Unlike the dissonant interaction of public bulletin boards and chats, Reality Check tried to create online spaces where the participants have a sense of ownership and belonging.

To do this, it was designed with comparatively high barriers to entry: Prospective members were asked to fill out a 35-question form, then had to wait a day or so before being able to take part. They also had to accept a lower level of anonymity.

Finally, rather than allowing people to jump into the middle of an asynchronous conversation, each discussion group was begun at a specific time with a fixed number of participants. You were there at the beginning or not at all.

The study drew several conclusions, beginning with the assertion that as the size of a group goes down, the quality of interaction goes up. It is important, said Web Lab's supervising producer, Barry Joseph, that groups be small enough so people know who is there and so they notice who isn't participating.

"They need to be able to track each other," Joseph said.

And while anonymity may be a strong drawing card in the online world, the Reality Check study shows that decreasing it leads to better and more trusting interaction between chatters.

"There's nothing inherently wrong with anonymity," Joseph said. "But when there's too much, accountability just flies out the window."

While participants in Reality Check weren't asked to give their real names, they were required to use a consistent login name and to provide honest and revealing biographies. "What we found," Joseph said, "was that people felt anonymous in a good way -- safe to engage in frank, personal dialogue, yet comfortable with the level of disclosure of the other participants."