In preparation for the HICSS conference which I’ll be attending next week, I have been thinking about the difference between using real-time data (like screen captures) and log files when analyzing online interaction. In the paper I will present (based on the same study that I presented in HUMlab not long ago), I argue that valuable information is lost if we only focus on log files. Since in this study I had access to information about current activities of the informant, I was able to find explanations for temporal aspects such as delay. Without this information I might have concluded that delay is the result of the informant being slow at typing, when, in fact, she is engaged in other simultaneous activities.

Apart from presenting the paper, I will also give a brief presentation at the Persistent Conversation workshop. This should be a valuable session, since participants are asked to apply their theoretical framework to the same chat system. Again, I am confronted with the question of what different types of analyses can be carried out depending on what data you have access to. Here, the real-time data does not capture the communicative environment of the informants in front of their computers, but rather the real-time interaction in the online environment. Nevertheless, there are some important differences, for example concerning the type of information you get regarding the communicative intention of the conversational partners (information about when others are typing is available in real-time but not in the logs). However, in this specific case, the log files are also quite useful, since they link the different modes (text chat and white board) and allow for detailed revisiting of the relationship between the two. Depending on your research agenda, a combination might be to prefer. I look forward to discussing this further!

Some other topics I hope to get an opportunity to discuss at HICSS include:

• Tools for analysis and visualization of complex data
• Ethnography and design
• Ethical and legal concerns
• The notion of communicative affordances and its applicability
• Plans for my future study

I am really looking forward to this trip, both because I think it will be valuable to my research, and, admittedly, because I long for some vacation in a warmer climate! Updates to be expected. Aloha for now!

The Department for Interactive Media and Learning at our university had many representatives at the NetLearning conference in Ronneby earlier this year. One of the papers which especially caught my attention is written by Peter Bergström and deals with student interactions in web based video conferencing. He has used Fahy et al.’s Transcript Analysis Tool (TAT) for his analysis, which, as far as I can make out, is a model for analysis, and not a specific software. In line with this model, utterances have been categorized as “questioning, statements, reflections, or interpersonal coaching and scaffolding” (English terminology borrowed from Fahy). Peter has followed two small groups and compared their interactions, and he has found variation in the types of utterances produced in the different groups. For example, one group used more “vertical” (closed) questions, which on the surface would indicate that they were less successful at creating dialogue. However, closer analysis reveals that these questions were often accompanied by reflections which might add a sense of openness. Peter claims that his findings are influenced for example by the medium and whether participants have used it previously, by differences in previous contacts between group members, as well as by group size. In both groups, he concludes, dialogue was achieved.

For the article I’m currently working on (at least in theory), I want to look at the form and function of feedback in student interaction online, but I will have to consider whether including information about other types of utterances would also be relevant.

If you know Swedish, you should check Peter’s paper out.

Tomorrow I’ll be teaching a course in HUMlab on transcribing and analyzing video. The main focus of the course will be on Transana, and I will demonstrate how I’ve used this free software in my own transcription and analysis work and let the course participants try it out for themselves.

In Transana, as in most tools of this kind, transcription is done manually. I have been looking around for examples of software attempting to automatically transcribe sound files, but what I have found so far are tools used by one individual, training them to understand his or her speech. I suspect these existing tools would not be able to transcribe more complex interactions.

If anyone knows of tools with automatic transcription which might be able to handle more complex material, please let me know. Even if such tools would exist, of course one would have to go over the automatically generated transcripts in relation to the sound files manually anyway, both in order to check the transcription, and not least also in order to find the subtle nuances conveyed in extralinguistic cues. No matter how advanced technology might have become, I suspect the computer would not be able to detect these types of cues. Not yet anyway.

A while back, my supervisor pointed me towards a post in Infocult on 3D vs. Video, relating to my research on communication in these types of environments as well as to previous posts here on the topic. Finally, I have now gotten around to check it out. In this post, Bryan refers to a list in the Oddcast blog giving the top 10 reasons why avatars are better than video representations, and presents some counter arguments. While I agree that there are some advantages with avatars over video (mainly relating to the relative anonymity of the avatars and the strong sense of co-presence in 3D), one of the claims on the list I would like to challenge is the first one:

1. Avatars are interactive. Video is a one way street–a broadcast medium. You can program an avatar to respond to different mouse clicks, for example, or control it in real time, like in Second Life. Try doing that with video.

To me, this is a question of what you use video for – video conferencing, for example, is very much about interaction and not about broadcasting at all. And not having to use mouse clicks, but instead being able to directly respond and give feedback via body language, could rather be seen as an advantage here… One thing I sometimes get annoyed with when using video, though, is that I can’t control the visual input on my screen in the same way as I can in a 3D environment. (I recall having mentioned this frustration earlier here in the blog… Ah, yes, here it is… ).

Over the last few days, I have spent some time in Second Life. I have long planned to try it out, but have never really gotten around to it. Now I have to get acquainted with the environment as a course requirement, and so I have been pushed to start explore a bit. Being used to working with ActiveWorlds, with its often relatively confined spaces, I have to admit the sheer size of this environment is a bit overwhelming. Therefore, in order to finally get started for real, I contacted my friend and colleague Mia, who kindly showed me around and answered some of my many questions.

There seems to have been a bit of a crisis in Second Life over the last few days. There has been an upgrade to fix previous problems, and while these appear to have been fixed, instead some new bugs emerged in the upgrade (difficulties to teleport, account balances not showing, inventory missing, for example). This was a hot topic when I entered yesterday evening, and one resident I met even referred to it as “the apocalypse of Second Life”. When the grid was down on Wednesday, it was interesting to see how some residents instead gathered on the Official Linden Blog to interact (this entry now has 1428 comments). Apparently there were also some problems later as everyone tried to enter the grid at once… It all seems to be working better now, though, and so I’m looking forward to exploring some more. More reflections to be expected, either here or in the HUMlab blog.