Figure 5. Effects of rule-breakers on popularity-based networks.

Individuals using rule-breaking strategy did not significantly affect either (A) their own social position (H(2) = 2.99, p = 0.22) or (B) the social position of others (H(2) = 0.24, p = 0.89). However, as their presence in the population increased, the behavior of these individuals will start to make an impact, affecting (C) the overall group organization (H(2) = 284.55, p<0.0001) even though there is no concurrent measurable effect on the individual values. The boxes show medians, quartiles, minima and maxima. Results significantly different from a relevant uniform network with no rule-breaking behavior (p<0.05 in Dunn's multiple comparison test for comparing each group with control) are designated with *.