Four-fifths rule calculator

How the calculation works

The impact ratio is the simplest defensible measure of whether a selection process treats groups differently. Three steps:

1. Selection rate for each group = number selected divided by number of applicants.
2. Reference group = whichever group has the highest selection rate. It becomes the denominator and its own ratio is 1.00 by definition.
3. Impact ratio = each group's selection rate divided by the reference rate.

Under the EEOC's Uniform Guidelines on Employee Selection Procedures, a ratio below four-fifths — 0.80 — is generally regarded as evidence of adverse impact. The guideline dates to 1978 and long predates machine learning; what is new is that automated tools now produce these disparities at scale and, increasingly, that the law requires you to measure and publish them.

What the calculator does not do

This tool computes impact ratios for the categories you enter. A compliant Local Law 144 bias audit requires more:

• Intersectional categories. Not just sex and race separately, but their combinations — Hispanic women, Black men, and so on. This is where disparities frequently appear that single-axis analysis misses entirely.
• Scoring rate analysis for tools that produce scores rather than pass/fail outcomes, measured as the proportion of each group scoring above the median.
• Treatment of unknown-category individuals, which must be counted and disclosed.
• Independence. An audit you run yourself does not satisfy the statute. The auditor must not hold an employment or financial relationship with you that compromises independence.

Use this to understand your exposure before commissioning an audit — not as a substitute for one. The full compliance guide covers what a defensible audit actually requires.

Statistical significance and small samples

Impact ratios computed on small applicant pools are unstable. With twenty applicants in a group, a single hiring decision can move the ratio by 0.15 or more. The Uniform Guidelines acknowledge this, and practitioners typically supplement the ratio with a significance test — a two-proportion z-test or Fisher's exact test — when group sizes are small.

A practical rule: treat ratios computed on fewer than about thirty applicants per group as directional rather than conclusive, and say so in whatever you publish.

This calculator is provided for general educational use. It is not legal advice and does not constitute a bias audit under NYC Local Law 144 or any other law.