Evaluation of UThink -- Existing Design Feature Explanation

This paper evaluates the features of UThink, the blog service of the University of Minnesota that is designed to generate on-line discussions among the university community (see http://blog.lib.umn.edu)∞. As the digital library project group at the University of Michigan is building a similar blog service for the UM community to encourage on-line interactions among faculty, staff and students, an evaluation of UThink will provide a reference for the design of UM blog service.

PaulResnick: I'm trying to identify the design feature(s) you're discussing. In the anonymity/privacy section, I presume it's the real-name/pseudonym distinction. If so, you should lay out the design space (real name, one-time pseudonym, changeable pseudonym, anonymous), say what the literature (in particular, the Friedman/Resnick article) says about the implications of those alternatives generally, and then apply those insights to UThink, and use it to explain empirical facts about how the sysem is working. In the Org Memory section, I couldn't figure out what design feature you were discussing. The Sociability/Usability section seems to bring in lots of different design choices.

UThink Overview

UThink is a blog service designed to support teaching and learning, scholarly communication, and individual expression for the U of Minnesota community. It was officially launched one year ago and is only available to the faculty, staff, and students at the University of Minnesota Twin Cities. The initiators of UThink envisioned a system where

[faculty could have blogs for their classes and attach the students in those classes to the blogs as authors to encourage discussion and debate; faculty could also use the blogging system to track a research initiative, or even publish the drafts of papers they are working on; a student could also have a blog for every class, or just use blogs to express opinions and viewpoints about world events; a student could also create a club blog, or a blog for his or her friends, and also attach as many authors to those blogs as he or she deems necessary. ]

Currently, UThink has 952 registered users including 167 faculty, 125 staff, 323 undergraduate and 338 graduate students . In all there are 1192 separate blogs. Although a few blogs are frequently updated and involve some communication between bloggers and readers, overall this on-line community has not had enough participation to meet the University?s goals for the system as discussed above. This paper will explain this phenomenon through relevant theories.

UThink as an Organizational Memory System
Argote defines organizational memory as the means by which knowledge from the past is brought to bear on present organizational activities (1999). Ackerman points out written records are only one form of organizational memory and that organizational memory can be retained in many other places, including organizational culture, processes, and structures (1996). Blogging involves many notions: self-expression, personal publishing, diary, on-line community, alternative media, etc. Blogs can be created for multiple purposes, such as social, educational and professional. Our survey of UThink reveals that the content in this community includes several categories, which are mainly class and class project activities, political and sport events discussions, research ideas and projects, library information, and self-reported student life. This clearly reflects that UThink is an organizational memory system and a shared repertoire (Wenger 1998) that has been adopted as part of the development of the University of Minnesota?s on-line community.

Concerns of Anonymity and Privacy

Friedman and Resnick (2001) point out that anonymity is desirable in certain on-line communities although disallowing anonymity would make participants more trustworthy. A number of other researchers have done research on encouraging participation and collaboration through anonymity. UThink does not allow anonymity. The Bloggers login in the blog service with their unique names affiliated with the University of Minnesota and post entries with their unique name appearing with the entries. It is easy to find out who users are through their unique name and searching university directory. It is tempting to assume that posts and discussions on political and other sensitive issues require some level of anonymity to increase participation. To address the issue of whether it is appropriate (in a university on-line community) to allow anonymity ( through the use of pseudonyms) and on what level to allow it needs more research.

Many UThink registered users express they would like to have a private blog where access is restricted by a username and password and hence only certain group of people could access the entries and comments posted on a private blog. UThink does not have any options for this concern. (Similar concerns appeared in our eCommunities class when the instructors intended to set up a class blog to encourage on-line discussions and information sharing.) Over the past year, among the more than one thousand blogs in UThink, only 25 have been set up for class discussion purposes and the majority of them were set up in last fall and then closed once the semester ended. Within this limited number of class blogs, many are used more for administrative announcements than intellectual discussion. An investigation of the class blog for ?Love, Sex and Marriage? (Sociology 4521) reveals that not a single entry involved intellectual content discussion. Some class blogs immediately closed down after only a couple of entries with the last entry stating that email is easier than blogs to maintain the class discussion. It should be clear that more issues beyond privacy are involved in the inefficiency of class blogs. However, to answer the privacy concern, UThink administration simply writes ?Keep in mind, though, that just because you write something does not mean people will read it.?, which is unlikely to satisfy in the least people who are concerned about the privacy issue while at the same time it discourages other people who want their entries read.

Problems of Sociability and Usability

Sociability and usability are two key issues leading to success of online communities and they are closely related to each other (Preece 2003). UThink uses Movable Type as its platform, which is one of the most popular blog software packages with many more functions than were available on previous blog services. However, UThink did not take advantage of these new functions and add some good features to the system. UThink is indeed a very heterogeneous and big on-line community. The participants form sub-groups based on common bonds (i.e. tied together by interpersonal connections) or common identity (i.e. tied together by a common goal or purpose) or the combination of the two ( Sassenberg 2002). For instance, the blog titled ?Greet Machine?, one of the most popular blogs, has discussions on only sports; whereas many less popular blogs are mainly places for self-expression and/or information sharing with close friends. The system does not provide an easy way for participants to search by topic among the many blogs. Instead, all blogs are arranged alphabetically. The function ?Categories in UThink? is based on the subjects of the entries entered by individual bloggers when they create new entries. Many subjects appear with tags like ?a story?, ?about this site?, ?a book?, etc. that do not help participants navigate to specific topics among the over one thousand blogs. To build a good categorization and navigation function on the interface would help participants efficiently locate the sub-community they might like to join.

To increase participation involves many aspects: getting more people in the university start blogging on UThink, encouraging more entries and more comments, enlarging readership, etc. Among the top 50 blogs with the most entries and the top 50 blogs with the most comments on UThink (these two categories overlap), 11 blogs had on average between one and two comments per entry; one class blog scored on average 4 comments for each entry, which presented good conversations among the members of this sub-community. However, the majority of blog entries received no comments. Someone might argue that for a personal blog, not having comments does not necessarily mean there is no participation, and people who read a blog but do not post responses are still participants.

A brief survey of other blog services reveals that beyond putting the most popular blogs on the homepage, randomly choosing different blogs that are frequently updated to appear on the homepage is another kind of explicit incentive to bloggers. This seems efficient to attract outsiders to jump in and retain members of community in sustainable social structures (Bulter 2001). Explicit incentive mechanism is described by Chen (2003) as one of 12 factors that might lead to increased content contributions to on-line communities.

To conclude, since UThink is a more complicated organizational memory system than many other blog services, enhancing overall participation in UThink requires more research on the key components of sociability (i.e. the purpose of the community, the needs of participants, and the policy that develops) and usability (i.e. dialog, social support, information design, navigation and access) and how they are related to each other (Preece 2003). Similar on-line communities, such as the design of the University of Michigan blog service, would be well informed by this type of research.

Reference
Argote, Linda (1999) Organizational Learning: Creating, Retaining and Transferring Knowledge, Springer

Ackerman, Mark S. (1996). Definitional and Contextual Issues in Organizational and Group Memories. Information Technology and People
Wenger, Etienne (1998) Community of Practice

Friedman, E. and Resnick, P. (2001). The Social Cost of Cheap Pseudonyms. Journal of Economics and Management Strategy, 10(2): 173-199.

Preece, J. and Maloney-Krichmar, D. (2003) ?Online Communities.? In J. Jacko and A. Sears, A. (Eds.) Handbook of Human-Computer Interaction, Lawrence Erlbaum Associates Inc. Publishers.

Sassenberg, K. (2002). Common bond and common identity groups on the Internet: Attachment and normative behavior in on-topic and off-topic chats. Group Dynamics, 6(1), 27-37.

Chen, Yan et.al (2003) Designing on-line communities to enhance participation -- bridging theory and practice. A proposal to the National Science Foundation.

Bulter, B.S. (2001) Membership Size, Communication Activity, and Sustainability: A Resource-Based Model of Online Social Structures. Information Systems Research. 12(4): 346-365.