Co-authored by Steve Shardonofsky and Rebecca DeGroff

Last week, in Sirko v. IBM, a federal district court in California rejected the plaintiffs’ efforts to use a rudimentary survey to establish Rule 23 class certification because the survey — designed and administered by plaintiffs’ counsel — “lack[ed] basic indicators of reliability.”  The case is yet another example of the trend by class action plaintiffs to use surveys and statistical evidence to avoid issues of individual proof and establish common liability at the certification stage.  More importantly, it reaffirms (like the recent Duran decision from the California Supreme Court discussed here and here) that any such surveys must be sufficiently reliable to be considered as part of the court’s rigorous analysis under Rule 23.  Striking the survey, the court denied the plaintiffs’ bid for class certification, because resolving the misclassification issues in the case would require too many individual inquiries.

Surveys and Other Types of Statistical Evidence at Class Certification Must be Reliable

The Sirko plaintiffs were employed by IBM to service Kaiser Permanente’s health information computer systems.  They sought to certify a Rule 23 class of exempt California-based IT employees who were allegedly misclassified and denied overtime by IBM.  To support their bid for class certification, plaintiffs’ counsel designed and sent a 47-question survey regarding work duties to the putative class members.  This, according to U.S. District Judge Dolly M. Gee, was “their foundational evidence in support of their motion for class certification.”  Judge Gee struck the survey, however, because it “lack[ed] basic indicators of reliability.”

The survey’s fatal flaws included:

• Plaintiffs’ counsel, created and administered the survey, but admittedly was not an expert in survey methodology;
• The vast majority of questions allowed only for “yes” or “no” responses, yet included questions (like one regarding the exercise of discretion and independent judgment) that required “a less binary response;”
• The cover letter accompanying the survey told recipients the purpose of the survey was to support a class action seeking overtime wages for employees who had been misclassified as exempt.  As Judge Gee observed, this information “undermine[d] any possible inference that the survey responses were objective,” because the survey recipients “had to have been aware they would be potential beneficiaries of such a lawsuit.” (Consider, for example, how the workers’ answers may have differed if they had been told the survey would be used to determine merit pay raises or promotions.)

These flaws rendered the survey unreliable, leading the court to strike portions of plaintiffs’ experts reports based on the survey.

Individual Inquiries Doom Class Certification Yet Again

Having struck the plaintiffs’ survey, Judge Gee turned to analyze the putative class members’ job duties based on declarations submitted by the parties.  This evidence revealed that while putative class members shared a job code, they actually performed a variety of duties in different positions.  For example, one named plaintiff spent all of his time assembling teams that responded when Kaiser’s computer systems experienced outages.  By contrast, the other named plaintiff spent most of his time making upgrades and installing updates to IT systems.  In addition, the declarations submitted by IBM also revealed differences in the work duties and responsibilities of putative class.  Thus, Jude Gee found, to determine whether the plaintiffs’ work satisfied the requirements of the administrative exemption could “only be answered on an individualized basis.”  Absent “common proof on a class-wide basis,” Judge Gee held that Plaintiffs could not establish commonality or superiority under Rule 23(b).

Implications For Future Wage/Hour Cases

Like the recent Duran case, the Sirko ruling is another example of class action plaintiffs misusing surveys and statistical evidence to establish Rule 23 certification.  Yet, in another welcomed ruling for employers, the case reaffirms that any such evidence must be reliable, and carefully evaluated and scrutinized by the court, before it can be considered even at the class certification.  Using surveys is somewhat common in wage and hour cases.  Surveys require time: time to carefully construct, time to properly test, and time to challenge and dispute.