Nudges Ineffective Against Online Hate

Seven nudges aiming to reduce hateful speech online all failed—but the nudges unexpectedly succeeded in increasing engagement with harmless and wholesome content. Controlling hate speech is an ongoing challenge for online communities. In a pre-registered experiment, Tatiana Celadin and colleagues compared the effects of seven "nudges," messages designed to promote prosocial behaviors: reminding posters of descriptive norms, injunctive norms, or personal norms; cooling down negative emotions; stimulating deliberation or empathy; and highlighting reputation. Over 4,000 Americans recruited through the online platform Prolific were asked to interact with a duplicate of Facebook's newsfeed page. Each participant was shown one of the nudges or, in the control group, no message at all before engaging with the website, which featured 14 posts. The posts were of varying degrees of harmfulness, as measured by study participants who were asked to assess how "abusive" and "hateful" each post was. Unexpectedly, a major effect of most of the nudges was to increase engagement with harmless posts, as compared to controls. Engagement with extremely harmful posts did not significantly differ between the groups receiving nudges and the control group. Also, none of the nudges reduced the spread of harmful content. According to the authors, nudges may be worth pursuing even if nudges only increase engagement with harmless content as an increase of harmless content can dilute and drown out the harmful content.