Boehme, Bibi (2014) How trustworthy is your voice? The effects of voice manipulation on the perceived trustworthiness of novel speakers. University of Glasgow. (Unpublished)
|
Download (418kB) | Preview |
Abstract
A person’s voice is not only loaded with cues to age, sex and emotional state, listeners also readily form personality impressions of novel speakers. Based on research on face perception, suggesting that people rapidly and reliably evaluate faces on personality traits which can be summarized in a two-dimensional space, with one dimension emphasizing warmth/likability/trust and the other emphasizing strength/dominance, a similar model has been proposed for personality impressions from voices. The present study builds upon these findings, investigating how trustworthiness is perceived in voices and conveyed by novel speakers. For both male and female voices, morph continua were created between voices previously rated low/high on trustworthiness to examine whether the manipulation towards an averaged un-/trustworthy voice would shift perception of the voices towards un-/trustworthiness, respectively, and whether vocal caricatures of these prototypical voices would enhance the effect. Through an online rating experiment, 422 participants rated 18 voices on their trustworthiness. Akin to a ‘zero acquaintance’ scenario, the stimuli were sub-second vocal utterances of a single word and no contextual information was provided. Repeated measures ANOVAs and correlation analyses showed that, for male voices, there was a positive linear relationship between the responses and manipulations towards a trustworthiness caricature. For female voices, on the other hand, the results were inconclusive due to an error in the generation of the stimuli. The findings contribute to the empirical bases for understanding personality judgments from brief speech signals and are discussed in light of previous research, implications, potential applications, limitations and suggestions for further research.
Impact and Reach
Statistics
Additional statistics for this dataset are available via IRStats2.