Accessibility Compliance Audit Sample Size Estimator

This estimator calculates a statistical sample size for a finite population of accessibility audit items. It uses the selected confidence level, margin of error, and expected issue rate to estimate how many records, pages, transactions, locations, or other units should be tested when the goal is to estimate a population proportion rather than inspect every item.

The result is a planning number for random or probability-based sampling. It does not create an audit standard, guarantee detection of every problem, or replace a required sample method specified by a regulator, contract, accessibility methodology, environmental program, or internal audit plan. A smaller margin of error, higher confidence level, or an expected issue rate closer to 50% generally increases the calculated sample.

Sampling assumptions

Population size

items

Confidence level

Margin of error

Expected issue rate

Result

—

items in suggested statistical sample

Unadjusted large-population sample—

Sampling fraction—

Population size—

Target margin of error—

How to use

1. Enter the population size
Count the total number of eligible accessibility audit items in the population from which the sample will be drawn.

2. Choose a confidence level
Select 90%, 95%, or 99% according to the statistical assurance level appropriate for your audit design.

3. Set the margin of error
Enter the maximum sampling error you are willing to accept for the estimated proportion, expressed as percentage points.

4. Enter an expected issue rate
Use a prior estimate when one is defensible. If the true rate is unknown, 50% is the conservative choice for this formula because it produces the largest variance.

5. Review the rounded sample
The estimator rounds up to a whole item and applies a finite-population correction so the sample cannot exceed the population.

6. Confirm the sampling method
Use random or otherwise statistically valid selection if you intend to interpret the confidence and margin-of-error outputs statistically.

Formula

The estimator first calculates the large-population sample and then applies a finite-population correction.

n₀ = z² × p × (1 − p) ÷ e² n = n₀ ÷ [1 + (n₀ − 1) ÷ N]

Where: n₀ is the unadjusted sample; n is the finite-population sample; z is the z-score for the selected confidence level; p is the expected issue proportion as a decimal; e is the margin of error as a decimal; and N is the population size. The displayed sample is rounded up to the next whole item.

The formula estimates a population proportion under a probability-sampling model. It does not account for design effects, clustering, stratification, nonresponse, mandatory minimum samples, or risk-based judgmental sampling.

What the result means

The result is the rounded number of accessibility audit items suggested by the selected statistical proportion-estimation assumptions.

This calculator is a planning aid. It does not determine legal obligations, establish a required compliance period, or replace advice from qualified legal, accessibility, environmental, procurement, or compliance professionals.

Example calculation

Given: a population of 1,200 items, 95% confidence, a 5% margin of error, and an expected issue rate of 50%.

Calculation:
n₀ = 1.96² × 0.50 × (1 − 0.50) ÷ 0.05² = 384.15.
Finite-population n = 384.15 ÷ [1 + (384.15 − 1) ÷ 1,200] = 291.18.
Round up = 292 items.

Result: Suggested statistical sample: 292 items.

Interpretation: If items are selected under an appropriate probability-sampling design, this sample size targets the selected confidence and precision for estimating the issue proportion. It does not guarantee that every type of defect will be represented.

FAQ

Why does a 50% expected issue rate often produce the largest sample?

For a proportion, statistical variance is highest at 50%. When there is no defensible prior rate, 50% is therefore a conservative input for this specific sample-size formula.

Can I use this sample for targeted high-risk items?

You can test high-risk items, but a judgmental or targeted sample does not have the same statistical interpretation as a random probability sample. If both objectives matter, consider separate statistical and risk-based components.

What happens when the population is small?

The finite-population correction reduces the required sample as the proposed sample becomes a meaningful share of the full population. The calculator also prevents the rounded sample from exceeding the population.

Does the margin of error mean the audit will miss no more than that percentage of problems?

No. Margin of error describes uncertainty around an estimated population proportion under the statistical model; it is not a defect-detection guarantee for individual issue types.

When should I follow a different sample-size rule?

Use a mandated or formally adopted method whenever a regulator, contract, audit standard, testing protocol, or internal methodology specifies one. This calculator is for a general proportion-estimation model when no controlling method has been supplied.

Related tools