Measuring satirical uptake using word sentiment

Abstract Satire is a discursive strategy used to convey criticism, yet such criticism is indirect and subject to individual interpretation. There is thus a great deal of variability in the comprehension of satirical meaning, which raises unique methodological challenges for researchers studying the cognitive processing and comprehension of satirical discourse. Because existing behavioral research suggests research participants are reluctant or unable to restate the satirical meaning of a text, other means of assessing satire comprehension are needed. In this short research note, we report on an analysis assessing whether computationally-derived valence of participants’ one-word descriptions of minimally-different satirical and non-satirical news texts index awareness of satirical intentions (as assessed by perceptions of authorial sincerity). Results from multilevel ordinal regression models indicate negative valence is significantly associated with perceptions of authorial insincerity, highlighting a clear correlation between the way texts are described and how they are interpreted. Crucially, this effect was found for both satirical and non-satirical texts, which may be an artifact of the method used to create non-satirical versions of the satirical news texts. Regardless, these results provide further empirical evidence for theoretical models of satirical comprehension and also give future researchers an additional strategy to quickly tap into inferential processes associated with satirical discourse.