Transgender Listeners Show Reduced Visual Bias When Classifying Voices

Post by Anastasia Sares

The takeaway

While we usually draw on multiple senses and general predictions to inform our perception, this can sometimes backfire, introducing bias into our judgments. This study found that transgender and nonbinary people were less susceptible to visual bias, and better able to classify a person’s vocal range while watching videos of them singing or speaking. This trans advantage could come from more extensive experience with voice-body mismatch in daily life. 

What's the science?

The brain is constantly trying to fuse information from its different senses and make predictions based on that information. Unfortunately, this can sometimes lead to biases when we are only asked to judge based on one sense alone, or if we are confronted with something that doesn’t match our ingrained predictions. One example of this is the McGurk effect, where misleading visual information causes people to perceive a different syllable than the one they heard: a “da” played over audio combined with the visual of someone saying “ba” can result in people reporting that they heard “ga” instead. Visual bias is especially problematic for voice-body mismatches in the context of opera. A person’s body size and shape doesn’t necessarily indicate what range they can comfortably sing, but the stereotypes are strong and can (consciously or subconsciously) influence the roles that opera singers are cast in. This can affect their long-term vocal health and be detrimental to their careers.

Visual biases are not set in stone, however. There is some evidence that they can be mitigated through training, like musicians learning to resist the McGurk effect. One group of people who may have natural sensitivity to voice-body mismatches are the transgender and nonbinary communities, since voice is a strong gender cue and often a source of insecurity or fear of being outed.

Recently in Frontiers in Psychology, Marchand Knight and colleagues showed that, when asked to judge vocal ranges of different speakers, trans and nonbinary people are more resistant to visual biases than their cis peers, making their judgments more accurate.

How did they do it?

The authors conducted an online experiment including a cis group of participants as well as a trans group, which was composed of a mix of trans and nonbinary identities. Participants started by learning about different voice categories used in opera (from low/dark to high/bright: bass, baritone, tenor, alto, mezzo, soprano) and next used this voice-typing scale to rate clips of people speaking and singing. Participants first got the audio-only (no video) versions of the clips, then the visual-only versions (guessing voice type purely based on looks), and finally the full clips with both video and audio. The researchers intentionally chose some actors who they thought might show stronger voice-body mismatches to better measure the effect of visual bias.

What did they find?

Participants were fairly successful at classifying voice type based on hearing the voices in the audio-only condition, but in the visual-only condition they tended to revert to a gender binary (rating videos of female- presenting people around the “mezzo” voice range and male-presenting people around the “baritone” range). The highest and lowest voice types had the most discrepancy between their audio and visual ratings.

When audio and visual were presented together, ratings fell somewhere in between the two previous conditions, showing that the visual information was influencing participants even though they had been asked to rate solely based on the audio. However, the trans participants were better at resisting the visual biases, so their ratings in the audiovisual condition were closer to the audio-only condition. Cis participants’ ratings were pulled more toward the visual information, 30% more so than trans participants. This difference in ability did not seem to be strongly related to demographic differences between the groups or to gender views in general, as far as the researchers could measure. 

What's the impact?

These findings highlight a strength of the trans and nonbinary community, in a time when most research is focusing on the disadvantages they suffer. It also brings up a crucial issue that can affect the vocal health of opera singers, and calls for it to be addressed.

Access the original scientific publication here. 

[Disclosure: The writer of this BrainPost summary is also a collaborator on the publication]