Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Division
Browse By Session Type
Search Tips
Personal Schedule
Sign In
Online communities differ dramatically in their membership size. The largest online communities have millions of participants while most communities attract very few. One popular explanation for "long tail" distributions like this is "preferential attachment", a model whereby people decide whether to join a community based only on its popularity. While preferential attachment models approximate empirical distributions, they do not provide a believable account of human decision making. On the other hand, models of community joining from HCI are more plausible but are rarely validated empirically. We use agent-based simulation to test one such model. We show that while simple versions of the model produce outcomes that are very different from real-world distributions an extension of the model produces simulated community distributions which are similar to those observed empirically. Finally, we show how social learning can lead to self-reinforcing outcomes. We provide our simulation code so that others can build on our work.