Class Certification Denied Despite Prior OFCCP Audit

The case is Randall v. Rolls Royce, decided on March 12, 2010. The facts of this case are hardly extraordinary. Two named plaintiffs at Rolls Royce sued for systemic sex discrimination under Title VII and unequal pay for equal work under the Equal Pay Act on behalf of a class size approximated at 537 women. District Court Judge Sarah Evans Baker of the Southern District of Indiana rejected class certification on failure to show commonality, typicality, and adequate representation. Essentially, the plaintiff’s expert did a regression analysis by aggregating men and women from five job classifications, each with four pay grades, and failed to consider important covariates, most notably distinct jobs with pay grades. The defendant’s expert included the covariates, and the rejection of class certification was based on his analysis.


Clearly, this happens a lot. One of the reasons we report this case is that Rolls Royce had previously entered into a sex discrimination settlement with the OFCCP based on a compliance evaluation. In commenting on the prior OFCCP audit, which the plaintiffs offered as evidence in the form of a press release, Judge Evans ruled:

Both the relevance and the admissibility of the press release is questionable in part because we do not know whether some, none, or all of the putative class members were actually employed by the Defendants' predecessor at the time of the OFCCP audit, let alone what compensation-setting polices were being followed by that company at the time. Moreover, the content of the press release does less to show the existence of discriminatory salaries at the start of the time period covered by the class definition than it does to suggest that any gender-based discrepancies may have been rectified by the nearly half million dollars in payments made by Defendants' predecessor as a result of the OFCCP compliance program. In any event, without greater detail as to what was involved in the audit and what the resulting agreement purported to cover, we can not assign any significance or weight to the press release.


Here are some other excerpts from the judge’s ruling, they are revealing.

Rolls-Royce conducts annual reviews of every individual employee's performance, in part to determine whether the employee's performance merits a raise. As part of the evaluation process, employees are assigned a Performance Development Review ("PDR") score by their manager. This score governs the amount of a merit increase the employee will receive. In addition to this annual review process, managers at Rolls-Royce have discretion to award various types of adjustments to an employee's base salary, such as an "equity adjustment" to modify salary to make it commensurate with the market and reflective of the scope of the employee's job duties. Managers can also make a critical skills adjustment ("CSA") for top performing employees who have skills critical to the business and are top performers. [the plaintiff expert’s]' report omits any consideration of these additional discretionary compensation tools in terms of gender; rather, it addresses only the base salary differentials comparing the two genders.


[the plaintiff expert] has reached his conclusions regarding the common gender effect on annual salaries within the defined class based on a regression analysis, which takes into account certain individual neutral factors (independent variables) such as seniority, time-in-grade, performance ratings, personnel area and education level. Defendants take issue with the way in which [the plaintiff expert] defines certain of those independent variables. Nevertheless, based on the neutral factors as he defines them, [the plaintiff expert] concludes that the salaries of men are typically somewhere between 4.27% to 5.84% higher during the time period of 2004 through 2007 throughout the various salary grades.


In addition to challenging the manner in which [the plaintiff expert] defines some independent factors, Defendants also take issue with [the plaintiff expert's] failure to include other neutral factors. Defendants contend that there are additional factors which are clearly significant and must be included in comparing salaries, such as the particular jobs held within a pay grade or an employee's experience in a particular job. They also contend that the interactions between and among these variables are crucial and must be incorporated into any statistical analysis which attempts to explain the reasons for particular salary differences.


Additionally, in one of the footnotes, Judge Evans makes the following distinction between adverse impact and pattern or practice claims as it relates to intentional discrimination:

Plaintiffs assert a disparate impact claim in addition to a pattern and practice claim. Both claims include the assertion of systemic discrimination. However, most courts, including this one, view a pattern and practice claim to require a showing of intentional discrimination (proof of which may include evidence of disparate impact), whereas a disparate impact claim involves identifiable facially neutral policies which, when applied, lead to discriminatory results. See Mozee v. American Commercial Marine Service Co., 940 F.2d 1036, 1051 (7th Cir. 1991); see also 1 Charles A. Sullivan & Lauren M. Walter, Employment Discrimination Law and Practice § 3.04(E) (4th ed. 2009).


The bottom line here is that the plaintiff’s expert relied on a simple analysis failing to control for critical covariates (on independent variables).


Interestingly enough, both the appropriateness of employee grouping structures and of multiple regression analyses were critical factors in this Title VII case. These factors are described in detail in the OFCCP systemic compensation discrimination Standards published in 2006. These Standards are currently in the process of being rescinded, and will likely be replaced by more general Title VII standards. Having said that, it is hard to imagine a Title VII systemic compensation case where these two factors would not be probative (assuming reasonable sample size).


by Art Gutman Ph.D., Professor, Florida Institute of Technology

    Stay up-to-date with DCI Alerts, sign up here:

    Advice, articles, and the news you need, delivered right to your inbox.


    Stay in the Know!