An empirical approach to measuring accent familiarity: Phonological and correlational analyses

Accent familiarity can influence listeners' perception of second language speech. However, the field lacks empirical evidence for defining and measuring accent familiarity. The present study seeks to better understand accent familiarity a) by developing an evidence-informed, phonology-based approach to measuring accent familiarity and b) by validating its relationship with existing measures. Six L1 Chinese speakers of English with varying degrees of accent recorded scripted and conversation tasks. Seven experienced teachers listened to the recordings and identified 18 phonological features common in Chinese-accented English. Audio recordings containing these features were included in an online questionnaire, where 69 listeners a) rated their familiarity with these target sounds and b) completed eight different measures of accent familiarity used in previous studies (e.g., accent identification). Results from exploratory factor analysis (EFA) suggested that listeners' familiarity with the 18 features was single-factored. Thus, a phonology-based measure of accent familiarity was developed by a compositing listeners' familiarity with the 18 features. Correlation analysis and EFA of this phonology-based accent familiarity measure and different previous measures suggested two factors: a) measures without audio input and b) audio-prompted measures. By implication, future studies should consider including audio recordings to measure listeners’ accent familiarity in addition to traditional measures.