudes, value orientations, and country-level determinants of environmental awareness. Table 1 presents the results of the multilevel regression. The first column shows the so-called null or 'empty model' to assess whether multilevel modeling is appropriate for the data analysis. The Intraclass Correlation Coefficient (ICC), which measures the degree of clustering in the data and indicates the variance explained by differences between countries (Snijders & Bosker, 2012), is 0.079. This means that country-level clustering explains about 8% of the total variance in climate change awareness.

Table 1 Determinants of climate change awareness in Europe. Two-level linear regression models with random intercepts, ESS data, Wave 8
Full size table

Model 1 in Table 1 supports several hypothesized relationships between values and climate change awareness. Specifically, right-wing political orientation is associated with lower climate change awareness, while trust in political parties correlates with higher awareness. Self-transcendence, reflecting a focus on concerns beyond one's personal interests, and openness to change are positively related to climate change awareness. These results also suggest a relationship between post-materialistic values and climate change awareness, as the items for operationalizing the openness vs. conservation dimension can be used as parallels for Inglehart's measure of post-materialism/materialism (Schwartz, 2003: 272). The hypothesis that conservatism negatively impacts climate change awareness is thus supported. However, media consumption related to politics and current affairs does not exhibit the anticipated relationship with climate change awareness. Notably, missing survey data provides additional insights, as respondents who did not disclose their political ideology or trust in political parties scored particularly low on the climate change awareness scale.

Model 2 in Table 1 incorporates sociodemographic characteristics that may impact climate change awareness among individuals in Europe. As hypothesized, younger age, female gender, higher educational attainment, belonging to an ethnic minority, higher household income, and religiosity are significantly associated with higher climate change awareness. For example, women score on average 0.13 points higher on the environmental awareness scale than men. However, parenthood does not show a significant association with climate change awareness. Interestingly, respondents facing financial hardship exhibit higher climate change awareness, which contradicts the "finite pool of worry" theory. Additionally, those who are more satisfied with their country's economy display lower climate change awareness. It is also noted that missing responses on educational variables and perceptions of the country's economy are not random; they are systematically related to lower climate change awareness.

The next model in Table 1 includes all individual-level variables. The results remain largely consistent with some modifications. For example, the coefficient for women remains positive and significant, but its magnitude is nearly halved in Model 3, indicating that the effect can be largely explained by gender differences in political attitudes and value orientations. While belonging to an ethnic minority and the missing values indicator for satisfaction with the country's economy no longer reach statistical significance, the coefficient for media consumption becomes positive and significant when controlling for sociodemographic factors. The positive main effect of education on climate change awareness persists, but the interaction term reveals that highly educated individuals with a right-wing political orientation exhibit significantly lower climate change awareness. These findings align with Czarnek et al. (2021), who observed that in the European context, the positive effects of education on climate change awareness are diminished by right-wing ideology. Figure 3 graphically displays the coefficients and 95% confidence intervals of the variables as estimated in Model 3.

Fig. 3
figure 3

Level-1 predictors of climate change awareness. Plots of regression coefficients (dots) and 95% confidence intervals (whiskers), as estimated in Model 3, Table 1. Source: Authors’ estimates from the European Social Survey, Wave 8

Full size image

Next, we estimated a series of models with aggregate level-2 independent variables, adding them one by one to the model that already contained level-1 variables. Model 4 (Table 1) incorporates the logarithm of GDP per capita for the countries involved. While the parameter estimate is positive (b = 0.175, p = 0.084), it does not reach conventional significance levels, indicating that the data do not support the assumption that climate change awareness is significantly higher in more affluent European countries when individual factors are controlled.

National risk factors were assessed using temperature change and the Climate Risk Index (CRI). Model 5 in Table 1 shows that a country's exposure to temperature changes in 2016 does not significantly influence its citizens' climate change awareness. However, Model 6 reveals a significant negative coefficient, indicating that higher climate change awareness is associated with countries that experienced extreme weather events, measured in terms of fatalities and/or economic losses. This finding partially supports Hypothesis 3a, suggesting that exposure to extreme weather events may enhance climate change awareness.

The coefficient for carbon footprint, measured by annual CO2 emissions per capita, is negative and highly significant. This supports the cost–benefit hypothesis, suggesting that climate change awareness is lower in countries with higher per capita CO2 emissions, where the transition to a more environmentally friendly economy is perceived as more costly. The cultural dimension of individualism vs. collectivism does not show a significant effect on climate change awareness. However, Model 9 indicates that individuals in post-socialist countries have significantly lower climate change awareness compared to those from other European countries.

Determinants of climate change denial

In the next stage, we examined the factors influencing climate change denial using a binary dependent variable in logistic multilevel analysis. The ICC in the null model (Table 2) indicates that country clustering explains nearly 18% of the total variance, signifying substantial variations in climate change denial among European countries, more so than in the level of climate change awareness.

Table 2 Determinants of climate change denial in Europe. Two-level logistic random intercepts models, ESS data, Wave 8. Odds ratios are reported
Full size table

In the next step, level-1 variables were included in the models. Results are reported as odds ratios (OR), which indicate the change in the odds of climate change denial associated with a one-unit change or presence of each variable, holding other independent variables constant. Model 1 (Table 2) reveals that political orientation and media consumption are not significantly related to climate change denial. However, trust in political parties reduces the odds of denial by 5% for every one-unit increase on the scale. The openness vs. conservation value orientation is linked to lower odds of denial for those emphasizing openness. Conversely, missing values indicate significantly higher odds for individuals who cannot or choose not to answer questions about this value orientation. Additionally, self-transcendence, which emphasizes concern for others and collective welfare, has a large, negative, and highly significant effect on the odds of climate change denial.

Age, parental status, household income, satisfaction with the state of the economy in one's country, belonging to an ethnic minority, and religiosity did not significant affect the odds of climate change denial (Table 2, Model 2). However, women and better-educated individuals are significantly less likely to deny the existence of climate change, whereas financial hardship is associated with an increase in the odds of denial. Although religiosity is not statistically significant, respondents who refused to answer questions on religiosity have odds of climate change denial increased by 83%. In the joint estimation (Model 3), both sociodemographic and value-related variables largely retain their significance. Log-odds coefficients and 95% confidence intervals are depicted in Fig. 4.

Fig. 4
figure 4

Level-1 predictors of climate change denial. Plots of log-odds coefficients (dots) and 95% confidence intervals (whiskers), as estimated in Model 3, Table 2. Source: Authors’ estimates from the European Social Survey, Wave 8

Full size image

Models 4 through 9 extend Model 3 by incorporating level-2 country-level variables one by one. Among these, only two country-level characteristics significantly impact climate change denial. The negative odds ratio (OR) for GDP per capita indicates that individuals in less affluent countries have higher odds of denying climate change. Furthermore, climate change denial is notably higher in post-socialist societies, with the odds being 2.4 times greater compared to other countries.

Conclusions

This study is the first to simultaneously investigate both climate change denial and the level of climate change awareness among individuals who acknowledge the possibility of climate change, while also accounting for a broad range of individual-level value orientations, socio-demographic characteristics, and country-level factors. This comprehensive approach provides valuable insights into these phenomena across European populations.

Our measure of climate change awareness, conceived as a multidimensional concept encompassing beliefs about the existence of climate change, its causes, environmental concern, impact, and personal responsibility, reveals significant variation both among individuals and across countries. For instance, Spain exhibits the highest average climate change awareness score, while most post-socialist countries, except for Slovenia, display below-average scores. Key determinants of climate change awareness include value orientations and political attitudes. Higher awareness is associated with left-wing political positioning, trust in political institutions, a strong inclination towards self-transcending values, and high openness to change. In simpler terms, individuals who embrace post-materialist values, value independent thinking and action, and prioritize the well-being of others are more likely to demonstrate higher climate change awareness. Conversely, those emphasizing self-interest, valuing order and resistance to change, possessing right-wing political orientations, and exhibiting low trust in political parties tend to show lower climate change awareness.

Our findings corroborate many previous studies linking socio-demographic characteristics with climate change attitudes. Higher educational levels, being female, and having trust in political parties are consistently associated with both the acknowledgment of climate change and higher climate change awareness. Higher education generally implies greater knowledge and understanding of climate change and its consequences (Droste & Wendt, 2021; Ranney & Velautham, 2021; van der Linden, 2017), and likely fosters greater trust in science and scientific warnings. We also examined whether the effects of education vary depending on political ideology, as reported by McCright and Dunlap (2016) for the United States. Our results align with Czarnek et al.’s (2021) findings that in more developed countries, the positive effects of education are attenuated by right-wing ideology. The interaction term between education and political orientation in our analysis indicated that highly educated individuals who align with right-wing politics exhibit lower climate change awareness. This may be due to conservative parties' tendencies to oppose climate science, leading their well-educated supporters to question prevailing knowledge and scientific evidence of global warming.

Interestingly, we did not find a significant relationship between political left–right orientation and climate change denial in Europe. This finding aligns with the observation that nationalist ideology has a greater influence than traditional left–right political orientation in Europe (Kulin et al., 2021). This contrasts with the strong influence of the conservative movement and anti-environmentalism in the USA, which is driven by think tanks, media figures, and contrarian scientists (Dunlap & McCright, 2010). In Europe, climate change denial is more closely associated with dissatisfaction with one’s financial situation (but not with household income levels), low trust in political parties, and value orientations that emphasize conservation and self-enhancement while showing low openness. Specifically, being a man with lower education levels who values power and achievement (emphasizing self-interest, relative success, and dominance over others, which corresponds to the self-enhancement pole on the human values scale), combined with dissatisfaction with one’s financial situation and high distrust in established social institutions, creates a fertile ground for climate change denial.

Multilevel modeling highlighted the existence of country clustering, indicating significant variation in climate change awareness across countries. On average, climate change awareness is notably higher in countries that have experienced extreme weather events before the survey, suggesting that such events make environmental issues more salient to the population. Conversely, a higher national carbon footprint is associated with lower public awareness, aligning with previous research (Knight, 2018; Lübke, 2022). Heavy reliance on fossil fuels and high CO2 emissions may lead to resistance to environmentally conscious attitudes, perceiving them as threats to a country's industrial development and economic interests.

Climate change denial, however, does not appear to be significantly influenced by national risk factors or carbon footprint. Instead, it is more prevalent in poorer countries, with a particularly high prevalence in post-socialist societies. This finding underscores the most pronounced climate-related attitude differences observed in cross-country comparisons, particularly between Eastern and Western European countries. This confirms earlier literature indicating that post-socialist countries maintain distinctive attitudes toward environmental issues. Chaisty and Whitefield (2015) attribute this phenomenon to the persistence of political culture, which causes citizens to view environmental issues through ideological lenses inherited from the socialist era, coupled with shared negative experiences during the post-socialist transition.

Our findings underscore the importance of considering both individual and country-level factors when examining climate change awareness and denial. The results suggest that fostering self-transcending values, enhancing trust in political institutions, and promoting openness to change may help improve climate change awareness. Educational efforts on environmental issues should also be increased to bolster public understanding. Additionally, addressing socio-economic disparities and post-socialist legacies may be crucial in tackling climate change denial and promoting broader climate action across diverse European contexts.

Finally, the inclusion of indicator variables for missing data not only enabled us to retain a larger number of cases for analysis but also uncovered systematic relationships. Respondents who did not answer questions about education, the left–right political scale, and trust in political parties were more likely to exhibit lower climate change awareness. Those unwilling or unable to respond to questions about human values (openness vs. conservation) and religiosity demonstrated a higher propensity to deny the existence of climate change. These findings strongly suggest that non-response to survey questions merits further investigation. Discarding missing data could lead researchers to overlook important theoretical insights related to non-response patterns and political orientation. Additionally, non-response may reflect ideologically motivated refusals and a broader skepticism toward scientific institutions, such as the ESS, and climate science itself. This area warrants more theoretical exploration to better understand the implications of non-response on climate change research.

In conclusion, our study indicates that climate change awareness and denial are influenced by distinct determinants and do not exist on a single continuum from denial to increasing awareness. Instead, these phenomena are driven by different factors. Climate change awareness is significantly shaped by sociodemographic characteristics, values, and political orientation, and is more receptive to educational efforts aimed at enhancing scientific understanding. In contrast, climate change denial appears to be less influenced by sociodemographic and political factors and is more closely associated with personal worldviews and a detachment from mainstream society. This distinction underscores the need for targeted strategies that address the unique drivers of each phenomenon to effectively promote climate action and understanding.

However, our findings should be interpreted in light of several limitations. First, the analysis relies on cross-sectional data, which precludes definitive causal conclusions. Second, using country-level aggregates as the unit of analysis may not be ideal for certain indicators, especially cultural aspects such as individualism versus collectivism, as within-country heterogeneity can often exceed differences between national averages (Kaasa et al., 2013). Additionally, systematic associations with missing data related to environmental issues were identified—an area not previously explored in depth. The absence of responses is frequently linked to lower climate change awareness and a higher likelihood of climate change denial. This raises concerns about the representativeness of individuals with specific value orientations and ideological motivations in existing surveys, even those with rigorous methodologies and sampling procedures like the ESS. It is possible that some individuals systematically avoid participating in scientific studies, suggesting that indifference or opposition to environmental issues might be more widespread than indicated by survey data. These limitations underscore the need for further research to better understand the complexities surrounding climate change perceptions.