General Mathematics · bivariate data · manipulative

r is not the slope

The least-squares gradient is b = r × (sᵧ / sₓ). Hold r fixed and change how spread y is compared with x — the line tilts, r does not.