Carnegie Mellon University

Students Respond to Teacher Comments: A Comparison of Online Written and Voice Modalities

Author: Loel Kim Robinson
Degree: Ph.D. in Rhetoric, Carnegie Mellon University, 1998

Composition students and teachers claim that spoken comments are more understandable, useful, and "personable" than written comments. At the same time studies in human-computer interaction (HCI) suggest that "media-rich" voice modality--which conveys more paralinguistic cues than written modality--may enhance commenting by supporting both semantic and affective understanding. Thus, this dissertation asks: Can voice modality's increased expressivity improve student reception of teacher responses? I hypothesized students would: (1) generally prefer voice over written modality; (2) rate voice comments higher than written ones; (3) prefer voice modality for comments which address high-level writing problems; and (4) perceive teachers more favorably in voice modality.

In a 2x2x4 factorial design, thirty-nine students received two sets of online comments, each for a different student draft, and both made by one of four teachers. One set was in written modality and one in voice. Drafts were seeded with ten writing problems: five high- and five low-level. Comments were made and received using PREP Editor software. After each set, students gave quality ratings for each comment, and filled out five-point Likert scales measuring teacher persona. Post-session interviews were conducted.

Although students generally liked receiving comments online, they were divided about modalities: forty-six percent preferred voice and 41% preferred written modality. Upon closer examination, student preferences varied for Instructor 1 in particular: Whereas students did prefer voice modality over written when receiving comments from Instructors 3 and 4, and were evenly split for Instructor 2, students who received comments from Instructor 1 preferred written modality.

When examined by modality, variances in persona ratings (30.3/60 to 43.0/60) were small for Instructors 2, 3, and 4 (.78 to 2.71), but significantly larger for Instructor 1 (13.3). Finally, over 79% of the students did not recognize their commentor as the same teacher in both modalities. These findings suggest that characteristics of persona may differ dramatically between computer-supported communication modalities, adding complexity to our understanding of the technologies, themselves, and of their implications for future pedagogical applications. They also remind us that media-rich modalities indiscriminately convey both positive and negative cues and, thus, cannot promise uniform improvements to teacher commenting.