One of the major trends during the COVID-19 pandemic was an uptick in the volume of research being posted as preprints prior to formal peer review. Mareike Fenja Bauer and Maximilian Heimstädt explore one instance of how a preprint was integral to the construction of conspiracy theories and suggest how better platform governance might mitigate these risks.

The prevalence of conspiracy theories has been amplified by social media platforms. Conspiracy theorists—individuals who consume, discuss, and share such narratives—use social media to connect and find audiences. During the COVID-19 pandemic, it became clear that the popularity of conspiracy theories on social media is intertwined with one innovation in scholarly communication: preprints. Preprints help scientists communicate their preliminary findings quickly, but they can also be misused to provide legitimacy to conspiracy theories.

In our recent study Clickbait or Conspiracy? How Twitter users address the epistemic uncertainty of a controversial preprint we (alongside Carlos Franzreb and Sonja Schimmler) followed up on this observation and systematically examined the relationship between one preprint and the conspiracy theories that grew out of it.

A controversial preprint circulates on Twitter

In early 2020, there was an urgent need for information, guidance and advice regarding COVID-19. To satisfy this need for sound and fast knowledge, many scientists published their preliminary results on the virus as preprints. At the same conspiracy theories spread around the globe, providing seemingly simple explanations for its origin—for example, that the virus was produced as a bioweapon in a laboratory and deliberately released.

On January 31, 2020, a biomedical preprint was uploaded to the preprint server bioRxiv. The authors of the preprint claimed to have found an allegedly “uncanny similarity” between the new coronavirus and HIV that is “unlikely to be fortuitous in nature”. Whether intentional or not, the preprint’s choice of words had clear parallels with the lab leak conspiracy theory. The preprint triggered a heated discussion in the comments section of the preprint server, which came to the unanimous conclusion that the preprint did not meet scientific standards. After a few days, the authors withdrew the preprint. At this point, however, the document and its debate had already spilled over from the small arena of the preprint server to the much larger arena of Twitter.

How scientists and conspiracy theorists engage with the preprint

On Twitter, two groups in particular were interested in the controversial preprint: scientists and conspiracy theorists. Using a social network analysis, we took a closer look at the composition and activities of these groups. In line with the principles of open science, our research data and code are openly accessible. This analysis produced three significant observations:

1. The number of conspiracy theorists and scientists engaging with the preprint on Twitter are roughly the same. This is surprising, as it is often assumed that conspiracy narratives spread in spheres where there is little pushback from scientific experts. (Fig.1)
2. Conspiracy theorists and scientists are “alone together”: Although members of both groups comment intensively on the preprint, there is very little interaction between the groups. This is surprising, as social media platforms are often described as arenas for open and interactive conflicts of expertise. (Fig.2)
3. The groups disseminate the preprint in different ways. Scientists prefer to retweet colleagues’ comments on the preprint. Conspiracy theorists prefer to send original tweets about the preprint. (Fig.2)

The need for platform governance

The growth of preprints presents advantages for science. Some even see preprints as the core of a new paradigm across all disciplines of academic knowledge production—the so-called “Publish, Review, Curate” model. However, as our study has shown, preprints also entail risks. Preprints are perceived as legitimate contributions to the scientific debate by large parts of the public, even though they have not yet been certified as such through the peer review process. This characteristic of preprints explains, as our study has illustrated, why the spread of a problematic preprint does not come to a halt when scientists evaluate it as erroneous in comment sections of preprint servers or on social media platforms. In order to exploit the benefits of preprints while minimizing its risks, we therefore believe it is necessary to integrate them more strongly than before into the debate on platform governance. This involves two types of platforms: Preprint servers and social media platforms.

An important question for preprint servers is how they deal with withdrawn preprints. In the case we studied the authors withdrew the preprint, but the document is still available via bioRxiv. The withdrawal is only noted by an info box. Even if texts cannot be completely “depublished” from the Internet once they have been posted, it seems appropriate to us for preprint servers to completely remove withdrawn preprints from their platform. This is important because a preprint server’s brand may still confer legitimacy on a text in certain contexts even after it has been officially withdrawn.

Another important question concerns the “sanity checks” for newly uploaded preprints. Many preprint servers have a screening process when uploading new preprints to ensure that the texts meet minimum scientific standards. Only if a preprint passes the screening process will it be published. This screening process is often carried out by a community of volunteers. As preprint servers are increasingly becoming an important infrastructure for science, it is time to think about how this form of platform governance can be sustainably financed and expanded. While it is conceivable that the screening process for preprint servers could be financed by grants from universities or learned societies, currently the burden lies mostly with the efforts of individual volunteers.

Preprints also raise governance questions for social media platforms. To date, much of the debate on the governance of these platforms has focused on the moderation of information that can be clearly identified as false. Preprints are characterised by their “epistemic uncertainty”. From a scientific perspective, they are neither clearly valid nor invalid knowledge. On preprint servers and in journalistic articles, this special feature of preprints is explicitly stated. It therefore seems appropriate for social media platforms to automatically display a corresponding classification of the linked source when a preprint is shared.

Originally, preprints were a niche interest used exclusively for communication within the scientific community. However, preprints are increasingly being recognised by groups outside the scientific community, including in the media where recognition of their epistemic uncertainty has gained some traction. They therefore represent an important building block for the transformation from science for society to science in society. The prerequisite for this, however, is that through means of platform governance, the genre of preprints is carefully embedded in non-scientific forms of communication.

The content generated on this blog is for information purposes only. This Article gives the views and opinions of the authors and does not reflect the views and opinions of the Impact of Social Science blog (the blog), nor of the London School of Economics and Political Science. Please review our comments policy if you have any concerns on posting a comment below.

Image Credit: Reproduced with permission of the authors.