A Study of Voice-Recognition Software as a Tool for Teacher Response

Abstract

Voice-recognition technology (VRT) promises ease of use in responding to student writing, but its impact on writing processes and the quality of teacher commentary is unclear. This article details the results of a study undertaken to evaluate the effectiveness of VRT in responding to student writing. Over the course of two semesters, in response to texts written by students in his first-year college composition course, one of the authors composed 58 end comments, alternating between two methods of composition: typing on a keyboard and dictating directly into text by means of VRT. While his writing processes in the respective modalities differed dramatically, particularly in terms of revision, the quality of the resulting texts appeared roughly the same: A detailed content analysis of the comments, using Richard Straub and Ronald Lunsford's (1995) typology, revealed significant variation in only 1 of 25 variables, measured as a proportion of total words. Meanwhile, students surveyed indicated they found few differences between the typed and dictated comments in terms of their usefulness, clarity, and tone. These findings, along with a comparison of time on task and user impressions of the two modalities, indicate that VRT represented a valuable tool for producing end comments that the user was able to dictate fluently, but that the technology was ineffective for the limited editing and revising attempted within the design of the study.