To apply it, calculate the selection rate for each group at a given step: the number who passed divided by the number who applied. Then divide each group's rate by the highest group's rate. If 50% of one group pass a CV screen and 35% of another do, the ratio is 0.7, below the 0.8 threshold, which flags the step for closer analysis.
The rule comes from the US Uniform Guidelines on Employee Selection Procedures and is a rule of thumb, not a legal test on its own. Small samples can mislead, and regulators may also use statistical significance tests. It is still a practical first check: teams running automated screening can monitor ratios by stage, investigate any flagged step and keep records of what they found and changed.