Fig. 2

Proportion of network informed by each timestep in simulated information spread on a network with τ={τ ¹,τ ²}, and \(q^{\tau ^{1}} = q^{\tau ^{2}} = \frac {1}{2}\). Simulation parameters set to x=2, p _strong=1, and p _weak=.5. When homophily = 0, random ties will not change neighborhood compositions on average, so the jump effect dominates and increasing density strictly improves information diffusion. At greater values of homophily, increasing density does change neighborhood compositions and strictly impedes information diffusion