-- MaudPof - 27 Nov 2021

Poster

Video

-- LéaMartinez - 06 Feb 2023

Tracing trust in a truth-less world? Comparing three scientific controversies on TikTok and Twitter

Team members:

Warren Pearce, Louise Jens, Anneke Claessens, Maxime le Cotey, Paola Leijssen, Sylvia Hayes, Daniel Joinson, Achraf Ramoudi, Grégoire Cotton, Léa Martinez, Emma Cachin, Rachid Aoutil, Priyanthini Kularajah, Cheikhou Oumar Cisse, Mohammed Islam Tibesse

Introduction

Questions about the role of science in decision-making have come into sharp focus over the Covid-19 pandemic. In particular, concerns over the online spread of misinformation and disinformation have reached new heights, with the World Health Organisation warning of a so-called ‘infodemic’. However, scholars within STS and digital media studies highlight the need for more critical approaches to the issue which take into account not only changing political contexts but the epistemic changes wrought by social media platforms that are truth-less by design (Marres, 2018). Covid-19 demonstrated that social media provides fertile ground not only for disinformation, but also for the construction and reconstruction of trusted knowledge, sometimes from sources outside of the scientific establishment (Pearce, 2020).

This project will trace the sources, content and networks of trusted science in three cases of contemporary controversy: monkeypox, precision breeding and low traffic neighbourhoods. These three cases have been selected for their diversity in media attention, cultural reach and spatial scale. For example, monkeypox has been a global event encompassing science, culture and politics (Paparini et al., 2022). Precision breeding is the newest terminology used for public communication about genetically modified organisms, a long-running controversy rooted in concerns around public trust together in different ways in different countries (Helliwell et al., 2017). Low traffic neighbourhoods are a series of locally implemented transport policies informed by evidence from environmental science (Dudley et al., 2022).

In all three controversies, scientific knowledge is uncertain and sometimes at odds with other public concerns such as corporate control, economic instability and social inequalities. Platforms such as Twitter and TikTok are places where these conflicts surface, what Callon et al (2009, p.15) call overflows, where “expected connections are established between what should have been a very simple technical project and a plurality of stakes that are anything but technical”.

Research questions

Who are the key opinion leaders in three areas of modern scientific controversy: monkeypox, precision breeding and low traffic neighbourhoods?

What are these opinion leaders’ preferred sources of evidence and opinion, and how are they presented as trustworthy?

What themes are visible within the evidence being shared online, and do they suggest consensus or dissensus around the available scientific evidence?

Methodology and datasets

Monkeypox methodology

The controversy about Monkeybox has exploded since May 2022. Thousands of videos were available on Tik Tok. So, we chose the hits that the platform allows us to see. We also found that most of these videos were about symptoms, but this project we are doing in the DMI week is about science and science behind monkeypox. That is why we decided to focus on the terms ‘Monkeypox’ and ‘Science’ so we could have a better idea about the monkeypox science. Thanks to the triangulate tool, we found that from the 957 hits that we got from ‘monkeypox’, 17 out of them contain the hashtag ‘science’. We made a short description of the content, the comments and then we figured out from every video if they are scientific or not. We found out that there are two people who have come up multiple times in our analysis. We decide to focus on them because they are both medfluencers.

For Twitter we created a dataset by scraping around 100 tweets from the platform that contains the words science and monkeypox in one tweet. We looked at the month of May because that was a peak month of tweets about monkeypox. We then did a qualitative research of the top 50 tweets that include the search terms ‘Monkeypox and science’. We decided to look at the top 50 tweets that have the most retweets. We made a short description of the content and the comments. Then, we figured out from every tweet if they are scientific or not. There were a lot of scientists between these 50 tweets.

LTN methodology

To analyze the LTN in Tiktok, we build a dataset of Tik Tok videos which are relevant to the scientific argument around these programs. These will enable us to explore the terms most relevant to these schemes, and how users engage in debate around them. We used Zeeschuimer for the data collection, we had a total dataset of 834 videos.

For Twitter, Two Twitter datasets were pre-made but it was found that the keywords CAZ and LTN were returning many Tweets completely unrelated to the topic area. Thus, it was decided that we would generate a new Twitter dataset, with the use of more specific keywords. These keywords were identical to the ones used in the Tik Tok dataset; Climate Lockdown, ULEZ, Ultra Low Emission Zone, Clean Air Zone, Low Traffic Neighbourhood, 15 minute city.

This analysis left us with the 100 most frequency relevant hashtags, which were used in the network analysis. We then used Gephi to produce “communities”, where hashtags are grouped and clustered according to how often they appear together. This allowed us to identify key topics and themes being used in the debate about clean air zone, and led us to find some key users in each space.

We used the prominent themes identified in the network analysis to identify the range of arguments against the schemes. This was used to perform qualitative analysis of Tweets and Tik Toks which cover content within these themes. We used this to identify categories of opinion leaders and content against the scheme. Our qualitative analysis of specific users involved a) analysing the overall style and tone of their content, including whether they typically post about low traffic neighbourhoods or not; and b) analysing in more depth specific videos identified as being about the low traffic neighbourhood issue.

GMO methodology

Data Collection

To answer the research question, we chose to scrape data from Twitter and TikTok. For our study, we chose to collect two datasets for TikTok and Twitter with the keywords "GMO" and "precision breeding". We used the Zeeschuimer tool which allowed us to collect videos and associated data to create our databases of posts that mention the selected keywords.

For TikTok, we also used the tool triangulate to compare two datasets from different accounts by selecting the “url” column. Our main challenge was to decide whether to select data from an old and regularly used account or from a newly created account. Finally, we chose to select data from a new account because it allows us to avoid biases we can have with an old account.

Method

We will create a network where the nodes are instances of authors and hashtags. These nodes will be connected via co-occurrence within the graph. This will be a weighted graph, we will create a new variable based on the calculation of engagement. We calculate engagement rate by:

Engagement Rate = (Total # of Post Engagements (likes, views, comments) /

# of Followers) x 100.

This calculation can be found at the above link, Grin is a well-documented social media tool used to measure the success of influencers across different social media platforms. Post engagements are considered to be, likes, comments, shares, and plays. These were weighted in the formula according to importance. Likes having the lowest importance, while shares and comments were considered to have the highest importance in the formula.

What is noticeable when studying the GMO network is that there are several small communities which are specialized on a particular subject; these communities exist in isolation of the large component in the center of the graph. Thus, the subject of GMOs is being addressed in several ways. For example, there is a community specifically for dogs, this community addresses the issues around genetic modification of these animals. Based on this we can say that the graph of GMOs on Tik Tok, is a distributed graph with many communities operating in isolation of one another.

As the first steps toward achieving this goal we needed to get The Data in the right shape, The followings illustrate the process:

Staging Data with help of Pandas (Python Framework)

Aggregating Data by authors and losing the extra data keeping only the necessary columns for our calculations.

Calculating the Engagement rate using "likes", "comments", "shares".

Aggregating the authors by the hashtags incorporated in the tweets.

Exporting the Data in order to establish our graph using authors as nodes and Hashtags as edges.

Findings

Monkeypox

We distinguish here different types of influencers, first of all the medinfluencers coming from the scientific or medical world are mainly influencers whose first goal is to educate their communities about these issues. In this study, we focus on the analysis of medinfluencers on tiktok and on twitter.

In the study of opinion leaders in this file, the case of influencers on the issue of Monkeypox is largely influenced by the algorithm of the platform. Particularly on TikTok where there is no correlation in terms of scientific quality of content and the number of views. For example, in the opinion leaders on the subject of Monkeypox there are accounts with high engagement rates that are not credible. Studying tiktok we see that the subject of Monkeypox is very much the idea of fake news around its impact in the LGBTQUIA+ community.

The peculiarity of twitter is to have more formed opinion leaders. The credibility of their words is based on the very intention of their accounts and their professional background.

Low Emission Zones

In this part we have made maps that show the links that exist between hashtags both in Twitter and in TikTok.

On Twitter we can see the most important networks that emerge with the pink color and on the TikTok the most connected nodes are represented by the blue color.

Also, on TikTok, another community materialized by the green color often talks about the conspiracy theory.

Thus, some hashtags create more links than others in both networks.

However, the discussion forums on both networks differ from each other, there are hashtags that are discussed locally than others.

Thus, the data we analyzed on twitter and Tiktak allowed us to isolate two main dimensions.

These are sets of arguments against these devices, in these arguments we have verifiable or sometimes non-verifiable contents.

Finally, this classification is summarized in the figure with individual groups of different status: politicians, local exaggerators, Conspiracy theorists in the UK, international conspiracy theorists, international opinion news, Opinion pundits, local activists, concerned local residents.

GMO

Nowadays, the use of genetically modified organisms seems to be an advantage for some, and for others, the production and use of GMOs carry considerable dangers for individuals and the surrounding environment.

However, this issue has provoked intense and successive debates in the public space.

Today, social media has become the privileged space of public debate. Thus, social networks have taken over the debate on GMOs,

Genome editing is a method of making specific changes to the DNA of an organism. It is often used in research and has the potential to treat or cure genetic diseases.

TikTok is a social media platform primarily used for sharing short videos. It can be used to introduce and educate people about the basics of genome editing, as well as to showcase ongoing research and its potential applications. Researchers can also use TikTok to share animations or short videos that explain complex scientific concepts in an easy-to-understand format. Additionally, TikTok can be used to engage with a younger audience and educate them about the potential benefits and risks of genome editing. However, TikTok is less commonly used for scientific research and collaboration compared to platforms like Twitter and LinkedIn.

We find different communities very different, for example, The globalcitizensunited account has 2093 followers, 10.2K likes and 124 posts on TikTok. The account publishes content related to Genetically Modified Organism (GMO), artificial intelligence and transhumanism. The account shares conspiracy theories about Covid and the vaccine in particular. It also promotes biohacking and smart cities through transhumanism and nano technologies. The profile bio is #geneediting #biohacking #crisprcas #transhumanism #future #AI. The account uses hashtags such as geneticallymodifiedhumans, geneticallymodified, gmo, geneeditingembryo, transhumanism, biohacking, nanotechnology.

In other way, The doublecloverfarm has 45,3K followers, 954,3K likes and 340 posts. The account is run by Loren Lynch, an American vegetable farmer who lives in Pennsylvania. She grows vegetables and chicks for cooking. She also delivers her products. She makes videos about everyday life on the farm. It promotes its products and its way of life. She also shares recipes. She defends local and organic products. She defends hybridisation, but is against large farms and GMOs. The profile bio is Small farm life How to grow, preserve, & cook food Plus knitting, baking Zone 6b. This account uses hashtags such as gmo, gmofood, bioengeneering, wordsmatter, growingvegetables, geneticallymodified.

It is noticeable when studying the GMO network is that are several small communities which a specialized on a particular subject these communities exist in isolation of the large component in the centre of the graph. Thus, the subject of GMOs is being addressed in several ways. For example, there is a community specifically for dogs, this community addresses the issues around genetic modification of these animals. Based on this we can say that the graph of GMOs on Tik Tok, is a distributed graph with many communities operating in isolation of one another.

The great diversity of the profiles present in this network is partly due to the very wide impact that LDAs represent on our productions. This scientific innovation transforms in a sustainable way the way of producing and conceiving the raw materials and living organisms used for our economic and productive models. Thus, we find through this network the great diversity of themes in which LDAs are present.

The presence of conspiracy communities is not surprising since LDAs are widely integrated into the rhetoric and the "anti-system" imaginary.

Discussion

Previous studies of controversy have highlighted the importance of overflows. What is notable about the overflows is that these issues all have their roots in scientific evidence, from virology, biosciences and environmental science. Yet while scientific content plays a role in issue representation on TikTok, the analysis here shows how these controversies overflow into what Callon et al. (2009) called the ‘anything but technical’. These overflows manifest themselves in two different ways: verifiability and situatedness.

Verifiability

We found that many claims being made within the controversies were not easily categorized as misinformation or disinformation, as they were not verifiable or falsified. As discussed by Marc Tuters in the Winter School’s keynote lecture (2023), key claims from accounts typically described as conspiratorial were not amenable to being debunked, rather their intention was to establish an alternative interpretation of ‘ground truth’ events. For example, the introduction of Low Traffic Neighbourhoods was heavily criticized by accounts such as @london.by.lononders as part of a wider ‘anti-woke’ agenda that particularly focused on London’s Mayor Sadiq Khan. Within genome editing, criticism of the technology is linked to cultural and historic themes of ‘small farms’ which are beyond scientific framings of whether or not genetically modified organisms are safe to consume. Within monkeypox, there was a larger role for verification, but even here there were non-scientific arguments entangled, such as the controversy over which communities should be prioritized for vaccines.

Situatedness

The Political, place-based experiences come to the fore that lie outside the scientific frame of the issues; for example, the link between social inequality and traffic pollution in urban areas. These arguments are sometimes overshadowed by posts from international conspiracy theorists, but are likely more germane to conflict resolution.

The embodiment of knowledge in TikTok users is also key to understanding scientific controversies on the platform. For example, in monkeypox science, medfluencers take centre stage but take different approaches, either focusing on traditional one-way science communication or posting replies to users alongside other non-science videos. This, perhaps, represents a different kind of overflow: showing how the production and dissemination of scientific knowledge is embedded in everyday life, providing a challenge to notions of disembodied objectivity as a scientific norm.

Conclusion

To go further in our study, it would be a good thing to check the sources that the medfluencers were using like URLs on Twitter.

We couldn’t find the communities that the medfluencers were embedded in - clue from Dr Noc’s passing comment about % LGBT who had MPX.

It would be also important to look at who is following these medfluencers - who are the people commenting? That will tell us more about where these ideas are being circulated.

The way we target our opinion leaders is key to studying the trends that emerge around these controversies on Twitter and Tiktok. Other methods of selecting accounts and items could be tested for potentially more relevant visualization results.
Topic revision: r2 - 06 Feb 2023, LéaMartinez
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback