Evaluation of poetic creativity: Predictors and the role of expertise—A multilevel approach

Poetry is one of the most creative expressions of language, but how we evaluate the creativity of a poem is not properly characterized. The present study investigated the role of various subjective qualities – clarity, aesthetic appeal, felt valence, arousal, and surprise – in predicting the creativity judgment of English poems. Participants (N=129) were presented with a broad range of English poems; they rated each poem on six characteristics: clarity, aesthetic appeal, felt valence, felt arousal, surprise and overall creativity. Linear multilevel analysis showed that aesthetic appeal was the strongest predictor of poetic creativity, followed by surprise and felt valence. Multilevel mediation analysis indicated significant mediation by surprise and felt valence on the relationship between aesthetic appeal and creativity at both within and between-participant levels. Further, expertise in English literature was found to significantly moderate the effects of all three predictors on the evaluation of creativity. The study simultaneously captured the surprise-evoking line(s).
Using the semantic distance computing approach, we have shown the objective validation of the subjectively chosen line(s) of surprise. Altogether, our findings suggest a parsimonious model of evaluation of creativity of poems and its interaction with expertise.