Revisiting listener evaluations of creaky voice

Abstract

In U.S. English, nonpathological creaky voice is a common and natural voice quality that contributes meaningful linguistic information. However, it is also a target of gender-based prejudice, with studies suggesting that speakers with creak sound less hirable and less pleasant. Unfortunately, the stimuli in these studies were often problematic (e.g., creak present in both creaky and modal stimuli; unusually high proportions of creak). To address these problems, we carefully selected naturallyproduced, low predictability sentences. Half of these contained creak on the final word (a common location for creak in both men and women speakers) and half did not. Listeners rated half of the sentences on a hirability scale and half on a pleasantness scale. They then completed an implicit association task regarding gender and career roles. Consistent with previous research, sentences with creak were rated as being produced by speakers who were less hirable than those without creak. Similar to previous findings, older listeners rated stimuli with creak as less pleasant than stimuli without creak. Implicit association scores did not predict ratings on either rating scale. Contrary to our expectations, less favorable evaluations of creak were found despite using naturallyproduced sentences with only sentence-final creak.