The studies
The empirical material investigated in the thesis was gathered during field studies inspired by ethnographic methodology. This resulted in four case studies investigating affordances on different levels, in accordance with the model below.
Study 1:
Disrupted turn adjacency and coherence maintenance in Instant Messaging conversations
(Published in Language@Internet)
In this study, I investigated 120 Instant Messaging (IM) conversations involving six people forming eleven dyads (7474 words, 1689 messages). The material was gathered during an ethnographic field study at a design school in the spring of 2007. The main focus of the study concerned instances of disrupted turn adjacency. This is a phenomenon whereby the system causes logically related messages to appear separated on the screen. I looked into how the links between these separated messages became clear, and concluded that explicit strategies, such as lexical repetition, substitution and conjunction, were strong indications of links, whereas also more ambiguous strategies, such as anaphoric reference and elliptical feedback were used. Further problems identified were related to difficulties in determining the whereabouts of the others and difficulties in resolving instances of miscommunication in the written medium.
Study 2:
Multimodal student interaction online: An ecological perspective
(Published in ReCALL)
In this study, focus is again on the communicative affordances of a specific tool, but this time a multimodal desktop video conferencing environment. During the spring semester of 2006 I observed and filmed a group of students taking part in a course of English at a distance. The investigation concerns their employment of the different modes available (voice, text, video, graphical elements) with a specific focus on how these modes are used when delivering conversational feedback. The material consists of five recorded sessions, and the analysis concerned the first and the last session. The results showed that the students often did not provide each other feedback, and reasons for this were found both in the design of the task and in the design of the tool.
Study 3:
Multiplex conversations afforded by technology
(Published in proceedings from HICSS’40)
Here, focus is on the conversational multitasking of one informant by the computer. In May 2006 I filmed the screen of the informant at her workplace for two hours (out of ethical concerns, I filmed with a blurred focus), and got access to the log files from her online conversations. This material was combined in a transcript which I used to investigate whether she was able to keep all of her conversations going, and whether there were conversation internal causes for switching tools. In both cases, I looked at sequencing patterns, and found that the informant was able to reply in all instances where her reply was expected, and in that respect her multitasking did not lead to conversational breakdown. I also found that the reason why the informant switched to a specific tool did not relate to whether she had asked a question to which she expected a reply, but rather to the alerts she got from ICQ that she had received new messages. Based on the results of the study, I presented suggestions for design of unobtrusive conversational management.
Study 4:
Navigating complex media landscapes
(In review)
In this study, I observed a group of students at a design school over three weeks in March 2007. During six days, I conducted more detailed observations, shadowing three different students. During these days, I also got access to all log files from in-group instant messaging, and conversation diaries from the students not individually observed. After the observations, I also conducted semi-structured interviews with the shadowed informants. This substantial material has been used to investigate how the participants initiate conversations in a coherent manner, keeping considerations concerning privacy, urgency and availability in mind. The results showed that IM was mainly used for private topics whereas work-related topics were the most common in face-to-face interaction. Furthermore, conversation initiation often occurred in connection with movement in face-to-face interaction, whereas windows of opportunity were more difficult to detect in IM (despite status messages). It was also seen that the participants immediately checked new messages upon receiving alerts, leading to constant interruptions in ongoing tasks.