Petit test
Guess first, no penalty. An AI screen passes 30% of male applicants and 20% of female applicants. Is that a warning sign?
When a tool screens applicants, check how each group does at that step. A selection rate is the share of a group's applicants who pass.
The US Uniform Guidelines on Employee Selection Procedures give a rule of thumb. A group's selection rate that is "less than four-fifths (4/5) (or eighty percent)" of the highest group's rate is generally treated as evidence of adverse impact (29 CFR 1607.4(D)). Adverse impact means a step screens out one group much more than another. It's a signal to dig in, not proof of discrimination, and small numbers can mislead.
Work it out
Et si…
Compare two groups at one screening step. The starting numbers are an example.
Group with the higher rate: applicants
200
Group with the higher rate: passed
60
Other group: applicants
150
Other group: passed
30
Other group's rate as a share of the higher rate
Other group's rate as a share of the higher rate: 66.7%
Une estimation à partir de vos chiffres, pas une prévision.
Here's the note you'd send with the example numbers:
Women's selection rate at the AI screen: 20% (30 of 150). Men's: 30% (60 of 200).
20 divided by 30 is about 67%, which is under four-fifths, so we're reviewing the screen.Complétez les trous
Fill the gaps in this note to the hiring manager.
Note
Women's selection rate at the AI screen: . Men's: 30%. 20 divided by 30 is about , which is .
Try it
Vos étapes, environ 2 minutes
- 1.Open the Screening bias check lab. Change one screening rule and watch each group's selection rate move.1 min
- 2.Put one real screening step's numbers into the calculator above, or use the example. Add the note to your fair hiring kit. Your kit now holds a fair job post, structured interview questions with a scoring guide, a summary prompt and a fairness check.1 min