Google settles class action pay disparity lawsuit

On June 10, 2022 Google agreed to settle a class action pay discrimination lawsuit for $118 million.  The plaintiffs in Ellis vs. Google, LLC alleged compensation discrimination against women.  In their September 14, 2017 complaint filed with the Superior Court of California, plaintiffs claimed “systemic discrimination” where “Google pays women less than men for comparable work.”  Although plaintiffs sought a jury trial, after five years the parties ultimately arrived at this settlement.

I was quite curious to find the circumstances leading to this settlement.  After all, Google has presented itself as a progressive, inclusive organization and a great place to work.  As I was reading the complaint, I was startled to see under the FACTUAL ALLEGATIONS section that the Office of Federal Contractor Compliance Programs (OFCCP) had initiated a compliance review of Google’s affirmative action plan in 2015.  OFCCP is the office in the US Department of Labor empowered to audit affirmative action plans as well as to bring enforcement action against contractors who are out of compliance.  

Plaintiffs in Ellis write: “OFCCP performed a statistical regression analysis of the compensation data for all approximately 21,000 employees at Google’s Mountain View office for the year 2015” and “‘found systemic compensation disparities against women pretty much across the entire workforce.’  OFCCP’s analysis showed six to seven standard deviations between pay for men and women in nearly every job classification in 2015.” [My italics]

In an analysis of the type performed by OFCCP, a finding of greater than two standard deviations is considered a statistically significant difference.  With a finding of six to seven standard deviations difference the probability of this degree of disparity occurring by random chance is “one in 100 million”. Which is to say we can be confident that this disparity did not happen by chance, but rather there are factors or practices in the compensation system which led to a disparate impact upon women, and a quite dramatic impact at that.

Affirmative action plan analysis can warn of gender disparity

As a federal contractor, Google is required to develop an affirmative action plan (AAP) and update it annually.  A compensation analysis is a required component of an AAP.  The level of mathematics employed in a compensation analysis can be found in an elementary statistics textbook (albeit in the later chapters).  Given the number of people at Google who are mathematically inclined, I have to believe the math was not beyond them.  Even in the unlikely event (p-value < 0.05) that Google didn’t know how to compare two sample means or do a regression, there are plenty of vendors who provide software to perform this analysis.  But isn’t it difficult compiling data to analyze 21,000 employees?  Having conducted such an analysis myself for 3,000 employees (using software), I can say, “Yes, it is.”  But isn’t Google a pioneer in Big Data and Artificial Intelligence?

Google had the obligation to produce and update an AAP.  Google certainly has the wherewithal to do this, so they knew or should have known of the pay disparity among their female* employees and should have taken action to rectify the situation.  The OFCCP compliance review and resulting enforcement action culminated in a $3.5 million settlement.  Let us remember that when there is a settlement, there is no admission of guilt.  But let us also remember that the force of statistical analysis compelled Google to write a check for millions of dollars.

Know what your affirmative action plan is telling you.

*OFCCP found statistically significant disparity in pay and hiring among women and Asians.