The objective of WP1 is to develop a new, empirically grounded conception of pro-social bystander reactions to online political hostility that transcends crude dichotomies and ensures the highest possible ecological validity and real-life relevance. 

To achieve this aim, WP1 sets up a novel data stream containing first-hand experiences of pro-social bystander reactions to online political hostility. This data stream will triangulate data sources, including political conversations scraped from Twitter and Facebook and interviews with online bystanders who reacted pro-socially to political hostility. Computational methods will be used to identify pro-social bystander reactions and map their prevalence on different social media platforms.

Based on the ‘Consequential Model’ of pro-social bystander reactions, WP2 will formulate and test specific hypotheses about the core drivers of bystander decision-making and pro-social reactions accordingly.

To test the causes of pro-social bystander reactions, the project employs a novel combination of survey and behavioral experiments. This includes population representative panel and factorial survey experiments, using a mock social media platform to investigate behavioral outcomes. The main case countries of the project are Denmark, the US, UK and Germany.

In this context, the project will develop and implement a new ‘Mock Social Media Bystander Platform’ for doing behavioral experiments, which can be combined with large-scale, online surveys. In doing so, we increase ecological validity of our studies, as interaction with the mock platform resembles actual social media sessions, and enables tracking actual behaviors (e.g. likes, comments, reports), while having full control over what content participants are exposed to.

WP3 builds on and extends the experiments conducted in WP2 to test short- and long-term consequences of exposure to pro-social bystander reactions. Specific hypotheses regarding the potential short-term cumulative effect, effect on metaperceptions of outpartisans and long-term civilizing effect of pro-social bystander reactions will be formulated and tested. One hypothesis regards the effect of exposure to cross-partisan, pro-social bystander reactions as a way of mitigating affective polarization.

By randomly exposing participants to stimuli containing/not containing pro-social reactions from e.g. out-partisans, we will measure the effect on subsequent bystander behavioral intentions and perceptions of outpartisans. By embedding some of these experiments in national representative panel surveys, the project aims to estimate causal effects of pro-social bystander reactions across time.

To further explore the potential of pro-social bystander reactions in mitigating the negative impact of online political hostility on society, WP4 tests whether and how pro-social bystander reactions can be encouraged. WP4 aims to establish proof-of-concept of the effectiveness of interventions encouraging pro-social bystander behavior online.

WP4 will test extant bystander campaign material in new contexts (i.e., regarding political hostility online) and develop innovative material to be tested. Informed by findings of WP2 and WP3, a sample of interventions will be tested that vary in terms of type of pro-social bystander reaction to be encouraged (e.g. reporting, counter-speech, supporting victim) or in terms of motivational cues. The effect of exposure to such interventions on bystander situational assessment and online behaviors will be estimated through population-based survey experiments, including cross-time panel experiments.