This project challenges the assumption that narrative fiction is valuable because it creates understanding of other people’s experiences and lived realities through empathy. Instead, the project addresses the main hypothesis that failed empathy is a strategically created effect of narrative fiction that activates affective, emotional, and cognitive processes by foregrounding differences between fictional characters and recipients. To date, popular perception frames failures of empathy also as failures of narrative fiction to engage recipients affectively, emotionally, and cognitively. In contrast, this project reframes failed empathy as an invitation to reflect on personal, cultural, and political beliefs, attitudes, and values. The innovation of REFEMP is twofold. First, the project develops a novel methodological approach for studying its multimodal dataset. This dataset consists of 1) a transmedia sample of narrative fiction in literature, graphic novels, films, television shows, and videogames released after the turn of the twenty-first century, and 2) a selected corpus of written online reviews of the primary texts as a shorthand for recipients’ responses to and empathic engagement with characters in narrative fiction. Second, the project is theoretically innovative by making the concept of failed empathy productive as an analytical framework for narrative fiction. The project's objectives are 1) providing a more refined vocabulary for talking about medium-specific possibilities of representation and narration that can facilitate, hinder, or block recipients’ empathic engagement with narrative fiction, and 2) theorizing the cultural effects of failed empathic engagement with narrative fiction based on recipients’ reviews. To this end, REFEMP highlights the value of moments of difference, subversion, or rupture – of being unable to feel with someone else – in narrative fiction to reflect on culturally shared beliefs and worldviews.