Fig. 2. Population segregation.

Red = high endowment and high productivity (HH), light grey = low endowment and high productivity (LH), dark grey = high endowment and low productivity (HL), blue = low endowment and low productivity (LL). a Spatial grid displaying the frequency of observed types in each possible pairing configuration in the assigned partner (left) and partner choice condition (right). Each dot represents one observation per type and pairing. For example, the top right block corresponds to a pairing configuration of an LL type interacting with an HH type: The blue dots represent the number of LL types, and the red dots represent the number of HH types that were in this pairing configuration in the assigned partner condition (left) or in the partner choice condition (right). All dots together reflect the number of participants that were part of an HH-LL pair. The frequency of dots along the diagonal shows that partner choice led to a segregation of the population into pairs of similar types (higher frequency of dots on the diagonal in the partner choice condition than in the assigned partner condition). b Stacked bar graph illustrating the relative frequency of observed pairs in the assigned partner (left) and partner choice condition (right). Each bar represents the relative frequency with which a participant type was paired with their own or another type. c Average length of consecutive interactions between different pairs (as a measure of pair stability) in the partner choice condition (n = 21 groups). Whereas HH-HH pairs interacted 9.7 consecutive rounds on average, HH-LL pairs were least stable and only interacted 1.4 consecutive rounds on average before breaking up. Error bars indicate the standard error of the mean.