3D Belief Correction: Undergraduate Lecture Hall

Lecture Hall

Course-exposure rate 14 / semester

Evidence quality 0.35

Initial belief level 70%

Cohort belief

–

Semesters elapsed

–

Mean log-odds L

–

Below 10% belief

–

How it works

Every seat is a student holding a belief p (probability the false explanation is correct), tracked as log-odds L = ln(p / (1−p)). The podium fires a pulse of formal evidence every simulated interval — a lecture, a lab result, a derivation — and because that evidence is costly to produce and high quality, its likelihood ratio reliably favors the truth. Each pulse updates every student's log-odds by a drift-diffusion step:

λ  = exposureRate / semester
ΔL = −q·λ·dt·engagement + σ·√(λ·dt·engagement)·ξ
p  = 1 / (1 + e^−L)

Seat color interpolates from red (high belief in the false explanation) to gold (converged toward truth) as each student's p falls. Because engagement varies seat to seat, the hall visibly desaturates unevenly — front-row, high-engagement students converge first even though everyone hears the same lecture.

Course-exposure rate — evidence pulses per semester from the podium.
Evidence quality — average log-odds pull per pulse; better-produced evidence corrects harder.
Initial belief level — starting probability across the whole hall.

This is correction, not contagion: belief moves toward truth because the evidence stream is asymmetrically informative, not because students are copying each other's opinions.

High belief in false claim Converged toward truth Podium evidence pulse

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.

3D Lecture Hall: Belief Correction via Costly-Accurate Signals