Gender stereotypes drive perceptual differences of vocal confidence

Abstract

One's ability to express confidence is critical to achieve one's goals in a social context—such as commanding respect from others, establishing higher social status, and persuading others. How individuals perceive confidence may be shaped by the socio-indexical cues produced by the speaker. In the current production/perception study, we asked four speakers (two cisgender women/men) to answer trivia questions under three speaking contexts: natural, overconfident, and underconfident (i.e., lack of confidence). An evaluation of the speakers' acoustics indicated that the speakers significantly varied their acoustic cues as a function of speaking context and that the women and men had significantly different acoustic cues. The speakers' answers to the trivia questions in the three contexts (natural, overconfident, underconfident) were then presented to listeners ( N = 26) in a social judgment task using a computer mouse-tracking paradigm. Listeners were sensitive to the speakers' acoustic modulations of confidence and differentially interpreted these cues based on the perceived gender of the speaker, thereby impacting listeners' cognition and social decision making. We consider, then, how listeners' social judgments about confidence were impacted by gender stereotypes about women and men from social, heuristic-based processes.

Gender stereotypes and social perception of vocal confidence is mitigated by salience of socio-indexical cues to gender

“…Once these initial edits to the original sound files were completed, we then adjusted the intonation contour for each talker's productions to include additional categories of confident (declining intonation) or not confident (rising intonation) (based on Roche et al, 2022 ) using Praat. To create these pitch contours, the talkers' flat/neutral pitch contour (i.e., based on their f 0 ) was used halved (+/– the f 0 ) at the last pitch point in the natural utterance.…”

“…Additionally, listeners were presented with the yeah and no files manipulated by the intonational contour (flat intonation, rising intonation, and declining intonation). Based on results from other studies, it was assumed that sound files with rising intonation would be perceived as less confident, while sound files with declining intonation would be perceived as more confident (Jiang and Pell, 2015 ; Roche et al, 2019 , 2022 ). We had no a priori assumptions about the natural/flat productions and included them mainly to increase the number of trials so that listeners would not guess the purpose of the task (i.e., consistent with the matched guise technique, Ball and Giles, 1982 ), and these trials acted as fillers.…”

“…Warren and Britain ( 2000 ) suggested that rising intonation may also be used as a way to back-channel or a way to check in with a communication partner to assess understanding. We may also use rising intonation to communicate uncertainty (Ward and Hirschberg, 1985 ) and/or a lack of confidence (Roche et al, 2019 , 2022 ). Roche et al ( 2022 ) found that both women and men produce rising intonation to convey a lack of confidence and declining intonation to convey increased confidence.…”

“…We may also use rising intonation to communicate uncertainty (Ward and Hirschberg, 1985 ) and/or a lack of confidence (Roche et al, 2019 , 2022 ). Roche et al ( 2022 ) found that both women and men produce rising intonation to convey a lack of confidence and declining intonation to convey increased confidence. However, women talkers tended to be perceived as less confident overall, especially when they used the rising intonation contour.…”

“…However, when many women and many men produce similar language and vocal cues suggesting a lack of confidence, then listeners may equally judge women's and men's confidence because there is no cognitive benefit to tracking separate cue distributions between men and women (Kleinschmidt, 2019 ). Language cues may also impact perceptions of confidence, such that negation would likely recruit higher ratings of confidence because affirmations are more likely to be a woman-typed cue that is associated with agreeableness (i.e., Roberts and Norris, 2016 ) and because women tend to be rated as less confident than men (Fisk and Ridgeway, 2018 ; Roche et al, 2022 ). To test this, we evaluated the interaction between language (i.e., affirmation and negation) and socio-indexical cues to affective expressions of confidence and speaker gender cues in single and multi-talker listening conditions in a social judgment task.…”