Pro-Kremlin German Twitter users overlap with anti-lockdown and anti-vaccine discourse

There is qualitative evidence of intersection between anti-vaccination, anti-lockdown and pro-Kremlin narrative in European Social Media, but no research quantifies the level of the overlap. Here I show that pro-Kremlin users are over 51 times more likely to be involved in both anti-lockdown and anti-vaccine clusters than anti-Kremlin users.

The aim of this blog is to evaluate the reaction of Social Media users in Germany and to correlate pro-Kremlin positions during the Russian invasion with antivaxx/supporting coronasceptic protest attitudes. Infodemiology is very useful in understanding social dynamics during epidemics acting a supplementary role to standard tools as surveys (Jarynowski, Wójta-Kempa, et al. 2020; Eysenbach 2020), infoveillance could be useful for public health decision makers (i.e. in early warning systems of prevalence estimation (Jarynowski et al. 2022) or burden of measures (Jarynowski et al. 2021)), but it should also be remembered that the COVID-19 pandemic is also a potential area of hybrid activities below the threshold of war. Thus, I deploy a social network approach to collected tweets related to three polarizing issues (in German language):

1) „Impfung” data set consisting of 1 160 941 vaccination related tweets with 171 542 unique selected users from the first half of 2021;

2) the biggest coronasceptic protest in Germany #B2908 with 389 217 tweets and 71 612 selected unique users taking place in Berlin in August 2020;

3) War related #IstandwithPutin 3 032 tweets with unique users 2 089 in first days of Russian invasion on Ukraine in February/March 2022.

Russia-sponsored traditional and social media have been marked by the European External Action Service (EU counter disinformation agency) as propagating dis-/mis-information during Covid-19 pandemic in Germany (EEAS 2021; EEAS 2020). According to surveys, the highest coronasceptic protest potential is mainly among far right i.e. AfD (59%) and to some extent far left i.e. die Linke (18%) part of the electorate (Lamberty et al. 2022) and similar mosaic can be found on Twitter (Jarynowski, Semenov, et al. 2020). Thus, both some fractions of the far right and far left side of the German political sphere did not support sanctions issued by the European Parliament on 01.03.2022 against Russia after invasions on Ukraine. Protesters (during the analyzed peak on 29.08.2020 on the streets of Berlin) were claiming (among others) that Germany was still an “occupied country” and demonstrators just wanted to “defend [their] freedom and [their] democracy” asking „Mr. Putin” for help (Loucaides 2021). Moreover, association between pro-Kremlin narration and vaccine diplomacy (Wiśniewska 2021) and hesitancy (Broniatowski et al. 2018) is not a new phenomenon. As COVID-19 vaccines uptake is promoted in state sponsored media inside Russia, anti-vaccination attitudes are fuelled to the international audience. AstraZeneca Covid-19 vaccine was potentially identified as the main target of the larger Kremlin campaign on Twitter aimed at discrediting the Western vaccines (Jemielniak & Krempovych 2021). Let’s note that German society reacted with the highest level of panic among European countries (i.e. comparing Google Trends search volumes of Thrombosis) rolling out the same vaccine to a more or less the same extent (Belik & Jarynowski 2021).

Pro-Russian/pro-Ukrainian propaganda with #IstandwithPutin in English was an inspiration for this analysis.

Classification of users to classes (pro/anti-Kremlin as well pro/anti-vaccination or pro/anti-protesters) is a difficult task (most social movements are accompanied by the opposition and phenomenon of hashtag hijacking can be observed) and various techniques for tweets/users identification were proposed (Helmus et al. 2018; Golovchenko 2020). Here, simple community detection algorithms (Jarynowski et al. 2019) were applied based on retweeting activity (Jarynowski & Płatek 2022). To assess overlapping sets, only accounts created before 2020.07.15 (with a history) have been selected to exclude obvious bots and trolls.

Thus 1890 accounts were classified as anti-Kremlin and 199 as pro-Kremlin, 42 314 users who were classified as protests supporters and 25 803 who were against protests, as well as 72 669 users who were classified as pro-vaccination and 26 792 anti-vaccination. Keywords frequencies analysis reveals that main discussion of Pro-Kremlin users are concentrating on Americans (i.e. building relativism stating that USA have been invading other countries) or energy (i.e. Germany needs Russian oil, gas and carbon). Sentiment of Pro-Kremlin is less positive than Anti-Kremlin, which suggests that Pro-Kremlin narration is less emotional and more calculated. I found that 66 (33%) of Pro-Kremlin users and only 18 (1%) of Anti-Kremlin users were involved in both anti-Vaccination AND pro-Protests discourse. Thus, 51-fold (p-Value<0.001) higher activity of Pro-Kremlin users in both anti-vaccination and anti-lockdown communities suggest strong cohesion and mobilisation of these accounts. 62% of pro-Kremlin users have been engaged in vaccines and 45% in coronasceptic protests (without distinguising sides). In a minority of Pro-Kremlin accounts, which have not been engaged in COVID-19 discourse at all, Middle-East and Serbs minorities or supporters were found.

Based on the analysis of protest material of the anti-lockdown Berlin demonstration in 2020 and COVID-19 vaccines, overlaps with accounts of pro-Kremlin attitudes (using a very specific hashtag #IstandwithPutin) can be identified, showing that they are significantly different from anti-Kremlin users. This is only signalling analysis while there is high uncertainty in user classification (as there is no perfect method) and further research is needed for method validation. Thus, pro-Kremlin agenda is different in each society, as in German speaking population fossils and liberty are the main frame of concern, in English speaking world anti-Western attitudes are highly present when in Polish Ukrainian genocide on Polish population during WWII is and anti-refugee attitude may be amplified. Here I wanted to discuss if parts of the liberal scripts activated during COVID-19 pandemic in European societies (i.e. the German speaking population) could be used and played by foreign intelligence. Especially, as this topic seems to be under-investigated in Western Europe in comparison to Eastern Europe or Anglo-American countries. However, Pro-Kremlin users’ motivation to engage in anti-vaccination/anti-lockdown communities may be either internal (to promote their own agenda) or external (to spread pro-Russian propaganda), but it is untraceable by my approach.

References

Belik, Vitaly & Jarynowski, Andrzej 2021 ‘Elucidating the interplay of COVID-19 epidemic and social dynamics via Internet media in Germany’ link

Broniatowski, David A et al. 2018 ‘Weaponized health communication: Twitter bots and Russian trolls amplify the vaccine debate’ American journal of public health 108/10:1378–1384

EEAS 2020 ‘Short assessments of narratives and disinformation around the Covid19 pandemic (UPDATE MAY-NOVEMBER 2020’ link

　 2021 ‘Short assessments of narratives and disinformation around the Covid-19 pandemic (UPDATE DECEMBER 2020 – APRIL 2021)’ link

Eysenbach, Gunther 2020 ‘How to fight an infodemic: the four pillars of infodemic management’ Journal of medical Internet research 22/6:e21820

Golovchenko, Yevgeniy 2020 ‘Measuring the scope of pro-Kremlin disinformation on Twitter’ Humanities and Social Sciences Communications 7/1:1–11

Helmus, Todd C et al. 2018 Russian social media influence: Understanding Russian propaganda in Eastern Europe Rand Corporation

Jarynowski, Andrzej et al. 2021 ‘Mild Adverse Events of Sputnik V Vaccine in Russia: Social Media Content Analysis of Telegram via Deep Learning’ Journal of Medical Internet Research 23/11:e30529

　 2022 ‘Analysis of perception of infectious diseases on the internet in Poland’ SVEPM link

Jarynowski, Andrzej; Paradowski, Michał B & Buda, Andrzej 2019 ‘Modelling communities and populations: an introduction to computational social science’ Methodological Studies 39:123–152

Jarynowski, Andrzej & Płatek, Daniel 2022 ‘Sentiment analysis, topic modelling and social network analysis.  COVID-19, protest movements and the Polish Tweetosphere’ in K Kopecka-Piech & B Łódzki, 2022 The Covid-19 Pandemic as a Challenge for Media and Communication Studies London: Routledge: 210–224

Jarynowski, Andrzej; Semenov, Alexander & Belik, Vitaly 2020 ‘Protest Perspective Against COVID-19 Risk Mitigation Strategies on the German Internet’ in S Chellappan, K-KR Choo, & N Phan eds. 2020 Computational Data and Social Networks Cham: Springer International Publishing: 524–535

Jarynowski, Andrzej; Wójta-Kempa, Monika & Belik, Vitaly 2020 ‘Trends in interest of COVID-19 on Polish Internet’ Epidemiol Rev 74/2:106–123

Jemielniak, Dariusz & Krempovych, Yaroslav 2021 ‘An analysis of AstraZeneca COVID-19 vaccine misinformation and fear mongering on Twitter’ Public Health 200:4–6

Lamberty, Pio; Holnburger, Josef & Tort, Maheba 2022 CeMAS-Studie: Das Protestpotential während der COVID-19-Pandemie link

Loucaides, Josef, Darren; Perrone, Alessio; Holnburger 2021 ‘How Germany became ground zero for the COVID infodemic’ link

Wiśniewska, Iwona 2021 ‘Sputnik over Europe’ OSW Commentary 387

Andrzej Jarynowski: computational epidemiologist with media coverage in Bloomberg and Washington Post among others. PhD candidate with accepted thesis on infectious disease modelling. An expert in infoveillance and infodemiology.