Content uploaded by Ralf Hohlfeld
Author content
All content in this area was uploaded by Ralf Hohlfeld on May 19, 2021
Content may be subject to copyright.
Communicating COVID-19 against the backdrop of conspiracy ideologies:
HOW PUBLIC FIGURES DISCUSS THE MATTER
ON FACEBOOK AND TELEGRAM
Disinformation Research Lab, University of Passau, Working Paper 1/2021
Author: Ralf Hohlfeld*, Franziska Bauerfeind, Ilenia Braglia, Aqib Butt, Anna-Lena Dietz, Denise Drexel, Julia
Fedlmeier, Lana Fischer, Vanessa Gandl, Felia Glaser, Eva Haberzettel, Teresa Helling, Isabel Käsbauer,
Matthias Kast, Anja Krieger, Anja Lächner, Adriana Malkanova, Marie-Kristin Raab, Anastasia Rech, Bianca
Scharnberger, Hannah Schmid, Malin Schmidt-Ott, Christine Triebelhorn, Franziska Vogl, Pia Weymar
*corresponding author (ralf.hohlfeld@uni-passau.de)
May 19, 2021
ABSTRACT
Social life has changed a lot in the course of the COVID-19 pandemic. The effects of the virus
influenced a wide variety of aspects of everyday life. As a matter of fact, the Internet with its
alternative media sources has a substantial influence on the spread of conspiracy ideologies and
right-wing populist perspectives and ideas alike. Especially in the context of social networks, diverse
content on these topics is shared and consumed by a vast audience. Important players on social
media are news networks and celebrities, because of their great impact on the opinions of their
followers. The study at hand examines the extent to which right-wing populist worldviews can be
identified in this context, as well as their connection with conspiracy ideology in statements made
by right-wing party AfD and prominent personalities on the social networks Facebook and Telegram.
Using a quantitative content analysis, more than 1700 social media posts with content related to
COVID-19 were analysed. The research period was set from March 11 to December 31, 2020, and
refers to event-related datasets. Amongst other findings, the analysis showed that Telegram is the
main platform for conspiracy ideologies: Posts from individuals who are devoted to conspiracy
ideologies and the AfD display their critique towards scientific facts concerning COVID-19,
especially on Telegram. Furthermore, Telegram may even be considered as the exclusive platform
for conspiracy ideologies used by these individuals.
keywords: conspiracy ideology || right-wing-populists || disinformation || Covid-19 || Coronavirus ||
pandemic || Facebook || Telegram || science skepticism || radical rhetoric || quantitative content
analysis
1 Introduction: COVID-19 – right-wing populist online communication of prominent actors and
alternative media coverage
The first infection with the new type of virus was recorded in Germany on Jan. 27th, 2020, shortly after cases were
detected in South-East Asia (WHO, 2020a). In the course of the next few weeks, detected infections increased
worldwide. Nevertheless, the virus was initially downplayed in the German media and compared to the seasonal
flu. After an exponential rise in infections, the World Health Organization (WHO) classified the newly discovered
respiratory disease as a pandemic in early March 2020 (WHO, 2020b). The seriousness of the situation then
became more apparent, and a number of nations imposed the first area-wide restrictions to contain the pandemic.
In Germany, the federal and state governments agreed to go into a nationwide lockdown starting on March 22,
consisting of stay-at-home orders and contact restrictions, as well as other strict measures (Tagesschau, 2020).
Due to fluctuating infection figures, there was a constant alternation between tightening and loosening restrictions
throughout 2020. The aim was to prevent the further spread of the pandemic and to still enable public life as far
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
2
as possible. The central role of the media as a mediator of information between politics, science and society
(Drentwett, 2009) became particularly relevant during this period.
Especially in times of crisis, alternative facts and associated worldviews find favor in society: The outbreak
of COVID-19 shows that primarily the Internet, and in particular social networks have a significant influence on
the dissemination of those perspectives and ideas (Weinert, 2018; Boberg et al. 2020; Au, Ho & Chiu 2021). Not
only do alternative news media play a major role, but also famous personalities who have a significant influence
on their followers’ opinion making (Krämer et al. 2021, Weinert, 2018). It is well known that especially during
the globally circulating pandemic, a remarkably large number of conspiracy ideologies or alternative voices
became apparent in society, spreading their opinions primarily on social networks (Roose, 2020). In this context,
the connection between right-wing populist orientation and conspiracy ideological statements are of special
interest, because the actual influence of alternative news media and populist voices on public confusion in this
crisis is still unclear (Precht, 2019; Bergmann, 2018; Boberg et al. 2020).
For the reasons stated before, it is highly relevant to examine this content, as it may influence the recipient
community and has the potential to steer public discourse and thus poses a threat to democracy. Therefore, this
paper and its analysis focus on accounts of political leaders and of celebrities that often address right-wing populist
ideas and are actively involved in pandemic communications. Right-wing populist content is disseminated
primarily via the platform Telegram, which is characterised as a messaging service inheriting a high reach and a
low regulation of content. The platform is known for to engage their audience with different news content (Barot
& Oren, 2015; Tandoc & Johnson, 2016). Its popularity as a chat app and instant messaging service rose over the
years (Lou et al., 2021). In contrast, Facebook serves as a comparative platform, with various regulatory
mechanisms, but its group structure also makes it a social medium used by right-wing populists (Amadeu Antonio
Stiftung, 2020). The data-set of this study is limited to posts by German-speaking actors in connection with the
COVID-19 pandemic. The period of investigation covers the entire pandemic’s time span in 2020, but it focuses
on significant events, such as the lockdown period described above. This study’s aim is to provide essential data
about public communication about COVID-19 on digital media platforms in Germany. In addition, it is of great
relevance to identify the extent to which right-wing populist worldviews can be linked with conspiracy ideologies
in statements by alternative parties and prominent personalities on Facebook and Telegram.
2 Definitions
In view of the research interest of the present study, it is important to define relevant constructs in order to
understand how they are interrelated. According to Holt et al. (2019) alternative news media represent the opposite
of the general tendency of public discourse arising from what is understood as the dominant mainstream media in
a given system. “Alternative news media can publish different voices (alternative content creators) trying to
influence public opinion according to an agenda that is perceived by their promoters and/or audiences as
underrepresented, ostracized or otherwise marginalised in mainstream news media”. They can provide “alternative
accounts and interpretations of political and social events (alternative news content), rely on alternative publishing
routines via alternative media organizations and/or through channels outside and unsupported by the major
networks and newspapers in an alternative media system” (Holt et al., 2019, 862). Alternative media have
developed as multitudinous and serve as a communication tool between the decentrally organised, heterogeneous
groups and projects of the counterculture (Wimmer, 2015).
The contributions of such alternative media, which are mainly published in social media, are not infrequently
characterised by disinformation (Schweiger, 2017). Disinformation is regarded as a form of conscious
communication. Although this usually contains empirically false information about certain facts, it always has a
claim to the truth - i.e. it is often perceived as true by the communicators. Accordingly, disinformation does not
necessarily involve an intent to deceive, although this is not an exclusion criterion for defining this concept
(Kohring & Zimmermann, 2020). It is a logical conclusion that this form of spreading disinformation poses a threat
to the credibility of verified facts and even leads to their denial. In the context of the current pandemic, this can be
observed in the attitude towards scientific facts that are discredited by populists (Nuissl & Popović, 2020).
Denialism is the systematic rejection of empirical evidence to avoid undesirable facts or conclusions. According
to Liu (2012), science-skeptical claims are, additionally, often supported by statements from so-called experts who
want to portray established experts as uninformed or corrupt. Plus, empirical data is repeatedly consulted to support
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
3
their claims by so-called “cherry-picking data”, i.e., excluding certain aspects and emphasizing others. This also
includes putting phenomena in relation to each other, which actually have no connection (Liu, 2012).
Especially on the side of science opponents, an obscure mixture of right-wing populists to right-wing
extremists dominates the public perception (Flügel-Martinsen, 2020). In general, populism can be described as a
form of an organization oriented toward a charismatic leader, as a style of communication without intermediary
instances, directly to the people, or, for example, as a form of discourse with certain rhetorical patterns (Von
Nordheim & Rieger, 2020). Populism in particular, however, describes above all a particular connection between
a political actor and the targeted public (Dubiel, 1986). A sense of “we” is created, which is directed primarily
against the elite or the establishment as “the others” (Albertazzi & McDonnell, 2008). The elite is portrayed as
corrupt, whereas the people are seen as a good, homogeneous society (Decker, 2000). As compared to populism
in general, right-wing populism is characterised by its nationalist, anti-semitic, racist and xenophobic nature. It
contains an “us versus them” narrative. Where “us” are people which represent a homogeneous ethnic group, and
the “them” is usually a group of minorities, such as migrants, which according to right-wing actors, are supposedly
favoured by the (corrupt) elites (Greven, 2016). Stöss (2010) claims that right-wing populism is above all a form
of communication. Therefore this paper investigates if such right-wing rhetoric was used during the pandemic.
A part of populist communication is the conspiracy ideological denunciation of the actions of the elites
(Precht, 2019), especially in concern of hostility toward journalism, science and the political establishment
(Krämer et al., 2021). According to the Cambridge Dictionary, conspiracy ideologies are “a belief that an event or
situation is the result of a secret plan made by powerful people”. In sociologist Armin Pfahl-Traughbers view, it
is an entrenched, monocausal and stereotypical attitude, which assumes the result of a conspiracy behind certain
events, pointing to the existence of real groups of people allegedly engaged in conspiracy (Pfahl-Traughber,
2002). The special characteristic of conspiracy ideologies lies precisely in this specification of real groups as the
perpetrators of conspiratorial activities and in the refusal to accept other views. In comparison with other scientific
definitions it becomes clear that one of the main aspects of conspiracy ideologies is the rejection of official
explanations of certain events. Furthermore, they usually include the assumption that there are well-kept secrets,
which are accompanied by negative intentions (Pfahl-Traughber, 2002).
3 Research Interest and Research Question
The research interest is focused on the comparison between the types of communication on Facebook and on
Telegram. This comparison is considered relevant because the published content on these platforms is regulated
differently. In fact, Facebook has regulation and deletion mechanisms for right-wing extremist content or
misinformation. Therefore, authors of posts about conspiracy ideologies face consequences and can be blocked
according to Facebook’s guidelines (Facebook, 2020).
This leads to the assumption that posts about COVID-19 published on Telegram are more likely to contain
conspiracy ideologies than those on Facebook (Hypothesis 1).
This assumption is further supported by the fact that neglecting to review shared content within a group gives
senders the feeling of being more unobserved by critics (Amadeu Antonio Stiftung, 2020). In order to be able to
give even more detailed information about this subject, the following sub-hypotheses were formed:
H1.1: Representatives of populist parties use conspiracy ideologies in their posts related to COVID-19
more frequently on Telegram than on Facebook (Silva et al., 2017; Spieß et al., 2020).
H1.2: Prominent personalities use conspiracy ideologies in their posts related to COVID-19 more
frequently on Telegram than on Facebook (Amadeu Antonio Stiftung, 2020; Bergmann, 2018).
Facebook has stricter regulation and deletion mechanisms than Telegram, which is also reflecting in the rhetoric.
According to Hoffmann (2020) the demonstrably radical and negatively connoted language of right-wing populist
activists intends to mobilise and convince followers, especially in consideration with posts. As studies have shown,
conspiracy ideology arguments distance themselves from the common narrative. According to Klinker et al.
(2018), this deviation manifests itself primarily in an increased use of vocabulary related to revealing and
disclosing alleged facts.
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
4
Hence it is assumed that those who post something about conspiracy ideologies tend to use negative and
disparaging rhetoric more often on Telegram than on Facebook (Klinker et al., 2018; Amadeu Antonio Stiftung,
2020; Hoffmann, 2020) (Hypothesis 2).
Populist actors trigger significantly more interactions and engagement on social media than the pages of other
parties (Dittrich, 2017). Statements by actors, which doubt or negate, generate a particularly large audience in the
digital space (Klinker et al., 2018). As a study from 2020 shows, this also applies for the German party Alternative
für Deutschland (AfD), which has been elected to represent a right-wing populist party in Germany. Von Nordheim
and Rieger (2020) found a correlation between the party’s clearly populist selection of topics and the success on
Twitter, in terms of likes and retweets, in comparison to the topics of other parties. Success on social media is
highly relevant to populist movements and parties. According to Precht (2019), a basic element of populist thought
and argumentation structures is the conspiracy-theoretical slander concerning elite’s activities. This proves the
connection between right-wing populist rhetoric and conspiracy ideologies. Therefore, the following hypothesis
was formulated:
Statements by AfD and its members which were published on Facebook and which refer to conspiracy ideologies,
generate more reactions than statements that do not (Hypothesis 3).
Right-wing populism makes room for a polarization of “people” and “elite”, whereby science as a factually
differentiated source can be located on the side of the “elite” and thus, has negative connotations from a right -
wing populist perspective (Nuissl & Popović, 2020). Science and related individuals are portrayed as not
trustworthy by right-wing populists (Nuissl & Popović, 2020). Fear and doubt form the basis for the spread of so-
called right-wing populist “alternative” facts (Hartung & Sentker, 2020; Hohlfeld, 2020). With regard to the
machinations of the elites, Precht (2019) refers to this as conspiracy-theoretical denunciation and lists it as a basic
element of populist thought and argumentation structures. Aforesaid content spreads rapidly, especially in social
networks. Rutjens and Lee (2020) report in their study that equivalent conspiracy thinking is a consistent predictor
of a generally low belief in science. Domain-specific evidence is provided by Hornsey et al.’s (2018) study that
linked vaccine skepticism to beliefs in conspiracies.
This suggests that science skepticism is evident in right-wing populist pandemic communication on both Telegram
and Facebook (Hypothesis 4).
With the pandemic’s outbreak, right-wing populist parties initially assessed the German government’s measures
to contain the virus as insufficient. However, their opinion suddenly changed and this counterposition resulted in
protests among the population (Gensing, 2020a). Therefore it is presumed that the right-wing populist actors’
negative point of view towards the federal government’s measures is also detectable in their posts on social media.
Effects of this can also be observed in reader comments, which is the reason why articles with populist key
messages not only result in more comments, but also lead to a more negative narrative in these (Blassnig et al.,
2019).
Resulting from this, it can be concluded that right-wing populist actors tend to express negative reactions on social
networks to governmental measures that intend to contain the pandemic, which then affects their readership and
results in negative comments (Hypothesis 5):
H 5.1: Right-wing populist actors disseminate predominantly negative statements in their posts about the
government’s actions to contain the Covid-19 pandemic (Gensing, 2020a).
H5.2: The more negative right-wing populist actors assess the government’s actions in their posts, the
more negative readers respond to the actions in the comment section (Blassnig et al., 2019).
In their study, Gadinger and Simon (2019) report on the use of radical rhetorical devices that right-wing populists
like to employ. They refer not only to right-wing populist parties as a whole, but also to individual actors.
Correspondingly, for example, members of the German right-wing populist party AfD regularly attract attention
with anti-semitic statements. This is particularly interesting because the party as a whole describes itself as pro-
Israeli and pro-Jewish (Pfahl-Traughber, 2017). For these reasons, it is assumed that there is a different degree of
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
5
radical rhetoric in communication between the official party account of the AfD and the accounts of individual
politicians. Hypothesis 6 therefore reads:
Individual political actors differ in their communication from political parties on social media. Their statements
contain more radical rhetoric than those by official party appearances.
A central feature of populist communication is strong personalization, which is reflected in an emotional
relationship between the people and the leader (Gadinger & Simon, 2019). This evokes a sense of closeness and
representation among citizens, leading to strong popular participation. Comparably, Siri et al. (2012) found that
there is a preference for personal messages, specifically on Facebook. Furthermore, their research led to the
conclusion that on the platform, posts by members of the Bundestag sharing private or emotional content and
referring to the person behind the official role receive a higher number of likes than posts with merely factual
content. Based on this, Hypothesis 7 is:
On Facebook, posts with personal content elicit stronger reactions (comments, likes) than posts with purely factual
content.
4 Method
To investigate whether and how right-wing populist world views with conspiracy ideology are being transported
through messages of right-wing populist politicians and celebrities on social media, a quantitative content analysis
of Facebook and Telegram posts was conducted. Quantitative content analysis is one of the most common text
analytical methods and has proven to be an effective research method to collect data (Bortz & Döring, 2006). The
latter not only allows to quantify word material in terms of certain aspects in this case content, but also to analyse
large amounts of material (Bortz & Döring, 2006). To ensure a rule-guided and intersubjectively verifiable
analysis, the paper at hand followed steps which include, amongst others, the development of a codebook and
testing of intercoder reliability through pretests, as described in the following section.
Especially social media posts and comments are examined through content analysis in regards to their
populist messages. Previous quantitative content analysis with social media content has for example addressed the
radicalization of political social media discourses (Riebe et al., 2018; Krämer et al., 2021), conspiracy ideologies
in regards to viruses (Wood, 2018; Chen et al., 2020), or the portrayal of vaccines on social media (Guidry et al.,
2015). To the best of our knowledge, the combination of content on Facebook and Telegram has yet to be analysed
in the content analysis context our study is proposing.
4.1 Codebook & Pre-Test
In order to answer the research question and to test the formulated hypotheses, a codebook was constructed. The
research group formulated a total of 38 variables, each of which can be classified as either a formal variable or a
content variable. The first category includes variables that aim to define posts’ features, such as the platfor m on
which they were posted (Facebook or Telegram) or the actor that posted them. Further variables such as the coder’s
name, variable ID or screenshot ID were created to simplify the tracking process. Variables belonging to the
second category, content variables, serve to analyse the coronavirus-related content posted by selected right-wing
oriented populist actors on Telegram and Facebook. They make up the main part of the codebook and include
variables that not only focus on the overall content, but also on specific aspects of the content. Since these aspects
are the centre of our research, they are depicted in the following.
First, since the research is focused on the identification of conspiracy ideology in the Telegram and Facebook
posts of selected actors, a variable dedicated to any mention of conspiracy ideologies was created. Moreover, it
was also analysed whether the conspiracy ideologies are coronavirus-related or not, or if the actor blamed specific
groups for the pandemic. These variables aim to test hypotheses 1.1 and 1.2. Second, variables that represent
aspects of the actors’ rhetoric serve to prove or disprove hypotheses 2, 3 and 6. For example, it was investigated
whether the actors’ communication was characterized by populistic or specifically disparaging rhetoric towards
the elite, science or media. Third, the research group defined several variables that aimed to identify actors’
attitudes towards the pandemic or towards actors connected to it. For instance, variables were constructed in order
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
6
to analyse whether the actors’ posts were personal or objective, if informational sources were named, or to which
extent scepticism towards science (hypothesis 4) or towards the German government (hypothesis 5) was present.
Finally, further variables were meant to analyse the readers’ reactions to the selected posts, for example, in the
Facebook comments of posts in which the actor criticised the pandemic-related decisions of the German
government (hypotheses 7 and 5.1).
After these variables were defined in the codebook, it was tested if they were sufficiently formulated and
whether all coders could code them in the same way. For this reason, a total of three pre-tests were carried out and
the codebook was adapted accordingly. The first two pre-tests were carried out in the period from December 18,
2020 to December 22, 2020 and from January 8, 2021 to January 10, 2021. Five posts were randomly selected for
this. After coding, the intercoder reliability value was below 0.70 for some variables. For this reason, the codebook
was revised and a coder training was conducted. After the revision, a third and final pre-test was carried out in the
period from January 14, 2021 to January 16, 2021. A total of 25 Telegram posts and 25 Facebook posts were
coded by each coder. The posts were selected randomly within the selected time periods. This ensured a varied
and accurate sampling for the period March 2020 to December 2020. After coding was completed, intercoder
reliability calculations revealed that there were still seven variables where less than 50 percent of the coders coded
the same. For this reason, further coding training was conducted, and incomprehensibilities were solved.
4.2 Sample & Dataset
In this research, we focus on German specific actors associating and propagating right-wing oriented and
conspiratorial ideologies on Facebook and Telegram who try to steer the discourse (Boberg et al., 2020; Hohlfeld,
2020). The data set starts on March 11, 2020, just before Angela Merkel announced a contact ban and first stores
and public facilities were urged to close their doors (ZEIT ONLINE, 2020a) and ends on December 31, 2020,
shortly after the first vaccination against COVID-19 in Germany was implemented (ZEIT ONLINE, 2020b). The
samples for the relevant posts were drawn within this period of time on specific events, as well as three days before
and three days after, always corresponding to seven days in total. At this point, the study criteria correlate with
certain COVID-related events that represent government restrictions that have caused dissatisfaction among the
German population. These events are, for example the first and second lockdown in March and December 2020,
as well as times when the German government imposed different measures to contain the pandemic, like wearing
face masks or social contact regulations. In addition to this, further relevant events, such as demonstrations against
the sanctions, are selected.
The posts for the analysis are extracted from two different social media platforms: Facebook and Telegram.
On the one hand, the social media platform Facebook is chosen because it is primarily used by political actors
whose posts contain right-wing oriented populist content (Spieß, et al., 2020; Häusler & Niedermayer, 2017). On
the other hand, the short message service Telegram is selected due to the fact that it is mainly used by actors or
celebrities to build social movements and groups. Related to this, a new digital marketplace for conspiracies and
right-wing oriented propaganda is getting to form (Holt, 2018; Baumgartner et al., 2020; Fischer, 2020; Vice,
2020).
As research objects, we picked up nine German actors, namely four politicians, four celebrities and one
political party. Specifically, these actors are the AfD, its most known politicians Alice Weidel, Beatrix von Storch,
Björn Höcke and Stephan Brandner, as well as the German public figures Attila Hildmann, Eva Herman, Michael
Wendler and Ruediger Dahlke. These authors are chosen mainly because in their Facebook posts or Telegram
channels, they tend to include conspiracy narratives regarding the COVID-19 pandemic (Spieß et al., 2020;
Häusler & Niedermayer, 2017). It is also worth noting that their channels and accounts have a high number of
followers. Moreover, both politicians and public figures are well known to the general German public because
they attract attention through their right-wing populist and conspiracist statements. The purpose of the sample
generation was to achieve a representative number of posts, which was used for the coding. Based on the codebook,
Facebook and Telegram posts are analysed in terms of content and form. Thus, they are a fundamental component
of the research project.
In order to be able to draw the sampling uniformly on both platforms, an overriding pick-up criterion is
defined. This includes the following points: in their posts, the celebrities and politicians must 1) have their own
contribution in text form, so that pure reposts and hyperlinks are excluded from the sampling, 2) create a content-
related reference to COVID-19 and 3) have at least one of the defined COVID-related keywords. The keywords
are based on an inductive selection, which was made after extensive research of various Telegram and Facebook
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
7
posts. The sampling method varied between the two platforms due to platform specific differences. On the one
hand, Facebook posts could be automatically drawn by CrowdTangle using the keyword list as the search basis.
After the search query, CrowdTangle was able to directly generate an Excel list for download that includes all
posts. Therefore, in the end only a manual check of the posts had to be carried out with regard to the own actors’
contribution and the COVID-19 reference. Telegram posts, on the other hand, had to be pulled manually, as no
automated tool has been developed here so far. Hence, the keywords had to be entered independently via the search
bar in Telegram and the respective posts in the relevant time periods had to be saved manually as screenshots. For
this platform, the pick-up criterion was additionally used as a selection factor for the posts.
The CrowdTangle and manual Telegram data set includes all relevant Facebook and Telegram posts during
the respective time period with a total of 1817 posts, which are divided into 1488 Telegram posts and 329 Facebook
posts. The distribution on both platforms is significantly higher for the celebrities (1255 posts) than for the
politicians (562 posts). Attila Hildmann takes up the largest share of the sampling with 791 Telegram posts, while
Björn Höcke is the smallest with 20 posts. In the case of Michael Wendler, Eva Herman and Attila Hildmann, only
Telegram posts and no Facebook posts are collected, as they do not operate an active Facebook account.
4.3 Methodological feedback
The collection of Facebook posts with CrowdTangle proved to be very helpful in many application areas. Due to
the automated query, all keywords can be entered at once and do not have to be searched individually one after
the other. One main advantage is that word stems can be entered. Thus, for the search query vaccin- forms such as
vaccination or vaccine obligation are included and do not have to be entered separately. However, CrowdTangle
also has negative aspects. For one, not every user’s posts can be crawled via this tool. Only accounts from 110.000
likes a page can be viewed via CrowdTangle. Lutz Bachmann, who was part of the research subject at the
beginning, had to be replaced by Ruediger Dahlke, because he was below this benchmark. On the other hand, it is
not possible to align the search in order to display no posts that are a pure repost. Therefore, posts that did not
show their own contribution had to be manually sorted out afterwards.
Sampling was notably more complex on Telegram since the keywords can only be searched individually one
after the other. This results in the problem that posts containing more than one keyword are displayed multiple
times. Therefore, it is essential to check the screenshots for duplicates afterwards. In addition, the keyword search
for a word stem is not possible as on Facebook. If the stem vaccin(-) is entered, no words with this root appear, so
words such as vaccination etc. have to be searched separately, which is time-consuming. A final aspect that makes
sample generation on Telegram more difficult is, that it is not always evident how a post is alike to be evaluated
in terms of content and form. For example, the question of the own contribution occurred, which could not always
be clearly determined in connection with reposts or quotes, since Telegram presents a rather new form of
interaction compared to the established post process on Facebook. In addition, the authors sometimes write
coherent statements as individual messages. However, if it is clearly recognisable that the content builds on each
other, the individual messages were counted as one post and the view count of the main text was taken.
5 Results
5.1 Frequencies and General Findings
When evaluating the posts, it can first be noted that an overwhelming majority (81.9%) was taken from the
Telegram platform. The rest (18.1%) of the posts can be attributed to the digital network Facebook. Regarding the
type of post, it is very interesting to observe that about a third of the posts (36.5%) are not the users’ own content,
but merely reposts or links. In addition, about two-thirds (67.1%) of the content is about topics such as fighting
the coronavirus-pandemic, the consequences of the pandemic, and the restriction of personal freedom or
constitutional rights due to the pandemic. Only one tenth (10.7%) of the posts deal with the pandemic itself. It is
also apparent that almost half (43.8%) of the authors identified the government as the main culprit for the
pandemic. About one-third (31.5%) of the content shows no attribution of blame. Nearly two-fifths (39.5%) of the
posts also serve up explicit conspiracy ideologies in or beyond the pandemic. 37.1% of the content does not include
a source. It is also worth highlighting that about one-third of the posts refer to the coronavirus as a cultivated
control tool. A clear majority (72.9%) of the content examined features disparaging and/or negative rhetoric. More
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
8
than half (52.3%) of the posts even exhibit right-wing populist rhetoric. An overwhelming majority (86%) of the
content has negative connotations. 28.7% of the posts also explicitly doubt scientific findings.
5.2 Evaluation of Hypotheses
Hypothesis 1: Correlation between the platform and the author and the mention of conspiracy ideologies.
As shown earlier, it was found that, firstly, more posts were posted on the Telegram platform overall and, secondly,
that all posts repeatedly contained conspiracy ideologies. In the following, the relationship between the platform
and the mention of conspiracy ideologies was examined. In a second step, the relationship with the author of the
post, the platform, and the use of conspiracy ideologies was considered. The results showed that there was a
medium effect between the variables V2 (author of the post) and V13 (mention of conspiracy ideologies): χ²(4) =
134.70, p ≤ .001, Cramer’s V = 0.272. Overall, 662 out of 1471 posts (45%) on Telegram contained conspiracy
ideology-related content associated with the Corona pandemic, while on Facebook it was a mere 35 out of 326
(10.74%). 11 posts (0.75%) on Telegram contained conspiracy ideology content that could not be linked to
COVID-19, on Facebook only 3 (0.92%). To support these findings, additional research was conducted to
determine the extent to which the platforms (V2) blamed alleged conspiracy actors (V12), such as Bill Gates, for
the pandemic. Here, a slight significant effect χ²(4) = 67.06, p ≤ .001, Cramer's V = 0.197 could be detected. On
Telegram, 10.19% (n=873) assigned blame to conspiracy subjects, while on Facebook, only 0.42% (n=240) did
so. Looking at the authors of the posts on the different platforms, both representatives of populist parties and
celebrities post more conspiracy ideological content related to the Corona pandemic on Telegram than on
Facebook (see figure 1).
Figure 1: Correlation between the platform and the author and the mention of conspiracy ideologies
It can also be observed that celebrities spread much more conspiracy ideological content, both on Facebook and
on Telegram, than members of the AfD. V12 (blamed conspiracy ideological actors) shows the same tendency: On
Telegram as mentioned before, the blame is more often placed on actors of alleged conspiracy than on Facebook
and this exclusively by celebrities. Celebrities blamed subjects of alleged conspiracy for the Corona pandemic in
12.07% of their blame posts on Telegram (n=729), and 2.04% on Facebook (n=49). Representatives of the AfD in
none of their posts, neither on Telegram (n=112), nor on Facebook (n=139).
Hypothesis 2: Correlation between the platforms Telegram and Facebook and negative and disparaging rhetoric
Based on former research, all of which have been elaborated on before, it was clear that Facebook has developt a
regulation and deletion mechanism of right-wings populist content. In contrast, the content on Telegram is harder
84,9%
45,5%
93,1%
81,5%
15,1%
53,5%
5,8%
18,5%
1%
1,2%
AfD party members (n=152)
celebrities (n=1155)
AfD party members (n=174)
celebrities (n=81)
Telegram Facebook
Platform
Correlation between platform, author and conspiracy ideology
No conspiracy ideologies mentioned
Conspiracy ideologies mentioned in connection with the Corona pandemic
Conspiracy ideologies mentioned without the connection to the Corona pandemic
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
9
to regulate. This led to the second hypothesis that assumes that authors of posts about conspiracy ideologies use
negative and disparaging rhetoric more frequently on Telegram than on Facebook. A chi-square test of
independence was calculated comparing the platforms Telegram and Facebook regarding the negative and
disparaging rhetoric. Even though all criteria of the chi-square test are fulfilled, no significant interaction was
found, χ²(1) = 0.078, p = .780, n=1809, further not supporting the hypothesis. There is no significant correlation
between Telegram and Facebook in terms of negative and disparaging rhetoric.
Hypothesis 3: On Facebook, posts with conspiracy narrative statements by the AfD and its members generate more
reactions than statements without.
The third hypothesis focuses solely on posts on Facebook. In addition, only the posts of the AfD and its members
are considered here, which is why the posts of the celebrities were subsequently filtered out; these filter factors
result in a total of 242 posts examined. The central feature of the investigation here is whether the posts with
content related to conspiracy ideologies generate more reactions than those that do not contain conspiracy narrative
content. In total, 21 posts with conspiracy narratives were detected, which makes about 8.7% of all posts by AfD
and its members on Facebook. Three types of reactions are distinguished in this study: likes, shares, and comments.
These are all surveyed individually. Among the surveyed reactions, equalities of variances could be detected.
Nevertheless, for each of the likes, t(242) = .419, p = .715, the shares, t(242) = .860, p = .625, and the comments,
t(242) = .649, p = .784, no statistically significant difference could be detected in the number of reactions of the
posts with conspiracy narratives and without. Thus, hypothesis three is rejected.
Hypothesis 4: Correlation between right-wing populist pandemic communication and science skepticism
The fourth hypothesis was tested by comparing the posts with and without right-wing populist rhetoric and
simultaneously examining them for possible science-skeptical content. With n=1817, science-skeptical statements
could be identified in 522 posts overall (29.1%), of which 182 (34.9%) doubted scientific findings and 340 (65.1%)
even rejected or denied them. Only 42 posts (2.3% of 1817 posts in total) accepted the results of scientific findings
as valid. 21 Cases could not be classified.
Table 1: Correlation between right-wing populist rhetoric and science skepticism
No science skepticism Science skepticism
n
No populist rhetoric
Telegram
%
73.3
26.7
673
Facebook
%
88.8
11.2
170
total
%
76.4
23.6
843
Populist rhetoric
Telegram
%
63.2
36.8
789
Facebook
%
80.9
19.1
157
total
%
66.2
33.8
946
Total
Telegram
%
67.8
32.2
1469
Facebook
%
85.0
15.0
327
total
%
70.9
29.1
1796
The percentage of rejection or denial of scientific findings was 29.1% in total (see table 1). Posts that did not
include right-wing populist rhetoric (n=843) contained science-skeptical statements in 23.6% of the cases. Posts
that exhibited right-wing populist rhetoric (n=946) had doubts about and denial or rejection of science in 33.8%
of the cases. Regarding the platform used by the communicators, a significant difference between Telegram and
Facebook was identified. 32.2% of all right-wing populist Telegram posts contained doubt, rejection, or denial of
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
10
science, while the equivalent characteristic for Facebook was a rate of 15%. Under the assumption of right-wing
populist rhetoric, important differences in the specific expression of science-skeptical statements were found (see
table 2). In 26.7% of the right-wing populist Telegram posts, scientific findings were ‘rejected or denied’, while
the comparative score on Facebook was vanishingly small at 1.9%. In contrast, the rate of ‘doubts’ about science
in right-wing populist rhetoric was higher on Facebook at 17% than on Telegram at 10.1%. Pearson’s chi-square
test was applied to identify the statistical dependence between right-wing populist rhetoric and all possible
expressions of the variable science skepticism. In the presence of right-wing populist rhetoric, significant values
concerning science skepticism resulted with χ²(2) = 47.58, p ≤ .001, Cramer’s V = 0.224, which proved a statistical
dependence of both variables. The effect was at a medium level.
Figure 2: Correlation between right-wing populist pandemic communication and science skepticism: Specific
expressions of science skepticism
Hypothesis 5: Right-wing populist agents express themselves mostly negatively in their posts about the
government’s actions regarding the containment of the COVID-19 pandemic.
As the pandemic has progressed, right-wing populist agents have undergone a transition from reproachful opinions
about the government’s measures to a clear, radical counter-position. Hence, it can be assumed that their posts on
social media platforms such as Facebook and Telegram are predominantly negative toward the measures. Taking
a closer look at the variable “author’s reaction to the government measures” in combination with the right-wing
populist agents, it turns out that the purely negative reactions account for 93,1% of all coded posts. A total of 1732
posts were coded as part of this work. In order to find out whether most of the posts by right-wing populists are
negative with regard to the government's actions concerning the containment of the Covid-19 pandemic, all posts
that do not comment on the government's actions and for which the reaction was not clearly classified must be
subtracted in advance. This leaves 952 contributions that will be analyzed in the context of this hypothesis.
Table 2: Author's reaction to the government measures by right-wing populist agents
reaction
AfD party
(n=106)
AfD member
(n=249)
Celebrities
(n=597)
Right-wing
populist agents
(n=952)
Positive reaction
4,7%
2,4%
1%
1,8%
Neutral reaction
6,6%
2,4%
6%
5,1%
Negative reaction
88,7%
95,2%
93%
93,1%
Hypothesis 5.1: The more negative the right-wing populist agents are about the government’s measures in their
post, the more negative the reader comments are toward these measures.
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
11
Moreover, it can be assumed that the negative comments of right-wing populist players also negatively influence
the reader comments under such posts. In other words, the more negative a post is, along with the reaction to the
government measures, the more negative the reader comments will be. Based on the following table 3, it can be
shown in percentage terms that this hypothesis is accepted. Negative reactions to the government measures always
entail at least 90% negative reader comments. In order to test the hypothesis "The more negative the right-wing
populist agents are about the government's measures in their post, the more negative the reader comments are
towards these measures", all reactions of the right-wing populists to the government's measures were looked at.
Since on Facebook, unlike Telegram, users are provided with the comment function, for all Facebook posts the
first 5 comments were also analyzed. Attention was paid to whether they reacted positively, neutrally, or negatively
to the government's measures. This way, the relationship between the reactions of the right-wing populists and the
reader comments on Facebook can be evaluated.
Table 3: Correlation between right-wing populists, the government’s measures and the reader reactions
Hypothesis 6: Correlation between communicator and radicality
The sixth hypothesis assumes that there is a different degree of radical rhetoric in the communication between the
official party account of the AfD and the accounts of individual politicians, since deviations from the party line
and working with negative emotions are noticeable in the communication of individual AfD politicians (Kirchner,
2019).
To test the hypothesis, the variable “author” (V3) was first recoded, so that it was available in three
dimensions. The first group represents the AfD, the second group the individual politicians, and the third group
includes all other authors who neither belong to the AfD nor function as politicians. The dependent variable
"radicality" was also recoded, into the expressions “yes” and “no”. Subsequently, to test the hypothesis, a Pearson’s
chi-square test was performed between the author variable and radicality. No expected cell frequencies were less
than 5. There was a statistically significant relationship between authorship and radicality, χ²(2) = 33.16, p < .001,
φ = 0.14. This is a weak effect. However, it showed that the AfD party made more radical statements in their posts
than did individual politicians. Out of a total of 523 radical statements, the AfD accounted for 16.06%, while
individual politicians accounted for 10.06%. Most radical statements were attributable to the other actors
(73.80%). Furthermore, it was examined in more detail which individual other authors accounted for the most
Response to government’s
measures
1. Comment (n=135)
Positive
Neutral
Negative
Positive reaction (n=5)
16%
17%
67%
Neutral reaction (n=2)
0%
0%
100%
Negative reaction(n=127)
2%
3%
95%
2. Comment (n=137)
Positive
Neutral
Negative
Positive reaction (n=6)
33%
17%
50%
Neutral reaction (n=3)
33%
0%
67%
Negative reaction (n=128)
6%
4%
90%
3. Comment (n=133)
Positive
Neutral
Negative
Positive reaction (n=4)
25%
50%
25%
Neutral reaction (n=3)
0%
33%
67%
Negative reaction (n=126)
3%
4%
93%
4. Comment (n=133)
Positive
Neutral
Negative
Positive reaction (n=5)
20%
20%
60%
Neutral reaction (n=2)
50%
0%
50%
Negative reaction (n=126)
1%
5%
94%
5. Comment (n=124)
Positive
Neutral
Negative
Positive reaction (n=5)
20%
0%
80%
Neutral reaction (n=3)
0%
33%
67%
Negative reaction (n=116)
1%
3%
96%
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
12
radical posts: Here, Attila Hildmann accounts for the majority with 62.72%, followed by Ruediger Dahlke, who
accounts for just under five percent. Furthermore, it was examined whether the use of radical statements depends
on the respective platform. It was found that 29.8% of the posts on Telegram were radical in nature, while the
number of radical posts on Facebook was slightly lower at 24%. The relationship between these two variables was
looked at in more detail using a Chi Square test. No expected cell frequencies were less than five. Using Pearson’s
Chi Square test, it was found that there was a significant relationship between platform and radicalness, χ²(1) =
4.46, p = .035, φ = -0.50. This is a strong effect. In addition, a Pearson’s chi-square test was used to check whether
there was a statistically significant correlation between the author and the use of right-wing populist rhetoric. No
expected cell frequencies were below five. Here, a statistically significant relationship between author and right-
wing populist rhetoric was found, χ²(2) = 37.34, p < .001, φ = 0.14. This relationship has a weak effect size.
Looking more closely at the individual authors, the AfD has a share of 8.42% of the total radical right-wing posts,
whereas the individual politicians collectively have a share of 19.47%. More than half of all radical right-wing
posts (52.84%) are attributable to Attila Hildmann. The next actor among the other authors is Eva Herman, who
accounts for just under 9% of all radical right-wing posts. Michael Wendler has the smallest share of posts with
radical right-wing rhetoric among the other actors, at just under 2%.
Hypothesis 7: On Facebook, posts with personal content elicit stronger reactions (comments, likes) than posts with
purely factual content.
In order to test hypothesis 7, a one-factor ANOVA was first calculated with the dependent variables ‘likes’,
‘comments’ and ‘shares’ and the independent variable ‘content’. First, the normal distribution of the data was
tested and interpreted using the Shapiro-Wilk test. Only a normal distribution for ‘factual content’ could be
confirmed concerning ‘likes’, ‘comments’ and ‘shares’. For the most part, there was no normal distribution (p ≤
.001). The examination of the homogeneity of variance was asserted within the Levene Test, according to which
no equality of the variances can be assumed (likes: p =.007, comment: p ≤ .047, shares: p ≤ .001). Since the
prerequisites for the one-factor ANOVA were not assumed, a Welch-ANOVA and the non-parametric Kuskal-
Wallis test were subsequently performed. The Welch-ANOVA shows statistically significant differences of
content for ‘likes’ F(3, 38.1) = 3.34 p = .029 , ‘comments’ F(3, 42.44) = 5.59, p = .003, and ‘shares’ F(3, 47.89)
= 5.29, p = .013. In addition, a post-hoc test was performed using the Games-Howell test to see to what extent the
groups are different. This showed no significant differences regarding the variable ‘likes’. For ‘comments’,
however, a significant difference (p = .013) in the number of comments was shown between the expressions
‘fundamental personal proposition’ and ‘personal and factual content’ (-433.36, 95%-CI[-799.03,- 67.69]). On
average, posts with ‘fundamental personal proposition’ have fewer comments than posts with ‘personal and factual
content’. In addition, the Games-Howell post-hoc test showed a significant difference (p = .003) in the number of
comments between the ‘factual content’ and ‘personal and factual content’ expressions (-652.94, 95%-CI[-
1116.3,- 189.57]). Posts with ‘factual content’ have on average fewer comments than posts with ‘personal and
factual content’. For ‘shares’, the Games-Howell post-hoc test showed a significant difference (p = .039) in the
number of shares between the expressions ‘fundamental personal proposition’ and ‘personal and factual content’
(-931.63, 95%-CI[-1831.05,- 32.20]), a significant difference (p = .030) in the number of shares between the
expressions ‘factual content’ and ‘fundamental personal proposition’ (-1423.47, 95%-CI[-103.26,- 2743.67]), and
a significant difference (p = .009) in the number of shares between the expressions ‘factual content’ and ‘personal
and factual content’(-1311.97, 95%-CI[-264.79,- 2359.16]).
Since the prerequisites for the single factor ANOVA were infringed, the non-parametric Kuskal-Wallis test
was also performed for comparison. There is a statistically significant difference between the distribution of the
number of likes regarding content, H(3) = 8.7, p = .034, as well as the distribution of the number of comments
regarding content, H(3) = 13.18, p = .004 and between the distribution of the number of shares regarding content,
H(3) = 17.99, p ≤ .001.In the pairwise comparison, the post hoc Dunn-Bonferrroni test showed a significant
difference between ‘fundamental personal proposition’ and ‘personal and factual content’ in terms of ‘likes’ (z =
-2.64, p = 0.5), this difference was also evident in ‘comments’ (z = -3.59, p = .002) and ‘shares’ (z = -4.24, p ≤
.001).To obtain clearer results for the hypothesis, only factual content and personal content were then tested against
each other. For this purpose, a t-test and a Mann-Whitney U-test were used, since the prerequisites for parametric
calculation were not given. The data were not normally distributed according to the Shapiro-Wilk test (p ≤ .001).
When testing for variance homogeneity according to Levene, equality of variances could be assumed for ‘likes’
and ‘shares’ (likes: p = .820, share: p = .962), but not for ‘comments’ (p = .027). The t-test showed no significant
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
13
difference from the number of ‘likes’ regarding content, t(324) = -.98, p =.329, and no significant difference from
the number of ‘shares’ regarding content, t(324) = -1.31, p =.192. However, there was a significant difference
from the number of ‘comments’ regarding content with a small effect, t(138,390) = -2.3, p =.023, d = -0.29. The
mean values are shown in Table 4.
Table 4: Mean values of Likes, Comments and Shares on facebook regarding to personal assertions or factual
content
N
M
SD
Likes
personal assertions
237
3478.65
4531.45
factual content
8
4023.39
4360.07
Comments
personal assertions
237
766.08
923.11
factual content
89
1065.20
1086.43
Shares
personal assertions
237
1754.99
2803.57
factual content
89
2205.60
2680.91
The non-parametric Mann-Whitney-U test also showed no significant differences in terms of ‘likes’ and content
U = 9141, Z = -1.85, p = .064. However, there was a significant difference between the number of ‘comments’
and the content (personal: Mdn = 456, factual: Mdn = 676), U = 8285, Z = -2.98, p = .003, and between the
number of ‘shares’ and the content (personal: Mdn = 698, factual: Mdn = 1309), U = 7744.5, Z = -3.7, p ≤ .001.
Since the distribution of the two groups did not differ according to Komologrov-Smirnov (p = .145), no
conclusions can be drawn about the mean values.
6 Discussion
Hypothesis 1
The results show that the hypothesis can be accepted. Reasons for this have already been addressed in the
derivation of the hypothesis. Due to the strict Facebook guidelines, the social network only offers a limited space
for conspiracy ideologies (Facebook, 2020). Moreover, according to Hoffmann (2020), the public nature of the
platform could lead to less extreme postings than the counterexample Telegram. Overall, it can be said that the
Telegram platform has developed into “the platform for conspiracy ideologies”. This supposition is further
supported by the fact that neglecting to review shared content within a group gives posters the feeling of being
more unobserved by critics (Amadeu Antonio Stiftung, 2020).
Hypothesis 1.1
Politicians also post more conspiracy ideology content on Telegram than on Facebook, but overall, unlike
celebrities, they are more reluctant to comment on conspiracy ideologies. Blame only goes to the government or
the WHO, not to “Bill Gates”. Anti-semitic accusations were also completely omitted. The reason for this could
be the super election year 2021 (a new state parliament is elected in six German states plus the election of the
national parliament), in which the AfD politicians do not want to attract too much negative attention. While the
party often speaks out against the government’s actions on Corona and criticises its actions and restrictions, it
rarely goes into the realm of conspiracy ideologies that focuses on alleged conspirators like Bill Gates. Probably,
the danger would be too great to lose voters who are more likely to belong to the right-wing fringe, but who do
not want to put themselves on the same level as “conspiracy theorists”. Because the party line of the AfD itself
does not openly stand by conspiracy ideologies, this could also be a reason why individual party members are
reticent in this area.
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
14
Hypothesis 1.2
Celebrities were found to post conspiracy ideology content almost exclusively on Telegram. Michael Wendler
only uses his Facebook account for concert announcements or music videos, Attila Hildmann has even deactivated
his Facebook account. The reason for this could be the target audience of the two platforms. On Facebook,
celebrities are followed by fans and people who are primarily interested in the page because of music, recipes, etc.
If Wendler were to post a conspiracy ideology there, for example, fans of his music would no longer follow him.
Or many of the comments under his posts would be negative. On Telegram, on the other hand, people follow the
celebrity primarily because of the posts about conspiracy ideologies and the like. The celebrity can feel ‘safe’ in
this group and spread his messages without facing critics (Amadeu Antonio Stiftung, 2020).
Hypothesis 2
In addition to the view of hypothesis one on the platforms and conspiracy ideologies, the subject of hypothesis
two is the interaction between the platforms and disparaging rhetoric. Furthermore, the results of our analysis
indicate that there is no association between platform and rhetoric. The results revealed that the populists do not
differ in their negative vocabulary among the two platforms. The overall rhetoric for both platforms was
noteworthy negative and disparaging. According to Hirschmüller & Hoffmann (2020) the right-wing populist uses
radical and negative language to mobilise and convince supporters. The evaluated politicians and the celebrities
all communicate negatively on Facebook and Telegram, which implies that they perceive the two platforms as
equally relevant to the type of audience addressed. The platform Telegram is growing rapidly and gained
importance and popularity especially in populist groupings. The Amadeu Antonio Stiftung (2020) assumed a
rhetorical difference for Telegram as the communicators allegedly feel more unobserved. In contrast to that, our
results indicate that the actors perceive the two platforms as equally suitable and therefore communicate in the
same negative rhetorical manner.
Hypothesis 3
According to Klinker et al. (2018), interaction with posts by populist groups on social networks is higher than with
posts by other parties. Posts that mainly refer to fact-based content and maintain transparent source work generate
fewer reactions. Here, the AfD occupies a central position for the German-speaking area: it deliberately curates
content that strikes a chord with the populist zeitgeist and, in doing so, invokes emotionalism more than actual
facts, which enables it to generate high click numbers (Von Nordheim & Rieger, 2020). Precht (2019) relates this
success to the association of the content with an anti-elitist position, which is why conspiracy narrative content
strongly corresponds to this schema. Based on this, it was investigated whether a significant difference in the
number of reactions to posts with conspiracy narratives could be demonstrated within the postings of the AfD and
its members. This could not be proven. The rejection of Hypothesis 3 may be due to the low number of AfD posts
with conspiracy narratives. On the one hand, in order for a real difference to be observed, this study focused only
on the Facebook platform. On Facebook, more moderate users of the platform can also access the content, creating
a better comparison for this study. However, this fact could also influence the content of the party and its members,
who are less likely to post extreme content here in order to secure a broader audience. Overall, conspiracy
narratives have shown up on Facebook in only 21 posts by the AfD and its members, which is why a direct
comparison to the 221 posts without conspiracy narrative content proves to be suboptimal. On the other hand, only
posts by the AfD were additionally examined and the distinction is made within the party. Dittrich (2017) and Von
Nordheim & Rieger (2020) refer in their research to the distinction of reactions of populist content and content of
other parties. It is possible that the AfD has a general user base that reacts to postings in the same way, as they see
themselves bound to the party by loyalty. Even the fact-based posts of the AfD and its members are saturated with
polarized content and misinformation, so attention is generated equally here and the users do not express any
particular interest in the conspiracy narratives in comparison.
Hypothesis 4
The study shows that in right-wing populist communication in the context of the COVID-19 pandemic, science-
skeptical content is strongly shared via Telegram and Facebook. These results are in line with the general
observation that scientific findings and people associated with them do not seem trustworthy in right-wing populist
environments (Nuissl & Popović, 2020). Furthermore, relevance arises from the fact that scientific findings are
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
15
being continuously produced in the COVID-19 pandemic, while politics and society are equally dependent on
them. In consequence, scientific research is forced to be omnipresent. The high proportion of sceptical or
dismissive attitudes toward science within right-wing populist rhetoric on Facebook and Telegram posts, in
particular, speak to the localization of science among the elite to be fought (Precht, 2019). The right-wing populist
communication on Telegram is far more science-critical than on Facebook, as concrete research results are rejected
or denied to a high level here. The degree of radicalization is higher on Telegram, although science skepticism
occurs increasingly in both social networks as a classic tool for distributing right-wing populist content. To
conclude, the fourth hypothesis can be accepted – with the differentiation that science is more likely to be doubted
than rejected on Facebook. In contrast, both doubts and, to a much greater extent, rejection and denial, come into
play on Telegram.
Hypothesis 5
According to Gensing (2020), right-wing populist agents, which not only include selected politicians but also
celebrities, expressed mainly negative opinions about government actions during the Corona pandemic. Celebrities
with a tendency towards conspiracy ideologies contributed a great deal to the confirmation of this hypothesis.
Hypothesis 5.1
Moreover, according to Blassnig et al. (2019), a clear negative positioning towards government measures by right-
wing populist actors in the reader comments also becoming more negative. In order to verify this, the relationship
between readers’ comments and politicians’ and celebrities’ attitudes has been examined. The results are
unambiguous and confirm the hypothesis. Yet only the Facebook comments were analysed, since Telegram does
not offer a comment function.
Hypothesis 6
Having discussed in more detail the negative tenor of the posts towards government measures, we now interpret
the finding that the communications of the right-wing populist authors feature radical rhetoric. A significant
correlation between the authors and the radicality of the posts was found. However, the posts of the official AfD
account show radical rhetoric more often than the individual AfD politicians. This shows that the use of radical
rhetoric seems to have reached the party base itself. However, it must be noted that other actors are by far the most
likely to use radical rhetoric. This is not surprising, as right-wing populist activists, such as Attila Hildmann, have
been shown to use radical language to convince supporters of their opinions (Hoffmann, 2020).
Hypothesis 7
Different than expected in hypothesis 7 (in terms of the t-test to evaluate) it only showed a significant difference
from the number of ‘shares’ in terms of content. ‘Personal assertions’ (M= 1754.99) were shared less frequently
than ‘factual content’ (M=2205.60). In terms of ‘likes’ and ‘comments’, there were no significant differences
whether the content was personal or factual. This means that hypothesis 7 cannot be accepted. However, looking
at all specific expressions, it is clear that ‘factual content’ has significantly fewer comments and shares than posts
that contain both personal and factual content. In contrast, “fundamental personal propositions” received fewer
comments and shares, i.e., posts in which the author makes an assertion for which he or she cites explicit sources
of experience than personal and factual content. The sources for ‘fundamental personal proposition’ do not have
to be factual or from recognised media, but can also be the author's own sources or sources from third parties. In
the case of posts with ‘personal and factual content’, the post contains statements by the user as well as usage of
one or more sources that go beyond personal experiences and evaluations. There is also a significant difference
between the categories ‘factual content’ and ‘unsubstantiated personal assertion’. Therefore, ‘factual content’ is
shared less frequently than posts with ‘unsubstantiated personal assertions’. This is, however, consistent with
hypothesis 7. In their study, Siri et. al (2012) focused particularly only on ‘likes’. Nonetheless, in the present study
no differences between ‘factual’ and ‘personal’ content could be found for ‘likes’. The distinction of content in
terms of ‘personal’ vs. ‘factual’ is also different here than in Siri et al. (2012). In the case of Siri et al. (2012),
‘personal content’ focuses particularly on private content. In other words, content showing the person behind the
role as a politician. In this study, the focus was on all posts in the context of the Corona pandemic from the AfD,
AfD politicians and prominent right-wing populists. Accordingly, not only data from people with an official role
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
16
were collected. The content, however, was already pre-filtered by the context of COVID-19. Therefore, it is
possible that private content in the sense of Siri et al. (2012) does not occur here, as it is not relevant for this
investigation.
7 Limitations
The present study is subject to a number of limitations that must be taken into account when classifying the results.
A number of these can be attributed to research economic framework conditions, others to theoretical and
methodological decisions. There is a limitation with regard to the sampling procedure and the resulting data basis.
For the content analysis, only text posts and pictures from the party, politicians and celebrities were evaluated,
whereas videos, for example, were excluded from the analysis. Furthermore, our analysis focused on two specific
Social Networking Sites, Facebook and Telegram, the latter being primarily a messaging service. Future studies
should include other SNS used for populist messages including Instagram, YouTube or the new platform
Clubhouse.Our study focuses on one single and specific country, namely Germany. Additional research on
comparative data from different countries may complement the valuable insights from our field studies in terms
of the communications pattern of populists on Telegram and Facebook about the Corona crisis. Additionally, our
study was conducted with a discrete population, the party Alternative für Deutschland (AfD), five politicians of
the right-wing-party AfD and five celebrities from Germany, who are known for spreading conspiracy ideologies.
Since we were interested in the difference in the communication style of populists on Facebook and Telegram, we
did not exclude players with a low posting-frequency or without a Facebook/Telegram account. Among other
things, this contributes to an unbalanced ratio between Telegram and Facebook posts. A broader design with a
higher number of populists on Facebook and Telegram could give more insights in order to better determine under
which conditions populist communication about the Corona pandemic is differing between platform and players.
An additional limitation of the sample concerns the selection of the comments analysed. We only included the first
five comments of each Facebook post, sorted by most relevant. Current ranking factors include most likes and
replies, as well as comments from verified pages and friends (Facebook, 2021). Based on the prerequisite to be
logged into Facebook to see comments, we cannot determine if the non-transparent and continuously changing
Facebook algorithm reduces the power of our results related to comments.
Those limitations notwithstanding, our findings contribute to contemporary research on populist
communication online by shedding light on how these individuals communicate on Telegram and Facebook during
the Corona pandemic.
8 Conclusion
As already mentioned in the introduction, social networks are attributed with a great influence in times of crisis
and uncertainty (Rauchfleisch et al., 2020). Those make many people vulnerable to half-truths, and some seem to
find it easier to accept bizzare explanations than to believe official sources or even scientific results (Duffy et al.,
2020). Since the outbreak of the Corona pandemic in early 2020, we are facing a period of absolute uncertainty.
As in other times of crisis, parts of the population tried to find their own “truths” and sometimes did find them in
questionable social media profiles. The aim of this study was to analyse the communication of profiles that could
conceivably be used to exploit this climate of uncertainty, thus providing substantial data for public communication
about COVID-19. The focus on the AfD, a number of politicians from this party, as well as other prominent
personalities was chosen because these profiles are precisely maintained by those people who purposefully
promoted a division of society and still continue to do so (Lengfeld & Dilger, 2018). For the dangers arising from
this division and the potential for conflict that results from it, one does not have to look as far as the USA – on the
contrary, it may be sufficient to look at the “demonstrations” of PEGIDA in Saxony or the “storming” of the
Bundestag in August 2020, through which the cornerstones of democracy seemed to be shaking (Gensing, 2020b).
This is precisely what marks the explosive nature of a network like Telegram, in which “like-minded people” can
spread their fake news in groups and exchange information in isolation from other parts of the population. At the
same time, it stresses the importance of scientific attention to this topic, because only topics receiving sufficient
attention can be counteracted.
The importance of Telegram regarding the topic COVID-19 is also supported by the enormous number of
posts that were found on this platform. It seems as if opinions are easier to spread and advocate when opposition
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
17
to them is seemingly non-existent and one does not have to deal with counter-arguments. This can also be deduced
from the higher frequency of posts spreading conspiracy ideologies, even though Attila Hildmann can be named
as the absolute main driver here, while the AfD and its members acted more restrained in this regard. The possible
reasons for this outcome have already been explained in the previous chapter. In summary, we can state that the
radical political discourse in the pandemic has shifted from the classic social network sites to the messanger
services. Radical positions on Covid-19 and the containment of the pandemic prevail where there is no fear of
regulation, sanctions and not even opposition. Telegram is a separate communication space that intensifies the
tendency towards polarization and the formation of hermetically sealed ideological spheres and thus deepens the
division in society.
However, what could be observed among AfD, politicians and prominent personalities alike, was the
dismissive attitude in their posts on Facebook as well as on Telegram. Thus, the pandemic itself was only directly
addressed in a few posts, while countermeasures such as the lockdown and the restrictions on fundamental rights
and freedoms that allegedly accompanied the lockdown were the subject of discussion instead. Surprisingly, it can
be stated that a higher amount of radical statements regarding COVID-19 could be recorded on the profile of the
AfD than on the profiles of its politicians. As has already been addressed, this can be taken as an indication of the
party's further drift to the absolute far right. If one considers the rejection of monitoring the party by the Federal
Office for the Protection of the Constitution's, decidet by the Kölner Verwaltungsgericht at the beginning of March
2021, this result highlights the danger that will continue to emanate from this party. Looking at the forecasts for
the Bundestag elections, a slight loss of votes for the party can be seen, but they are still high enough to continue
spreading their statements and opinions in the Bundestag (see Bundestagswahl 2021). In addition, they remain
strong in the new German states, especially Saxony, Saxony-Anhalt and Thuringia, paving the way for the
preceding split in society that can still be derived from these poll figures. However, the influence of prominent
personalities such as Attila Hildmann or Eva Hermann, who are at the forefront with regard to the spread of
conspiracy ideology, should not be underestimated.
How can communication be countered, nevertheless, when above all the elites – and thus, also science – are
seen as “opponents” by the communicators? The most important starting point may be to provide information in
the form of background information and transparent communication from the official side – unfortunately, trust is
further damaged by scandals and (attempted) enrichment by a large number of CDU/CSU politicians in the context
of the COVID-19 pandemic. Another approach would be to build and strengthen media literacy, as the results of
hypothesis 7 show that it only takes an indiscriminate source to increase sharing numbers and with that, the spread
of the represented opinions. Due to the high social importance of the topic of COVID-19 and communication on
social media, which has already been mentioned, a high value should be placed on scientific observation of this,
in order to recognize possible radical tendencies more rapidly and to be able to avoid or at least contain possible
“real” effects of those.
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
18
References
Albertazzi, D. & McDonnell, D. (2008). Introduction: The Sceptre and the Spectre. In: Albertazzi, D. &
McDonnell (editors.): Twenty-first century populism. The spectre of western European democracy. New
York: Palgrave Macmillan: 1-14.
Au, C.H., Ho, K.K.W. & Chiu, D.K. (2021). The Role of Online Misinformation and Fake News in Ideological
Polarization: Barriers, Catalysts, and Implications. Information System Frontiers
https://doi.org/10.1007/s10796-021-10133-9.
Amadeu Antonio Stiftung (2020). Alternative Wirklichkeiten. Monitoring rechts-alternativer
Medienstrategien. amadeu-antonio-stiftung.de.
Barot, T., & Oren, E. (2015). Guide to Chat Apps. Tow Center for Digital Journalism, Columbia University.
https://academiccommons.columbia.edu/doi/10.7916/D8377N0W.
Baumgartner, J., Zannettou, S., Squire, M. & Blackburn, J. (2020). The Pushshift Telegram Dataset. Preprint
Paper, Researchgate: 1–7.
Bergmann, E. (2018). Conspiracy & Populism. The Politics of Misinformation. Switzerland: Springer Nature.
Blassnig, S., Engesser, S., Ernst, N. & Esser, F. (2019). Hitting a Nerve: Populist News Articles Lead to More
Frequent and More Populist Reader. Political Communication, 36(4): 629-651.
Boberg, S., Quandt, T., Schatto-Eckrodt, T. & Frischlich, L. (2020). Pandemic News: Facebook Pages of
Mainstream News Media and the Coronavirus Crisis – A Computational Content Analysis. Muenster
Online Research (MOR), Working Paper 1/2020, p. 1–21.
Bortz, J., & Döring, N. (2006). Forschungsmethoden und Evaluation. Wiesbaden: Springer Verlag.
Bundestagswahl 2021 (2021). Umfragen, Prognosen und Projektionen. bundestagswahl-2021.de
Chen, L., Ling, Q., Cao, T. & Han, K. (2020) Mislabeled, fragmented, and conspiracy-driven: a content
analysis of the social media discourse about the HPV vaccine in China. Asian Journal of Communication,
30(6): 450-469.
Decker, F. (2000). Parteien unter Druck. Der neue Rechtspopulismus in den westlichen Demokratien.
Wiesbaden: VS Verlag für Sozialwissenschaften.
Dittrich, P. (2017). Social Networks and Populism in the EU. Four Things you should know. Policy Paper,
19(2): 1–20.
Drentwett, C. (2009). Vom Nachrichtenvermittler zum Nachrichtenthema. Metaberichterstattung bei
Medienereignissen. Wiesbaden: VS Verlag für Sozialwissenschaften.
Dubiel, H. (1986). Vorwort. In: Dubiel, H. (editors): Populismus und Aufklärung. Frankfurt a. Main:
Suhrkamp Verlag: 7-11.
Duffy, A., Edson T. & Rich L. (2020). Too good to be true, too good not to share: the social utility of fake
news. Information, Communication & Society, 23(13): 1965-1979.
Facebook (2021): Policies. facebook.com
Fischer, G. (2020). Licht und Schatten in der akademischen Medienindustrie. In: Simon Schrör, Georg Fischer,
Sophie Beaucamp & Konstantin Hondros (editors): Tipping Points. Interdisziplinäre Zugänge zu neuen
Fragen des Urheberrechts. Baden-Baden: Nomos Verlag: 223–240.
Flügel-Martinsen, O. (2020). Zeit der Pandemie, Zeit der harten Wissenschaften? Über einen fatalen
Fehlschluss und die Perspektiven einer kritischen politischen Theorie der Pandemie. In: Arnold, Flügel-
Martinsen, Mohammed, Vasilache (editors): Kritik in der Krise. Perspektiven politischer Theorie auf die
Corona-Pandemie. Baden-Baden: Nomos Verlag: 181-196.
Gadinger, F. & Simon, E. (2019). Kalkulierte Ambivalenz, mobilisierte Ängste und volksnahe Inszenierung:
Rechtspopulistische Erzählstrategien in Wahlkampagnen und Regierungspraxis. Zeitschrift für
Politikwissenschaft, 29: 23-52.
Gensing, P. (2020a). Kritik an Corona-Maßnahmen - Radikaler Wandel. tagesschau.de
Gensing, P. (2020b). Mit gezielten Falschmeldungen aufgehetzt. Mit gezielten Falschmeldungen aufgehetzt.
tagesschau.de
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
19
Greven, T. (2016). The Rise of Right-wing Populism in Europe and the United States. Friedrich Ebert
Foundation.
Guidry, J. P., Carlyle, K., Messner, M., & Jin, Y. (2015). On pins and needles: how vaccines are portrayed on
Pinterest. Vaccine, 33(39): 5051-5056.
Hartung, M. J. & Sentker, A. (2020). Raus, raus, raus! Eine Wissenschaft in der Vertrauenskrise muss sich der
Gesellschaft öffnen – viel radikaler als bisher gedacht. In: Schnurr, J & Mäder, A. (editors): Wissenschaft
und Gesellschaft: Ein vertrauensvoller Dialog. Positionen und Perspektiven der
Wissenschaftskommunikation heute. Berlin: Springer VS: 129-137.
Häusler, A. & Niedermayer, O. (2017). Debatte: Ist die Alternative für Deutschland eine rechtspopulistische
Partei? bpb.de
Hoffmann, T. (2020). Die Sprache der Populisten. In: Fleck, Marina, Hirschmüller, Tobias & Hoffmann,
Thomas (editors.): Populismus – Kontroversen und Perspektiven. Ein wissenschaftliches
Gesprächsangebot. München: AVM – Akademische Verlagsgemeinschaft München: 55-74.
Hohlfeld, R. (2020). Die Post-Truth-Ära. Kommunikation im Zeitalter von gefhlten Wahrheiten und
Alternativen Fakten. In: Ralf Hohlfeld, Michael Harnischmacher, Elfi Heinke, Lea Sophia Lehner &
Michael Sengl (editors): Fake News und Desinformation. Herausforderungen für die vernetzte
Gesellschaft und die empirische Forschung. Baden-Baden: Nomos Verlag: 43–60.
Holt, K. (2018): Alternative Media and the Notion of Anti-Systemness: Towards an Analytical Framework.
Media and Communication, (6): 49-57.
Holt, K., Ustad Feigenschou, T. & Frischlich, L. (2019). Key Dimensions of Alternative News Media. Digital
Journalism, 7(7): 860-869.
Hornsey, M. J., Harris, E. A., & Fielding, K. S. (2018). The psychological roots of anti-vaccination attitudes:
A 24-nation investigation. Health Psychology, 37(4): 307–315.
Kirchner, M. (2019). Wie wir mit der AfD-Rhetorik umgehen sollten. deutschlandfunknova.de
Klinker, F., Scharloth, J. & Szczęk, J. (Hrsg.) (2018). Sprachliche Gewalt. Formen und Effekte von
Pejorisierung, verbaler Aggression und Hassrede. Stuttgart: J.B. Metzler.
Kohring, M. & Zimmermann, F. (2020). „Fake News“: Aktuelle Desinformation. In: K. Marx, H. Lobin & A.
Schmidt (editors): Deutsch in Sozialen Medien. Berlin, Boston: De Gruyter, p. 147-162.
Krämer, B., Fernholz, T., Husung, T., Meusel, J. & Voll. M. (2021). Right-wing populism as a worldview and
online practice: Social media communication by ordinary citizens between ideology and lifestyles,
European Journal of Cultural and Political Sociology, DOI: 10.1080/23254823.2021.1908907
Lengfeld, H. & Dilger, C. (2018). Kulturelle und ökonomische Bedrohung. Eine Analyse der Ursachen der
Parteiidentifikation mit der „Alternative fr Deutschland“ mit dem Soziooekonomischen Panel 2016.
Zeitschrift für Soziologie, 47(3): 181-199.
Liu, D. W. C. (2012). Science Denial and the Science Classroom. CBE—Life Sciences Education, 11: 1 29–
134.
Lou, C., Tandoc, E. C. Jr., Hong, L. X., Pong, X. Y., Lye, W. Y. & Sng, N. G. (2021). When Motivations
Meet Affordances: News Consumption on Telegram, Journalism Studies 22(7): 943-952. DOI:
10.1080/1461670X.2021.1906299
Nuissl, E. & Popović, K. (2020). Populismus und Bildung. Zeitschrift für Weiterbildungsforschung, 43(3):
339–355.
Pfahl-Traughber, A. (2002). Bausteine zu einer Theorie über Verschwörungstheorien: Definitionen,
Erscheinungsformen und Ursachen. In: Reinalter, Helmut (editors): Verschwörungstheorien: Theorie –
Geschichte – Wirkung. Innsbruck: StudienVerlag: 30-45.
Pfahl-Traughber, A. (2017). Die AfD und der Antisemitismus. bpb.de
Precht, J. (2019). Verbreitung populistischer Narrative in den Kommentarspalten des sozialen Netzwerks
Facebook. In: Müller, Michael & Precht, Jørn (editors): Narrative des Populismus: Erzählmuster und -
strukturen populistischer Politik. Wiesbaden: Springer VS: 93-114.
Communicating COVID-19 against the backdrop of conspiracy ideologies, Working Paper 1/21, May 2021
20
Rauchfleisch, A., Vogler, D., Eisenegger, M. (2020). Wie das Coronavirus die Schweizer Twitter-
Communities infizierte. In: Forschungszentrum Öffentlichkeit und Gesellschaft (fög). Qualität der
Medien. Schweiz – Suisse – Svizzera. Jahrbuch 2020. Basel: Schwabe: 61-75.
Riebe, T., Pätsch, K., Kaufhold, M.-A. & Reuter, C. (2018). From Conspiracies to Insults: A Case Study of
Radicalisation in Social Media Discourse. In: Dachselt, Raimund & Weber, Gerhard (editors), Mensch
und Computer 2018 - Workshopband. Bonn: Gesellschaft für Informatik e.V.. DOI: 10.18420/muc2018-
ws12-0449
Roose, J. (2020). Verschwörung in der Krise. Repräsentative Umfragen zum Glauben an
Verschwörungstheorien vor und in der Corona-Krise. kas.de
Rutjens, B. T. & Lee, R. (2020). Spiritual skepticism. Heterogeneous science skepticism in the Netherlands.
Public Understanding of Science, 29(3): 335-352.
Schweiger, W. (2017). Der (des)informierte Bürger im Netz. Wie soziale Medien die Meinungsbildung
verändern. Wiesbaden: Springer Fachmedien.
Silva, B. C., Vegetti, F. & Littvay, L. (2017). The Elite Is Up to Something: Exploring the Relation Between
Populism and Belief in Conspiracy Theories. Swiss Political Science Association, 23(4): 423-443.
Siri, J., Melchner, M. & Wolff, A. (2012). The political network: Parteien und politische Kommunikation auf
Facebook. kommunikation @ gesellschaft, 13(special edition): 1-29.
Spieß, E., Frieß, D. & Schulz, A. (2020). Populismus auf Facebook. Zeitschrift für Politikwissenschaft, 30(2):
219-240.
Stöss, R. (2010). Rechtsextremismus im Wandel (3. Aufl.). Bonn: Bonner Universitäts-Buchdruckerei.
Tagesschau (2020). Maßnahmen gegen Coronavirus. Einigung auf umfassendes Kontaktverbot. tagesschau.de
Tandoc, E., & Johnson, E. (2016). Most Students get Breaking News First from Twitter. Newspaper Research
Journal 37(2): 153–166. DOI:10.1177/0739532916648961
Vice (2020). Themenschwerpunkt Telegram. vice.com
Von Nordheim, G. & Rieger, J. (2020). Im Zerrspiegel des Populismus. Eine computergestützte Analyse der
Verlinkungspraxis von Bundestagsabgeordneten auf Twitter. Publizistik, 65(3): 403-424.
WHO (2020a). Situation report - 39. Coronavirus disease 2019 (COVID-19). who.int
WHO (2020b). WHO erklärt COVID-19 Ausbruch zur Pandemie. euro.who.int
Weinert, F. (2018). Populismus und Öffentlichkeit. Promis und Populismus in den Sozialen Medien.
Communication Socialis, 51(2): 181-191.
Wimmer, J. (2015). Alternative Medien, Soziale Bewegungen und Medienaktivismus. In: Hepp, Andreas,
Krotz, Friedrich, Lingenberg, Swantje & Wimmer, Jeffrey (editors): Handbuch Cultural Studies und
Medienanalyse. Wiesbaden: Springer VS: 191-202.
Wood, M. J. (2018). Propagating and debunking conspiracy theories on Twitter during the 2015–2016 Zika
virus outbreak. Cyberpsychology, behavior, and social networking, 21(8): 485-490.
ZEIT ONLINE (2020a). Bund und Länder vereinbaren Schließung von Geschäften. zeit.de
ZEIT ONLINE (2020b). 101-Jährige hat erste Corona-Impfung in Deutschland erhalten. zeit.de
