Improving Transcription of Qualitative Research Interviews with Speech Recognition Technology
American Educational Research Association Annual Meeting,
The recent development of high-quality voice recognition software greatly facilitates the production of transcriptions for research and allows for objective and full transcription as well as annotated interpretation. Commercial speech recognition programs that are appropriate for generating transcriptions are available from a number of vendors, with varying degrees of difficulty in use. There are two fundamental approaches to using speech recognition to produce transcriptions: (1) real-time; and (2) batch. The real-time approach uses speech recognition while the interview is in progress; the batch approach relies on an audio recording to make it possible to process several interviews in a batch. The use of speech transcription still requires the use of a human transcriptionist, and the best that can be achieved for transcription speed is a mere doubling of the interview time. However, when the user is not a skilled typist, considerable savings of time can be achieved. If a researcher finds voice recognition software to be superior to conventional typing approaches for transcribing interviews, he or she is likely to find it useful for other tasks as well. (Contains 21 references.) (SLD)
Fogg, T. & Wightman, C.W. (2000). Improving Transcription of Qualitative Research Interviews with Speech Recognition Technology. Presented at American Educational Research Association Annual Meeting 2000.