How Frames Diffuse to Create Media Storms: Relations Between Legacy, Social and Web Media in the Hybrid Media System

In the current age of a hybrid media system (Chadwick 2011), the circle of actors exercising influence over the media, public, and political discourse has dramatically grown. In older media logics, a few (media) elites heavily shaped how information was portrayed, which meant that interpretations of events usually trickled down from top to bottom (Entman 2003). In today’s hybrid media system, there are many more actors who may come up with, promote and disseminate their interpretations of events and issues. At first sight, this makes the process of how events and issues are framed in public discourse more hectic and confusing, as different seemingly coherent interpretations spring up in different online and offline public spaces, which make up our fractured public sphere (Bruns 2023). Yet, the process of frames being created, contested and spread is also no longer done by a small number of editors, but negotiated openly in interlinked public venues. Some studies have already traced how selected issues and topics spread between legacy and social media (e.g., Gilardi et al. 2022; Langer and Gruber 2021). However, a study that assesses the specific content, where it originates and how it diffuses is so far missing. We use data that consists of traditional (newspaper articles and TV transcripts) and new types of news media (digital native, alternative, and social) about the discourse in the United Kingdom about the "Windrush scandal". This appears an ideal case, as Langer and Gruber (2021) have already postulated that this high profile immigration scandal was pushed into the agenda by an assemblage of different actors. We combine state-of-the-art computational methods for text and inferential network analysis to trace where different ways to frame the scandal originate, as well as where and when they spread among actors. We first use the technique by Gruber (2023) to identify and the code the most prominent frames in the corpus of Twitter posts, online news and in legacy outlets. To assess when and in which direction emerging frames central to the overall discussion transition between actors, and importantly between actors in different spaces, we employ relational event models (REM) (Butts 2008). This technique, used in longitudinal network analysis to pinpoint how events influence each other, can be used to measure how a public statement was influenced by previous contributions to the discussion. With this approach, we answer three research questions: 1. Which frames are most prevalent within the discourse on the Windrush scandal over time, between media types and between actor types (politicians, crow-sourced elites, general users and journalists) 2. What are the dynamics of the diffusion processes between legacy and social media? 3. What drives diffusion of arguments within and between media types?