Credibility at stake. A comparative analysis of different hate speech comments on journalistic credibility and support on climate protection measures

Abstract

The contentious nature of certain forms of climate activism has led to a proliferation of hate comments targeting activists in online news forums. This study examines the ‘Nasty Effect’—the impact of uncivil language—on attitudes towards climate protection news. Specifically, we investigate how various types of hate speech within user comments influence the credibility of articles on climate protection and activism. Employing a between-subject design (7x1), we conducted an online survey in Austria during autumn 2023 (N = 486). Our findings reveal that as the hate speech intensity increased, the perceived credibility of related news articles declined. While overall means did not yield significant differences, individual assessments compared to the control group were significant. Explicit hate speech exhibited a more pronounced negative impact on article credibility. The absence of a significant ‘Nasty Effect’ underscores the intricate nature of this phenomenon and its nuanced influence on public attitudes towards climate protection measures.

KEY POLICY HIGHLIGHTS

• The increasing influence of implicit hate speech on online discourse, which has the potential to undermine the deliberative function of user forums, must be addressed. Measures are needed to combat implicit forms of hate speech as well.

• Hate speech has the potential to harm joint social endeavours such as combating climate change. Awareness of this danger must be created in society and politics.

• Recognise the link between perceived journalistic credibility, user comments and topic reception. Measures need be taken to mitigate the impact of incivility in online forums to increase journalism’s credibility.

Keywords:

• Hate speech
• news credibility
• climate protection
• between-subject design
• ‘Nasty Effect’

Reviewing Editor:

• Guangchao Charles Feng, School of Communication, Hong Kong Baptist University, Hong Kong

Subjects:

• Environmental Communication
• Mass Communication
• Political Communication
• Political Behavior and Participation

1. Introduction

Combating climate change is one of the most significant challenges of the 21st century. The Sustainable Development Goals (SDGs) of the United Nations represent global guidelines for climate goals. For instance, goal 13 is ‘take urgent action to combat climate change and its impacts’ (United Nations, 2015). This aim can only be achieved through a collective social effort. Public communication plays a central role in this achievement. In a mediatised society, news media bear a unique role as they are fundamental in the discourse and framing of crises (Koerber, 2020). Accordingly, it is vital for democratic societies that people can trust the news media (Strömbäck et al., 2020), as otherwise, they will turn to alternative bodies of knowledge, some of which purposefully spread disinformation and thus weaken support for democratic structures (Kalogeropoulos et al., 2019). This can only be countered by comprehensive news media coverage perceived as credible (Prandner, 2022). The credibility of information plays a significant role in what information is shared and disseminated (Yang et al., 2022). However, current data show a downward trend for trust in news media worldwide (Newman, 2023) and especially in Austria (Gadringer et al., 2023, p. 111).

The media also play an important role in reporting climate change because people living in a mediatised society learn about climate change from news media (Schäfer, 2015), as the most essential source for information about climate related events, such as protest actions carried out by environmental movements. Meaning makers and important political actors in a democracy, who highlight emerging and pressing issues and introduce new ideas (Cammaerts, 2024, p. 108), actively seek media attention through protest actions to reach a wider audience (Bugden, 2020, p. 2). Particularly in relation to sensitive and controversial topics, a thematic polarisation of Austrian society becomes evident. This phenomenon is illustrated by Melischek and Seethaler (2023, p. 19) in the context of the COVID-19 measures, and a similar trend appears in the case of climate protests, which can lead to polarisation between liberal and conservative partisans (Bugden, 2020, p. 2). In her discourse-analytical study for the USA, Morris (2021) points out that supporters of progressive environmental policy are constructed as a threat and as the ‘others’. A similar phenomenon can be observed in German-speaking countries. The German tabloid ‘Bild’ directly disparages demonstrators and their movement. By reframing protest movements such as Fridays for Future as a debate about truancy, the tabloid eliminates the political implications of climate change by discussing the protest as a politico-educational issue. Conservative quality press remains subtler in its disparagement. Nevertheless, its episodic coverage reveals a variety of frames that undermines both the protesters and their political agenda (Bundtzen, 2023; von Zabern & Tulloch, 2021, p. 41).

This could be the reason why attitudes towards the demonstrations are rather reluctant in Austria and why hate comments against individual activists and the group frequently accompany coverage on social networking sites (SNS) or news media forums. Studies show that these forms of user-generated information reduce the credibility of news articles (Naab et al., 2020).

This paper aims to investigate the impact of hate speech in online comments on news articles related to climate activism and climate protection measures. We focus on Austria, because it is an interesting case to understand how negative and uncivil communication – probably based on scientific scepticism – and climate attitudes are related (Huber et al., 2021, p. 1123). We examine the effects of hate speech on the reception of news coverage and attitudes towards the discussed topic by building upon the concept of the ‘Nasty Effect’ (Anderson et al., 2014). These reveal the impact of uncivil language on public perceptions of scientific topics in journalistic content. Our study aims to dissect various intensities of hate speech, as categorised by Bahador (2023). We seek to explore how these different categories of hate speech affect both the credibility of information and individuals’ attitudes towards climate protection measures. This leads to the following research questions:

(RQ1): How does hate speech in user comments affect the credibility of news articles related to climate protection measures and activism?

(RQ2): What is the impact of hate speech comments under news articles on climate protection and of activism on individuals’ attitudes towards climate protection measures?

(RQ3): Which individual political attitudes influence the perceived credibility of climate protection measures’ reporting and attitudes towards climate protection measures and activism?

The paper is structured as follows. First, a basic understanding of climate coverage and activism is presented, followed by a presentation of current developments in attitudes towards climate protection and hate against climate activists. Then, hate speech and its effects are discussed in greater detail and our hypotheses are introduced. The operationalisation of the variables is presented in the methods part. This is followed by the description of our findings. The paper concludes with a discussion of the results, as well as a critical reflection on the procedure, the limitations and we offer an outlook for future research.

2. Theoretical background

2.1. Climate coverage

In Austria, scientists and politicians have not communicated substantially and publicly about climate change for a long time and more scientific attention should be given to the research fields of public communication, media representation of climate change and framing and its effects (Rhomberg, 2016). The Digital News Report 2020 shows that a large part of the global population informs itself about climate change through journalistic mass media. TV seems to be the most important media source (35%), followed by online news sites from major news organisations (15%). Only 7% of respondents to the Digital News Report 2020 state that they do not pay attention to climate change at all. It was shown that people on the left political spectrum consider climate change as more serious problem than right-wingers (Andı, 2020, p. 53). This underlines that political attitude seems to moderate the perception of climate change. Of all respondents, around half believe that the media reports well on climate change. However, almost half of those who do not consider climate change to be a problematic issue believe that the media do not report on it well (Andı, 2020, p. 54f.). These results illustrate that people’s opinion about how well the media reports on climate change is influenced by their overall view of how serious the issue of climate change is. Climate change sceptics are more likely to use alternative media sources, especially SNS, to inform themselves about climate related issues (Andı, 2020, p. 55).

Overall, the news media seem to be considered an important and trustworthy source for climate change information, which is more credible than the personal environment. Thereby, climate coverage shows agenda-setting effects and increases the knowledge about climate change (Schäfer, 2015, p. 858). Schäfer and Painter (2021, p. 14) also identify too much focus on researching climate reporting in print media and a research gap for (among others) TV and online sectors. The content of climate reporting thus varies significantly between countries (Schäfer, 2015, p. 858). Neverla and Schäfer (2012, p. 18) also note that climate change is not primarily seen as a scientific issue but is treated in the context of particular political or popular events, extreme weather events or natural disasters. One of the most striking series of events related to climate change in the 2010s is probably the climate protest by movements such as Fridays for Future, Extinction Rebellion, or Last Generation. This is mainly due to their ‘historically large’ mobilisation potential, especially for females and pupils, their emphasis on ‘disobedient’ forms of protest (De Moor et al., 2021, p. 623f.), and the integration of social media in their activities (see, e.g., Belotti et al., 2022). In contrast to earlier forms of climate protests, their actions are primarily directed at ‘local or national governments’ and not towards multinational industries, e.g. the ‘fossil fuel industry’ (De Moor et al., 2021, p. 623f.).

2.2. Coverage of climate activism

In general, protest movements and activists seek to change certain societal circumstances. They aim to trigger societal change, tackle societal issues, and rectify inequities in laws, institutions, and social interactions (Feinberg et al., 2020, p. 1). One prominent protest organisation active in the German-speaking region that focuses on combating climate change is Last Generation.

The protest forms primarily staged by this movement are (according to their website) non-violent direct actions, including road blockades and vandalism using paint. The activists state that they have tried other forms, such as petitions or protests against the government. By using ‘some colour on a glass pane in a museum’ they might more easily get onto the front pages of newspapers than by other forms of protest (Last Generation, 2023).

Rödder and Pavenstädt (2023, p. 38) have found that the aims of the current climate activism movements are mainly based on evidence from mainstream climate research. The authors see protesters and climate scientists as ‘allies’. While the scientists develop and disseminate ‘simulations and scenarios’ of climate change, the protesters bring them ‘to broad public attention’.

When discussing the reception of climate activists and their aims, Hayes and O’Neill (2021) highlight that visually, the climate activists, including Fridays For Future and Extinction Rebellion, are primarily portrayed in a young, female, powerful and hopeful way. The authors also call for more reception research in the field of climate activism. Feinberg et al. (2020) pinpoint what they refer to as an ‘activist’s dilemma’. This dilemma revolves around the apparent impact of (what they call) extreme forms of protest, such as road blockades, vandalism, and aggressive rhetoric, on public support. These tactics may diminish public backing, affecting the protest movement and the causes it advocates for. Paradoxically, these protest methods seem particularly effective in garnering attention for the movement and its objectives and compelling organisations to act (Feinberg et al., 2020).

Scheuch et al. (2023) have found evidence that legal actions and protest forms are more often covered than illegal ones – which is contrary to the ‘activist’s dilemma’. Moreover, they show that climate action receives more attention in both liberal and conservative newspapers compared to neutral titles. Climate activism tends to generate more extensive coverage in partisan media than in non-partisan papers.

2.3. Attitudes towards climate protests and climate change

A recent survey illustrates that 60% of Germans said they would completely disagree with the attempt to enforce more climate protection with street blockades. Right-wing and conservative voters rejected the actions almost entirely – while green voters only rejected the actions by around 53%. Younger people are more inclined to support the protesters than older people (SZ, 2023). Huber et al. (2021) outline that people who follow populist political actors are particularly inclined to deny climate change. Even if the right-wing Freedom Party of Austria (FPÖ) represents environmental (Voss, 2020) and isolated climate protection policy positions (Oswald et al., 2021, p. 187), the party is mobilising against the current protests of the Last Generation (e.g. FPÖ, 2023; Kurier, 2023), which are mainly supported by young activists and framed as the ‘terrorist’ ‘others’ (Morris, 2021). Cammaerts (2024, p. 125) can summarise for the UK that the public anger at forms of protest, as portrayed by the media, means that society is far less supportive of the movement’s repertoire of actions than of its concerns about climate change. A similar observation can be made for Germany (Mau et al., 2023, p. 215, 241). Shanaah et al. (2024) can show that pro-climate action is not only supported by the political left, but the motivations of right-wing partisans lie more in ecofascism, which is rooted in a rejection of multiculturalism and ethnocentrism rather than concern for the environment.

However, despite the relatively low support for climate activists, climate change seems to be considered a severe problem within the European Union. Around 77% of EU citizens consider climate change a serious problem at the moment – though only 62% of Austrians agree. Moreover, most EU citizens have a positive attitude towards climate protection measures and they are seen as an opportunity (European Commission, 2023). However, another recent study for Austria points out that right-wing voters (around 50%) think Austria is doing too much for climate change (Der Standard, 2023a).

Gallup data illustrate that only slightly more than one fifth of Austrians demonstrate understanding for the climate activists. Support for the protests is also greater among younger people in Austria: 40% of those under 30, but only 15% of those over 50 express understanding for their ways of protesting. Party preference also seems to influence attitudes towards climate activism in Austria. 55% of Green party sympathisers support them, while most supporters of the other parliamentary parties are sceptical (Gallup, 2023). These data seemingly support the theory of Feinberg et al. (2020) that some forms of protest lead to rejection. This rejection can also be observed in online newspaper forums reporting on these movements. The NGO Global Witness investigation found that in 2019, about 219 ‘environmental defenders’ were killed (Global Witness, 2020, p. 5). Half of the killings took place in Colombia and the Philippines. Hate against climate or environmental activists can also be observed in Europe. The above-discussed rather negative attitude towards climate activists in society can be observed as hate speech against them. Media ethicist Paganini highlights the brutalisation of language by politicians directed against the activists. Terms like ‘climate terrorists’ can lead to direct hate speech against individuals. This also leads to a harsh tone towards climate activists on SNS (ORF, 2023). Surprisingly, there are only few studies on online hate speech against climate activists (e.g. Arce-García et al., 2023; Katsarou et al., 2021). However, a recently published study from Germany illustrates that the comment sections of young and high-reach climate activists are also affected by extreme speech, such as forms of hate speech. This can alienate young supporters, intimidate them and make hate speech seem acceptable. In addition to hate messages calling young activists ‘mentally ill’, hate is also specifically targeted at women through sexist and misogynistic comments. Climate activists are often compared to terrorists like the ‘Red Army Faction’ or the ‘Taliban’ (Jugendschutz.net, 2023).

2.4. Hate speech and its effects on credibility

While research in the field of environmental and climate communication has investigated, for example, the credibility of climate scientists (Friedman, 2024), informal communicators (Sleeth-Keppler et al., 2017), and environmental claims (Keilmann & Koch, 2023), the influence of hate speech on the credibility of online climate coverage is largely unresearched. Incivility is a central descriptive feature of contemporary online communication (Kümpel & Rieger, 2019, p. 6), whereby hate speech, according to the United Nations-definition, is directed at groups or persons in a group who are ‘hated’ because of characteristics that cannot be changed, such as origin, religion, sexuality, gender, ethnicity, among others. Hate speech is often used to describe dysfunctions of the digital public sphere (Sponholz, 2022, p. 161) and frequently serves as a synonym for forms of uncivil communication, such as cyber mobbing or shitstorms. Hate speech can be understood as various communicative acts (Frischlich, 2022, p. 100) that are universally directed at different groups in order to extend to other non-traditional groups (Bahador, 2023, p. 291), such as climate protection activists, in the sense of the threat situation outlined above.

In this context, the medial mediation and dissemination of explicit but also implicit hate messages play a relevant role. (Hateful) online comments influence the perception of online news contributions (Coe et al., 2014). Prochazka et al. (2018) were able to demonstrate that the quality of news articles on websites is rated worse if they are accompanied by hateful comments that lack justification. Abusive comments under news articles negatively affected the credibility of the respective journalist (Searles et al., 2020). When investigating the credibility of the journalists, Searles et al. (2020) used the concept of source credibility. This concept assumes that perceptions about the communicator influence the evaluation of news (e.g. Hellmueller & Trilling, 2012). Media credibility is based on source credibility and states that the publishing medium can influence the evaluation of messages (see, e.g., Metzger et al., 2003). Message credibility assumes that certain messages cannot be evaluated depending on the source, especially in social media, as the source often remains unclear on platforms. This ‘source confusion’ is considered when examining credibility with message credibility (Appelman & Sundar, 2016). Therefore, this concept argues that credibility is influenced by the characteristics of the message itself. Prochazka et al. (2018) partly use the concept of message credibility when investigating the perceived quality of news articles as depending on hateful user comments. Haim and Maurus (2021) test both (source and message credibility) when examining the influence of gender, topic and user comments on journalists’ credibility. However, message credibility has not yet been studied in the context of different hate speech comments.

In our study, we focus specifically on different types of hate speech comments: Here, we differentiate between more subtle forms of hate speech intensities, which are often perceived as even more ‘hateful’ than more offensive and explicit types (Benikova et al., 2018), such as agitative and defamatory hate speech types to assess the differences in their impact on the message credibility (Appelman & Sundar, 2016).

For this purpose, we would like to answer our research questions with the help of the Hate Speech intensity scale by Bahador (2023, p. 295f.). This established monitoring differentiates between a broader set of group categories. The first three of the six categories can be understood as an early warning system. The first early warning category includes (1) ‘disagreement’ with groups and confronts readers with constructions that portray the ‘others’ in a negative light. The rhetoric is likely not hateful, but slight misinformation against a group can already occur in this category. The second category blames the whole group for particular (2) ‘negative actions’, which are often carried out by group individuals. Speech includes calls for non-violent actions against the group. The third category includes (3) ‘negative characterizations’ or insults. This is worse than calls for negative non-violent actions, ‘as it makes an intrinsic claim about the group as opposed to a one-off action claim’ (Bahador, 2023, p. 296). Categories (4) ‘dehumanisation’, as extreme form of negative characterisation, including sub-human and superhuman characteristics, (5) ‘Call for violence’, which includes infliction of physical harm or metaphoric harm and death and (6) ‘Call for death’, which includes the literal killing of the group or group member correspond to more narrow and consensual hate speech classifications (see Table A1 for further descriptions). However, recent findings from effect and reception research suggest that the type of hate speech can have different effects on the reception of online content. Therefore, differentiating the type of hate speech and its consequences is useful (Benikova et al., 2018; Schwertberger & Rieger, 2021). Especially the distinction between explicit, agitative (call for actions) and defamatory forms and implicit, subtle hate speech seems necessary (Bahador, 2023). This is due, among other things, to the fact that automatic detection of online hate speech (by platform operators and academia) is ineffective due to linguistic means, such as metaphors, irony, ambiguities, intended misspellings, comparisons and new word constructions (e.g. ‘Klima-Nazis’ [Climate Nazis]’, ‘Öko-Totalirismus [Eco-totalirism]’, ‘Klimatizismus [Climatism]’, ‘Kohlenstoffdiktatur [Carbon Dictatorship]’, ‘Öko-Hassprediger [eco-hate preacher]’) (Scharloth, 2021, p. 8, 31, 79, 108). The particular danger with these implicit forms is, on the one hand – as part of a right-wing meta-politics – that they insidiously shift the discourse with ‘shameless normalisations of the unsayable’ (Wodak, 2020). The advantage of comparing these different hate speech intensities (from ‘disagreement’ to ‘call for death’) in comparison to common binary hate speech is that no hate speech conceptualisations can clarify the influence that even mild forms of hate speech can have on the reception of credibility and the attitudes of the content of a news article. This leads to the following hypothesis:

H1: The more intense the level of hate speech in the comments, the less credible the news article on climate protection measures is perceived to be.

2.5. Hate speech and the ‘nasty effect’

Anderson et al. (2014) illustrate that uncivil communication can have (weak) persuasive effects on attitudes toward a topic. Specifically, the so-called ‘Nasty Effect’ shows that uncivil comments increase risk perceptions of scientific topics. We will also explore this aspect for our aforementioned topics (hate speech comments among news articles on climate protection measures, especially on activism).

H2: The more intense the level of hate speech in the comments, the lower the perceived importance of climate protection measures is rated.

2.6. Political affiliation and cues on credibility and climate protection measures

There is evidence that attitudes towards climate change issues and the perceived credibility of the news article can be influenced by political attitudes (e.g. Andı, 2020). We have included two additional hypotheses to incorporate these critical factors into our study design and to find further explanations for potentially different credibility cues and differences in the perceived importance of climate action.

According to Prandner (2022, p. 112), people with right-wing political views tend to trust traditional media less. In addition, Fawzi (2019, p. 157) drew a connection between right-wing populists and negative attitudes towards the media. These findings are consistent with the current Digital News Report 2023: Especially right-wingers seem to mistrust news media (Gadringer et al., 2023, p. 113). Studies attest that Austria has a higher level of media and scientific scepticism than the EU average, which correlates strongly with right-wing populist voting intentions (Huber et al., 2021; Starkbaum et al., 2023, p. 372).

H3a: Supporters of right-wing populists perceive the credibility of the news article to be lower than supporters of other parties.

Political attitudes play a significant role in shaping perceptions and attitudes towards climate protection. In Austria, right-wing voters perceive that sufficient measures have already been taken for climate protection. On the other hand, social democrats, greens, and liberals advocate for additional climate protection measures. This highlights the divide among different political groups regarding their stance on climate change and the necessary actions to combat it (Bugden, 2020). Recent research shows that political partisanship in particular has a major influence. Specifically, people who vote right-wing populist tend to be more negative towards climate protection measures than left-wing or liberal-conservative voters (Forchtner & Lubarda, 2023; Huber et al., 2021; Ulmke et al., 2023).

H3b: Supporters of right-wing populist parties agree less strongly with statements on climate protection than supporters of other parties.

3. Method

3.1. Participants

We conducted an online survey with the open source programme SoSci Survey, distributed via Bilendi/Respondi, an online-access panel provider. The study was conducted between September 7^th and September 15^th 2023 in Austria with 486 cases (after data cleansing). Regarding ethical conformity, participants’ involvement was voluntary, they gave written consent to participate in our study. Quotas were determined for gender, age, highest formal education and voting intention. The sociodemographic characteristics in Table 1 illustrate that the collected data are not structurally identical to the respective populations. Only the voting behaviour corresponded to the election polls of the respective time, whereas 29% of the Austrian voting population indicated they would vote for the right-wing Freedom Party of Austria (FPÖ).

Table 1. Sample (Frequency, Percent, Population Quota in Percent).

Gender^1
Female	262	53.9	50.2*
Male	224	46.1	49.8*
Age
<30 Years	127	26.1	20.9*
30 to 39 Years	113	23.3	20.0*
40 to 49 Years	101	20.8	19.2*
50 to 64 Years	51	10.5	22.8*
65+ Years	94	19.3	17.1*
Highest formal education
Primary education	47	9.7	17.5**
Lower secondary school education (e.g. apprenticeship)	149	30.7	33.0**
Secondary technical school education (e.g. BMS)	131	27	30.4**
A Level (e.g. AHS-/BHS-Matura)	71	14.6
University education	88	18.1	19.1**
Political orientation
Left	100	20.6
Centre	276	56.8
Right	110	22.6
Voting Intention
Right-wing	138	30.3	29***
others	293	64.3	65***
Non-voters	25	5.5
Online News Usage
Several Times daily/daily	250	51.5
Several times the week	97	20
Once a week/ Less than once a week	93	19.2
Never	48	9.5
N	486

Note. Mean survey duration 8.27 minutes, 1 = 4 options: ‘female’, ‘male’, ‘other, namely’ and ‘I do not wish to provide any information’. *Reference data from Eurostat database; **Statistik Austria (16.05.2023): https://www.statistik.at/statistiken/bevoelkerung-und-soziales/bildung/bildung-in-zahlen; ***INSA (07.09.2023): https://politpro.eu/de/oesterreich/wahlumfragen.

3.2. Procedure and measures

For the implementation, we chose a between-subject design with vignettes, which was integrated into a survey (7x1). Participants were randomly assigned to one of six treatment groups, where everyone could read another user’s comment on an online news article (see Table A1 for the stimulus material). The news article (see Figure 1) did not vary across treatment groups TG1 to TG6 and control group (CG). The experimental group comparison focuses on typical user forum comments under online news articles and their influence on the credibility of the article and the perception of climate protection measures.

Figure 1. Stimulus (Translated Version).

Source. Der Standard (10.07.2023). Letzte Generation begann am Wiener Schwedenplatz neue Protestwelle [Last generation started new wave of protests at Schwedenplatz in Vienna]. https://www.derstandard.at/story/3000000178190/letzte-generation-protestiert-am-schwedenplatz-in-w.

3.3. The stimulus material

The original stimulus article, on which the between-subject design is based, comes from the Austrian quality daily ‘Der Standard’, which has one of the largest news user forums in the German-speaking countries. The article ‘Letzte Generation begann am Wiener Schwedenplatz neue Protestwelle’ [‘Last generation started new wave of protests at Schwedenplatz in Vienna’] was shortened and linguistically modified for the experimental setting. In order to ensure internal validity, typical colouring or fonts were not used and the participants did not see that the article originated from ‘Der Standard’ (Appelman & Sundar, 2016, p. 68). The participants did see a typical article design, which indicates the clear structure of the topic, headline, subheader, date, article structure and transition to the discussion section of the forum. They were not able to go back to see the article when answering the credibility questions.

3.4. Hate speech

In order to determine the differences in the impact of different types of hate speech, we followed the intensity scale by Babak Bahador, which conceptualises hate speech on a six-level monitoring system from (1) ‘disagreement’, (2) ‘negative actions’, (3) ‘negative characterisation of the target group’, (4) ‘demonising and dehumanising’, (5) ‘calls for violence’ to (6) ‘death’. The advantage of using this established system is that it can also detect hatred against non-traditional types of groups and that the early warning system includes both ‘traditional’ hate speech and escalating speech (Bahador, 2023, p. 291). According to this differentiation (see Table A1 in the Appendix), we collected original user comments on the topic of climate activism in Austria and Europe, adapted them accordingly and tested them prior to the study. We conducted a cognitive pre-test (N = 12) of the stimulus material to investigate whether the comments could be assigned to the hate speech types according to the intensity scale. After the pretesting, the comments were further sharpened to make them clearer.

After measuring the dependent variable (credibility of the news article about climate protection measures and in particular activism), we asked the participants to what extent they agreed with the statement whether the comment under the article contained non-appropriate language. This was measured on a 4-point scale from strongly disagree to strongly agree.

Respondents who agreed or strongly agreed with the statement regarding non-appropriate language (tend to yes/yes: n = 204, 48.9%) were forwarded to the question of whether the comment contains hate speech. We defined hate speech for the participants in an info box as follows; ‘[…] hate speech is a linguistic expression of hatred with the aim of disparaging and denigrating certain persons or groups of persons’. They were also able to answer the question from yes to no on a 4-point scale (tend to yes/yes: n = 156, 79.6%), as well as the alternative category do not know. These two questions serve as a check to see that a) the comments are recognised as hate speech and b) to measure the frequency of the messages, and therefore the intensity of the hate increases across groups from TG1 to TG6.

3.5. Credibility

The dependent variable was measured immediately after the stimulus. The credibility item battery by Appelman and Sundar (2016) was queried in the validated translated version by Thielsch (2017) on a 7-point scale from very poor to very good. We asked, ‘How well do the following adjectives describe the news article you have just read?’ The items ‘accurate’, ‘authentic’, and ‘credible’ were supplemented with ‘correct’ (Cronbach’s α = .906). This was based on comments from our pretest, which indicated that ‘accurate’ refers more to how well informed someone is and ‘correct’ refers more to truthfulness. The additional item improved the understanding of the scale and the Cronbach’s α value.

3.6. Political attitudes

With regard to general political attitudes, we asked about voting intentions in the introductory part of the survey, and in addition to the Austrian parliamentary parties Peoples party (ÖVP), Social democrats (SPÖ), Freedom party (FPÖ), Liberals (NEOS) and Greens (Die Grünen), we also asked about alternative options such as do not vote, do not want to give an answer and other parties. Furthermore, the participants were invited to rank themselves from left to right on the common 10-point scale (e.g. SOEP, ALLBUS) for their political orientation/ideology.

Following the AUTNES (2018) survey, we asked the participants how interested they were in politics on a 4-point scale from ‘very much’ to ‘not at all interested’. In order to measure internal and external political efficacy, we use the PEKS (Political Efficacy Short Scale) (Cronbach’s α = .640), which has been validated for the German-speaking countries and comprises four items on a 5-point scale from ‘do not agree at all’ to ‘agree completely’: ‘I can understand and assess important political issues well’; ‘Politicians care about what ordinary people think’; ‘I dare to actively participate in a conversation about political issues’; ‘Politicians make an effort to maintain close contact with the population’ (Beierlein et al., 2012).

3.7. Climate protection attitudes

We asked a series of individual questions and about attitudes towards climate protection (CP), climate protection measures such as demonstrations, and climate change in general. Adapted from Eurobarometer 92 (2019), we wanted to know on a 4-point scale from ‘very’ to ‘not at all’ how important is climate protection personally. Then we presented 7 statements on climate protection, which had to be answered on a 4-point scale. The two items (1) ‘Environmental problems have a direct impact on my daily life and health’ and (2) ‘My consumption habits have a negative impact on the environment in Austria’ were taken from Eurobarometer 97.5 and the five questions: (3) ‘Tackling climate change can provide new opportunities for innovation, investment and jobs’; (4) ‘Tackling climate change can help improve my own health and well-being’; (5) ‘Tackling climate change can harm our economy’; (6) ‘Tackling climate change now can help reduce the cost of major environmental damage in the future’ and (7) ‘Reducing oil and gas imports and investing in renewable energy are important for our overall security’ from Eurobarometer 92. After a confirmatory factor analysis, we created an index with six items (Cronbach’s α = .854), excluding (5) ‘harm for the economy’. In addition, we wanted to know how respondents feel about climate activism demonstrations and asked them to what extent the following statements are true (Cronbach’s α = .814): ‘I support demonstrations for climate protection’; ‘I would participate in demonstrations for climate protection’; and ‘It is acceptable to use violence to resist at demonstrations for climate protection’, were measured on a 5-point scale from ‘not at all’ to ‘absolutely true’.

4. Findings

4.1. H1: the more intense the level of hate speech in the comments, the less credible the news article on climate protection measures is perceived to be

An analysis of the comparison of means detects no significant differences in the perceived credibility of the article with regard to the differences between the various treatments (TGs) and the control group (Kruskal-Wallis-test p = .15). However, on a descriptive level, some differences can be observed; the credibility decreases with increasing hate speech level. An exception is TG4, ‘Demonising & Dehumanising’, in which the user comment dehumanised climate activists as a deadly cancer. This high mean value of 4.89 can be attributed to the sample characteristics, as the groups are not normally distributed. It also possible that the dominance of such user comments under climate protection articles is particularly common, and therefore, the article is perceived as more credible. In TG6, ‘Death’, however, it becomes clear that the perceived credibility differs significantly from CG (t(106) = -2.08, p = .04). Accordingly, hypothesis 1 can be partially accepted, due the outliers in certain TGs.

4.2. H2: the more intense the level of hate speech in the comments, the lower the perceived importance of climate protection measures is rated

Concerning the index values on the approval of climate protection measures and climate protection demonstrations as well as the personal importance of climate protection, the Kruskal-Wallis-test also shows no significant differences (see Table 2). However, particularly concerning the support for demonstrations, conspicuous features can be observed so that the approval values are generally low (m = 1.87). The value for support demonstrations in TG3 ‘Negative Character’ (m = 1.58) is significantly lower than for example TG2, ‘Negative Actions’ (m = 2.02 (t(137) = -2.58, p = .01)), or in TG4, ‘Demonising & Dehumanising’ (m = 1.88 (t(152) = -1.98, p = .02)). Despite a slight downward trend with regard to the approval of demonstrations from the first group TG1 (m = 2.07) to TG5 ‘Violence’ (m = 1.81), the approval in TG6 is (not significantly) higher than in CG or TG5. Also striking but not significant are the differences concerning the post-test questions on climate protection measures. TG6 ‘Death’ has the highest rating in supporting climate protection measures (m = 3.48; comparison CG m = 3.3). This cannot be attributed to the sample composition, as there are no significant differences between political attitudes (TG6 m = 1.97, CG m = 1.91), voting behaviour (TG6 m = 1.33, CG m = 1.33) or age (TG6 m = 45.11, CG m = 44.41), between the groups. But TG6 is more male dominated and there is less formal education in the composition of the group. This contrasts with the structure of today’s climate protection movements, which are heavily made up of female and more educated participants (Noth & Tonzer, 2022). It is possible that a kind of mobilisation effect for demonstrations can be observed here, which is triggered by the violent language of the comment. Hypothesis 2 can be rejected.

Table 2. Descriptive Statistics (Means).

	TG1	TG2	TG3	TG4	TG5	TG6	CG	m	H-Test
Credibility Index	4.89	4.81	4.71	4.89	4.61	4.45	4.98	4.74	.15
CP Personally	3.07	3.07	3.07	2.89	3.09	3.13	2.91	3.04	.49
CP Measures	3.33	3.37	3.31	3.3	3.36	3.48	3.41	3.36	.96
CP Demonstrations	2.07	2.02	1.58	1.88	1.81	1.91	1.83	1.87	.13
Non-appropriate Language	2.0	2.14	2.45	2.7	2.82	3.08	___	2.53	.00***
Hate speech	2.59	2.6	2.79	3.28	3.0	3.58	___	3.09	.00***
N	70	76	75	81	76	76	32	486

Note. Credibility Index measured by 4 items on 7-point scale from bad to very good; Climate Protection personally measured on 4-point scale from unimportant to very important; Support Climate Protection Measures, measured by 6 items on a 5-point scale: 1 = completely disagree, 5 = completely agree; Climate Protection demonstrations Index measured by 3 items on 4-point scale from not at all true to absolutely true; Non-appropriate Language; Hate speech: 1 = completely disagree, 4 = completely agree; +p < .10, *p < .05, **p < .01, ***p < .001.

Independent of this non-significant group comparison, however, it can be observed, that the participants also perceive hate speech more strongly as it increases in intensity in the groups (Kruskal-Wallis-test: p < .001). Respondents who agreed that the user comment contained non-appropriate language were directed through a filter function to the question of whether the comment contains hate speech. Agreement also increased significantly over the course of the group from TG1 to TG6. The differences are significant at a level of p < .001.

4.3. H3a: Supporters of right-wing populists perceive the credibility of the news article to be lower than supporters of other parties

In the next step, we focussed on individual attributes that influence both the credibility of reporting and the support for climate protection measures. For this purpose, we first ran regression models for the entire sample (see Table 3), including sociodemographic variables such as gender, age, education, political attitude variables such as voting intention, efficacy, general political interest and attitudes towards the personal importance of climate protection and demonstrations. In addition, the perceived credibility of the news article (DV) model included assessments of climate protection measures, news media usage frequency and whether the commentary contained inappropriate language. In the second model for the support of climate protection measures (DV), we integrated the personal importance of climate protection, efficacy and the credibility assessment.

Table 3. Regression models I and II.

	I DV: Credibility	II DV: support CP measures
(Constant)	4.1 (.57)***	1.24 (.34)***
Gender	−.10 (.13)	.09 (.08)
Age	.00 (.01)	−.00 (.03)
Education	−.06 (.05)	−.10 (.03)***
Political Interest	.07 (.09)	−.11 (.05)*
Efficacy	________	.09 (.06)
Voting Intention	−.28 (.16)+	−.36 (.09)***
Credibility	________	.06 (.03)*
CP Personally	________	.47 (.05)***
CP Measures	.30 (.09)***	________
CP Demonstrations	.07 (.07)	.186 (.04)***
Non-appropriate Language	−.07 (.06)	.09 (.03)**
News Media Usage	−.03 (.04)	−.02 (.02)
R^2	.114 486	.543 486
N

Note. Beta (SD). Gender: 0 = female, 1 = male; Age ascending; Education: 1 = max. primary education, 5 = university education; Political interest (1 = not at all, 4 = very interested); Voting Intention: 1=(ÖVP, SPÖ, NEOS, Die Grünen), 2 = FPÖ; Efficacy: 1 = low, 4 = high; Credibility of the journalistic article: 1 = bad, 7 = very good; Climate Protection Personally: 1 = not important, 4 = very important; Support climate protection measures: 1 = completely disagree, 5 = completely agree; Support climate protection demonstration, measured by 3 items: 1 = not true at all, 5 = is absolutely true; News Media Usage: 1 = never, 6 = several times daily, +p < .10, *p < .05, **p < .01, ***p < .001.

In model I on the factors influencing credibility, it becomes clear across the entire sample that only personal attitudes to the article topic, i.e. attitudes to climate protection and approval of the demonstrations, have a positive influence on credibility. In other words, the greater the personal interest in climate protection and understanding for climate protection demonstrations, the greater the perceived credibility of the article. This is true regardless of the user comments below the post. However, it also shows that the more the respondents perceived the language used in the comment as non-appropriate, the less credible they perceived the article to be.

A closer look at the experimental groups (see Table A2 in the Appendix) illustrates that these variables do not have the same effect sizes across all hate speech groups. For example, the effects of recognising non-appropriate language can only be detected for the first three TGs, with implicit forms of hate speech. In the groups with more explicit hate speech, the credibility assessment of the article increased with growing agreement (not significant). Similar changes in the sign of the effect sizes can be observed in the attitudes towards climate protection and its measures: According to this, credibility in TG5 ‘violence’ and TG6 ‘death’ decreases with increasing interest in climate protection. Due to this, hypothesis 3a can only be partially confirmed, as only a minor significant effect size can be observed across the entire sample.

4.4. H3b: Supporters of right-wing populist parties agree less strongly with statements on climate protection than supporters of other parties

In model II on support for climate protection measures, the results of the regression analysis for the entire sample (in contrast to model I on credibility) indicate stronger determinant effects for the sociodemographic characteristics. According to our sample analysis, men are significantly more in favour of support than women. Increasing formal education also has a significant negative effect. Like in the model for credibility, age has no significant effect. Voting intention has a strong influence on support for the measures. It can be seen that voters of the right-wing populist FPÖ support the measures significantly less than partisans of the other parties. Thus, hypothesis 3b can be accepted. Interestingly, the more credible the contribution is perceived to be, the more likely people would support the measures. However, attitudes towards the issue can best explain the climate measures’ support. The more the respondents are interested in climate protection and the more they support the demonstrations, the more likely they are to support the measures (Table 3). These results can also be demonstrated in the comparative analysis between the groups (see Table A3 in the Appendix).

5. Discussion, limitations & outlook

This study aims to explore the impact of different hate speech comments on the credibility of a news article and the topic covered in the article. For this purpose, the highly polarising topic of climate protection measures and, particularly, forms of demonstration by climate protection activists were chosen (see, e.g., Feinberg et al., 2020; Gallup, 2023; Scheuch et al., 2023). Similar to the findings of Searles et al. (2020), our study reveals an influence of hate speech on the perception of news articles: The results illustrate that the more hateful the language used in the comments, the less credible the article is perceived to be. The results concerning RQ1 How does hate speech in user comments affect the credibility of news articles related to climate protection measures and activism? are not significant by comparing the mean values of the treatment groups, but there are significant differences in the individual analysis compared to the control group. With regard to the more implicit hate speech comments, with increasing hate speech intensity, the perceived credibility decreased. For RQ2 What is the impact of hate speech comments under news articles on climate protection and of activism on individuals’ attitudes towards climate protection measures?, there are no significant differences between the groups with their different hate speech comments. Accordingly, the ‘Nasty Effect’, as in the study by Anderson et al. (2014), cannot be replicated in our study. However, there is a kind of ‘mobilisation effect’ in TG6 ‘Call for Death’, in which the approval ratings for both demonstrations and measures are higher than in other TGs and the control group as well.

With regard to RQ3 Which individual political attitudes influence the perceived credibility of climate protection measures reporting and attitudes towards climate protection measures and activism?, the regression analyses have shown that sociodemographic characteristics are less decisive than political attitudes, as in the case of support for climate protection measures. Our results are thus in alignment with studies and reports from SZ (2023), Gallup (2023) and Der Standard (2023a, 2023b), which indicate a connection between the political partisanship and support for climate protection measures and particular climate demonstrations. In addition to political attitudes and partisanship (Fawzi, 2019; Gadringer et al., 2023, p. 113; McCright et al., 2016), it is the interest in the topic itself that has an effect on both the media trust and the support for climate protection measures. The results show that the assessment of the credibility of information seems to be a very personal issue depending on certain personal attitudes (e.g. Prandner, 2022).

As explained above, there may be various reasons why the ‘Nasty Effect’ cannot be clearly confirmed in our study. One reason may be the lack of a normal distribution, whereby it can be assumed that certain views and attitudes in the population are not subject to a normal distribution. This is most evident concerning the approval of climate protection demonstrations, which is sometimes pathologized and defamed in Austrian media and political discourse so that a polarising bias can be assumed (Buzogány & Mohamad-Klotzbach, 2021).

Journalistic mass media are one of most important sources for information about climate change (e.g. Andı, 2020). They are pivotal when it comes to awareness, agenda-setting and knowledge about climate change (Schäfer, 2015, p. 1). However, on the one hand, media performance is evaluated depending on personal attitudes towards climate change. If people consider climate change as a serious threat to society, they are more likely to say that journalistic media are doing a good job in climate reporting and vice versa (Andı, 2020, p. 55). On the other hand, media are also important for scientists as public communication tools to communicate and influence the perception of scientific findings (Bonfadelli et al., 2017). Recent events covered by journalistic media include climate demonstrations by movements such as Last Generation. However, their protest forms (e.g., blocking roads) seem to generate great media attention but simultaneously diminish public backing (Feinberg et al., 2020; Scheuch et al., 2023). The low public support for climate activists (Gallup, 2023) thereby seems to be influenced by political attitudes – Greens party backers are more in favour of their actions than right-wing supporters.

The study also provides conclusions for German-speaking countries and presumably the online discourse on climate protests and climate protection measures in general. As mentioned above, the low level of public support for climate protests in Germany is similar to that in Austria (see, e.g., SZ, 2023; Tagesschau, 2023a). Furthermore, climate change and protests, hate speech and a general decline in trust and credibility of the media are global phenomena that could affect political and media systems worldwide. However, further research is needed to validate this - especially when comparing countries. While Austria and Germany belong to the same ‘central’ media system (Brüggemann et al., 2014), an analysis of reporting on climate activism and climate action is particularly necessary for Austria (for Germany, see, e.g., von Zabern & Tulloch, 2021). The lack of research on the impact of hate speech on public discourse and credibility limits the generalisability of our findings.

Apart from the limitations, such as the sample composition and the between-subject design, which could more suitably show the effect influences than a within-subject design, it would also have made more sense to present the participants with another stimulus article to validate the measurements. Furthermore, it should be critically mentioned that despite recognising non-appropriate language and hate speech by the participants, it is impossible to reliably state whether the credibility statements actually refer only to the article or to both the article and the user comment. Following on from this, it is also uncertain whether the participants would indeed classify the comments as hate speech types, as we have designed them.

Future research should make use of these indications to connect to established conceptualisations such as Bahador’s intensity scale in the sense of the appeal for more case studies and comparative hate speech and incivility research. The fact that hate speech is not only directed at groups and their members in the sense of the UN-definition is currently becoming clear in the socio-political discourse insofar as the mediatised hate can have real consequences, as recent attacks on activists based on their ‘group-identity’ illustrate (Stern, 2023; SWR, 2023; Tagesschau, 2023b).

Our results show that implicit hate speech can also increasingly influence the discourse. And thereby the connection between the perceived credibility of journalistic reporting, user comments and the reception of topics is evident. In order to preserve the deliberative function of user forums, the effect of (non-)justiciable speech and its consequences for journalism and democracy is pointed out. To achieve the Sustainable Development Goals (SDGs) by 2030, it is essential to give more emphasis to environmental communication. Currently, this aspect is only implicitly addressed in the SDGs, for example in goal four, which focuses on quality education, or in goal ten, which addresses the digital divide (Karmasin, 2020, p. 13). Nevertheless, our study illustrates a toxic discourse on climate protests and climate action can potentially hinder collective societal efforts against climate change. More attention is therefore needed at a policy level, as also called for by the Austrian Academy of Sciences, which sees social media and hate speech as a threat to democracy and recommends (among other things) the promotion of lifelong learning and media literacy in order to be able to react appropriately to hate speech (Karmasin et al., 2024, p. 36). At the EU level, in particular, there are several proposals to combat hate speech: The so-called ‘code of conduct on countering illegal hate speech online’ is in force, requiring social media platforms to take action against hate speech (European Commission, 2016). In addition, hate speech is also addressed in the European Union’s ‘Digital Services Act’ as an issue that needs to be addressed for a ‘safe, predictable and trustworthy online environment’ (European Union, 2022, p. 4).

Author contribution statement

Andreas Schulz-Tomančok and Florian Woschnagg jointly conceptualised the study design. Florian developed the state of research and the theoretical framework, while Andreas analysed and interpreted the collected data. The paper draft was prepared jointly. All authors agree to be accountable for all aspects of the work and approved the version to be published.

Ethical approval

The Ethics Council of the University of Klagenfurt approved the study on 18 August 2023. This was based on the briefing, letter of consent, debriefing, the stimuli, research questions and hypotheses as well as the listing of all relevant variables (application number 2023-056).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available in OSF at https://osf.io/evt9r/

Additional information

Funding

This research was funded by the Austrian Society of Communication (ÖGK).

Notes on contributors

Andreas Schulz-Tomančok

Andreas Schulz-Tomančok (MA, M.A.) is a junior researcher at the Institute for Comparative Media and Communication Studies at the Austrian Academy of Sciences (ÖAW) and the University of Klagenfurt. His research interests include political communication, right-wing populism, hate speech, gender and social media.

Florian Woschnagg

Florian Woschnagg (MA) is a university assistant at the Department of Media and Communications at the University of Klagenfurt. His research interests include political communication, credibility research, media and elections, and crisis communication.

Appendix A

Table A1. Stimulus Hate speech comments (According to the Intensity scale of Bahador, 2023).

	TG1: Disagreement	TG2: Negative actions	TG3: Negative character	TG4: Demonising & Dehumanising	TG5: Violence	TG6: Death
Characteristics	Responses include challenging claims, or trying to change their view.	Rhetoric includes negative non-violent actions associated with the group.	Rhetoric includes non-violent characterisations and insults.	Rhetoric includes sub-human and superhuman characteristics.	Responses include calls for literal violence/metaphoric aspirational phy-sical harm or death.	Responses include the literal death/elimination of a group.
Example	False, incorrect, challenge, persuade, change minds	Threatened, Stole, outrageous act, poor treatment	Stupid, Thief, aggressor, Fake, crazy	Rat, monkey, Nazi, Demon, cancer, monster	Punched, raped, starved, Torturing, mugging	Killed, annihilate, destroy
Stimulus in Experiment / Source	Ich bin gegen diese sinnlosen Klimakleber Aktionen. die rein gar nichts bringen und völlig an der Sache vorbei gehen.	Wahnsinn. diese ‘Klimaaktivisten’ fliegen alle nach Bali und sonst wo hin und hier blockieren Sie die Straßen der ehrlichen Arbeiter!	Diagnose ausgeprägter Narzissmus. Hinzu kommt inhaltsleeres in sich widersprüch-liches Geschwafel. Prognose ihr seid nicht normal!	Diese Klimaaktivisten sind ein tödliches Krebsgeschwür und zerstört unsere Gesellschaft.	Warum darf man da nicht zur Selbsthilfe greifen - und diesen Leuten einfach eine reinhauen?!	Es wird Zeit diese Klimaterroristen alle zu überfahren. Vergasen ist auch eine Lösung!
Translation	I am against these pointless climate sticker campaigns that achieve absolutely nothing and completely miss the point.	Madness. these ‘climate activists’ all fly to Bali and elsewhere and here you block the roads of honest workers!	Diagnosis of pronounced narcissism. In addition, there is empty contradictory gibberish. Prognosis you are not normal!	These climate activists are a deadly cancer and are destroying our society.	Why can’t we help ourselves and just punch these people in the face?	It’s time to run over all these climate terrorists. Gassing is also a solution!

Table A2. Regression Models on Article Credibility.

	TG1	TG2	TG3	TG4	TG5	TG6	CG
(Constant)	4.92 (1.2) ***	4.63 (1.3)***	4.57 (1.7)**	2.2 (1.5)	3.94 (1.4)**	4.16 (.16)*	−.32 (2.1)
Gender	−.12 (.36)	−.48 (.29)	−.02 (.41)	.06 (.3)	.46 (.34)	−.23 (.34)	.38 (.52)
Age	−.02 (.01)+	.03 (.01)**	.02 (.02)	.01 (.01)	.00 (.01)	.01 (.01)	.03 (.02)
Education	−.1 (.14)	−.08 (.12)	.05 (.16)	.06 (.11)	−.07 (.13)	−.12 (.15)	−.06 (.24)
Political Interest	.09 (.25)	.27 (.19)	−.11 (.28)	−.05 (.21)	−.14 (.25)	−.15 (.25)	.24 (.29)
Voting Intention	−.07 (.45)	−.06 (.32)	−.92 (.47)+	.15 (.43)	−.78 (.44)+	.13 (.43)	.42 (.49)
CP Measures	.51 (.2)*	−.18 (.21)	.21 (.25)	.64 (.22)**	.18 (.19)	.24 (.24)	.59 (.42)
CP Demonstrations	.09 (.19)	.24 (.13)+	.32 (.31)	−.11 (.16)	.13 (.17)	−.00 (.19)	−.00 (.28)
Non-appropriate Language	−.4 (.17)*	−.29 (.14)*	−.27 (.2)	.15 (.32)	.09 (.16)	.22 (.16)	____
News Media Usage	−.07 (.11)	−.01(.1)	−.02(.13)	.04 (.11)	.18 (.11)	−.23 (.12)+	−.02 (.19)
R^2	.32	.29	.23	.26	.28	.17	.31
n	70	76	75	81	76	76	32

Note. Beta values (SD). DV: Credibility of the Journalistic article, measured by 4 items: 1 = bad, 7 = very good. IV: Gender: 0 = female, 1 = male; Education: 1 = max. primary education, 5 = university education; Age ascending; Voting Intention: 1=(ÖVP, SPÖ, NEOS, Die Grünen), 2 = FPÖ; Efficacy, measured by 4 items: 1 = low, 5 = high; Political interest (1 = not at all, 4 = very interest); Political ideology: 1 = left, 10 = right; Climate Protection Personally: 1 = not important, 4 = very important; Support Climate Protection Measures, measured by 6 items: 1 = completely disagree, 5 = completely agree; Support climate protection demonstration, measured by 3 items: 1 = not true at all, 5 = is absolutely true; Non-appropriate Language: 1 = completely disagree, 4 = completely agree; News Media Usage: 1 = never, 6 = several times daily; +p <.10, *p <.05, **p<.01, ***p<.001.

Table A3. Regression models on support climate protection measures.

	TG1	TG2	TG3	TG4	TG5	TG6	CG
(Constant)	.17 (1.0)	1.96 (.79)*	.36 (.87)	2.5 (.85)**	1.02 (.99)	1.63 (1.03)	1.49 (1.2)
Gender	.01 (.24)	.19 (.18)	−.11 (.20)	.08 (.17)	.29 (.22)	−.09 (.22)	−.10 (.29)
Age	.01 (.01)	.00 (.01)	−.00 (.01)	.01 (.01)	−.01 (.01)	.00 (.01)	−.02 (.01)
Education	−.00 (.09)	.13 (.07)+	.12 (.07)	.08 (.06)	.18 (.08)*	.10 (.09)	.23 (.12)+
Political Interest	.27 (.17)	−.13 (.12)	−.23 (.13)+	.01 (.15)	−.11 (.16)	−.06 (.15)	.20 (.16)
Efficacy	−.08 (.17)	.09 (.14)	.20 (.15)	−.20 (.15)	−.17 (.22)	.26 (.17)	−.17 (.17)
Voting Intention	−.03 (.29)	−.25 (.19)	.04 (.24)	−.98 (.24)***	−.27 (.29)	−.36 (.25)	−.36 (.3)
Credibility	.22 (.09)*	−.19 (.09)*	.02 (.07)	.13 (.08)+	.12 (.09)	.07 (.08)	.21 (.13)
CP Personally	.36 (.16)*	.64 (.13)***	.71 (.12)***	.32 (.11)**	.63 (.15)***	.22 (.15)	.14 (.2)
CP Demonstrations	.15 (.13)	.15 (.08)+	.26 (.14)+	.11 (.09)	.1 (.12)	.27 (.11)*	.35 (.15)*
Non-appropriate Language	.19 (.11)+	−.04 (.09)	.14 (.09)	.07 (.09)	.07 (.1)	.07 (.1)	____
News Media Usage	−.04 (.07)	−.03 (.06)	−.02 (.06)	.02 (.06)	−.06 (.07)	−.05 (.07)	.11 (.11)
R^2	.54	.60	.69	.69	.61	.54	.73
n	70	76	75	81	76	76	32

Note. Beta Values (SD). DV: Support Measures, measured by 6 items: 1 = completely disagree, 5 = completely agree. IV: Gender: 0 = female, 1 = male; Education: 1 = max. primary education, 5 = university education; Age ascending; Political interest (1 = not at all, 4 = very interest); Efficacy, measured by 4 items: 1 = low, 5 = high; Voting Intention: 1=(ÖVP, SPÖ, NEOS, Die Grünen), 2 = FPÖ; Credibility of the journalistic article, measured by 4 items: 1 = bad, 7 = very good; Climate Protection Personally: 1 = not important, 4 = very important; Support demonstration, measured by 3 items: 1 = not true at all, 5 = is absolutely true; +p <.10, *p <.05, **p<.01, ***p<.001.