“Social cognition of voices: Perception of gendered cues in vocal communication“
Kelsey Neuenswander, University of California, Los Angeles
Abstract
Despite the important role of voices in social interactions, they are understudied in social cognition. Voices communicate similar information to faces and bodies, but have been largely overlooked. This dissertation addresses this in two ways. First, it offers a comprehensive review of social cognitive research on voices, identifying key gaps in the literature. Second, it advances empirical research focused on the perception of gendered vocal cues. The first set of studies examines voice-based evaluations of men and women, focusing on how recent perceptual exposure influences social evaluations of individual voices. The next set of studies explores group-level perception and tests how groups are evaluated based solely on the voices that compose them, including whether listeners are attuned to the sex ratio within groups, and whether group composition influences judgments of threat and belonging. In real-world groups, not only who is present affects perceptions, but also how much individuals participate. The final set of studies investigates perceptions of speaking time in groups, including examining whether listeners accurately estimate who speaks and for how long, and whether systematic misperceptions of speaking time give rise to an illusion of equity. Implications of these discovered biases for the reinforcement of social inequalities are discussed.
Project Summary
Background
Kelsey’s dissertation begins from a striking and important observation: although voices are central to human social life, they have been comparatively neglected within social cognition, a field that has historically focused on visual information such as faces and bodies. Her work directly addresses this imbalance by advancing a unified framework for understanding how vocal cues inform social perception. In doing so, she not only identifies a critical gap in the literature, but also provides a programmatic and compelling set of studies that establish voice perception as a fundamental domain of social cognition in its own right.
The dissertation is conceptually elegant and empirically rigorous, integrating theory and method across cognitive, social, and perceptual science. It opens with a comprehensive and deeply impressive review of the literature (Chapter 1), in which Kelsey maps existing work on voice perception onto core social cognitive mechanisms, including attention, categorization, evaluation, and memory. This approach does more than summarize prior research—it provides a novel conceptual framework for organizing a previously fragmented literature and, in doing so, identifies critical gaps and opportunities for future work.
Building on this theoretical foundation, Kelsey develops a series of empirical chapters that together articulate how listeners extract and use gendered information from voices across multiple levels of analysis. Importantly, this work is notable not only for its theoretical contributions, but also for its methodological breadth and innovation. Across Chapters 2 and 3 in particular, Kelsey adapts and extends paradigms that have been used almost exclusively in visual cognition—such as sensory adaptation and ensemble coding—and applies them with precision to the auditory domain. These studies are deeply rigorous and exemplify a hallmark of especially strong experimental science: the ability to translate sophisticated methods across domains in ways that are both technically precise and conceptually generative.
Methods and Results
In Chapter 2, Kelsey reports a series of experiments demonstrating that social evaluations of individual voices are dynamically shaped by recent perceptual experience. Using a sensory adaptation paradigm, she shows that exposure to gendered vocal contexts systematically shifts not only the categorization of subsequent voices (e.g., perceived gender typicality), but also downstream social evaluations, including judgments of traits such as trustworthiness, warmth, and threat. Voices that are identical in their acoustic properties are perceived differently depending on prior auditory context, providing compelling evidence that even seemingly stable social evaluations from voices are inherently context-dependent. This work, published in Royal Society Open Science (Neuenswander, Gillespie, Lick, Bryant, & Johnson, 2024), represents a powerful extension of adaptation principles into the auditory domain.
In Chapter 3, Kelsey presents a multi-study investigation of vocal ensemble perception, introducing a novel application of ensemble coding to vocal stimuli. Across these studies, she demonstrates that listeners rapidly extract aggregate information from groups of voices and use this information to form social judgments about the group as a whole. In particular, the sex ratio of vocal ensembles systematically shapes perceptions of group-level traits, including threat and social belonging, with groups containing more male voices perceived as more threatening and groups containing more female voices perceived as more affiliative or welcoming. These findings show that complex social evaluations can be computed from distributed auditory input, even in the absence of individuating information about specific speakers. This work, published in Scientific Reports (Neuenswander, Goodale, Bryant, & Johnson, 2024), significantly expands the scope of ensemble perception research and demonstrates that group-level social impressions can arise from purely auditory cues.
Finally, in Chapter 4, Kelsey reports multiple studies that combine tightly controlled experimentation with ecologically valid designs to examine how listeners perceive participation within groups. In one set of studies, she develops a sophisticated experimental paradigm in which vocal ensembles are constructed from recorded speech and systematically manipulated. Using standardized recordings of individual speakers, she generates controlled group conversations in which the numerical composition of men and women and their relative speaking contributions are independently varied while holding other acoustic features constant. Critically, in one of these studies, she orthogonally manipulates numerical minority status and speaking time by assigning specific speakers to contribute either substantially more or substantially less than others within the group, allowing her to disentangle the effects of being numerically rare from the effects of speaking more or less. This level of control enables precise causal tests of how numerical and social salience jointly shape perception.
Importantly, in this final chapter, she extends the broad paradigm from Chapter 3 (i.e., ensemble perception) to more naturalistic, multi-speaker conversations, demonstrating that these perceptual biases generalize beyond tightly controlled experimental stimuli. Even as the conversational structure becomes more realistic, she maintains careful measurement and control by selecting and annotating recordings in which the speaking time of men and women can be precisely measured. This progression—from highly controlled synthetic ensembles to naturalistic interactions while preserving analytic rigor—provides especially compelling evidence for the robustness and real-world relevance of her findings. Across these studies, she demonstrates that the speaking time of numerical minorities is systematically overestimated, particularly when those individuals speak less. This bias is especially pronounced for women and reflects the combined influence of numerical and social salience. Critically, her findings show that these perceptual distortions produce what she terms an “illusion of equality”—when women are underrepresented, overestimation of their contributions leads observers to perceive group interactions as more equitable than they actually are.
Discussion
This work makes a particularly important contribution by identifying a novel mechanism through which inequality may persist—not only through unequal participation itself, but through biased perception of that participation. Even when disparities exist, they may go unrecognized because observers systematically misperceive them. Moreover, the work highlights a consequential asymmetry: overestimation of women’s participation inflates perceptions of fairness in ways that do not occur when men are in the minority, underscoring how perceptual biases interact with gendered expectations to shape social judgments.
Taken together, this dissertation makes a compelling and cohesive theoretical argument: voice perception is not a peripheral or secondary aspect of social cognition, but a central mechanism through which people form impressions, evaluate others, and navigate social life. Importantly, Kelsey does not simply identify this gap—she fills it with a rigorous and programmatic body of work that already demonstrates substantial scholarly impact, as evidenced by multiple peer-reviewed publications in leading interdisciplinary journals.
Impact Statement
The broader implications of this work extend well beyond basic research. By demonstrating how auditory cues shape judgments of threat, belonging, and participation, Kelsey’s research speaks directly to real-world contexts in which voices are primary sources of social information, including workplace meetings, educational environments, political deliberation, and virtual communication. Her findings suggest that individuals and organizations may systematically misperceive participation dynamics, thereby obscuring inequities in influence and contribution. This work provides a powerful perceptual account of how inequality can be maintained even in the presence of increasing representation, with clear implications for organizational practices, diversity and inclusion efforts, and the design of communication technologies.
All research described in this dissertation was conducted in accordance with institutional review board (IRB) approval at the University of California, Los Angeles.
