Human Review Sample Size Calculator

Work out human review sample size instantly with clear inputs, formula shown and shareable results.

About Human Review Sample Size Calculator

Sample size for estimating a proportion is z^2 p(1-p) / e^2, then reduced by the finite-population correction when the sample is a meaningful fraction of the whole. Note what is absent: population size barely matters. Reviewing 1,000 items gives about the same precision whether the population is 100,000 or 100 million, which is why audit samples are far smaller than intuition suggests.

Formula

Sample size for a proportion

n0 = z^2 p(1-p) / e^2; n = n0 / (1 + (n0 - 1) / N)

z: Confidence z score
p: Expected rate
e: Margin of error
N: Population

Worked example

With the default values below, Human Review Sample Size Calculator returns Items to review of 384. Change any input above and every figure here is recalculated in your browser.

Inputs used

Items in the populationminimum 1: 250,000
Confidence levelaccepts 80–99.9 %: 95 %
Acceptable margin of erroraccepts 0.1–20 %: 3 %
Expected defect rateaccepts 0.1–99.9 %: 10 %

Results

Items to review: 384
Sample before finite-population correction: 385
z score for the confidence level: 1.960
Share of the population reviewed: 0.154%

Step by step

1Values used
2Sample size for a proportion
3Items to review= 384
4Sample before finite-population correction= 385
5z score for the confidence level= 1.960
6Share of the population reviewed= 0.154

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to 3 decimal places, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

What if I have no idea of the defect rate?

Use 50 percent. It maximises p(1-p) and therefore gives the largest, safest sample size.

Does the population size really not matter?

Only through the correction term, which is negligible once the population is more than about 20 times the sample. Precision is driven almost entirely by sample size.