Assessment of Personality Dimensions Across Situations Using Conversational Speech

Abstract—Prior research indicates that users prefer assistive technologies whose personalities align with their own. This has sparked interest in automatic personality perception (APP), which aims to predict an individual’s perceived personality traits. Previous studies in APP have treated personalities as static traits, independent of context. However, perceived personalities can vary by context and situation as shown in psychological research. In this study, we investigate the relationship between conversational speech and perceived personality for participants engaged in two work situations (a neutral interview and a stressful client interaction). Our key findings are: 1) perceived personalities differ significantly across interactions, 2) loudness, sound level, and spectral flux features are indicative of perceived extraversion, agreeableness, conscientiousness, and openness in neutral interactions, while neuroticism correlates with these features in stressful contexts, 3) handcrafted acoustic features and non-verbal features outperform speaker embeddings in inference of perceived personality, and 4) stressful interactions are more predictive of neuroticism, aligning with existing psychological research.

Introduction. I. INTRODUCTION U NDERSTANDING and modeling human personality is fundamental to the development of affective computing systems capable of personalized, adaptive, and socially intelligent interactions. Recent user studies have found that users are more engaged with and have greater trust in assistive technologies that reflect or adapt to their own personalities, thereby demonstrating the value of personality-aware systems [1], [2], [3]. Consequently, there is a growing field of research on automatic personality perception (APP), the task of inferring one’s personality as perceived by external judges. Unlike automatic personality recognition (APR), which focuses on inferring selfreported personality traits, APP captures perceived personality and better reflects the cues that affective computing systems are designed to interpret. Describing an individual’s personality is a complex task.

Discussion / Conclusion. tionships between speech features and perceived personality dimensions. However, the relationships do not generalize well across different conversation scenarios. These results showcase the importance of developing affective computing systems that are adaptable to varying contexts, as a system developed for one context may not perform well in a different context. We investigated the relationship between perceived personality and conversational speech using the UbImpressed dataset, which contains audio of the same participants in two distint conversation scenarios. Our experiments showed that perceived personality differs significantly for the same participants across a neutral and stressful interaction. We found that features related to loudness, equivalent sound level, and spectral flux are correlated with perceived extraversion, agreement, conscientiousness, and openness in the neutral scenario (interview interaction) and with perceived neuroticism in the stressful scenario (desk interaction).