Social media – Facebook, Twitter, etc., – are a challenge for science. Often considered frivolous or at least not serious, they are tempting to ignore. But it seems that many anti-scientific or pseudoscientific ideas a promoted by social media. What’s more, these new social media are very effective at promoting messages, especially in areas of social health, so we ignore the media at our peril. Social media are a fact of modern life and if we can’t beat them, perhaps we should use them ourselves. These are the messages I got from a recent study of the way public health misinformation is promoted via social media. The paper is:
Seymour, B., Getman, R., Saraf, A., Zhang, L. H., & Kalenderian, E. (2015). When Advocacy Obscures Accuracy Online: Digital Pandemics of Public Health Misinformation Through an Antifluoride Case Study. American Journal of Public Health, 105(3), 517–523.
Although limited to an anti-fluoride Facebook case study I believe that the findings are also applicable to other anti-scientific movements like anti-vaccination and climate change denial. Here are some of the findings in the paper.
Anti-fluoride Facebook groups are highly connected
The connectedness within and between anti-fluoride networks was measured to get an idea of their social influence. Facebook “friend” connections between members within groups and between groups was very common and overall the study sample showed a significantly higher degree of connection than in Facebook overall. This is probably the result of group members having similar ideology or other traits leading to a desire to form social connections. Social influences lead to group shaping of individual attitudes and behaviours related to health and not reliance on expert findings.
Social diffusion of scientific misinformation
Using an original source article the authors followed posts about the article to determine the degree of social diffusion. They found 60% of the time a reader would have to follow 2 to 3 links to arrive at the original source with such a search failing 12% of the time. This demonstrated that:
“on average, there was a high risk that antifluoride Facebook group members engaged in posts about the article would be forced to navigate through multiple pages to locate the original post or would never succeed in locating it at all, greatly increasing the likelihood for the spread of misinformation and misrepresentation of the scientific article’s content.”
Although not detailed in the paper I believe that key sources of primary information for such social networks is very often articles in the “natural” health or alternative health magazines and websites which have already misinterpreted and misreported the scientific literature.
Engagement and sentiment
The levels of engagement and sentiment of the most influential posts were analysed:
“to explore the user experience with social diffusion of information and to determine how these posts could potentially influence group member attitudes and behaviors. Considering the distance between reference posts and the original source, this is particularly important.”
The figure below illustrates their finding for the 2 most influential posts in the social diffusion map. Interestingly:
“The most frequent type of comment about the posts and the type of comment that received the most engagement were the science-based comments . . . Science-based comments overall received a positive (profluoride) total sentiment score of 36 (additive over all comments), whereas all other types of comments received an overall negative (antifluoride) sentiment score of –47.”
“These results demonstrate that the user experience, when engaging with these influential posts, is just as likely to be negative and irrelevant to the original source as it is to be positive and reference scientific information (accurate or not). Our results demonstrate a high probability (1 in 2 chance) of encountering negative and non-science-based information about fluoride that is unrelated to the original peer-reviewed scientific publication under discussion.”
Social media overriding traditional health communication?
The new social media have given a new power to social networks:
“Historically, naturally limiting factors such as geography and communication barriers inhibited opportunities for strengthening networks with outlying views. Risky behaviors as a result of shared moral evaluations, such as opting out of recommended childhood vaccination schedules and rejecting fluoridation, reverberated in existing small networks without necessarily scaling to dangerous magnitudes.”
“online social networking allows greater connectivity among networks through the increased visibility of group behavior; previously nonnormative behaviors can thus become normative through the use of social media.”
Now that minority ideas and behaviours have an online market they can spread through confirmation bias and the expanding social networks, and the strong ties in the networks reinforce the sentiments behind the spreading messages.
The group thinking and the “silo,” even protective, nature of social networks means that:
“Expert opinion grounded in evidence that contradicts the sentiments embedded in a socially diffused message will be quickly rejected; acceptance of this contradictory information would be socially detrimental to the network, challenging its very identity. Thus marks the beginning of digital pandemics of misguided and incomplete health information in which evidence becomes entirely secondary to the sociology of the networks diffusing it.”
Are there lessons here for pro-science groups?
The findings of this research will strike a chord with sceptics and others who fight anti-evidence-based approaches to health online. They will easily recognise the social norms of such groups which leads to rejection of any attempt to inject a pro-science discussion. They will have experienced the put-downs and banning or blocking used to reinforce and protect those within-group social norms.
But should we be drawing lessons from this? The authors conclude:
“Traditional vertical health communication strategies, such as broadcast diffusion through peer review publication and media reporting, may no longer be effective because of the existence and viral potential of social diffusion”
Perhaps we have to accept that:
“the sociology of networks is perhaps just as influential as, if not more influential than, the information content and scientific validity of a particular health topic discussed within and between certain networks via social media.”
Rather than trying to beat this new technology perhaps we should try to use it ourselves more effectively. The authors suggest:
“Empirical social strategies for health communication should focus not only on high-quality digital information production and dissemination but also on socially targeted and custom-designed messaging that conforms to the norms and values of specific target networks rather than challenging them. Developing an appreciation for the sociology of target groups could assist public health experts in increasing influence in problem networks and could provide the tools to predict, prevent, or reverse digital pandemics.”
The internet has forced changes on the traditional methods of information dissemination in other areas so it is really not surprising to expect it to change scientific and public health communication. As the authors say in their final words:
“The public nature of social media is at once a barrier to accurate information flow online and a tremendous opportunity for public health research, innovation, and intervention. In an age when negative digital pandemics can go viral, public health communication management strategies must go social.”