Emotional Dialogue Predicts the Success of Prestige Television Episodes

Whether a narrative succeeds in satisfying its audience is of paramount importance, but predicting this success is difficult. Although previous research has found that the textual characteristics of novels can predict how they are received, what about multimodal narratives like film and television? In a television show, the only text available is the dialogue, with so much of the story conveyed non-verbally (e.g., facial expressions and actions, cinematography, editing, music). Although dialogue is just one aspect of a television episode, might the emotional nature of the dialogue still be able to explain some of an episode's appeal? We gathered all of the dialogue for 45 prestige television series that vary widely in terms of genre, setting, themes, and characters. The ability for the dialogue to explain the success of these episodes was then examined within a series, to hold constant many other factors (e.g., writers, producers, characters, themes, marketing), using bidirectional stepwise linear regression models. The emotional nature of the dialogue spoken within an episode managed to successfully predict audience ratings of these episodes within a series. Moreover, these models proved to be robust and able to predict data "out of sample." This result demonstrates that the emotional characteristics of dialogue alone can help explain the audience appeal of multimodal narratives, highlighting the importance of language even in the context of rich audiovisual information. Our findings have implications for those tasked with crafting compelling television stories.