Huge Win for Federal Contractors: Statistical significance alone is not enough to prove a claim of discrimination
In an earlier blog in this alert, Art Gutman summarized the April 21, 2016 Department of Labor Administrative Review Board (ARB) ruling related to a long-standing set of OFCCP allegations against Bank of America (BOA). This ruling serves as a warning to the Office of Federal Contract Compliance Programs (OFCCP) and other federal agencies to rely on more than just statistical significance when pursuing employment discrimination claims.
This ruling trumps the Administrative Law Judge (ALJ) ruling back in 2010 that supported the use of statistics demonstrating a disparity of two or more standard deviations as enough to establish a prima facie case of unlawful discrimination (see our previous blog for more details on that issue). In 2010, the ALJ found BOA in violation based on statistical evidence during the time periods of 1993 and 2002-2005. Why did they rule differently for the 2002-2005 period this time around?
Statistical versus practical significance
A key reason for the 2002-2005 period ruling of insufficient evidence is due to the extremely small shortfalls. Shortfalls give the number of expected hires for the disadvantaged group if selection rates across all groups were equal. The larger the shortfall, the more practical impact a difference in selection rates has. The ALJ noted “in 2003, 44 African Americans were offered a job instead of the expected number of 47.9; in 2005, 32 instead of 34.5. Those are small shortfalls.” In other words, even though the OFCCP found a statistically significant difference, the actual discrepancy in terms of hires was less than five people in a given year. Our takeaway from this is a precedent is set for courts to not rely on statistical significance in isolation: practical significance is another key piece of evidence.