What is viral about the virus?
Digital social research in the age of covid-19
Aside from the obvious tragedy, the disruptions, and the socio-political chaos spurred by the ongoing outbreak, global events such as covid-19 also play right into the hands of our datafied times. Aside from everything else that it surely is, it is definitely our first deeply mediatised pandemic. It plays out in a digitally transformed society. It raises issues of truth, trust, trolls, and technology, as much as it raises medical and economical ones.
As the current and urgent state of affairs gradually subsides, there will be a pressing need for critical research to be carried out by digital social scholars with the ability to tie data science skills together with a critical analytical sensibility.
Do you see what I see?
The last few chaotic weeks have underlined that, while we should of course not take this virus lightly in any way, the ways in which we metrify and visualise reality is always arbitrary. As a very simple example, one can make the European spread of the disease appear quite differently by just playing around with different zoom levels:
Don’t get me wrong, I am not trying to downplay the crisis. The point rather being that we indeed are in the midst of a truly challenging balancing act: a huge communicative predicament.
Tell me that story where all will be well
As Swedish epidemiologist and science communicator @EmmaFrans has put it, we must somehow manage to not scare people too much (especially those in risk groups, or those suffering from health anxiety 🙋♀️), but still make everybody understand the seriousness of the situation. This means that we may sometimes have to use certain sets of images and narratives, and sometimes others, depending on what story we want to tell, with what purpose, to whom, and when.
The Swedish Public Health Agency (and others globally), when making recommendations about shutdowns and other strategies, have repeatedly stated that we must implement the right measures and sanctions at the appropriate stage of the pandemic. I guess then, that we accordingly are faced with a situation where we sometimes need to tell some stories (“Relax. We are not all going to die. This shall pass.”), and sometimes others (“This is one of the biggest crises in modern times, so please take it seriously”). Because they are both true, and they are needed at different points in time.
The warm embrace of what once may have been
Interestingly, huge events such as the covid-19 crisis may mark a kind of ‘revenge’ for old-media. This is because TV, radio, and newspapers gain from these dramas being so conducive to Public Service broadcasting, to the trusty old proven channels, and to the Great Voices of the Nation. People (myself included) seem to, at least partly, retrogress back into getting their information from where they did when they were last in any similar situation, which was probably some decades ago for many of us well-protected and comfortable Westerners.
Similarly, in this age of distrust and the broad proliferation of ‘anti-establishment’ ideas, what is happening now may also mark a sort of ‘revenge’ for the ‘establishment’. Dramatic times call for dramatic measures, and maybe one such measure is to retreat back into what may feel as the warm embrace of the experts, our elected officials, or even the Nation.
What sticks?
So, what is a digital sociologist in self-imposed home office isolation to do with all that is going on? Naturally, I am continuously streaming for tweets though Twitter’s Public API, and will most likely able to make quite interesting analyses of those data, once the relevant research questions begin to appear.
I begin here with a modest analysis of the proportional relationships between original tweets (tweets crafted by a user; machine or human) and retweets (tweets simply echoed in their original form by a user other than that who originated it). We can see this as a snapshot over the last two days of which aspects, feelings or topics that relate to the virus, get acknowledged to larger or lesser degrees. How does the virus go viral? Which images and narratives seem to stick and take off ?
When tweets about the disease as such are posted (the four pink bars at the far left of the graph), they are generally met with a high level of acknowledgement. This is illlustrated by the high level of the red curve above these bars. As an example: In the category ‘health’, 30% are original tweets, and 70% are retweets of those original tweets.
Naturally (and this goes for the entire analysis), we don’t know whether we should interpret any such acknowledgement as ‘amazement’, ‘panic’, ‘ridicule’, a ‘neutral’ sharing of information, or something else. What we do know, is that a large number of users (N for the whole thing is ~2 million) deemed the tweet worthy of being echoed forward, uncommented (because the analysis does not include commented retweets or so-called quoted tweets) to their list of followers, and to the world (as 88% of Twitter accounts are public, and only such accounts are included in this analysis).
[Conclusion 1] Tweets about the disease as such, and especially those about deaths get a strong response, and much attention, which seems reasonable.
When covid-19 related tweets that have a prominent emotional element to them are posted, we can observe two interesting patterns.
[Conclusion 2] Tweets that have a strong affective element, especially if they are ‘angry’ and use strong language, get a significant boost by being retweeted by others. This would indicate that the twittersphere is fuelling such flames. Seventy percent of these feelings are represented by the mere amplification of them through others echoing them further.
[Conclusion 3] It is interesting — especially in relation to conclusion 2 — that the ‘softer’, if we will, emotional responses, through posting about general negative emotions, and prominently ‘anxiety’ and ‘sadness’, fall a bit flatter to the ground than do the more aggressive emotional responses. The ‘sadness’ category, for example, consists of more original content than retweets.
A preliminary observation here may be that social media audiences have a stronger tendency to encourage 😡 than 😢 responses, even if both are put out there. This might, pending further analyses of course, contribute to hindering crucial processes of collective coping and mutual emotional support in trying times. The darker, more open-hearted, feelings of distress may be more difficult for the twittersphere to deal with.
Moving on, tweets about close social relations get strongly acknowledged, as evidenced by conclusion 4:
[Conclusion 4] Tweets about friends and family in these times of #StayAtHome and #SocialDistancing get a large number of retweets, which might be interpreted as a form of recognition and identification with others and with the shared situation.
Finally, looking at the right end of the graph, at the responses received by posts in more factual, analytical and strategic categories, it is notable that covid-19 related tweets about numbers, risks, future strategies, analyses, causes and certainties are not something that particularly stick on social media. The tweets are put out there, but retweet proportions are very modest. The only exception here is that when tweets get more political and talk in terms of ‘power’ (which admittedly, of course also may be about having the ‘power’ to beat the virus), the response is stronger.
[Conclusion 5] Tweets that are about facts, causes, and routes of action in relation to the virus are not a main driver of retweeting activity, and one might hypothesise that this has to do with such issues being so abundant in mainstream news media, that the twittersphere as a consequence does not prioritise dealing with them.
Note that all of these analyses are very, very preliminary.
Concluding words
A colleague of mine aptly pointed out that the covid-19 crisis is indeed “a metrics crisis for our metrics time”. We can be sure that while the virus hopefully, eventually, subsides, there will be an explosion of digital social research that deals with the infodemic accompanying the pandemic. Some of those studies will of course be about disinformation, some about social support and collaboration in crisis management at both individual, group and societal levels, yet others will be about how knowledge and information about the virus was constructed and disseminated. Other possible topics abound.
As the director of DIGSUM —the Centre for Digital Social Research at Umeå University, Sweden, I envision lots of future, and urgent, projects about social, legal, technological, pedagogical, economical, informational, and other aspects of how this virus has struck society.
