Migration narratives and media framing: an input–processing–response perspective of media–audience associations

Abstract This study investigates associations between media narratives and public discourse on immigration in Italy using an input--processing--response framework in which media outlet posts are the inputs, audience cognition and interaction constitute processing, and social media comments are the responses. We combine Structural Topic Modeling, Fightin' Words, Regularized Association Analysis, and sentiment/emotion measures to examine topic transmission and keyword alignment between outlet posts and comments. We document systematic variation in framing across outlets—ranging from humanitarian to security emphases—and observe that comment threads display broader thematic dispersion than outlet posts. Regularized models indicate that specific post topics (e.g., humanitarian and migration management) are associated with distinct comment–topic mixtures and shifts in lexical salience. Sentiment and hate indicators vary by outlet and topic; hostile and extreme content is more prevalent around humanitarian and migration-management themes. Taken together, these results characterize how media inputs relate to audience responses in this observational setting and provide evidence relevant to media regulation, platform governance, and media literacy.