When sample sizes are small, a change of only a few individuals could result in different adverse impact outcomes. Practical tests help guide decisions regarding the existence of adverse impact when samples are small. They help balance the Type I error associated with the 4/5ths rule and the power limitations of statistical test as a result of small samples. There are two practical tests of adverse impact that can be used to determine the extent to which adverse impact is a result of fluctuations due to small samples.
The N of 1 rule (or flip-flop rule; cf., the Uniform Guidelines Question and Answer #21 and Roth, Bobko, & Switzer, 2006) calculates an adjusted impact ratio assuming one more person from the minority group (where minority refers to the group with the smallest selection ratio) and one less person from the majority group were hired (and, consequently, one less minority and one more majority were not hired). If the resulting selection ratios are such that the minority selection ratio is now larger than the majority selection ratio, selection rate differences may be attributed to the small sample size.
The one-person rule (cf., Roth, Bobko, & Switzer, 2006) is computed by taking the difference between actual minority hires (again, where minority refers to the group with the smallest selection ratio) and the expected frequency of minority hires (rounded down to the nearest whole number). If the difference is less than 1, selection rate differences may be attributed to the small sample size.
Use Adverse Impact Analysis to estimate adverse impact using practical tests now.