In the first experiment, they recorded 161 college students reading a passage aloud; this first recording captured baseline acoustics. The participants were then randomly assigned them to play a specific role in an ensuing negotiation exercise.
Students assigned to a “high” rank were told to go into the negotiation imagining that they either had a strong alternative offer, valuable inside information, or high status in the workplace, or they were asked to recall an experience in which they had power before the negotiation started. Low-rank students, on the other hand, were told to imagine they had either a weak offer, no inside information, or low workplace status, or they were asked to recall an experience in which they lacked power.
The students then read a second passage aloud, as if they were leading off negotiations with their imaginary adversary, and their voices were recorded. Everyone read the same opening, allowing the researchers to examine acoustics while holding the speech content constant across all participants.
Comparing the first and second recordings, the researchers found that the voices of students assigned to high-power roles tended to go up in pitch, become more monotone (less variable in pitch), and become more variable in loudness than the voices of students assigned low-power roles.
And the students’ vocal cues didn’t go unnoticed. A second experiment with a separate group of college students revealed that listeners, who had no knowledge of the first experiment, were able to pick up on these power-related vocal cues to determine who did and did not have power: Listeners ranked speakers who had been assigned to the high-rank group as more likely to engage in high-power behaviors, and they were able to categorize whether a speaker had high or low rank with considerable accuracy.
In line with the vocal changes observed in the first experiments, listeners tended to associate higher pitch and voices that varied in loudness with high-power behaviors. They also associated louder voices with higher power.