Self-certification: A novel method for increasing sharing discernment on social media

Abstract

The proliferation of misinformation on social media platforms has given rise to growing demands for effective intervention strategies that increase sharing discernment (i.e. increase the difference in the probability of sharing true posts relative to the probability of sharing false posts). One suggested method is to encourage users to deliberate on the veracity of the information prior to sharing. However, this strategy is undermined by individuals' propensity to share posts they acknowledge as false. In our study, across three experiments, in a simulated social media environment, participants were shown social media posts and asked whether they wished to share them and, sometimes, whether they believed the posts to be truthful. We observe that requiring users to verify their belief in a news post's truthfulness before sharing it markedly curtails the dissemination of false information. Thus, requiring self-certification increased sharing discernment. Importantly, requiring self-certification didn't hinder users from sharing content they genuinely believed to be true because participants were allowed to share any posts that they indicated were true. We propose self-certification as a method that substantially curbs the spread of misleading content on social media without infringing upon the principle of free speech.

Fig 1. An example of a false social media post.

Note that in the above image the photographs was replaced by a copyright-free photograph depicting similar content https://www.rawpixel.com/image/8742382/photo-image-public-domain-people.

Fig 2. Results from Experiment 1.

Subplot A shows the perceived accuracy, defined as the proportion of participants who indicated that each news post was true, for the accuracy prompt (AP) and self-certification (SC) conditions. Subplot B shows the share rate for the baseline (BL), AP and SC conditions. Each dot shows the data averaged over a single social media post. The bars show the averages over all social media posts that were actually false (red) or actually true (green).

Fig 3. Results from Experiment 2.

The conditions are described in Table 1. Subplot A shows the perceived accuracy, defined as the proportion of participants who indicated that each news post was true. Subplot B shows the share rate. Each dot shows the data averaged over a single social media post. The bars show the averages over all social media posts that were actually false (red) or actually true (green).

Fig 4. Results from Experiment 3.

The conditions are described in Table 1. Each dot shows the data averaged over a single social media post. The bars show the averages over all social media posts that were actually false (red) or actually true (green). The four conditions were baseline (BL), accuracy prompt (AP), media literacy (ML) and crowd self-certification (CSC).