An interactive version of the opinion model in Eun Lee, Petter Holme, and Sang Hoon Lee, “Modeling the dynamics of dissent,” Physica A 486 (2017) 262–272. Watch whether a dissenting opinion held by low-authority agents can spread through a scale-free network.
The society
Dissenting (σ=1)Affirmative (σ=0)Node size ∝ log authorityArrow: opinion can flow by comparison
Time (MC sweeps)0
Dissent fraction m–
Pr[q ≥ 0] measured–
Pr[q ≥ 0] Eq. 10–
Followship SGC–
⟨k⟩ / kmax–
Dissent over time
m(t), the fraction of agents holding the dissenting opinion. The run stops at consensus (m = 0 or 1).
Phase diagram
Final m over an 11×11 grid of p (across) and α (up) on the current network, after at most 300 sweeps, 4 runs per cell. Compare with Figs. 2 and 4. Click a cell to use its α and p.
The update rule
Each update picks a random agent i and a random neighbour j. If j dissents, i adopts dissent with probability p, regardless of authority. Otherwise (or with probability 1 − p) i compares authorities using the impact function q = α si − (1 − α) sj: if q < 0, i copies j's opinion; if q ≥ 0, i keeps its own. Authority scores follow p(s) ∼ s−γ with s ≥ 1, degrees follow p(k) ∼ k−γ with kmin = 2 (configuration model, largest component kept). One sweep is N updates. With α = p = 0 the model reduces to the voter model.