What your voice tells people

People form impressions from a few hundred milliseconds of speech, and while those impressions are frequently inaccurate, they are consistent enough to be worth understanding.

What listeners judge

The findings, stated with appropriate caution.

Studies of voice perception find that listeners form rapid and highly consistent impressions of traits such as trustworthiness, competence and dominance from very brief samples.

The consistency is the striking part: different listeners agree with each other substantially.

The accuracy is another matter entirely, and is generally poor.

Which means these are reliable stereotypes rather than reliable perceptions — and they operate anyway.

The features involved

What drives the judgements.

Pitch: lower pitch is associated with perceived authority and competence in many studies, with the effect found for both men and women.

Pitch variation: more variation is associated with warmth and engagement; monotone with disinterest.

Speech rate: moderate to slightly fast is associated with competence; very fast with untrustworthiness; slow with either authority or hesitancy depending on other features.

Breathiness, creak and roughness, each of which carries associations.

And volume steadiness, where trailing away is read as uncertainty.

What this is worth doing about

The practical part.

Some of these features are habits rather than fixed characteristics, and habits can change.

Pitch variation is the most improvable and has the largest effect on engagement.

Speech rate is entirely controllable.

Volume steadiness through the end of a sentence is trainable.

And relaxed production — jaw released, throat open — lowers habitual pitch slightly without any deliberate lowering.

What not to do about it

The important limit.

Deliberately speaking at an artificially low pitch strains the voice and sounds artificial.

Adopting a voice that is not yours is detectable and is tiring to maintain.

Which means the useful adjustments are to tension, pace and variation rather than to the fundamental characteristics of your voice.

And a relaxed voice at its natural pitch outperforms a manufactured one at any pitch.

Voice-based bias

Which should be named rather than accommodated silently.

The judgements described above apply to accent, to voice quality and to features associated with gender and age, and they affect hiring, evaluation and credibility.

They are stereotypes with poor predictive accuracy operating as though they were information.

Which is a problem for the listener to fix rather than the speaker, though the speaker bears the cost.

And it is worth being aware of as a listener, since knowing that you form these impressions instantly is the only defence against acting on them.

Vocal fry and similar

A specific and contested case.

Creaky voice at the ends of phrases occurs across speakers and has been the subject of considerable and disproportionate criticism, directed particularly at younger women.

The research on whether it affects perceptions is mixed and depends heavily on listener and context.

What is clearer is that the volume of commentary about it exceeds any measured effect, which suggests the criticism is about something other than the acoustics.

Which is a reason to be sceptical of advice to eliminate it, and a reason to work on it only if it is accompanied by strain.

Warmth versus authority

The trade-off people navigate.

Features associated with warmth — variation, higher pitch, faster rate — are partly in tension with those associated with authority.

Which means the useful skill is moving between them rather than choosing one.

Lower and slower when stating a conclusion; more varied and warmer when explaining or engaging.

And a speaker who can do both is read as more credible than one who can only do either.

Your own perception

Which is unreliable.

You hear your own voice partly through bone conduction, which adds low frequencies that a recording does not.

Which means your sense of your own pitch, volume and pace is systematically wrong.

Recording is the only correction, and the discomfort of listening back fades with repetition.

And a single recording, listened to carefully, tells you more than any amount of theory about what listeners are actually receiving.