A new study published in Evolution and Human Behavior has found that while facial and bodily attractiveness tend to align in observers’ judgments, a person’s voice provides an independent cue about traits such as health and dominance. The research, led by Tobias L. Kordsmeyer of the University of Goettingen in Germany, tested how independent raters perceive attractiveness, health and physical dominance from faces, bodies and voices. The study involved 165 men and 155 women as targets and a panel of 400 raters assessing the stimuli.
The researchers frame their investigation around two competing theories. The first, known as the “one ornament hypothesis,” posits that different physical traits are shaped by the same underlying biological factors, meaning facial and bodily features could function as a single signal of health or genetic fitness. The second, the “multiple messages hypothesis,” argues that distinct body parts convey separate pieces of information, such as the face signalling immune system strength while the voice signals physical dominance. The study also explored whether sex hormones like testosterone and estradiol might synchronise development of these cues.
To conduct the experiment, participants underwent several data collection steps. Standardised facial photographs and three‑dimensional body scans were obtained, with the body models rendered in grey and stripped of skin colour and texture to focus raters on shape and proportion. For voice samples, male targets spoke for one minute about positive aspects of their lives, while female targets read a standard rainbow passage. Five‑second clips from the men and the women formed the basis of final evaluations.
A separate group of 400 independent raters evaluated the stimuli, divided so that no rater assessed more than one type of stimulus (for example, male faces or female bodies). Raters scored attractiveness, health and physical dominance using standardised scales, ensuring that judgments of one trait did not contaminate perceptions of another.
The study also collected biological data, including saliva and hair samples to measure hormone levels. In men, baseline testosterone and reactive testosterone—taken shortly after a competitive task—were recorded. For women, baseline testosterone and estradiol were measured across different phases of the menstrual cycle.
The findings showed positive correlations between how raters judged faces and bodies. Targets rated highly for attractiveness and health on their faces tended also to receive high body ratings for those same attributes. This pattern appeared as a medium-sized effect across both male and female targets.
When it came to physical dominance, a different picture emerged. The coordination between face and body ratings of dominance was observed only in male targets; men whose faces were judged as physically dominant also tended to have bodies judged as dominant. Female targets did not show a consistent face–body linkage for dominance.
These results support the one ornament hypothesis for facial and bodily cues, suggesting observers draw on redundant information about health and genetic fitness from both facial and bodily signals. In contrast, the lack of robust association between voices and visual cues points to the multiple messages hypothesis, with voice assessments offering a separate channel of information about an individual.
The hormone analyses yielded unexpected results. Adult hormone levels did not mediate the correlated perceptions of faces and bodies. While baseline testosterone in men showed some positive associations with facial dominance and vocal attractiveness, the hormones did not underlie the synchronized ratings across body parts.
Adjusting for factors such as body mass index or relationship status did not alter the overall pattern, indicating that the observed correlations were not simply artefacts of these variables. The researchers emphasise that they measured adult baseline and reactive hormone levels rather than developmental hormones, which are known to shape physical traits during puberty and adolescence. They caution that adult hormone levels may not reflect the hormonal milieu that originally moulded the body’s form and voice.
Another limitation noted by the team is that removing skin colour and texture from the body scans, while isolating shape, may have altered perceptions of health cues. The study thus assesses subjective impressions derived from form rather than objective health indicators, and the authors call for future research to include developmental hormones and additional sensory cues such as odour or dynamic movement.
The study is titled “Correlated perceptions and underlying hormonal factors of bodily, facial, and vocal attractiveness, health, and physical dominance” and was authored by Tobias L. Kordsmeyer, Philine A.C. Rasokat, Julia Stern, Christoph Schild and Lars Penke.
