People often predict how others might speak based on the speaker’s traits, such as gender. However, most research into the relationship between gender and listener perception assumes a gender binary, which fails to capture the full spectrum of gender diversity.

Merritt et al. measured listener perception of speakers with a range of gender identities to investigate how listeners organize speakers and the acoustic features that inform them.

“Studies on gender attribution generally limited listeners’ response options to categories like ‘man’, ‘woman’, or ‘don’t know,’” said author Brandon Merritt. “These decisions may have obscured listeners’ true attribution of talker characteristics. We wanted to allow listeners to make their own decisions in classifying speakers with minimal bias from words/terminology imposed by the researchers.”

The team recruited cisgender, transgender, non-binary, and agender speakers to record a sentence. Listeners grouped anonymous recordings of the talkers based on “general similarity” and “gender identity,” and rated them on auditory perceptual characteristics like masculinity or age.

Data analysis indicated fundamental frequency, or pitch, and formant frequencies, associated with voice quality, informed classifications, consistent with prior research. However, this study revealed a more multidimensional story of these features.

“In one dimension, listeners tended to organize speakers from low to high frequencies that were strongly correlated with auditory-perceptual ratings of masculinity/femininity,” said Merritt. “For the second dimension, we found that listeners tended to separate speakers with fundamental frequency and formant frequencies at the extreme ends of the scale away from speaker values in the more intermediate range. We interpreted this dimension as ‘Gender Prototypicality.’”

This research can help speech-language pathologists offer better gender-affirming communication training for transgender patients.

Source: “Auditory free classification of gender diverse speakers,” by Brandon Merritt, Tessa Bent, Rowan Kilgore, and Cameron Eads, Journal of the Acoustical Society of America (2024). The article can be accessed at