Working with an imperfect medium: An exploratory case study of adult learners using speech recognition-based reading software
Gunizi Kartal, Teachers College, Columbia University, United States
Teachers College, Columbia University . Awarded
This study focuses on the read-aloud and attendant behaviors of learners working with an interactive computer-based program in its development stage. The software makes use of speech recognition technology to provide reading practice with immediate feedback for early readers.
The study draws on data from thirteen participants, who used the software as part of their twice-weekly schedule for 6 to 8 weeks at an adult basic education site in New York. The participants consist of 7 users enrolled in an English as a Second Language class, and 6 users attending Adult Basic Education classes. All utterances recorded by each user were retrieved, and analyzed using various interrelated coding and analytical schemes developed by the researcher. Observations and interviews were conducted in order to glean information that could not be captured by the audio records.
The analyses were done in terms of three broad categories: patterns of interaction and shifts in patterns over time; influence of reading practice afforded by the software; and the function of speech recognition technology. The findings showed that the patterns of interaction occurred at three levels: metacognitive, decoding, and mechanics. The most frequently observed pattern of interaction was attempts at self-correction/change. Persistence in interaction as well as shifts in patterns over time were also common. In terms of the influence of reading practice, examination of the read-aloud behaviors following the computer feedback that invited users to try again showed that reading/pronunciation tended toward the target as practice ensued. In terms of speech recognition technology, the particular phonetic features of the users' speech were described as these pertain to speech recognition issues. In sum, the findings indicate that the users were engaged in the interaction, they attended to how the system responded to their input in categoric ways, and made progress regarding specific words.
The conclusions offer understanding of the relation between reading practice and speech recognition technology. The recommendations include ways in which the software can be altered in order to better meet the needs of similar user populations, as well as suggestions for how speech recognition technology can be used in reading practice in general.
Kartal, G. Working with an imperfect medium: An exploratory case study of adult learners using speech recognition-based reading software. Ph.D. thesis, Teachers College, Columbia University.
Citation reproduced with permission of ProQuest LLC.
For copies of dissertations and theses: (800) 521-0600/(734) 761-4700 or https://dissexpress.umi.com