Robert Kraut is the Herbert A. Simon Professor of Human-Computer Interaction in the Business School at Carnegie Mellon.
It’s interesting to study online communities because the interactions are exposed and documented.
– Success is multidementional: transactional (did your question get answered?, were resources exchanged?); individual (was commitment developed?); and group (did it successfully recruit and retain members, and persist over time?).
Commitment develops ov er time, with early phase especially fragile, and it’s a bi-directional process. There’s a cost-benefit analysis.
It’s rational for groups to be skeptical of newcomers. Newcomers take resources from existing members. The group is more likely to be welcoming if they perceive newcomers as”deserving”. Thesis: individuals may use self-revealing introductions to signal both legitimacy and investment.
He’s looking at research questions about whether groups ignore newcomers and whether conversational strategies encourage group members to pay attention to newcomers. Looked at 99 Usenet groups, around 40k messages. They’ve seen that groups respond less to newcomers across the board, particularly in political and hobby groups. They then used machine learning to analyze messages to find self-introductory messages. Attempt to predict whether a given message will get a reply. Found that newcomers with a self-introduction are treated as well as old-timers without one. They found that messages with a group-oriented introduction (“I’ve been lurking here for a while…”) almost doubled the chance that a message would get a reply.
I wonder how this connects with the public profiles like in Facebook?
When will indivduals “join”?
Individuals evaluate potential benefits from the group. The reactions they get from initial attempts to engage the group will be especially meaningful. Hypothesis is that people will be more likely to continue to participate if people respond to them, and if the reply comes from people with higher status in the group and if they are positive in attitude. Found that only 20% of newcomers who don’t get a reply to their initial message are seen again, while 40% of those that do get a reply are active subsequently. The idea of the “welcoming committee”, like Wikipedia has, is very useful in developing commitment. The more central the replier is to the group, the more powerful it is for developing commitment of the newcomer. The tone of the welcoming language also has an effect.
I asked whether they’ve done any work in looking at the formation of new online communities and what factors might lead to success. It’s hard to research the formation of new communities because it’s hard to catch communities at the moment of formation. The problem with starting new groups is that there’s no content, so no reason for people to go there – a chicken and egg problem. One thing that helps is to find niche markets where people have a very high need for information sharing and will accept relatively low returns as worthwhile. Working with an existing organization might help.
Facebook is a fantastically successful community. Some of its success can be attributed to it having started with a small handful of communities (universities) that provide a pre-existing connection (students at the same institution), and then built on the early success, with the latest example being opening up the API to allow other people to build new services for the community.
In his courses they’ve been using Drupal because it offers lots more flexibility than course management systems. Even then they’ve had a hard time getting students to participate – so they’re learning how to issue challenges, use reputation-building systems, and other techniques to encourage participation. In on-campus communities, the hostility towards newcomers is less of a problem because people already consider themselves part of an existing collective.