Online Class Attention & Dropout Simulator

Session Setup

Class size 30 students

Quiz checkpoint every 6 min

Session speed 4×

Live Session

Elapsed

00:00

Class attention

100%

Predicted quiz score

—

Left the call

Session starting…

How it works

Each remote student's attention A(t) ∈ [0,1] relaxes toward a personal boredom floor between engaging events, following the same exponential relaxation used to model vigilance decay in sustained-attention research:

dA/dt = -k · (A − floor)
A(t+Δt) = floor + (A(t) − floor) · e^(−kΔt)

Every student has their own decay rate k (± random spread), so a class never drifts down in lock-step — some tune out fast, others stay locked in.

Quiz checkpoints — an active-recall prompt boosts attention toward the class ceiling (the "testing effect": retrieval re-engages attention far more than passive listening). A student already deep in the floor has a real chance of quietly leaving the call instead of re-engaging — a stand-in for the well-documented online-course drop-off curve.
Breaks — restore attention further than a quiz and also claw back accumulated fatigue.
Fatigue — the achievable ceiling erodes slowly over the session (screen fatigue), so the same quiz gives a smaller boost late in a long, break-free session than early on.
Predicted quiz score is a simple monotone mapping of live class-average attention onto a 0–100 scale — a proxy for how comprehension tracks engagement, not a literal grading formula.

Tune checkpoint frequency and class size and watch the grid: green tiles are engaged, red tiles are checked-out, dark tiles have left the call — a visual case for why short, interactive online sessions outperform long uninterrupted ones.

Engaged Drifting Checked out Left the call

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