Quantifying dominance in the Multisimo corpus

Abstract

This paper is concerned with investigating conversational (verbal and non verbal) features related to conversational dominance and discovering evident relations between these features, personality traits and perceived dominance scores. Ordinal regression models were applied to find associations between dominance scores, both as a predicted value and as a predictor, and personality traits and verbal features. Results report the association of high dominance scores to the large number of words a speaker utters per minute, as well as to high extraversion and high openness traits. Low dominance scores have been associated with high agreeableness.