Homogeneity check

Duplicate results for 10 items: sw, ss and the 0.3 σpt verdict. Free, no sign-up. Everything runs in your browser; nothing you enter is sent or stored.

Try a homogeneity check

One line per item: the two replicate results. Change a value and watch s_s and the verdict. Decimal commas work too.

Duplicate results, one item per line

σ_pt

ISO 13528 asks for at least 10 items, chosen at random after packaging.

Items, g10

Grand mean5.6025

s_x (SD of item means)0.0266

s_w (within items)0.0254

s_s (between items)0.0196

0.3 σ_pt0.0750

0.5 σ_pt (for s_w)0.1250

VerdictPass

s_w = √(Σw² / 2g) = 0.0254 · s_s = √(max(0, s_x² − s_w²/2)) = 0.0196

sₛ ≤ 0.3 σpt: the items are sufficiently homogeneous.

How it works

A homogeneity check (ISO 13528 Annex B) asks whether the PT items differ from each other enough to affect scores. Pick at least 10 packaged items at random, measure each in duplicate in random order, and enter the pairs.

The check estimates the within-item SD s_w from the duplicate differences and the between-item SD s_s from the item means, and passes when s_s ≤ 0.3 σ_pt. If the method is imprecise (s_w near or above 0.5 σ_pt), the check cannot see real differences.

For learning and for checking your own arithmetic. Your scheme's report and rules are what count.