Academic Opinion Polarization Simulator

Bounded-Confidence Debate

Confidence threshold ε 0.15

Convergence rate μ 0.30

Academics 120

Opinion clusters

–

Polarization (var)

–

Interactions / s

–

Top strip: every academic's current opinion on the axis (0 = skeptical, 1 = convinced). Below: a scrolling history of the opinion distribution — brighter cells mean more academics hold that opinion at that moment.

How it works

Each academic holds a continuous opinion o ∈ [0, 1] about a contested claim. Every tick, two academics i and j are picked at random and compared. If their opinions are already close, they nudge toward each other; if not, they simply ignore one another — no interaction happens at all. This is the Deffuant–Weisbuch bounded-confidence model (Deffuant, Neau, Amblard & Weisbuch, 2000), a standard opinion-dynamics model distinct from simple contagion, where every contact changes the receiver's state a little.

if |oᵢ − oⱼ| < ε:
    oᵢ ← oᵢ + μ·(oⱼ − oᵢ)
    oⱼ ← oⱼ + μ·(oᵢ − oⱼ)
else:
    no change  (opinions too far apart to interact)

Confidence threshold ε — how open-minded each academic is. Small ε means only near-identical peers can influence each other, so the population fractures into several stable opinion clusters that never converge — real, persistent polarization. Large ε (above ≈0.5) lets almost everyone interact, and the population collapses to a single consensus.
Convergence rate μ — how far two agreeing academics move toward each other per interaction. It changes how fast clusters form, not how many clusters end up existing — that is governed almost entirely by ε.
Why clusters, not just contagion — unlike an SIR/SI contagion model where any contact can flip a belief, bounded confidence means an extreme skeptic and an extreme believer never move each other at all, no matter how many times they interact. That is what lets multiple rival camps coexist indefinitely instead of one belief simply spreading through the network.
Polarization is the variance of the opinion distribution: near 0 means everyone has converged to one point, near 0.25 means the population has split into camps at opposite ends of the scale.

0 — Skeptical1 — Convinced

Opinion (hue = value) Distribution history →