Date created: | Last Updated:

Category: Project

Description: This study tests multiple behavioral science interventions to reduce the toxicity of online communication – various simple changes in the user interface design (prompts) and an educative intervention targeting users’ competences (boost). Following the theoretical turn towards the role of empathy for democratic deliberation, this study focusses on affective empathy and cognitive perspective taking. After assessing participants’ attitudes towards different policy relevant topics that bear conflict potential within the public discourse in Germany and the US (Climate change, refugees, gender-neutral language and abortion), participants are allocated to opinion statements as they could typically be found on social media. By deliberately allocating participants to statements they likely disagree with, we aim to maximize conflict potential. Before participants comment on the statements, one group is prompted to take the perspective of the author and to think about possible reasons the author thinks this way (cognitive perspective taking) and one group is prompted to try to create a sense of empathy with the author and reflect on how the author might feel (affective empathy). The first control group does not receive a prompt, whereas the second control group receives a friction placebo, a prompt to take a deep breath before answering. Prior to reading the provocative posts as a form of educative boosting intervention, a fifth group is educated about the role of empathy and perspective taking for constructive discussions and the distinction between legitimate opinion and harmful content online. All participants are then asked to comment on the controversial statements. Those free text answers are analyzed using an established classifier to measure the toxicity of online comments. Furthermore, as a second outcome, the length of comments will be considered as approximate measure for argumentation. If the interventions prove effective, they could easily be implemented in the context of polarized discussions on social media to enhance the quality of the online public discourse while mitigating destructive escalation on controversial issues. In future iterations of the project, the design could be extended, for example, by using ingroup salience markers to improve the effectiveness of the empathy and perspective-taking interventions through mechanisms of social categorization.