In the recent Tenth Circuit Court of Appeals case, Apsley v. Boeing Co., the Court rejected statistical evidence offered by a class of approximately 700 older workers because the evidence inappropriately lumped together thousands of hiring decisions.
Plaintiffs’ statistical evidence analyzed the multi-facility workforce together as a whole and concluded there were five standard deviations (extremely high in the world of statistics) between the number of older workers who should have been recommended by Boeing for rehire and those actually rehired. The Tenth Circuit, however, disagreed with the plaintiffs’ aggregated approach.
Unfortunately, most courts (like many attorneys) freeze up when presented by Plaintiffs with big numbers and arrive at a conclusion without fully understanding the implications of the concept that “Big Numbers Are Bad Numbers.”
Statistically, when large numbers of selections (e.g., hiring, promotions or terminations) are grouped together, even small differences in selection rates tend to show statistically significant adverse impact, which the Plaintiffs in this case tried to rely on to establish an inference of systemic discrimination. The Court, however, understanding the implications of this concept, concluded Plaintiffs statistics were “inaccurate or insignificant” and did not support a finding of discrimination.
It remains to be seen whether judges in other jurisdictions will follow suit and likewise consider the implications of data aggregation in employment cases, but this decision is a good sign that – in this new age of robust HRIS and applicant tracking systems and the resulting massive amounts of data that can be analyzed – at least some courts will not easily accept plaintiffs’ aggregated statistics.