Regression to the Mean (2D)

Each dot is one individual measured twice: a true value plus independent random noise each time. Extreme first measurements tend to be followed by less extreme second ones.

Controls

Noise level: 1.00 Sample size N: 250

Highlight top 10% extreme

Live statistics

Correlation r—

Regression slope—

Extreme group: mean M1—

Extreme group: mean M2—

Regression amount—

Measurement 1 = true value + noise₁. Measurement 2 = true value + noise₂ (independent draw). Lower noise → higher correlation r → regression slope closer to 1. Yellow points are the top-10% most extreme scorers on Measurement 1; the red arrow shows how far their average Measurement 2 falls back toward the population mean.