Vaccine discourse in white nationalist online communication: A mixed-methods computational approach

Research has indicated a growing resistance to vaccines among U.S. conservatives and Republicans. Following past successes of the far-right in mainstreaming health misinformation, this study tracks almost two decades of vaccine discourse on the extremist, white nationalist (WN) online message-board Stormfront. We examine the argumentative repertoire around vaccines on the forum, and whether it assimilated to or challenged common arguments for and against vaccines, or extended it in ways unique to the racist WN agenda.

We use a mixed-methods approach, combining unsupervised machine learning of 8892 posts including the term "vaccin*", published on Stormfront between 2001 and 2017. We supplemented the computational analysis with a manual coding of randomly sampled 500 posts, evaluating the prevalence of pro- and anti-vaccine sentiment, previously identified pro- and anti-vaccine arguments, and WN-specific arguments.

Discourse was dynamic, increasing around specific events, such as outbreaks and following legal debates about vaccine mandates. We identified four themes: conspiracies, science, race and white innovation. The prominence of themes over time was relatively stable. Our manual coding identified levels of anti-vaccine sentiment that were much higher than found in the past on mainstream social media. Most anti-vaccine posts relied on common anti-vaccine tropes and not on WN conspiracy theories. Pro-vaccination posts, however, were supported by unique race-based arguments.

We find a high volume of anti-vaccine sentiment among WN on Stormfront, but also identify unique pro-vaccine arguments that echo the group's racist ideology.

As with past health-related conspiracy theories, high levels of anti-vaccine sentiment in online far-right sociotechnical information systems could threaten public health, especially if it 'spills-over' to mainstream media. Many pro-vaccine arguments on the forum relied on racist, WN reasoning, thus preventing the authors from recommending the use of these unethical arguments in future public health communications.