Interpretations of spoken utterance fluency in simulated and face-to-face oral proficiency interviews

Abstract

Research examining test taker fluency in simulated and face-to-face oral proficiency interview performances has primarily focused on quantitative spoken utterance fluency data alone, with further qualitative investigation of test taker processing fluency’s effect being omitted. This study compared four test takers’ spoken utterance and processing fluencies in output retrieved from a computer-based Aptis General speaking test and a purposively developed identical face-to-face direct oral proficiency interview using a counterbalanced research design. Speed, composite, breakdown, and repair utterance fluency measures were analyzed from test performances using Praat speech analysis software and fully-coded transcribed spoken data, with processing fluency qualitative data retrieved through post-test stimulated recall verbal report interviews and questionnaires, and co-constructed, semi-structured interviews. Macro-level quantitative analysis results demonstrated that test takers’ spoken utterance fluency was broadly similar between delivery modes. The breakdown measures of filled and unfilled pauses were salient devices responsible for the minimal difference encountered between modes. The test takers’ qualitative data reflecting on their performances revealed aspects of processing fluency’s effect on test takers’ utterance fluency. In turn, processing fluency was influenced by both test takers’ cognitive and affective fluencies, through test taking strategy use, test taker characteristics, and the face validity of the test’s delivery mode.