Figure 2

The segregation reached at the end of each game in the experiment compared to the predictions from the simulations. Segregation is measured as the average percent similar neighbors. The solid circles show the segregation reached by the groups in the experiment and the gray areas show the proportion of the simulation replications that predicted that particular segregation level for each group size. In the simulation, agents are situated on a two-dimensional 6 × 6 grid. At every time period, an agent is selected uniformly at random from the population and given the opportunity to relocate. The agent evaluates the nearest four available locations in the up, down, left and right directions, as well as its current location. If the agent identifies a single strictly best available location among these, it moves to (or stays at) it; if it identifies multiple best locations, it randomly chooses to relocate to one of them. Simulations were run for 100,000 time periods and replicated 1000 times.