Behavioral Indicators of Dominance in an Adversarial Group Negotiation Game

• Published in: Detecting Trust and Deception in Group Interaction pp. 99 - 122
• Main Authors: Pentland, Steven J., Spitzley, Lee, Chen, Xunyu, (Rebecca) Wang, Xinran, Burgoon, Judee K., Nunamaker, Jay F.
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing 2021
• Series: Terrorism, Security, and Computation
• Subjects: Automated feature extraction; Deception; Dominance; Language; Voice
• ISBN: 9783030543822; 303054382X
• ISSN: 2197-8778; 2197-8786
• DOI: 10.1007/978-3-030-54383-9_6

Adversarial group negotiations often involve contentious strategies such as deception and dominance. Understanding characteristics of language and voice associated with deception and dominance helps negotiators identify the use of these strategies and achieve higher self-interests in various use cases, including business negotiations and law enforcement. We aim to expand traditional research on these characteristics from dyadic interactions to group communication. This study follows the same group experiment, a modified Mafia game, as described in the other chapters. Linguistic and vocalic features were extracted from the recorded audio files, and several models were built to examine the predictors of perceived dominance and deceptive cues. Our results show that some features are significantly correlated with dominance as suggested by the existing research, such as lower fundamental frequency, greater variability in loudness, higher voice quality, longer turn-at-talk duration, larger dominance ratio and a greater number of words. However, some features turn out not to be significantly correlated with dominance, such as mean level of loudness, polarity of emotions, hedging ratio and disfluency ratio, though the existing research predicts these relationships. Among our explored features, only turn-at-talk duration is significantly correlated with deceptive status. Our research shows preliminary evidence for the similarities and differences of signals of dominance and deception in dyads and groups and has implications for negotiators in real life.