Abstract
Recent studies documented alarming growth in antiscientific discourse among extremist groups online and especially the relatively high anti-vaccine attitudes among White Nationalists (WN). In light of accelerated politization of COVID-19 containment measures and the expansion of containment to lockdowns, masking, and more, we examine current sentiment, themes and argumentation in white nationalist discourse, regarding the COVID-19 vaccines and other containment measures. We use unsupervised machine learning approaches to analyze all conversations posted in the “Coronavirus (Covid-19)” sub-forum on Stormfront between January 2020 and December 2021 (N = 9642 posts). Additionally, we manually analyze sentiment and argumentation in 300 randomly sampled posts. We identified four discursive themes: Science, Conspiracies, Sociopolitical, and Containment. Negative- sentiment was substantially higher than what was found in prior work done before COVID-19 regarding vaccines and other containment measures. The negativity was driven mostly by arguments adapted from the anti-vaccine movement and not by WN ideology.
Keywords: COVID-19, Conspiracy Theories, Political Extremism
1. Introduction
Efforts to contain COVID-19 in the U.S. were hampered by misinformation and conspiracy theories [1]. Resistance to public health measures, such as lockdowns, masking and vaccines was particularly strong among conservatives and Republicans [2], driven by elite cues from media and politicians, including former President Trump [3], [4]. This partisan asymmetry continues a decades-old trend of politicization of science [5], [6], [7], that was accelerated during the pandemic, but influenced health behaviors that are not related to COVID-19 as well [8].
Recent studies have warned about the far-right’s use of science for justification of extremist, racist, and misogynist ideologies, which have the capacity to spill-over from fringe websites into mainstream American right-wing discourse [9], [10]. These include, for example, conspiracy theories like the Great Replacement, according to which non-white elites, including Jewish people, liberals, and feminists are working in concert to eliminate the white race and its dominance through ideologies of diversity and policies of immigration, and the promotion of medical procedures like abortion [9] or vaccines [10]. Such conspiracies are often being used to justify extreme violent actions, including mass-shootings, and could also erode trust in science and compliance with public health efforts [11].
Our main goal in this study is to examine a White Nationalist (WN) sub-forum dedicated to COVID-19 on the platform Stormfront, between January 2020 and December 2021. It substantially extends prior work that found very high levels of anti-scientific sentiment on Stormfront, but was limited to vaccines and was conducted before COVID-19 [10]. We believe studying the sub-forum dedicated to COVID-19 is required due to two reasons. First, Politicization of science accelerated during the pandemic [8]. Second, politicized resistance to COVID-19 containment went beyond vaccines, and was directed at lockdowns, masking, mandates, and testing as well [12].
Rejection of official explanations for and recommendations in regard to COVID-19 were prevalent among the American Right [13]. Republicans were significantly more likely to believe in at least one falsehood about COVID-19 [14], with many rejecting the existence of the pandemic altogether [15]. Resistance to public health containment efforts was not limited to the public, as many Republican officials and politicians refused to wear masks in public [16], often directly and explicitly contradicting and dismissing public health experts, scientists, and governmental agencies such the Centers for Disease Control and Prevention (CDC) [17]. Congruently, many Republicans refused to comply with public health recommendations [18], [19]. At times, resistance was violent in nature, some mild, such as Trump supporters’ refusal to wear masks on airplanes [20], and some extreme, as when a group of individuals was arrested for plotting to kidnap the governor of Michigan, Gretchen Whitmer, over her plan to impose lockdowns [21].
Prior studies on far-right forums found that resistance to vaccines was based on arguments similar to the ones employed by the broader anti-vaccine movement [22], including misbeliefs about physical harms caused by vaccines, their alleged ineffectiveness, and the ability to treat infectious diseases using alternative medicine. These were often rationalized via extremist lenses, for example when WN argued that toxic vaccines are being promoted by non-whites, and specifically Jewish people, for purposes of financial gain and demographic control. When WN expressed support for vaccines, it was often driven by racist accusations of non-whites as carriers of disease, or by collective narcissism, as when they celebrated vaccines as an achievement of white scientists. Overall, 52.5% of discussions coded expressed anti-vaccine sentiment, a share much higher than what was found in prior research in mainstream social media.
Since its establishment in 1995, Stormfront had served as a main online channel for WN, neo-Nazis, and other far-right extremists [23]. By 2015, the website served over 300,000 users [24], [25]. Multiple domestic terrorism attacks had been inspired by and even encouraged or coordinated over Stormfront [25], [26], [27]. Pertinent to our study, concerns were voiced about the use of science and other non-political topics to attract and recruit newcomers to white nationalism [28], [29]. While anti-immigration, racism, antisemitism and misogyny run across most of its pages [24], Stormfront consists of sub-forums dedicated to specific topics. The “Coronavirus (Covid-19)” sub-forum was created on February 2020, and at the time of data collection on December 9th, 2021, it included 9642 posts in 1557 threads. We used a multi-method approach to examine the sub-forum’s discussion of not only vaccines, but also other public health measures, such as lockdowns, masking, testing, and mandates.
Specifically, we asked:
RQ1: What themes dominated online WN discourse around COVID-19?
RQ2: How did theme dominance change over time?
RQ3: What sentiment towards containment measures and vaccines was expressed in online WN discourse around COVID-19?
RQ4: How did sentiment change over time?
We hypothesized the following:
H1: Sentiment towards vaccination will be more negative compared to previously measured sentiment towards vaccination pre-COVID on the same platform.
H2: WN resisting COVID-19 containment efforts will rely on adapted arguments used by the anti-vaccine movements.
H3: WN supporting COVID-19 containment efforts will rely on WN-related argumentation.
H4: Generally, discourse around containment measures will be guided more by non-WN argumentation than WN-related argumentation.
2. Methods
2.1. Data
The stormfront sub-forum was directly scraped at the end of December 2021. All the sub-forum content posted between January 18, 2020 to December 9, 2021 was downloaded. In total, 9642 posts and comments were downloaded from 1557 threads. After removal of extremely short documents (less than 6 characters, n = 165) and duplicated documents (n = 207), we ended with 9270 documents in 1552 threads.
2.2. Analysis
We used the inductive Analysis of Topic Model Networks or ANTMN framework [30] in three steps: Topic modeling, topic networking, and community detection. Topic modeling (using LDA) is an unsupervised machine learning method for identification of topics in large textual data [31]. Topics are distribution lists of word probabilities based on co-occurrences of words in the same documents. We pre-processed the corpus, removing stop-words, punctuation, numbers and symbols, and converted all text to lowercase. We removed words appearing in more than 90% of the documents or less than 5 posts (0.05% of all documents). For hyper-parameter tuning [7], we utilized 5-fold cross validation iterating over a range of topic numbers (from k = 5 to k = 100, in ‘skips’ of 5) and various levels of the alpha hyperparameter (from α = 50/k to α = 1/k). We found the model with k = 36 and α = 2/36 to offer optimal results, based on perplexity scores. Topics were labeled qualitatively by examining the words, unique words, and top 30 documents most representative of each topic (see Appendix 1 for top unique words). Eight “boilerplate” or mixed topics containing no consistent linguistic meaning was excluded from analysis [30].
We calculated a network of topics in the second step, based on the co-occurrence of topics in the same posts. We used cosine similarities over the document-topic matrix as edges in a network where each topic serves as a node, and their co-occurrence serves as edges. We used a backbone method [32] to discard non-significant edges. In the third step, we utilized a community detection algorithm, Louvain [33], to group topics into broader themes. Finally, we estimated the prevalence of themes over time.
2.3. Manual coding
Two authors adapted the codebook from Walter et al., 2022 [10] to include containment measures beyond vaccines. The revised codebook was tested using small samples of posts. Once the codebook was finalized, the two coders estimated their agreement on a sample of 60 posts. The coders reached an intercoder reliability of Brennan and Prediger Kappa greater than 0.79 for all items, aside from item 2 which had the value of 0.67 (full details in Appendix 2). After reaching reliability, the two authors manually coded a random sample of 300 posts. The full codebook can be found in Appendix 3.
3. Results
Our model identified 28 topics clustered into four themes in WN discourse on COVID-19. Based on our qualitative reading of the most representative words and full posts associated with each topic we named the themes: Science (in purple), Conspiracies (in blue), Sociopolitical (green) and Containment measures (gray) (See Fig. 1 ).
Fig. 1.
(Left Panel)Topic network of 28 topics (8 removed); Edges represent co-occurrence of topics in documents; Color represents community membership (Louvain). Network is weighted, and undirected. Edges filtered using the ‘backbone extraction’ method. (Right Panel) The monthly prevalence of the four themes inductively emerging from ANTMN on Stormfront’s COVID subforum 2020–2021 (smoothed Loess Curves).
The Science theme focused on the biology of COVID-19, its potential causes (many of which misinformed), and the severity of it, often dressing up conspiracy theories in ‘scientific’ language to justify their attack on containment measures and vaccination efforts. Two prominent topics in this theme discussed the dangers of vaccines and the benefits of alternative medicine. For example, a post relying on the former posted updates about the development of the vaccine, stating that “scientists believe they found potential coronavirus vaccine”. Another post however, dismissed the effectiveness of the vaccine and argued that they are filled with “poisonous substances” and are developed by “a system which has already proved that it fails us when it comes to keeping us safe”. Another user called the COVID-19 injection “the Jew vaccine”. Other posts relying on this theme discussed the origins and severity of the pandemic. These often included conspiracy theories that contradicted one another.
These discussions were extended further in the second theme, dedicated almost entirely to promoting the claim that the pandemic was a hoax or fake. Posts in this theme focused on refuting official recordings of fatalities, as exemplified in a post by a user stating that COVID-19 was “BS from the left… COVID isn’t even real, and no one is dying from it”. Another post accused public health institutions of inflating death numbers, saying “Every-one else died of something else, like a heart attack, cancer, car accident, and were called a covid death. Pretty much every-one who dies in a nursing home is called a covid death”. The alleged lie, according to Stormfront users, was sometimes promoted by non-whites: “Just like all other fake things Jews do, they make fake money backed by nothing, and turn it into something to collect interest on”. Others called the disease “a hoax”, “a fake pandemic”, and a “fraud”, and those believing in it “stupid” and “credulous”. Many users have pointed out a finger towards the mainstream media: “I have never seen such an overblown story except for 9/11. All news outlets are running it 24/7. Crazy”.
The Sociopolitical theme was driven by WN racist ideology. This is where Stormfront users were attacking non-whites, blaming them for spreading diseases and speculating about political and financial corruptions, mostly aimed against whites. Those acknowledging that the virus was “real” often blamed it on non-whites: “Yeah, this is what you get when you add negros to the mix” says one user, and another claimed that “Areas infested with nigra --one marker, aside from the general run down condition and blight, is you find spit on the sidewalks. Groids spit a lot. This spreads the virus, and the groids walk in it, and track it back to their hovels”. Many users echoed the arguments made by Trump and other conservative leaders and media personalities, that the virus was created intentionally by the Chinese in laboratories as part of a planned warfare against white people, potentially, as one user argued, at the “Wuhan Institute of Virology and the Wuhan National Biosafety Laboratory. Research was conducted here on bat coronaviruses, and a manipulated virus was created.” Users argued that officials were hiding “conclusive evidence that the virus was released by China to destroy the world”. Some implied it was part of a political plot against Trump himself: “This is a bio weapon, when the impeachment failed they released the virus”.
Finally, the Containment theme was dedicated to discussions of the necessity of masks, lockdowns, and vaccines, and how they threaten liberty and freedom of choice. In other words, many topics focused on resistance to local and federal mandates. Vaccine mandates, for example, were discussed as “unconstitutional, unlawful, and unwise”. Others explicitly warned that “We have got to stop the Covid cult before the damage becomes irreversible”. Lockdowns were described in the sub-forum as offensive action by “totalitarian bastards”, and people who faced a request from their employers to comply with public health restrictions were advised that “no job is ever worth having your health deliberately ruined by a system of tyrants.” Some argued that lockdown were a step towards the deployment of martial law, supporting their allegation with the claims that “the National Guard enforce curfew laws.” Others elaborated on this argument of COVID-19 restrictions as a slippery slope towards totalitarianism, a “trial run” and an “incremental martial law”, claiming for example, that cities were preparing to “ban sale of guns and alcohol while addressing coronavirus”.
Unlike in previous explorations of public health discourse on Stormfront, the prominence of the four themes changed drastically over time. For example, the theme of science was the most prominent over a large part of the timeframe studied, and showed a clear upwards trajectory, as usage of scientific language (even if faulty) increased over time. On the other hand, the conspiracy theme showed a downward trajectory with mild peaks (as denying the toll of COVID-19 became more challenging). The containment theme reached a peak around August 2020, and then started to decline gradually, rising again a year later at July 2021. Lastly, the sociopolitical theme remained largely constant with a slight decline towards July 2021.
To examine sentiment (RQ3) and changes in it over time (RQ4), following the computational analysis, we hand coded a sample of 300 posts for whether they included specific argumentation, both non-WN and WN-related, and for the sentiment exhibited in these posts towards vaccines and other containment measures. Sentiment was overwhelmingly negative towards both, with the share of anti-containment posts being 68%, and anti-vaccination share reaching 86%. These levels are not only higher than what was found in other platforms, including the high levels on Pinterest [34], but, as hypothesized in H1, also higher than what has been found in Stormfront pre-COVID [10].
The inflection points for containment and vaccines sentiment can be seen in Fig. 2 , which presents changes to the share of anti-vaccination and anti-containment content over time (grouped into units of 3 months due to data sparsity in some earlier months). As Fig. 2 shows, the share of anti-containment posts was (relatively) low at first, but skyrocketed at the second half of 2020. Anti-vaccination sentiment increased substantially at the start of 2021. The difference in the time of change is likely the result of vaccines being introduced later into the pandemic, after other containment measures from masks and social distancing to lockdowns have already been deployed. In other words, resistance to vaccines seem to increase as they turned from a hypothetical solution to an available, and at times mandatory, one. At around the same time when the vaccine became available, Joseph Biden replaced Donald Trump as the President of the United States. We conducted an interrupted time series analysis and found that the segment of January-March 2021, the beginning of the Biden administration and the most extensive rollout of the vaccine, served as a significant (p < .05) interruption point (see vertical dashed line in Fig. 2). Our data cannot disentangle these two possible explanations for the increased resistance, as the increase in vaccine rollout took place at about the same time as Biden’s inauguration. Due to the nature of our data, we cannot reject alternative explanations, such as the progress of the epidemic itself. Nevertheless, our analysis does point to an increase in resistance, where January-March 2021 serve as a key moment of change.
Fig. 2.
Share of anti-vaccination and anti-containment measures posts in Stormfront’s COVID subforum 2020–2021 based on manual content analysis (n = 300 posts). The vertical dashed line on the vaccine pane corresponds to the interruption point of January 2021 used in our time series analysis.
To gain additional insight into resistance to containment, we conducted an ANOVA to test whether the type of containment (vaccine, social distancing/lockdowns, and other) was associated with sentiment. The test indicated significant differences (p < .001) between the categories, with vaccines yielding the most antagonism (share of posts containing negative attitude was M = 0.89, SD = 0.31), followed by social distancing/lockdowns (M = 0.66, SD = 0.47), and finally other types of containment discussions (M = 0.48, SD = 0.50). All differences were statistically significant (p < .05).
Next, Fig. 3 presents the various argumentation types and their prominence. As hypothesized in H2, these followed adapted versions of anti-vaccine staples. As can be seen, infringment of individual freedoms was the most frequent argumentation category, followed closely by elite conspiracies, and then the necessity of containment measures and the biological possible harms of such measures. This is in contrast with the aforementioned previous study showing biological harms to be the most prominent argument-type pre-COVID-19. However, as in the previous study and in accordance with H4, non-WN argumentation was much more prominent than WN-specific argumentation with the top-3 non-WN argumentation (H3) being used in more than 25% of the posts (and the fourth argument over 20% of the posts), while the most prominent WN-related argumentation appearing in 15% of the posts.
Fig. 3.
Prevalence of various content argumentation features in posts in Stormfront 2020–2021 manually analyzed content (n = 300 posts).
To gain additional insight into argumentation, we broke the share of anti and pro/neutral sentiment based on the argument used in the post. Fig. 4 shows the arguments adapted from the anti-vaccine movement [22] and WN-related argumentation side-by-side. The non-WN argumentation types were found to be highly negative, aside from the rare category of debunking anti-vaccine claims (as expected). All other arguments were overwhelmingly negative, with the least antagonistic category being Elite Conspiracies (87% of posts are negative). In contrast, WN-specific argumentation was largely positive with only arguments connecting non-whites to conspiracies around COVID-19 being negative. This category has risen substantially from previous studies on Stormfront. Posts containing the other 3 categories (non-whites spread diseases, white narcissism, and anti-containment delegitimizes white nationalists) were mostly non-negative in their stance towards containment.
Fig. 4.
Share of positive and negative posts in various non-WN argumentation (Left Panel) and WN argumentation (Right Panel) according to hand coded materials from the COVID subforum of Stormfront.
4. Discussion
The decades-old partisan politicization of science and medicine have reached a zenith during the tensed sociopolitical time of COVID-19, with conservative media and Republican politicians promoting distrust in the scientific community as a whole, public health organizations and their representatives like Dr. Anthony Fauci, and the science behind containment efforts, from masking to vaccines [3], [4], [17]. As a result, rejection of COVID-19 science and noncompliance with science-consistent behaviors were disproportionately high on the American right. The resistance to containment resulted in multiple cases of mild and severe violence [21], often guided by misinformation and conspiracy theories that were created and propagated on fringe extremist online websites.
This study substantially extends prior work that unveiled unprecedented levels of antagonism towards science and medicine among white nationalists (WN), as expressed on social media. We examined all discourse in a Stormfront sub-forum dedicated to COVID-19 between January 2020 and December 2021. We found that volume of activity in the sub-forum peaked at the early stage of the outbreak as the US began considering, and soon thereafter deploying, restrictions and public health measures aimed at slowing down the spread of the virus.
We identified four themes; the most prominent, which only became more prevalent as vaccines were developed was Science, which focused on the biology of the virus and its potential vaccines, with users expressing unsubstantiated worries about the pandemic being intentionally exaggerated and that a vaccine was either unnecessary or dangerous. The alleged-panic and push for vaccines was put on Jewish people and other elites; These arguments were promoted further in the Conspiracies theme, that promoted conspiracy theories, blaming Jewish people and their supposed allies (e.g., the media) for propagating a “fake pandemic”, and promoting containment efforts, from lockdowns to vaccines, for financial gains and control. The Sociopolitical theme echoed WN racist ideology and tropes, accusing non-whites of creating or spreading the disease, either due to their alleged inferior morality (in the case of Black people) or their sinister plan to use it as a bioweapon (China). Finally, the Containment theme attacked the use of public health measures like masks and lockdowns (and later on, vaccine mandates), portraying them as infringing liberty and freedoms. Users speculated that lockdowns and mandates would be a first step in a slippery slope towards a totalitarian coup by the elites, liberals, and non-whites.
Sentiment towards vaccines and other measures was extremely negative. Sentiment became more negative over time, as more containment measures were introduced and mandated and as Biden replaced Trump as President. It is also possible that the rise in negativity resulted, in part, from fatigue among those believing in the severity of the disease and therefore in the need for containment, who may have left the sub-forum over time. Our data indicate that the number of unique users declined over time, and it is possible that the stage was left to more extremist or at least antagonistic users. However, sentiment continued to increase in negativity even after the number of users stabilized. Importantly, while the level of negativity on Stormfront is remarkably higher than what was found in prior research on other social media platforms, we acknowledge that a direct comparison between our findings and national, and particularly conservative, discourse on mainstream platforms, would allow to better understand the deviance of extremist views from those held by mainstream conservatives. As such comparative analysis is beyond the scope of our work, we call for future studies to extend our findings by comparing the results to other conservative platforms, such as Parler, Truth-Social, or dedicated sub-Reddits, during COVID-19.
5. Conclusion
Taken together, the already anti-scientific sentiment on Stormfront that was identified in previous studies was vastly exacerbated during COVID-19, with antagonism rising further as time passed, and particularly after former President Trump was replaced by President Biden and when more restrictions and containment mandates were put in place. As hypothesized, resistance was not limited to vaccines, but was expressed in regard to other public health measures as well.
Though our study did not directly test for effects or dissemination outside of Stormfront, prior research did point to the potential of fringe conspiracies on the dark corners of the internet, such as Qanon, to spill-over into mainstream discourse [9], [35]. We therefore believe that scholarly and regulatory attention should be given to the rise of anti-scientific sentiment on far-right platforms, and the proliferation of misinformation and conspiracy theories. Future studies can test such spillover effects directly by comparing our findings to discourse on other platforms. If the anti-scientific and anti-medical sentiment and arguments we’ve identified here do spillover into mainstream discourse, regulators should consider monitoring and even removing such extremist content. We acknowledge that regulating extremism is a complicated task. On one hand, the existence of extremist platforms provides precious resource for those who fight domestic terrorism and radicalization, both in academia and in the intelligence community [36]. On the other hand, keeping the platforms active may risk allowing such dangerous and violent discourse to make its way to other, broader and more popular media environments. This could result in further widening the already concerning gap between Democrats and Republicans in trust in science, in general, and compliance with public health measures, specifically [37]. As politicization of COVID-19 science was not limited to the U.S., we urge scholars to extend our analysis to non-American, and non-Western contexts as well.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix.
Appendix 1
Top 10 Unique (Frex) Words Per Topic.
| # | Topic Name | Top 10 Unique Words |
|---|---|---|
| 1 | Fake pandemic | media, people, fake, news, real, like, one, believe, truth, just |
| 2 | Vaccine mandates | vaccine, vaccinated, vaccination, mandate, employees, covid, get, can, covid-19, court |
| 3 | financial Corruption | money, million, business, company, companies, pay, get, make, working, stock |
| 4 | International conspiracies and threats | jews, israel, world, people, war, america, jewish, white, china, just |
| 5 | Vaccine pop control conspiracies | world, people, great, control, think, population, one, global, reset, new |
| 6 | DELETED Videos media and links | video, youtube, thread, post, link, just, news, watch, read, please |
| 7 | Celebs and elites infections | coronavirus, said, new, news, covid-19, health, positive, virus, tested, york |
| 8 | Mixed or Health disparities | bill, gates, people, old, well, like, corona, just, see, can |
| 9 | Containment measures and restrictions | lockdown, people, country, countries, measures, restrictions, coronavirus, travel, australia, government |
| 10 | Research and biology | covid-19, coronavirus, health, disease, said, research, fauci, study, cdc, u.s |
| 11 | DELETED images | image, full, bar, original, view, click, sized, resized, toilet, paper |
| 12 | Curfews lockdowns and martial law | state, law, governor, states, order, county, martial, mayor, florida, texas |
| 13 | DELETED | like, donâ, itâ, know, good, just, going, time, get, can |
| 14 | Vaccine harms | vaccine, covid, vaccines, pfizer, covid-19, shots, people, shot, vaccinated, blood |
| 15 | Races | white, blacks, whites, black, africa, race, african, people, africans, many |
| 16 | Jewish conspiracy | jews, jew, jewish, white, like, kill, even, good, people, one |
| 17 | DELETED | people, get, now, just, like, going, can, think, time, work |
| 18 | Cases and deaths statistics | deaths, coronavirus, cases, new, people, covid-19, million, number, york, death |
| 19 | Toxics | like, water, air, since, can, live, years, area, fracking, animals |
| 20 | Natural Medicine and Shortages | food, get, good, take, eat, can, stores, store, water, also |
| 21 | Vaccine effectiveness | vaccine, vaccines, vaccinated, get, covid, people, take, just, virus, shot |
| 22 | Biology and medicine | virus, blood, cells, body, can, vaccines, viruses, dna, immune, spike |
| 23 | Wuhan virus | china, virus, chinese, wuhan, lab, world, pandemic, coronavirus, spread, bat |
| 24 | Education | school, children, students, kids, covid, schools, public, parents, child, must |
| 25 | DELETED Personal stories | just, one, know, got, get, people, like, think, now, time |
| 26 | Police | police, woman, man, children, video, mother, arrested, one, officers, kids |
| 27 | Politics | trump, president, biden, coronavirus, white, house, like, donald, news, americans |
| 28 | MIXED society | people, one, society, right, even, medical, must, another, control, may |
| 29 | Masks and social distance | masks, mask, wear, wearing, people, face, can, distancing, virus, even |
| 30 | Mixed | like, people, just, know, think, one, can, even, something, conspiracy |
| 31 | Causes and severity of epidemics | radiation, just, even, thing, virus, check, know, seems, health, get |
| 32 | Refuting COVID deaths | covid, death, people, died, die, covid-19, health, heart, age, deaths |
| 33 | Health systems and institutions | hospital, nursing, care, home, medical, patients, hospitals, doctors, homes, new |
| 34 | Comparisons to flu | flu, people, virus, rate, even, deaths, year, death, influenza, much |
| 35 | Alternative medicine | drug, treatment, ivermectin, chloroquine, patients, covid-19, hydroxychloroquine, drugs, use, treat |
| 36 | COVID tests | test, covid, positive, virus, tested, tests, testing, covid-19, coronavirus, can |
Appendix 2
Reliability Scores for Manual Coding Per Item.
| item | PERC | KRIP | BP | |
|---|---|---|---|---|
| 1 | thread_id | 1 | 1 | 1 |
| 2 | title | 1 | 1 | 1 |
| 3 | post_counter | 1 | 1 | 1 |
| 4 | user | 1 | 1 | 1 |
| 5 | time | 1 | 1 | 1 |
| 6 | text | 1 | 1 | 1 |
| 7 | index | 1 | 1 | 1 |
| 8 | date | 1 | 1 | 1 |
| 9 | date2 | 1 | 1 | 1 |
| 10 | X1_relevant.0.0.irrellevant. | 0.95 | 0.839982 | 0.9 |
| 11 | X2_Freedom | NA | 0.732394 | 0.791667 |
| 12 | X3_Elites | NA | 0.801134 | 0.84375 |
| 13 | X4_threat_non_white_cont | NA | 0.636015 | 0.875 |
| 14 | X5_threat_non_white_disease | NA | 0.296296 | 0.833333 |
| 15 | X6_white.narc | NA | 0.373626 | 0.875 |
| 16 | X7_cont.unnecessary | NA | 0.616162 | 0.666667 |
| 17 | X8_cont.dont.work | NA | −0.02151 | 0.875 |
| 18 | X9_alternative.med | NA | −0.88906 | NA |
| 19 | X10_bio.harm | NA | 0.717094 | 0.791667 |
| 20 | X11_psych.harm | NA | −0.04396 | 0.791667 |
| 21 | X12_debunk.antisci | NA | −0.01064 | 0.916667 |
| 22 | X13_stormfron.look.crazy | NA | 0 | 0.958333 |
| 23 | X14_anti.cont.0.1.anti. | NA | 0.621292 | 0.8125 |
| 24 | X15_antivax.0.1.anti. | NA | 0.832245 | 0.8125 |
| 25 | X16_which.cont | NA | 0.563555 | NA |
Appendix 3 – Codebook for Manual Content Analysis.
Containment Codebook: Stormfront COVID.
INTRODUCTION.
The health and social risks associated with outbreaks of infectious diseases, and the interventions to prevent and counter them, particularly vaccines, have always been sources of heated debate (Offit, 2011); however, recent studies have shown an uptick in the politicization of health and vaccine topics, with evidence of intentional weaponization of vaccine discourse in order to increase political polarization (Walter, Ophir, & Jamieson, 2020). A recent study identified very high levels of resistance among White Nationalists (WN) on Stormfront (SF), yet it focused exclusively on vaccines (Walter et al., 2022). This study extends the analysis by examining the general discourse of COVID-19 on SF, and specifically argumentation and stances regarding containment measures associated with COVID-19.
CODING SCHEMA
For each of the following items, mark 1 if the argument is represented in the content, and 0 if it does not appear in the content. For convenience, statements have been separated into groups. There are example statements representing example arguments below each Group.
In posts where there are two “originally posted”, only code the top part of the content before the second “originally posted”.
1. Does the item discuss containment directly or in an implied manner?
2. Does it mention specific containment measures.
3. Does it discuss the viability of fighting the epidemic in general.
4. Does it discuss the severity, causes or nature of the covid epidemic as serious/hoax.
0 = post is unrelated to containment (move to next item).
1 = post is related to containment.
GROUP 1: EXTERNAL THREAT.
Representations of out-groups (e.g. minorities, elites, etc.) within society as a “threat” to dominant groups, provoking “group level emotions”.
5. Individual freedom and liberty are under threat .
6. Civil liberties are violated by taking away choice on medical treatments, vaccinations, and other behaviors (e.g., stay-at-home orders).
7. Civil liberties are violated by taking away people’s bodily autonomy.
8. Those who do not follow public health recommendations are being harassed.
9. Mandates (e.g., vaccines, stay-at-home, COVID testing) are excessive government control.
10. Those who resist public health recommendations are being censored or villainized.
11. People know what’s best for them and their families, not the government.
12. Elites (Elites have to be specified) harm the masses for their benefit including financial gain or control.
-
a.
COVID policies are motivated by profit of individuals or industries like the pharmaceutical companies.
-
b.
Those promoting public health measures benefit from their detrimental effects (e.g., in a case where illness is caused by vaccines).
-
c.
Government protects doctors/manufacturers from liability (e.g. keeping information off media, censoring opinions and news).
-
d.
COVID policies and measures are motivated by profit of the deep state (e.g. New World Order— NWO— and Zionist Occupied Government— ZOG).
-
e.
The medical industry and/or public health establishment are corrupted, and findings or data are being withheld from the public.
-
f.
The damages of public health measures (e.g., vaccine reactions or lockdown depression) are underreported (i.e. they are suppressed by government, pharma companies and other interested parties).
-
g.
The government uses public health measures to control, brainwash, numb, sterilize or harm (e.g., vaccines are bioweapons)
-
h.
The government develops vaccines to track us (microchips)
-
i.Media can be elites
-
1.Threats from other out-groups: Non-whites specifically – Containment measures
-
1.
-
a.
Non-whites control public health organizations / vaccines / science / pharmaceutical companies / government
-
b.Non-whites benefit financially from public health measures (including vaccines)
-
2.Threats from other out-groups: Non-whites specifically – diseases
-
2.
-
a.
Non-whites spread diseases
-
b.
Non-whites are more likely to get sick
-
c.
Non-whites refuse to follow public health measures
GROUP 2: WHITE NARCISSISM.
“Collective narcissism— an emotional investment in an unrealistic belief about the ingroup’s greatness— aiming to explain how feelings about an ingroup shape a tendency to aggress against outgroups” (de Zavala et al., 2009, p. X). Common characteristics of collective narcissism include appeals to in-group authority and achievement, requirement for special treatment and getting what they feel they deserve, including appropriate respect, while at the same time seeing the in-group as commonly misunderstood and being angry about group criticism.
-
1.
White narcissism
13. Emphasizing white intervention and achievements in science and public health.
14. White race carries the burden of financing global and domestic public health efforts, benefitting and treating non-whites.
15. White culture has moral purity and doesn’t need containment (e.g., careful and responsible so don’t need masks or vaccines).
16. White supremacy, white race is better biologically.
17. Whites are more responsible than non-whites.
GROUP 3: CONTAINMENT EFFICACY (ANTI)
-
2.
Containment measures are unnecessary
18. Disease incidences naturally declined without public health measures (e.g. from improved hygiene, natural selection).
19. Dangers of COVID are exaggerated.
20. COVID cases are rare/not contagious/mild.
21. COVID doesn’t exist at all.
-
3.
Containment measures don’t work
22. The science that drives policy is outdated and/or have been disproven.
23. Measures like masks, vaccines, or lockdowns don’t work (i.e., don’t slow down the spread or severity of COVID).
-
3.Alternative medicine is as good / better than public health measures (e.g. Western medicine)
-
a.Alternative treatments (e.g. homeopathy, acupuncture, chiropractics, etc.) and/or alternative products (e.g. vitamins, essential oils, plant-based products, etc.) are as good as or superior to what the government recommends (including masks, vaccines, lockdowns, testing).
-
b.Established medical knowledge about COVID is wrong (e.g. germ theory is untrue).
-
c.“Natural” approaches are best, and nature should take its course (i.e. people should get and recover from COVID naturally).
-
d.Biomedicine is epistemologically wrong; there are other effective/accurate ways of “knowing” (e.g. intuition, instinct).
-
a.
GROUP 4: RISKS.
-
4.
Containment measures cause biological harm
24. Masks INCREASE contagion of COVID.
25. Masks cause rashes or respiratory issues.
26. Washing hands, vaccines, masks etc. reduce efficacy of immune system.
27. Vaccines cause idiopathic illnesses and developmental disorders (e.g. autism).
28. Vaccines erode immunity, weaken the body.
29. Rushing vaccine trials causes injury and adverse effects.
30. Taking multiple vaccines at once increases adverse effects.
31. Vaccines and other medical treatments cause long-term adverse effects on children (including autism, etc.).
32. Vaccines cause short- and long-term adverse effects on women’s reproductive system (e.g. miscarriage, infertility, etc.).
33. Vaccines cause diseases to mutate, making them more dangerous.
34. Contaminated vaccine lots have more adverse effects.
35. Vaccines contain toxins and contaminants.
36. Containment measures cause non-biological (Psychological/social/economic) harm.
37. Public health measures will cause depression, loneliness, suicides.
38. Public health measures will harm children’s development, social skills, stability.
39. Some children struggle mentally with masks, distancing, loneliness etc.
40. Containment measures harm the economy or individuals financially.
41. Containment measures harm social relations between people or groups in society.
GROUP 6: CONTAINMENT EFFICACY (PRO).
42. Debunking anti-containment narratives.
43. Claims made by those resisting public health measures are outdated and/or have been disproven .
44. Sources cited against public health measures are not used truthfully; false conclusions drawn.
45. No statistic/citations provided to support public health measures claims.
46. Unsupported claims against public health measures are made.
GROUP 7: SF RELIABILITY (PRO).
Anti-containment arguments and narratives make the Stormfront community/white nationalists look crazy.
GROUP 8: CONTAINMENT SENTIMENT.
48. Finally, examine each post and state whether the post exhibits anti-public health measures sentiment. Ignore quoted content in your coding.
1 = Only or mostly anti-containment efforts.
2 = Not anti-containment.
49. Examine each post and state whether the post exhibits specifically anti-vaccine sentiment. Ignore quoted content in your coding.
0 = post is unrelated to vaccines.
1 = Only or mostly anti-vaccination.
2 = Not anti-vaccination.
Contrasting pro containment arguments is considered as a anti argument. Arguments do not have to be substantive and can relate to groups rather the technology (e.g. “all anti vaxxers are crazy”). One way to decide is to ask yourself, would the author of the post refuse to be vaccinated if offered.
Data availability
Data will be made available on request.
References
- 1.Romer D., Jamieson K.H. Conspiracy theories as barriers to controlling the spread of COVID-19 in the U.S. Soc Sci Med. 2020;263:113356. doi: 10.1016/j.socscimed.2020.113356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Levin B. Thanks to Republican Anti-Vaxxers, the U.S. May Never Reach COVID-19 Herd Immunity | Vanity Fair. Vanity Fair. Published 2021. Accessed June 14, 2021. https://www.vanityfair.com/news/2021/04/republicans-anti-vaccine-herd-immunity.
- 3.Motta M., Stecula D., Farhart C. How Right-Leaning Media Coverage of COVID-19 facilitated the spread of misinformation in the early stages of the pandemic in the U.S. Can. J Polit Sci. 2020;53(2):335–342. doi: 10.1017/S0008423920000396. [DOI] [Google Scholar]
- 4.Yamey G, Gonsalves G. Donald Trump: a political determinant of covid-19. BMJ. Published online April 24, 2020:m1643. doi:10.1136/bmj.m1643. [DOI] [PubMed]
- 5.Bolsen T., Druckman J.N. Counteracting the Politicization of Science. J Commun. 2015;65(5):745–769. doi: 10.1111/jcom.12171. [DOI] [Google Scholar]
- 6.Mooney C. Basic Books; 2007. The Republican War on Science. [Google Scholar]
- 7.Walter D., Ophir Y., Jamieson K.H. Russian twitter accounts and the partisan polarization of vaccine discourse, 2015–2017. Am J Public Health. 2020;110(5):718–724. doi: 10.2105/AJPH.2019.305564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Harry E. Flu shots uptake is now partisan. It didn’t use to be. - CNNPolitics. CNN. Published 2021. Accessed August 10, 2022. https://www.cnn.com/2021/11/14/politics/flu-partisan-divide-analysis/index.html.
- 9.Ophir Y, Pruden ML, Walter D, Lokmanoglu AD, Tebaldi C, Wang R. Weaponizing reproductive rights: a mixed-method analysis of White nationalists’ discussion of abortions online. Inf Commun Soc. Published online June 27, 2022:1-26. doi:10.1080/1369118X.2022.2077654.
- 10.Walter D., Ophir Y., Lokmanoglu A.D., Pruden M.L. Vaccine discourse in white nationalist online communication: a mixed-methods computational approach. Soc Sci Med. 2022;298 doi: 10.1016/j.socscimed.2022.114859. [DOI] [PubMed] [Google Scholar]
- 11.Bodner J., Welch W., Brodie I. COVID-19 Conspiracy Theories: QAnon, 5G, the New World Order and Other Viral Ideas. McFarland. 2020 [Google Scholar]
- 12.Gelfand M., Li R., Stamkou E., Pieper D., Denison E., Fernandez J., et al. Persuading republicans and democrats to comply with mask wearing: an intervention tournament. J Exp Soc Psychol. 2022;101:104299. doi: 10.1016/j.jesp.2022.104299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Calvillo D.P., Ross B.J., Garcia R.J.B., Smelter T.J., Rutchick A.M. Political ideology predicts perceptions of the threat of COVID-19 (and Susceptibility to Fake News About It) Soc Psychol Personal Sci. 2020;11(8):1119–1128. doi: 10.1177/1948550620940539. [DOI] [Google Scholar]
- 14.Relman E. 94% of Republicans believe or are unsure of at least 1 falsehood about COVID-19 or the vaccines. Business Insider. Published 2021. Accessed August 15, 2022. https://www.businessinsider.com/94-republicans-believe-are-unsure-of-covid-19-vaccine-falsehoods-2021-11.
- 15.Bump P. Analysis | Conspiracy theories are common on the right — but few Republicans adhere to all of them. Washington Post. https://www.washingtonpost.com/politics/2021/07/19/conspiracy-theories-are-common-right-but-few-republicans-adhere-all-them/. Published 2021. Accessed August 15, 2022.
- 16.Sotomayor M, Wagner J. House Republicans refuse to follow new mask mandate, leading Pelosi to call McCarthy a ?moron? for his comments - The Washington Post. Washington Post. Published 2021. Accessed August 15, 2022. https://www.washingtonpost.com/politics/mask-mandate-reinstituted-in-house-of-representatives-by-congressional-physician-as-concern-increases-about-delta-variant/2021/07/28/22a4010c-efa4-11eb-ab6f-b41a066381df_story.html.
- 17.Young D.G., Rasheed H., Bleakley A., Langbaum J.B. The politics of mask-wearing: political preferences, reactance, and conflict aversion during COVID. Soc Sci Med. 2022;298 doi: 10.1016/j.socscimed.2022.114836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Grayer A, Fox L. Nearly half of House Republicans won’t say publicly if they are vaccinated - CNNPolitics. CNN. Published 2021. Accessed August 15, 2022. https://www.cnn.com/2021/07/22/politics/house-republicans-vaccination-rates/index.html.
- 19.McKelvey T. Coronavirus: Why are Americans so angry about masks? BBC News. https://www.bbc.com/news/world-us-canada-53477121. Published July 20, 2020. Accessed August 15, 2022.
- 20.Batchelor T. Southwest Airlines kicks Black Trump supporter off plane for not wearing mask. Newsweek. Published October 14, 20Accessed August 15, 2022. https://www.newsweek.com/southwest-airlines-kicks-trump-supporter-off-plane-not-wearing-mask-1539038.
- 21.Smith M. The case of four men accused of plotting to kidnap Michigan’s governor over Covid rules goes to the jury. - The New York Times. Published 2022. Accessed August 15, 2022. https://www.nytimes.com/2022/04/02/world/deliberations-begin-in-the-trial-of-four-men-accused-of-plotting-to-kidnap-michigans-governor-over-covid-rules.html.
- 22.Kata A. Anti-vaccine activists, Web 2.0, and the postmodern paradigm – An overview of tactics and tropes used online by the anti-vaccination movement. Vaccine. 2012;30(25):3778–3789. doi: 10.1016/j.vaccine.2011.11.112. [DOI] [PubMed] [Google Scholar]
- 23.Bowman-Grieve L. Exploring, “Stormfront”: a virtual community of the radical right. Stud Confl Terror. 2009;32(11):989–1007. doi: 10.1080/10576100903259951. [DOI] [Google Scholar]
- 24.Caren N, Jowers K, Gaby S. A Social Movement Online Community: Stormfront and the White Nationalist Movement. In: Earl J, Rohlinger DA, eds. Research in Social Movements, Conflicts and Change. Vol 33. Emerald Group Publishing Limited; 2012:163-193. doi:10.1108/S0163-786X(2012)0000033010.
- 25.Hern A. Stormfront: “murder capital of internet” pulled offline after civil rights action. the Guardian. Published August 29, 2017. Accessed July 2, 2021. http://www.theguardian.com/technology/2017/aug/29/stormfront-neo-nazi-hate-site-murder-internet-pulled-offline-web-com-civil-rights-action.
- 26.Bunch W. The Pittsburgh cop killer, the radio “hit list,” and the right-wing media. https://www.inquirer.com. Published 2010. Accessed July 2, 2021. https://www.inquirer.com/philly/blogs/attytood/Poplawski.html.
- 27.Neiwert D. Alt-America: The Rise of the Radical Right in the Age of Trump.; 2017. Accessed July 2, 2021. https://www.google.com/books/edition/_/vedqDwAAQBAJ?hl=en&gbpv=1&printsec=frontcover.
- 28.Hartzell S.L. Whiteness feels good here: interrogating white nationalist rhetoric on Stormfront. Commun Crit Stud. 2020;17(2):129–148. doi: 10.1080/14791420.2020.1745858. [DOI] [Google Scholar]
- 29.Meddaugh P.M., Kay J. Hate Speech or “Reasonable Racism?” The Other in Stormfront. J Mass Media Ethics. 2009;24(4):251–268. doi: 10.1080/08900520903320936. [DOI] [Google Scholar]
- 30.Walter D., Ophir Y. News frame analysis: an inductive mixed-method computational approach. Commun Methods Meas. 2019;13(4):248–266. doi: 10.1080/19312458.2019.1639145. [DOI] [Google Scholar]
- 31.Blei D.M., Ng A.Y., Jordan M.I. Latent Dirichlet Allocation. J Mach Learn Res. 2003;3:993–1022. [Google Scholar]
- 32.Serrano M.Á., Boguñá M., Vespignani A. Extracting the multiscale backbone of complex weighted networks. Proc Natl Acad Sci. 2009;106(16):6483–6488. doi: 10.1073/pnas.0808904106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Blondel V.D., Guillaume J.L., Lambiotte R., Lefebvre E. Fast unfolding of communities in large networks. J Stat Mech Theory Exp. 2008;2008(10):P10008. doi: 10.1088/1742-5468/2008/10/P10008. [DOI] [Google Scholar]
- 34.Guidry J.P.D., Carlyle K., Messner M., Jin Y. On pins and needles: how vaccines are portrayed on Pinterest. Vaccine. 2015;33(39):5051–5056. doi: 10.1016/j.vaccine.2015.08.064. [DOI] [PubMed] [Google Scholar]
- 35.Forberg P.L. From the fringe to the fore: an algorithmic ethnography of the far-right conspiracy theory group QAnon. J Contemp Ethnogr. 2022;51(3):291–317. doi: 10.1177/08912416211040560. [DOI] [Google Scholar]
- 36.McMinimy K., Winkler C.K., Lokmanoglu A.D., Almahmoud M. Censoring Extremism: Influence of Online Restriction on Official Media Products of ISIS. Terror Polit Violence. 2021:1–17. doi: 10.1080/09546553.2021.1988938. [DOI] [Google Scholar]
- 37.Hotez P.J. America’s deadly flirtation with antiscience and the medical freedom movement. J Clin Invest. 2021;131(7) doi: 10.1172/JCI149072. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data will be made available on request.