Abstract:

This study examined the use of automatic speech recognition (ASR) scored elicited oral response (EOR) tests designed to predict speaking ability, as well as the relationship between ASR-scored EOR and other language proficiency measures and the ability of the ASR to rate speakers without bias to gender and native language. To that end, 179 subjects were given an ASR-Scored EOR Test with 60 items followed by an oral interview and a battery of other language tests. Findings suggest that ASR-scored EOR results could be used alone to predict speaking ability in specific situations and for limited purposes such as initial placement of students in language training situations. However, if more certainty should be required, adding a listening component would improve the assessment. Analysis of the study results also suggests that differences in gender and native language had no significant effect on ASR-scoring of EOR. While EOR is not an authentic performance assessment of the specific construct in questions, it does have high reliability, good content validity and seems to be a practical estimate of speaking proficiency that could be used for initial placement of students in situations where speaking is not currently being assessed.

Author: Troy Cox, Randall S Davies