PCST Network

Public Communication of Science and Technology


Social Variables and Credibility Assessment of Scientific Content on Social Networking Sites

Arnon Hershkovitz  

Credibility assessment of scientific information is key to people's attitude towards, and acceptance of, science communication. Despite the central role social networking sites (SNSs) play in users' content consumption-scientific content included-the credibility assessment mechanisms utilized within these platforms are understudied. This paper reports on five pilot studies exploring "social variables" (e.g., number of Likes or comments on Facebook, or information poster identity) and how they affect people's credibility assessment of scientific information posted on SNSs. These studies are drawn from theories of online information assessment, specifically the Prominence-Interpretation theory and the Elaboration Likelihood Model. The five studies included both young and adult populations and used different methodologies, including questionnaires, eye-tracking data collection, and SNS data collection using NodeXL. In each group, participants were presented with scientific content, and were asked to rank its credibility and to explain their choice. The social variables appearing on these statuses were manipulated. In Studies 1-3 (N1=90 sixth-grade students, N2=91 fifth-grade students, N3=89 tenth-grade students), statuses were shown to two groups. For one group, the statuses had only few Likes, for the other - thousands of Likes. In study 4 (N4=19 undergraduate students), participants were shown a positively- and a negatively-designed statuses (differed by number of Likes/Comments/Shares, user name/profile image, etc.). In Study 5 (N5=88 undergraduate students), a status was shown to control and experiment groups, along with a fictitious figures of people who have favorited/retweeted it; in the experiment group, three of these figures were replaced with the participants' Twitter followees. In all cases, the positively/personally-designed statuses were statistically significantly associated with higher credibility assessment ranks, compared with the other statuses, except for the cases in which content was known to the participants. These results have important implications for both science communicators, consumers, and for broader understanding of credibility assessment within SNSs.

A copy of the full paper has not yet been submitted.