Nos. A125557, A126827.

203 Cal.App.4th 212 (2012)

137 Cal. Rptr. 3d 391

SAM DURAN et al, Plaintiffs and Respondents, v. U.S. BANK NATIONAL ASSOCIATION, Defendant and Appellant.

Court of Appeals of California, First District, Division One.

Attorney(s) appearing for the Case

Carlton DiSante & Freudenberger, Carothers DiSante & Freudenberger, Timothy M. Freudenberger , Alison L. Tsao and Kent J. Sprinkle for Defendant and Appellant.

Ellen Lake ; Wynne Law Firm, Edward J. Wynne and J.E.B. Pickett , for Plaintiffs and Respondents.



Following a bifurcated bench trial, defendant U.S. Bank National Association (USB) has appealed the resulting $15 million judgment in this wage and hour class action brought under Business and Professions Code section 17200 (section 17200). The plaintiffs in the class action are 260 current and former business banking officers (BBO's) who claimed they were misclassified by USB as outside sales personnel exempt from California's overtime laws, and were thus unlawfully denied overtime pay. In addition to arguing the case should not have been certified as a class action, USB contends the trial court's trial management plan deprived it of its constitutional due process rights in that the plan prevented it from defending against the individual claims for over 90 percent of the class. We agree the trial management plan was fatally flawed and reverse the judgment. We also conclude the case must be decertified, and reverse an order awarding certain expert witness fees to plaintiffs.1 We remand the two named plaintiffs' meal and rest break period violation claims for reconsideration in light of the California Supreme Court's ruling in Brinker Restaurant Corp. v. Superior Court (2008) 165 Cal.App.4th 25 [80 Cal.Rptr.3d 781], review granted October 22, 2008, S166350.2


In January 2005, when plaintiffs moved to certify this case as a class action, USB reportedly operated over 130 traditional bank branches in California with approximately 40 persons then currently employed as BBO's. Each BBO was typically assigned to work with one to four bank branches. BBO's report to sales managers who in turn report to regional managers. Regional managers report to a division manager who then reports to the executive vice-president.

USB applies a uniform wage exemption to BBO's, categorizing them as exempt outside salespersons to whom USB is not obligated to pay overtime and related wages.3 The outside salesperson exemption is set forth in Labor Code section 1171, which states: "The provisions of this chapter [(including Lab. Code, § 1194)]4 shall apply to and include men, women and minors employed in any occupation, trade, or industry, whether compensation is measured by time, piece, or otherwise, but shall not include any individual employed as an outside salesman . . . ." California Industrial Welfare Commission wage order No. 4-2001 (Wage Order No. 4) (codified at Cal. Code Regs., tit. 8, § 11040) defines an "outside salesperson" as a person "who customarily and regularly works more than half the working time away from the employer's place of business selling tangible or intangible items or obtaining orders or contracts for products, services or use of facilities."5 (Cal. Code Regs., tit. 8, § 11040, subd. 2(M), italics added.)

The BBO classification was created after USB merged with another bank in 2001, when an existing small business banker (SBB) position was redesignated with the BBO title. A May 1997 job description states that SBB's were responsible for "developing and managing small business banking customer relationships by advising on and selling a wide range of financial services to include: credit, deposit, cash management and other applicable bank products and services." Persons in this classification were also responsible for "growing the business through prospecting, networking, cross-selling and relationship management in order to develop and expand the relations with [USB]." SBB's were encouraged to "build strong customer relationships and analyze data to determine the customer's business needs and financial condition." This job description was essentially retained when the BBO position was created in 2001.6 However, a 2002 job description specifies that more than 80 percent of a BBO's time should be spent on "Outside Sales Activity."7 Of the five functions listed in the job description under this heading, several are required to take place outside branch offices.8


I. The Complaint

On December 26, 2001, Amina Rafiqzada, acting on behalf of herself and other similarly situated current and former USB employees, filed a class action complaint against USB. The complaint alleges Rafiqzada had been employed as an SBB, and that USB improperly classified her and other SBB's as "exempt" thereby denying them compensation for overtime hours in violation of Labor Code section 1194. The complaint contains three causes of action: (1) violation of the Labor Code for misclassification and failure to pay overtime, (2) violation of Business and Professions Code section 17200, and (3) conversion.

On February 26, 2003, a first amended complaint (FAC) was filed, removing Rafiqzada as the named plaintiff, and substituting Vanessa Haven, Abby Karavani, and Parham Shekarlab as the named plaintiffs. The amended complaint alleges the same causes of action as the prior complaint.

On March 28, 2003, USB filed its answer to the FAC. USB generally denied the allegations of the complaint and asserted 31 affirmative defenses. As its seventh affirmative defense, USB claimed the proposed class members were, at all relevant times, properly classified as exempt employees for overtime purposes under California law.9

II. Contested Class Certification

On January 6, 2005, plaintiffs filed a motion to certify the case as a class action.10 The motion was supported by declarations from 34 current and former BBO's who indicated they regularly worked overtime hours and spent less than half their worktime engaged in sales-related activities outside of branch offices.

That same day, USB filed a motion seeking to deny class certification. As part of its motion, USB claimed class certification should be denied because plaintiffs could not establish that common issues predominate. USB also contended the named plaintiffs were not typical of the class because all four of the current and former named plaintiffs had admitted, at depositions or in interviews, facts establishing they were properly classified as exempt employees under the outside salesperson exemption. In its motion, USB included declarations signed by 83 putative class members who described their job duties. Of these declarants, 75 stated they regularly spent more than half their time engaged in sales activities outside USB branch offices.

On March 14, 2005, plaintiffs' counsel filed a second amended complaint (SAC), substituting two new class representatives, Sam Duran and Matt Fitzsimmons, in place of the previously named plaintiffs. The SAC contains the same causes of action as the prior complaints, and adds new allegations asserting Labor Code violations for failure to provide meal and rest break periods. (See Lab. Code, § 226.7.)

On March 16, 2005, the trial court granted plaintiffs' motion for class certification and denied USB's motion to deny class certification.11

On April 29, 2005, the trial court filed an amended and corrected order after hearing regarding its certification rulings. The order defines the class as "`all current and former California-based salaried employees with the title "small-business banker" (SBB's) and/or "business banking officers" (BBO's) employed by defendant any time between December 26, 1997 and April 28, 2005.'"12 The order states the court found the proposed class to be both ascertainable and numerous, and that common questions of law and fact predominated over individual legal or factual disputes.13

III. Summary Adjudication of Two Exemption Affirmative Defenses

On September 30, 2005, plaintiffs filed a motion for summary adjudication on USB's affirmative defenses that the BBO's were exempt from overtime laws under the administrative and the commissioned salesperson exemptions.14

On February 17, 2006, the trial court filed its order granting plaintiffs' motion for summary adjudication as to the commissioned salesperson exemption. The court deferred ruling on the administrative exemption to allow USB an opportunity to conduct a limited amount of discovery relevant to its opposition.

On May 23, 2006, the trial court granted plaintiffs' motion for summary adjudication as to the administrative exemption, finding the evidence showed the performance of administratively exempt duties was atypical for BBO's.

IV. Pretrial Proceedings

A. The Trial Management Plan Is Selected

On June 22, 2006, plaintiffs submitted a trial setting conference brief. In response to USB's pretrial proposal to assign all class members to groups of 20 to 30 in order to conduct individualized evidentiary hearings before special masters, plaintiffs, citing to Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 337 [17 Cal.Rptr.3d 906, 96 P.3d 194] (Sav-On), asserted that USB "has no due process right to assert its affirmative defense against every class member in a class action."15 Instead, plaintiffs proposed a "phased trial plan using surveys and random sampling."

Under plaintiffs' plan, the trial would have proceeded in three phases: (1) Task identification and classification, (2) use of a classwide survey followed by the selection of a random sample of plaintiffs who would be made the subjects of the trial, and (3) damages. The survey instrument would "question the class members about the amount of time class members spent out of the branch and the amount of time spent on sales or sales related activities out of the branch." Plaintiffs included a declaration from their expert statistician Richard Drogin, in which he proposed a statistical methodology for assessing both liability and damages.

On September 7, 2006, the parties filed a joint case management and trial management statement in which USB noted its objections to plaintiffs' proposed use of a survey to gather data from which a representative sample of class members would then be selected to determine liability and potential damages for the entire class. USB claimed plaintiffs' proposed methodology was statistically unreliable, and was infeasible given the amount of time remaining before trial. USB also argued the proposed procedure would violate its due process right to assert its affirmative defense against each individual class member who sought to recover damages.16

On September 13, 2006, a case management conference was held in which the trial court, on its own initiative, proposed the idea of taking a sample of 20 plaintiffs selected on a random basis to testify at trial. The court suggested this method would be preferable to conducting a survey, due to the possibility that bias might infuse the survey process.

On October 6, 2006, USB submitted a brief in which it again argued that representative testimony could not be fairly utilized in this case because it would violate USB's due process rights. USB further noted that the only California case relied on by plaintiffs in support of the use of representative testimony in an overtime case (our opinion in Bell v. Farmers Ins. Exchange (2004) 115 Cal.App.4th 715 [9 Cal.Rptr.3d 544] (Bell III)), involved the issue of class action damages only, and not liability. USB stated that if a representative sample were to be used, USB must be permitted to call witnesses of its own choosing in order to meet its burden of proof. USB submitted another declaration from their statistical expert, Phillip Gorman, who pointed out the possibility of "sampling error" if a small number of randomly selected class members were to be used to determine whether the entire class had been misclassified as exempt. Gorman further noted that even after the liability phase it would still not be possible to determine which nonsampled class members would be entitled to compensation.

At a hearing held on October 11, 2006, the trial court confirmed its intent to use a random sample of 20 class members to testify as representatives for the class. To choose the representatives, the court proposed putting the names of all the potential class members into a "hat" and drawing 20 names, along with five additional names to serve as alternates in case any of the initially selected plaintiffs were unavailable. Plaintiffs' counsel then informed the court that he intended to dismiss the SAC's legal claims and proceed to bench trial under section 17200 only.

B. Second Opt-out Notice

After a hearing in which the trial court granted plaintiffs leave to dismiss the legal claims, the court required a second opt-out notice be sent to the class notifying it of the dismissal.17 The court clerk then drew from a batch of index cards containing the names of each class member and compiled a list of 20 class representatives and five alternates.

On November 9, 2006, plaintiffs filed a motion to amend the SAC to include allegations that USB failed to provide meal and rest break periods to class members, and to conform the complaint in light of the dismissal of its legal claims.

On November 28, 2006, the parties submitted a joint case management and trial management statement.18 USB asked that the second opt-out notice be sent without additional opt-out rights, noting that giving class members another opportunity to opt out "would [give] incentive [to] those who have been randomly selected to opt out based on their unwillingness to participate in the legal proceedings." On November 29, 2006, the trial court granted the motion to amend the SAC.

On November 30, 2006, plaintiffs filed their third amended complaint (TAC). The TAC asserts a single cause of action for violations under section 17200.19 The bases for the Business and Professions Code section 17200 claim are alleged violations of Labor Code sections 204 (payment of wages), 216 (refusal to pay wages due), 226.7 (meal and rest periods), 1194 (overtime), 1198 (maximum hours), and 1199 (failure to comply with labor laws).

C. Results of Second Round of Opt Outs

On January 9, 2007, USB filed a trial management conference statement. USB reported that Steven Dias, one of the 20 randomly selected class members, had since opted out of the lawsuit. USB asserted that any member of the 20-person group who chose to opt out after receiving the second opt-out notice should be required to provide deposition and trial testimony in order to preserve the representative nature of his or her testimony and to ensure statistical reliability of the extrapolation process.

On January 29, 2007, USB filed its answer to the TAC.

On February 6, 2007, USB filed a motion regarding the status of the 20-person random witness group (RWG) members. USB asked the trial court to allow four RWG members, who had by now opted out of the action, the opportunity to opt back in. USB claimed the witnesses had decided to opt out because they felt they were properly classified as exempt employees, and because plaintiffs' counsel allegedly encouraged them to avoid involvement in the lawsuit.20 USB argued the representative nature of the RWG was compromised because witnesses whose testimony would have supported its affirmative defense were now omitted from the representative group. A declaration provided by USB's statistical expert, Phillip Gorman, noted that while four of the 20 RWG members had elected to opt out, only five of the remaining 250 absent class members had done so. Thus, a much smaller percentage of non-RWG plaintiffs had opted out (2 percent), relative to the RWG, which saw a 20 percent opt-out rate. Gorman stated that removing these four witnesses would undermine the accuracy of any extrapolation process.

In opposition to USB's motion, plaintiffs submitted a declaration from Drogin, who opined that substituting the alternate RWG members for the four opt-out plaintiffs would be statistically acceptable as there was "no reason to infer that the sample is not representative, or that there is any bias in the sample." He remarkably concluded that "as long as the set of persons selected to testify at trial includes those in the original random selection made by the court, and is restricted to those in the class, the testifying group will be a random sample of the class. Under these conditions, the results determined by the court for those testifying can be reliably projected to the class as a whole." He deemed Gorman's statements to the contrary to be both incorrect and irrelevant.

On February 16, 2007, the trial court denied, without prejudice, USB's motion to allow Michael Lewis and Sean MacClelland to opt back into the lawsuit, noting that the issue of whether these witnesses would be allowed to testify should be addressed at trial.

On March 1, 2007, plaintiffs filed a motion for class certification of the meal and rest break claims. They also filed a motion to compel further responses to special interrogatories, requests for production of documents, and requests for admissions.

D. First Motion to Decertify the Class

On March 22, 2007, USB filed its opposition to plaintiffs' motion for class certification of the meal/rest break claims. USB also filed a motion to decertify the class action. In its motion to decertify, USB contended Proposition 64 requires each class member to prove "injury in fact."21 USB also argued that then recent case law affirmed the need for individualized inquiry in exemption cases. In connection with this argument, USB relied heavily on Walsh v. IKON Office Solutions, Inc. (2007) 148 Cal.App.4th 1440, 1456 [56 Cal.Rptr.3d 534] (Walsh), in which the Court of Appeal upheld an order decertifying a subclass of account managers in a wage and hour dispute. USB also included a declaration of Andrew Hildreth, an expert in the area of inferential statistics, who opined that "proceeding with the sample of 20 class members randomly selected by the court to determine the results for the entire population (class) would, from a statistical perspective, be highly inaccurate and unrepresentative."

On May 16, 2007, the trial court filed its order denying USB's motion for decertification and denying plaintiffs' motion to certify the meal/rest break claim. The order does not discuss either Walsh or Dunbar, and does not assess the issues raised by Hildreth's declaration.22 Instead, the court indicated it was unpersuaded by USB's claim that individualized inquiries predominated over common questions with respect to liability.

V. Motions in Limine

On May 18, 2007, the parties filed their motions in limine.

In their motion in limine No. 11, plaintiffs sought to exclude the introduction of testimony or declarations from, or evidence or argument related to, any non-RWG BBO. Plaintiffs claimed the introduction of such evidence "would debase the trial and destroy the court's representative witness trial plan." Motion No. 11 was granted over USB's objections. In granting the motion, the trial court indicated it would allow USB to call any percipient witnesses for impeachment purposes, but would not allow non-RWG class members to increase the "bank of data" from which expert witnesses would ultimately draw their conclusions. In arriving at this decision, the court appears to have relied on Bell III, noting it had reread that case with interest "because of its detailed discussion and approval of the trial management plan."

In its motion in limine No. 3, USB sought to require plaintiffs to rely on the testimony of all the witnesses originally selected for the RWG, to the extent they were willing to testify. The motion essentially claims the composition of the RWG members had been altered so much since the time of its original selection that it now lacked representative integrity. The court denied this motion.

In its motion in limine No. 4, USB sought to exclude testimony regarding the trial court's methodology of sampling. The motion reiterated USB's claims that the trial plan, particularly with respect to the issue of classwide liability, failed to comport with the standards set by our opinion in Bell III because it was unreliable, was not generally accepted in the scientific field, and was not supported by a report from any expert witness. The trial court denied this motion and declined to revisit its trial management plan. The court also ruled the two named plaintiffs would be allowed to testify and to be present during the entire trial.

VI. Phase I of the Trial

The liability phase (Phase I) of the trial began on May 24, 2007. All but one of the RWG members appeared at trial. They testified that they had spent more than half of their worktime inside bank offices and had worked varying amounts of overtime. We summarize the relevant testimony of some of these witnesses.

A. Plaintiffs' Case

1. Matt Fitzsimmons

Matt Fitzsimmons, one of the two named plaintiffs, was a BBO from April 2003 to March 2004.23 He testified that when he was hired he was not told of any expectation as to the amount of time he was to spend outside bank property. Nor was he told he was entitled to an off-duty meal break. During his time with USB, he never saw a job description for the BBO position.

Fitzsimmons typically arrived at the bank at 8:30 a.m. Two days a week he would leave by 5:30 p.m. to pick up his daughter from daycare, and the other three days he would leave between 6:00 and 6:30 p.m. On most weekends he would spend two or three hours folding flyers to send to prospective customers. He estimated he generally worked about 45 to 50 hours a week. He was never told to spend most of his time outside the bank. He estimated he worked inside branch offices 75 to 80 percent of the time, with the balance spent outside. There was never a week when he spent more of his time outside as opposed to inside. He also never received a 30-minute meal period in which he was relieved of all duties.

On cross-examination, Fitzsimmons testified that everything he did as a BBO was aimed towards selling USB's products and meeting his quarterly sales goals. He believed he was expected to work inside bank locations because the managers at two of the branches he was assigned to would ask him to come and work at their offices. His supervisor stated in a performance evaluation that he met his "ramp-up" sales goals in his first two quarters but met only 73 percent of his goals in the third, and final, quarter that he worked for USB. Fitzsimmons acknowledged he was one of the poorer performing BBO's in his supervisor's group during his third quarter. His supervisor advised him to make 15 client appointments per week, a strategy that normally results in three loan applications and one loan approval, thereby securing one loan funded per week. Two weeks after his evaluation he applied for a job at Citibank. One of the reasons he left USB was because he could not work the number of hours he believed the bank expected him to.

2. Chad Penza

Chad Penza was a BBO from October 2001 through March 2005. He testified that initially he worked from 7:00 a.m. to around 5:30 p.m. After a few months, he began working from 6:00 a.m. to 6:30 p.m. or 7:00 p.m. He also worked at least three weekends a month, on both Saturday and Sunday, from around 7:00 a.m. to about 12:00 noon. Most of his time was spent calling potential clients on the telephone. From January 2002 to the end of his employment, he estimated he was in the office at least 80 percent of the time. He was never told to spend most of his time outside bank property. After his initial ramp-up period, his quarterly sales figures ranged from $5 million to $10 million. At one point, he was the most successful BBO in the entire company.

On cross-examination, Penza acknowledged he had signed a declaration for USB shortly after he started working as a BBO. In the declaration, he stated he was outside of the office 75 percent of the time. He testified he felt compelled to sign the declaration because he was a relatively new employee at the time. He admitted he provided all the information in the declaration to the attorney who interviewed him. He also signed another declaration in May 2004 in which he reaffirmed his prior declaration. When he signed the second declaration he had several commissions at stake. He chose to work longer hours because he wanted to be the top producer in the country and achieve the numbers that his supervisors wanted.24 On redirect, he denied telling USB's attorney he had ever spent around 75 percent of his working hours outside of the bank.25

3. Troy Petty

Troy Petty worked for USB from 1997 to 2001. He became a USB employee after the bank he worked for merged with USB. At his former bank, his job title was "commercial loan officer." His job duties did not change when he became a USB employee. His employment with USB ended when he retired. During his time with USB, he was never told to spend most of his time outside the branch office.

Petty would make about five to 20 sales calls per week at customers' and potential customers' places of business. He estimated on average he spent 75 percent of his time inside the office. He spent no more than 35 percent of his time outside the office in any given week, or about two or three hours per day. At trial, he testified that he worked 50 to 60 hours per week. In his deposition, he had stated he worked 50 hours in a typical week.

On cross-examination Petty conceded he was expected to work eight hours a day only, and his manager never required him to keep certain hours and never told him what to do with his worktime. USB's counsel introduced the job description for the "commercial banking relationship manager" and attempted to show Petty actually held this position, and was not an SBB. Petty admitted he had never been classified as an SBB.26 He signed a release of all claims when he received his severance package.

4. Matthew Gediman

Matthew Gediman was a BBO from April 2005 until December 2006. At the time of trial, he was still employed by USB as a district sales manager. As a BBO, he was never told that he was required to spend most of his time outside the bank, and he spent the majority of his time inside his office. His hours ranged between 30 to 50 hours per week. He rarely worked more than 40 hours per week or more than eight hours per day.

5. Sam Duran

Sam Duran, the other named plaintiff, was a BBO from April 2003 to March 2004. He was not told he was required to spend most of his time outside of the branch office. In a typical day, he would go to the office, check his e-mail and voice messages, return phone calls, review loan packages, follow up with customers to get documents, and work on loan or line renewals for existing customers. He estimated that he worked 45 hours a week. He was never told that he was expected to arrive at a particular time and he was given flexibility to leave the office to meet with a client, go to doctor's appointments, or pick up his children from school.

When he was outside the office, Duran would meet with customers, pick up documents and review them for completeness, and conduct site inspections. He estimated he spent about 70 percent of his time inside the office. He would also work at least three hours on weekends. He understood that he could take a half-hour meal break when he wanted to. He would take 10- to 30-minute meal breaks, but believed he was expected to be available for work during lunchtime. He left USB voluntarily because he was frustrated with the difficulties in getting loans approved through the underwriting process.

On cross-examination, Duran stated he generally tried to have one meeting with a client outside the branch per day, and sometimes he had two or three such meetings. The meetings would last anywhere from five minutes to three hours, although three hours was very unusual. It would generally take between 15 and 30 minutes to travel from his branch to a customer's or a prospect's place of business. He did not recall telling USB's attorney in July 2003 that he spent, on average, 60 percent of his time outside the office.27

Duran received a signing bonus when he accepted the position with USB. The bonus was conditioned on his remaining employed with USB for 12 months. He left before the 12-month period was up and did not return the bonus. In early 2004, his supervisor counseled him about the deficit in his loan production. She told him that he was expected to conduct 15 quality in-person sales calls on a weekly basis. He did not recall if she told him that these calls should take place at prospects' places of business. He failed to meet this requirement and resigned shortly after a focused counseling session with his supervisor.

Duran also admitted he falsified his USB salary figure when he applied for his next job. He claimed his headhunter told him to lie. The headhunter, David Vallecillo, testified that he contacted Duran about an employment opportunity with another bank. He denied telling Duran to falsify his salary information on the job application.

6. Brett Lindeman

Brett Lindeman began working for USB in May 2005 as a BBO and was still employed at the time of trial. When he was hired he was told the job was a sales position involving new loans. He was never told that he would be expected to spend the majority of his time outside USB properties. He testified he typically works about 40 hours a week. There are times when he works more than 40 hours a week, but he was not able to give an accurate estimate as to how often this occurs. Once or twice a month he takes a call from a client on a weekend. These calls last anywhere from two or three minutes to 40 minutes. More than half of his meetings with clients take place inside his branch office. He estimated that he spends the majority of his time inside USB branches and offices.

Lindeman stated he does hit his sales goals, but not in every quarter. He has never been told that he is expected to work more than 40 hours per week. At his deposition, he said it was "completely plausible" that there were weeks when he spent more than half his time outside of the office. At trial, after viewing his mileage claim records, he stated that he believed he regularly spent more time inside USB properties than outside.

7. Adney Koga

At the time of trial, Adney Koga was employed by USB as a BBO. He was a BBO in California from July 2004 to April 2006. He left USB for a time and came back as a personal banker in October 2006. He became a BBO again in February 2007 in the state of Arizona. He was never informed that he was expected to spend most of his time outside the bank as a BBO.

After he was hired, Koga underwent a three-month training period. He spent about one month shadowing a BBO. During this period he would work from about 9:00 a.m. to about 6:30 or 7:00 p.m. As a BBO in California, he estimated he worked on average about nine hours a day and about 45 hours a week. He estimated that he spent most of his time inside the branch, as opposed to at a customer's place of business. He was never counseled by any of his supervisors that he should spend more of his time outside of the bank location.

Koga had signed a declaration for USB prior to trial. At trial he denied ever having told USB's attorney that he spent 55 percent of his time outside of the office. On cross-examination, Koga admitted he did not make any changes to his declaration even though a cover letter instructed him to correct any inaccuracies. He was not promised any benefit in exchange for signing the declaration and his job was not threatened in any way by anyone at USB if he refused to sign it. Deposition testimony was introduced in which he stated he told USB's attorney that he spent 55 percent of his time outside of the office making sales calls.

Koga testified he had discretion and control over his daily work schedule. He had no set time to arrive at or leave work and no one ever told him the number of hours that he needed to work in a given day or week. He did not report his daily working hours to his manager.

8. Steven Bradley

Steven Bradley worked for USB as a BBO from March 2003 to July 2004, and again from May to August in 2006. When he was hired, he was told the job involved sales and that he would be paid partly on a base salary, but the most important part of his compensation would be derived from the commissions he would get for booking products and services. He was never told he had to spend more than half of his time outside bank locations. His normal work schedule was from around 9:00 a.m. to around 5:30 p.m. Sometimes he worked at home after he left the branch. He estimated he worked between 40 to 45 hours during the week. He also worked once or twice a month on weekends, spending two to four hours folding flyers.

On cross-examination, Bradley acknowledged he had voluntarily signed a declaration for USB. He admitted the statements contained in the declaration were based on information that he provided to the attorney who interviewed him, and that the information was true and correct at the time he signed the document in June 2004.28 Also, at his deposition, Bradley had stated that on average he spent 60 to 65 percent of his time outside the bank, with as much as 80 or 90 percent of his spent time outside at the start of each quarter. After having reviewed his travel expense reports, he testified at trial that he believed he never spent more than 50 percent of his time outside bank property. However, he admitted the expense reports did not reflect the amount of time he spent performing his sales duties while outside the branch. He also acknowledged there were occasions when he worked outside of the office and did not submit a reimbursement claim. He admitted that on a typical day he worked eight hours, and in a typical week he worked 40 hours.

B. USB's Case

1. Ted Biggs

Ted Biggs was the western regional manager for USB's small business group at the time of trial. The small business group focuses on borrowers who need loans of between $100,000 to $2.5 million. The primary role of the BBO position is to contact prospects and customers at their places of business to develop new relationships or to expand the existing relationships, with the focus on securing deposits and loans.29 BBO's are expected to be physically outside of bank property when performing their sales duties because that is where the customers and prospects are located.30

Biggs developed a tool to communicate his performance expectations to BBO's. He used a stylized eye chart called "15-3-1-1." The chart was used to explain that if a BBO makes an average of 15 customer contacts a week he or she should come away with three applications. Normally, one loan approval and one funded loan would result from the three applications. By following this model, a BBO would make or exceed his or her sales goals. This sales strategy was communicated to all of his regional managers, sales managers, and BBO's. In Biggs's experience, in-person sales calls take from one and a half to two hours. This means that up to 30 hours per week should be spent with prospective customers at their places of business under the 15-3-1-1 model. A site visit could count as one of the 15 meetings, provided the BBO actually met with the customer during the visit.

Because USB has a small share of the market in California, BBO's need to obtain most of their business from new customers. In Biggs's experience, BBO's need to spend a majority of their time outside bank property engaged in sales activities in order to achieve their loan production goals. BBO's are also expected to engage in networking activities outside the bank. They are also required to do site inspections to make sure the customer's business is operational and to verify that any collateral is in good condition. BBO's were not required to work a specific number of hours per week or per day.

When Biggs was asked if he had ever concluded a particular BBO was not spending sufficient time outside the bank, the trial court barred this testimony because the incident did not involve an RWG member.31 Biggs did testify that several of the RWG members did not meet their sales goals. These included Duran and Fitzsimmons. He also noted that he was a family friend of Koga, who never said he felt pressured to sign a declaration for USB.

Biggs testified that mileage reimbursement records do not reveal the actual amount of time a BBO spent outside of bank property during any given day or week. The records only state how many miles a BBO traveled and possibly the destination. In his opinion, the 2002 BBO job description reflected that BBO's would be spending the majority of their time outside.

2. Michael Lewis

At the time of trial, Michael Lewis was a district manager for USB. He started out as a BBO in June 2004, and became a sales manager in June 2005. As we noted earlier, he was initially selected for the RWG, but opted out by the time of trial. He testified he never wanted to be a part of the case, and had opted out after plaintiffs' counsel telephoned and told him he had already been chosen as a witness but could opt out if he did not want to participate. When he opted out he did not know his testimony would have been extrapolated to the whole class, and he was upset when he learned he could not give his testimony.

As a sales manager, Lewis's duties were to support, train, and educate the BBO's under his supervision, and to assist them if they had questions about analyzing financial statements and submitting loan applications. He also instructed them on how to "drum up" more business. He went out on joint sales calls with his BBO's two to four times a week. These calls would last an hour and a half to two hours, excluding travel time. He specifically told his BBO's that it is more effective to go to a customer's place of business, rather than to make appointments at the branch.

Consistent with the trial court's ruling on plaintiffs' motion in limine No. 11, the court prevented Lewis from testifying as to the percentage of time he spent outside bank property when he was a BBO. Further, the court essentially prohibited him from testifying as to whether the non-RWG BBO's under his supervision spent most of their time engaged in sales activities outside the bank. The court also barred him from testifying as to whether he had ever disciplined a non-RWG BBO for failing to spend sufficient time outside of bank property: "The only evidence that is meaningful to the court on that subject [(discipline for something other than failing to meet sales goals)] because of the structure of the trial are members of the [RWG]. If this witness has percipient knowledge of disciplining members of the RWG, then I am interested in hearing. . . . [I]f not, no."

Lewis supervised Gediman for a little over a year. Lewis instructed Gediman to spend more time out of the office in order to be more successful, but he did not specifically say that he was expected to spend more than half of his time outside. During the first quarter he supervised Gediman, Lewis estimated Gediman spent 60 to 70 percent of his time outside the office. During the second quarter it was 55 to 60 percent. During the third quarter, as Gediman's production was starting to slip, Lewis encouraged him to engage in more sales-related activities outside of the office. Gediman told him that he never worked more than 40 hours a week.

Lewis also supervised Lindeman for approximately three months. Initially, Lindeman spent too much time telemarketing inside the branch and was not meeting his goals. He became more successful when he followed Lewis's advice and began spending more time outside bank property.

Lewis met Chad Penza in July 2004 at a BBO conference. At that time, Penza was the number one BBO in the nation. Penza told Lewis that his strategy was to meet with prospects at their place of business, and that BBO's should also meet with real estate brokers to get leads. Penza never said that the secret to success in the BBO position was to spend most of one's time cold-calling from the office.

On cross-examination, Lewis testified that when he was a sales manager he did not keep track of how many hours BBO's worked in a week. Nor did he track how much time any of his BBO's spent inside or outside USB property. BBO's are not evaluated based on the number of hours they work.

3. Pat Collins

Pat Collins was a sales manager from February 2003 to September 2005. During that time she supervised Duran.32 When she hired Duran, she told him he was expected to spend the majority of his time outside USB bank locations. She testified it normally takes three to four meetings with a customer to close a loan. At some point, Duran informed her that he had met only 14 percent of his quarterly sales goal. She told him he needed to start building his "pipeline"33 quickly and not wait until the last minute, meaning he needed to follow the 15-3-1-1 sales strategy and make more outside sales calls. Subsequently, she gave him weekly focused coaching sessions in which she told him to meet with 10 to 15 prospects or customers on a weekly basis and complete profiling forms for each sales call. Based on her conversations with him and his assigned branch managers, along with her review of his weekly activity reports, Collins believed he was out of the office a majority of the time. Eventually Duran was placed on a performance action plan. He resigned shortly thereafter. Collins testified she did not know how much time any BBO spent inside or outside USB property during any given workweek.

4. Sean MacClelland

At the time of trial, Sean MacClelland was a regional sales manager for USB's Northern California business banking section. He started with USB as a BBO and worked at that position for about a year. He then became a sales manager and supervised about seven or eight BBO's. About once or twice a week he would join one of his BBO's on an outside sales call. Sales calls with prospective customers would last from one to three hours on average. He encouraged his BBO's to conduct calls at a prospect's place of business. He told them their primary responsibility was to be out in the field generating loan sales. Like Lewis, MacClelland was prevented by the trial court from testifying as to whether he spent most of his time outside the office when he was a BBO.

MacClelland told his BBO's that they were expected to spend the majority of their time outside of bank property engaged in sales activities. This expectation was given to him by Biggs, and he reinforced this expectation in periodic contacts with BBO's under his supervision. Getting outside the bank builds a rapport with the customer, is a fraud prevention technique, and creates an opportunity to cross-sell products and generate more referrals from prospective customers.

MacClelland testified that Koga had some successful quarters, but had others where he did not meet his goals. MacClelland counseled him to be more focused on getting appointments, generating leads, and going to networking events during slow periods. Koga never told him he was uncomfortable interviewing with USB attorneys about the lawsuit or filling out a declaration.

5. Hector Zatarain

Hector Zatarain began working as a sales manager at USB in March 2001. He supervised Chad Penza. For at least the first year and a half of Penza's employment, he spent the majority of his time outside the office. After that time, he began spending the majority of his time inside the office. Zatarain did not know of other BBO's who spent the majority of their time inside their office and still made their production goals. He did not discipline Penza for being inside the office, but he did tell him that he was working too hard.

C. Motion for Judgment

On July 25, 2007, after plaintiffs had completed the presentation of their case, USB filed a motion for judgment under Code of Civil Procedure section 631.8. In the motion, USB claimed plaintiffs were required to have proved that each and every BBO had been misclassified and had worked unpaid overtime hours. USB also contended the evidence showed several class members were properly classified and/or did not work overtime.

On September 20, 2007, the trial court denied the motion for judgment. In its tentative ruling, the court stated its use of the RWG as the basis for trying the lawsuit had been "authorized" by our decision in Bell III.34

D. Posttrial Proceedings

On October 25, 2007, the trial court issued an order regarding closing briefs that, in part, prohibited USB from referencing the four former named plaintiffs' depositions as well as the declarations USB had sought to admit.

On November 5, 2007, USB filed a "due process motion for briefing and additional argument re phase one evidence," asking the trial court to issue a new ruling regarding the evidence excluded by the October 25, 2007 order.35

On December 4, 2007, the trial court filed its order rejecting USB's due process motion.36 At a hearing also held on this date, the court articulated its finding in favor of plaintiffs on the issue of liability. The court focused primarily on USB's expectations as to whether BBO's should spend the majority of their workday outside of the office: "[T]he Court is compelled to reach the conclusion that the employer, through its descending hierarchy of managers, did not care where the Class members spent their time as much as they cared about the goal of getting new business or, for that matter, retaining old business at the least cost to the bank. [¶] . . . [¶] It is striking to the court that if the bank's expectation was that more than 50 percent of the class members' time was to be spent outside bank property, that there is not clear specific expression of . . . this expectation in the record."

The trial court remained steadfast in its position not to hear proffered BBO's from the defense to challenge plaintiffs' version of the facts. As to USB's protests regarding the denial of its due process motion, the court again referred counsel to Bell III.

E. Proposed Phase I Statement of Decision

On April 8, 2008, USB filed its objections to the proposed Phase I statement of decision that the trial court had ordered plaintiffs to prepare. USB's papers include a declaration prepared by Hildreth, who raised several concerns regarding the proposed statement's computation of the average number of overtime hours worked per week by each of the testifying witnesses.

On May 7, 2008, plaintiffs filed an ex parte motion to amend its expert declarations to include, in part, testimony regarding a telephonic survey of class members that their survey expert Jon Krosnick had conducted after the close of Phase I. The trial court granted the motion, noting it was not making an advance determination as to whether the results of the survey would be admitted in evidence.

On July 10, 2008, USB filed a motion to exclude Krosnick's survey evidence.

On August 4, 2008, plaintiffs submitted their opposition to USB's motion to exclude the new survey evidence. The opposition includes a declaration of Drogin, stating the trial court's finding that the RWG members were improperly classified as exempt could be "reliably projected to the whole class" since the ruling was based on a "random" sample. He quantified the average weekly unpaid overtime hours at 11.87 hours, with a margin of error of plus or minus 5.14 hours under a 95 percent confidence interval.37 Though the margin of error was large, he deemed three other reliability factors worked in favor of plaintiffs: (1) the high response rate among the RWG members, (2) the absence of measurement error because "we are actually sampling the court's findings, which by definition, become a fact after the ruling is made," and (3) Krosnick's survey evidence, which showed the average overtime hours worked by RWG members appeared to be lower than that of the class as a whole. He then discussed the survey data obtained by Krosnick, which showed the average overtime hours of the surveyed class members to be 14.391 hours per week, with a margin of error of plus or minus 2.121 hours per week using a 95 percent confidence interval. He concluded Krosnick's study provided corroborating evidence that the RWG estimate was accurate.

On August 8, 2008, the trial court ruled the survey would not be admissible as affirmative evidence in phase II (Phase II) of the trial, but left open the possibility that the survey could come in for rebuttal purposes to impeach evidence that might be offered by USB.

F. USB's Second Motion for Decertification

On August 22, 2008, USB filed its second motion to decertify the class. USB argued decertification was mandated because the evidence adduced at trial demonstrated individualized issues predominated as to liability and restitution. USB also claimed the evidence showed class treatment was unmanageable and not superior to individualized proceedings. USB reiterated its reliance on the 70-plus sworn declarations from non-RWG BBO's attesting that they spent the majority of their worktime outside bank property. The motion was denied.

G. Statement of Decision for Phase I

After considering extensive briefing and argument regarding disputed issues, the trial court filed its statement of decision on September 22, 2008, a year after the conclusion of Phase I of the trial.38 The court observed introduction of evidence regarding non-RWG class members "would be inconsistent with the court's trial plan and its ruling on plaintiffs' motion in limine No. 11 and defendant's motion in limine No. 4." At the same time, the court noted USB bore the burden of proof on the application of the outside salesperson exemption.

The trial court concluded the two named plaintiffs and the 19 RWG members who testified at trial had been misclassified and had all worked overtime hours, finding their testimony on these issues "credible and persuasive." The statement of decision sets forth the amount of overtime each testifying BBO claimed to have worked, the dates of their employment, and whether they spent most of their worktime inside bank property.39 The statement concludes, "The court finds that the RWGs are typical and representative of the entire class and validates the viability of the use of the [RWG] process as part of the trial management plan of a wage and hour class action." The court found USB did not negate the reasonableness of plaintiffs' evidence "because it either failed to produce testimony to rebut the RWG witnesses or those witnesses that it did produce lacked relevant knowledge and there was no proper foundation." This finding is critical in that USB was prohibited from introducing evidence pertaining to any non-RWG members, evidence that arguably would have shown some class members were either properly classified or did not work overtime. The finding was based, in part, on USB's failure to obtain testimony from all of the immediate supervisors of the RWG members who testified. The order incorporates Drogin's assessment that the testifying witnesses worked an average of 11.87 overtime hours per week.

VII. Phase II of the Trial

A. Motions in Limine

For Phase II of the trial, plaintiffs filed their motion in limine No. 17, seeking to prevent USB from referencing any evidence regarding liability other than the trial court's Phase I statement of decision. The motion was based on the premise that the court had already determined the entire class was improperly classified as exempt, and therefore any contrary evidence would be irrelevant and would necessitate undue consumption of time. The trial court granted plaintiffs' motion.

USB filed its motion in limine No. 11, again seeking to exclude Krosnick's survey evidence from Phase II. The trial court granted this motion, subject to the qualifications in its prior orders regarding the potential use of this evidence for impeachment purposes.

USB's motion in limine No. 12 sought to exclude evidence of time BBO's spent in nonclass positions. This motion was based on USB's contention that two RWG witnesses had occupied nonexempt positions of banker trainee and BBO trainee during a portion of their employment with USB. Plaintiffs filed their corresponding motion in limine No. 20 to exclude evidence of any training time allegedly worked by the class members. In opposition, USB argued that to allow recovery for time spent in nonclass positions would violate its due process rights. The court granted plaintiffs' motion and denied USB's motion, except to the extent USB was to offer evidence that overtime wages had been paid, which would serve to offset restitution claims.

USB also filed motion in limine No. 13, again seeking to admit the depositions of the former named plaintiffs and the declarations of the approximately 70 non-RWG class members. The motion was denied.40

B. Plaintiffs' Case

1. Richard Drogin

Drogin was called as plaintiffs' expert witness in the area of statistics. As the basis of his opinions, he indicated he had relied on the trial court's Phase I statement of decision and the survey report prepared by Krosnick. USB's counsel made multiple objections to the references to Krosnick's report, based on the trial court having granted its motion in limine to exclude evidence of the survey. The trial court overruled the objections, without prejudice.

Drogin explained that random sampling is a process whereby a subset of a population is selected because "each person in the class had the same chance of being selected." He noted that the "random nature" of such a selection results in a representative subset that should reflect the traits and characteristics of the population as a whole.

Drogin defined various statistical terms. The "point estimate" is the number desired to be obtained from a sample, which is then used to predict what the number will be for the rest of the population. In the present case, the average number of overtime hours per week is the point estimate. The "ratio estimate" is a standard statistical estimator that uses two numbers, here the average number of overtime hours worked and the number of weeks worked by the BBO's. Another way to think of the ratio estimate is that it is a weighted average where the weights are the total weeks worked by each person.

Drogin testified the "confidence interval" reflects variations in data and random fluctuation due to sampling. The confidence interval is based on three components, the point estimate, the margin of error, and the level of confidence. Most statisticians use a 95 percent confidence level. Margin of error is a "plus or minus" amount that is associated with the point estimate at a given level of confidence. In this case, Drogin determined the absolute margin of error to be 5.14 hours of overtime per week, relative to a 95 percent confidence level.

Drogin explained "standard deviation" is a quantity used to measure dispersion in data. Standard deviation is used to calculate the margin of error. "Response rate" is the proportion of a sample from which results have been obtained. In this case, he calculated the response rate at 95 percent because only one RWG member (Borsay Bryant) failed to testify. "Measurement error" is the error that can occur in samplings if something is mismeasured in a systematic way. Drogin deemed there was no measurement error in the present case because the data he used was derived from the court's statement of decision.

Drogin testified his understanding was that the trial court selected the RWG by pulling their names from a container that held cards with the names of all the putative class members written on them. This procedure is consistent with a random sample.41 Drogin acknowledged the two named plaintiffs were not part of the random sample. He first calculated the average weekly overtime hour figure for the 19 testifying RWG members, and then for the RWG members plus the two named plaintiffs. The figure came out higher when the two named plaintiffs were excluded from the group.

Drogin said that when he was retained as the plaintiffs' expert in Bell III, his goal was to work with the experts on the opposing side to jointly develop a sampling plan that could be used to get an accurate estimate of overall damages, as liability had already been determined. He acknowledged that no such joint plan was developed in this case. The only people who testified in the damages phase in Bell III were the parties' three experts. No auxiliary report, such as the one prepared by Krosnick, was used. Also, there was no disagreement on the average number of overtime hours worked per week, and no dispute on how to compute damages using that number. The only disagreement was whether the mean or the median should be used for the purpose of calculating damages.

After reviewing the trial court's Phase I statement of decision, Drogin initially calculated the average overtime as 11.299 hours per week. He later increased the figure to 11.87 hours due to adjustments made for vacation time, and the exclusion of a period of time when one of the RWG BBO's was on leave.42 Drogin arrived at the 11.87 hour figure by totaling the number of overtime hours worked by the 21 BBO witnesses (as set forth in the statement of decision), and dividing that figure by the total number of weeks they all had worked. If a witness testified to having worked a range of overtime hours, Drogin picked the midpoint.

Drogin considered his calculations to be reliable based on his assumption that the sample was random, while noting "In some situations you may have a sample, but there are certain factors that could destroy it from being a perfect textbook random sample—for example, if there's any self-selection, if there's something that wasn't random of how the sample is selected." He concluded the 11.87 figure could be reliably projected to the absent class members for purposes of calculating the restitution owed. He noted that whether the confidence interval was too wide was a matter for the court to decide, but that his was the best estimate based on the available data.

On cross-examination, Drogin admitted he was aware that 20 percent of the originally drawn sample had opted out after they were selected for the RWG, whereas only 2 percent of the remaining class members had opted out. He indicated he had no way of knowing why anyone had opted out. When asked if, at a 95 percent confidence interval, there was a statistical probability that up to 13 percent of the class could be properly classified even if 21 out of 21 sample members were found to be misclassified, Drogin replied that the 13-percent figure would be "part of the range—the lower range of the confidence interval." He testified that "the 95 percent confidence interval ranges from 87 percent to 100 percent for the percentage of class members who are misclassified." When asked if, "statistically speaking," would not this mean that up to 13 percent of the class could possibly be properly classified, Drogin responded: "That's true." Drogin clarified that, at any given confidence level, he can make a projection based on variables but that he would not know "as a matter of fact or personal knowledge" whether 100 percent of the class was misclassified or not.

Drogin acknowledged that the sampling procedure he had proposed at the outset of this case was not used by the trial court. One element of his proposal would have included a survey of the amount of time each class member spent outside the office engaged in sales activities.43 He also acknowledged that he never recommended the inclusion of nonrandomly selected individuals in the sample. The addition of nonrandomly selected people to a random group could bias the sample in some way. But in this case, the effect of adding the two named plaintiffs was actually to lower the average overtime figure. As a statistician, he would not consider two individuals who are not randomly selected into a sample to be representative of the population as a whole.

On the subject of witnesses who testified they worked an uncertain range of hours per week, Drogin admitted there is no formula or rule in statistics that says the midpoint should be used if a subject responds with an interval instead of a specific number. In arriving at his overtime hour calculation, Drogin calculated weekly overtime only, not daily overtime. If a witness testified using an open interval rather than a range of hours, Drogin used 7.5 hours as his or her weekly overtime figure. Drogin explained that the absolute margin of error is expressed in the units of the estimate, and the relative margin of error is the ratio between the absolute margin of error and the estimate expressed as a percentage. Here, the absolute margin of error for the overtime worked by the sample plaintiffs is 5.14 hours per week, and the relative margin of error is 43.3 percent. On redirect, he stated that there is nothing wrong, statistically speaking, with choosing a sample without first determining the desired margin of error.

Drogin agreed with USB's counsel that, in general, the larger the sample size, the lower the margin of error.44 Drogin also acknowledged the two top workers in the sample group here had large values for their overtime hours. Penza's overtime was calculated at over 5,000 hours, or roughly five times more than the total number of overtime hours worked by the next highest RWG member (apart from Petty). Petty's figure was calculated at almost 3,000 hours, almost three times more than the next highest member. Drogin also agreed it is statistically improper to allow people to self-select or volunteer to become part of a sample. He opined that the sampling methodology used by the court in this case was statistically appropriate, even though the sample is not completely random and the resulting margin of error is 43.3 percent.

2. Paul Regan

Paul Regan testified for plaintiffs as an expert in accounting. His task was to calculate the restitution due to plaintiffs based on Drogin's figures. He testified that if Duran and Fitzsimmons were removed from the sample group, the total restitution owed to the class would have increased by about $411,000. On cross-examination, he testified he did not perform any calculations with respect to the margin of error. He also did not extrapolate to the class the two weeks in which Penza was properly classified. He did take off two weeks from Penza's recoverable time period. Regan did not exclude time during which USB claimed certain class members were working in nonexempt training positions. Nor did he make any deductions for holidays taken by class members in calculating their recovery.45 The total class recovery calculation, including prejudgment interest, was over $14 million. At the close of Regan's testimony, plaintiffs rested.

C. USB's Defense in Phase II

The trial court prohibited USB from presenting evidence showing that some class members had been classified in nonexempt positions during their tenure with USB. For example, USB offered to show that a class member, while working in a training position, signed a weekly timesheet verifying his hours and stating he did not work overtime. USB argued that he should not have been allocated overtime pay for the weeks in which he verified he did not work overtime. The court sustained plaintiffs' objections to these timesheets on the basis that it violated the trial plan. The court barred USB from introducing any evidence as to class members who may have held a BBO training position, or evidence that this position represented a distinct, nonexempt classification. The court also prohibited two non-RWG BBO's from testifying.

1. Joe Anastasi

Joe Anastasi testified as USB's expert in forensic accounting He was asked by USB to prepare a rebuttal analysis to Regan's projection of overtime recovery to the class. Anastasi reviewed USB's records with respect to holidays, vacation, sick leave, and leaves of absence taken by class members during the relevant time period. In preparing his analysis, he disagreed with Regan's calculations because Regan did not deduct holiday time for any individual in the class, which overstates the number of workdays per person by up to 10 days per year. He also disagreed with Regan's inclusion of hiring bonuses to determine each BBO's hourly rate of pay.

Anastasi prepared a damages calculation using the low end of the margin of error, which is 6.73 overtime hours per week. Using the lowest end of the ranges of hours given by RWG members, and using a simple average as opposed to a weighted average, the total overtime award would be $6,141,000. The amount would be reduced to $5,125,000 using the 6.73-hours-per-week figure. Using a 7 percent rate of interest, Anastasi calculated total class recovery using the 6.73-hour figure to be $7,226,624, or about half of the total Regan had calculated.

In Anastasi's opinion, undue weight was accorded to Penza in the calculation of a weighted average for the RWG group. Anastasi opined that the data in the trial court's statement of decision is so imprecise that it results in arbitrary projections. He testified that one could derive a total of 29 different "bottom lines" based on different assumptions and calculations of the data as provided in the court's statement of decision. The choice of which "bottom line" to pick is totally arbitrary due to the variability and lack of specificity in the data.

2. Andrew Hildreth

Hildreth testified as an expert witness in statistics for USB. In his opinion, the sampling plan chosen by the court was statistically unacceptable, in part because the size of the sample was insufficient for extrapolation purposes. He stated that simply drawing a random sample is not necessarily sufficient to provide an unbiased and accurate estimate regarding an underlying population. A sample has to be sufficiently large, and has to be free from bias caused by potential sampling or nonsampling errors. It also must be undertaken with a focus on the questions of interest.

Hildreth stressed the importance of defining a population before collecting a sample in order to gauge whether any subsequent changes to the sample are reflective of the underlying population. Hildreth believed selection bias occurred in this case with the second opt-out notice because the RWG members were told they could opt out or stay and testify at trial, whereas the remainder of the population was simply given the choice to opt out. This lead to a different behavioral response in that only 2 percent of the nonsample group opted out, whereas 20 percent of the sample group opted out. As a result the sample suffers from selection bias.46

Hildreth found the inclusion of the two named plaintiffs in the sample group was another example of selection bias. The two were placed into the sample on a nonrandom basis. Because they were not randomly selected, there is no statistical basis for considering their testimony as representative of the entire class. When he excluded the two named plaintiffs from the sample group, Hildreth calculated that the margin of error increased from 43.3 to 47 percent. He concluded selection bias and measurement errors rendered the sample unrepresentative.

Consistent with Drogin, Hildreth testified that even assuming there were no sampling errors and that every member of a 20-person random sample was found to be misclassified, from a statistical standpoint 13 percent of the class might nonetheless be properly classified under a 95 percent confidence interval. With a sample of 21, up to 12 percent of the class may be properly classified. With a sample of 19 (here, if the two named plaintiffs are excluded), up to 14 percent of the class could be properly classified at a 95 percent confidence interval, which would amount to approximately 36 class members. At a 99 percent confidence interval this number could increase to 20 percent of the class, or 52 class members. Thus, he opined there is no statistical basis to conclude that 100 percent of the class was misclassified.47 He also explained that the 43.3 percent margin of error at a 95 percent confidence interval means that if the sample study were repeated multiple times, one would expect to obtain average weekly overtime estimates as low as 6.7 hours or as high as 17.8 hours 95 percent of the time.

The trial court prohibited Hildreth from testifying as to his review of the 70-plus BBO declarations USB had obtained, notwithstanding USB's complaint that Drogin had been permitted to testify concerning the non-RWG survey responses contained in Krosnick's report. The court indicated it considered USB's declarations to be unreliable.

D. Closing Briefs and Statement of Decision for Phase II

After the close of Phase II evidence, the trial court permitted each side to file briefs and objections. A hearing before the court followed. The court issued its decision on May 20, 2009.

The trial court stated its initial finding as follows: "The court finds, with a 95% level of confidence, the best estimate based on all admissible evidence is that the class worked 11.86 overtime hours per week with an absolute margin of error of approximately plus or minus 5.14 hours and a relative margin of error of approximately 43%." The court found the total amount of overtime restitution owed to the class as of October 1, 2008, to be $8,953,832. The amount owed to named plaintiffs Duran and Fitzsimmons for meal and rest break violations was calculated at $25,373. The court included interest at a rate of 10 percent per annum at $5,980,360, for a total award of $14,959,565 owed as of May 15, 2009.

Additionally, the trial court stated that even though the margin of error as to the number of overtime hours is 43.3 percent, ancillary factors identified in our Bell III opinion were present and "serve to bolster the reliability of the estimate." The court found those factors to be "(1) random selection of the sample; (2) high response rate; (3) absence of measurement error; (4) probable distribution within the margin of error based on auxiliary and anecdotal evidence; (5) data that is not skewed; (6) the calculation of the average overtime was not an afterthought; and, (7) other procedures were considered."

E. Posttrial Proceedings

In its motion for new trial, USB argued the trial court's refusal to admit non-RWG class member declarations and deposition testimony, and its preclusion of all non-RWG witnesses in USB's defense case, deprived USB of its constitutional due process right to a fair trial. On July 17, 2009, the trial court denied USB's motion for new trial.48


I. The Trial Plan Erroneously Relied on Representative Sampling

USB raises several challenges to the trial court's rulings on class certification as well as to the conduct of the trial. We turn first to the claim that the trial plan violated USB's right to due process of law. USB contends the court's strategy of relying solely on the evidence derived from a 21-person sample to determine classwide liability and restitution violated principles of due process and resulted in a statistically invalid result, as evidenced by the 43.3 percent margin of error in weekly overtime hours. USB also asserts the court infringed on its due process rights when it rejected USB's efforts to introduce evidence challenging the individual claims of the 239 absent class members. We agree with USB that the trial plan employed here was seriously flawed and the judgment must be reversed.

A. Standard of Review

The general standard of review of a trial plan in a class action case is abuse of discretion. (Bell III, supra, 115 Cal.App.4th 715, 751.) Both parties agree we review de novo the legal issue of whether a trial plan violated a party's right to due process. (Hypertouch, Inc. v. Superior Court (2005) 128 Cal.App.4th 1527, 1537 [27 Cal.Rptr.3d 839].)

(1) We recognize our Supreme Court has encouraged trial courts to "think outside the box" when managing class action lawsuits: "For decades `[t]his court has urged trial courts to be procedurally innovative' [citation] in managing class actions, and `the trial court has an obligation to consider the use of . . . innovative procedural tools proposed by a party to certify a manageable class' [citations]." (Sav-On, supra, 34 Cal.4th 319, 339.) While innovation is to be encouraged, the rights of the parties may not be sacrificed for the sake of expediency. As we observed in Bell III, "In this area of litigation, the California Supreme Court has in fact `challenged the trial courts to develop "pragmatic procedural devices" to "simplify the potentially complex litigation while at the same time protecting the rights of all the parties." [Citations.]' [Citation.]" (Bell III, supra, 115 Cal.App.4th 715, 747, italics added, fn. omitted.)

B. Principles of Due Process

(2) The due process clause is "`flexible and calls for such procedural protections as the particular situation demands.' [Citation.]" (Mathews v. Eldridge (1976) 424 U.S. 319, 334 [47 L.Ed.2d 18, 96 S.Ct. 893].) In Connecticut v. Doehr (1991) 501 U.S. 1, 10 [115 L.Ed.2d 1, 111 S.Ct. 2105] (Doehr), a case involving a prejudgment attachment procedure, the United States Supreme Court set forth a test for determining whether a procedure by which a private party invokes state power to deprive another person of property satisfies due process: "[F]irst, consideration of the private interest that will be affected by the [procedure]; second, an examination of the risk of erroneous deprivation through the procedures under attack and the probable value of additional or alternative safeguards; and third, . . . principal attention to the interest of the party seeking the [procedure], with, nonetheless, due regard for any ancillary interest the government may have in providing the procedure or forgoing the added burden of providing greater protections." (Id. at p. 11; see also Hilao v. Estate of Marcos (9th Cir. 1996) 103 F.3d 767, 786 (Hilao).)49

Due process principles are designed to ensure a party is afforded his or her right to be heard during adversarial proceedings: "`As the rubric itself implies, "procedural due process" is simply "a guarantee of fair procedure."' [Citations.] Hence, we review cases involving adversarial hearings to determine whether, under the specific facts and circumstances of a given situation, the affected individual has had a fundamentally fair chance to present his or her side of the story." (In re Nineteen Appeals (1st Cir. 1992) 982 F.2d 603, 611.)

C. Affirmative Defense of the Outside Salesperson Exemption

Plaintiffs here contended they were entitled to overtime pay under Labor Code section 510 and that USB misclassified them as exempt from this overtime requirement. As an affirmative defense, USB claimed the class members were exempt from overtime laws, relying primarily on the outside salesperson exemption as set forth in Wage Order No. 4 and Ramirez v. Yosemite Water Co., supra, 20 Cal.4th 785 (Ramirez).

(3) The Supreme Court in Ramirez explained that the outside salesperson exemption generally turns on how an employee actually spends his or her time, as well as on the employer's realistic expectations of the job and the extent to which the employee diverges from them. (Ramirez, supra, 20 Cal.4th 785, 790, 802-803.) Ramirez involved a single employee suing to recover overtime wages after having allegedly been erroneously classified as exempt under the outside salesperson exemption. (Id. at p. 790.) The dispute centered on whether the plaintiff spent the majority of his time engaged in sales-related activities or in nonexempt delivery duties. (Id. at p. 803.) The court stated that in addressing this exemption, courts "should consider, first and foremost, how the employee actually spends his or her time." (Id. at p. 802.) The court observed that the issue of whether an employee is an outside salesperson "turns on a detailed, fact-specific determination of this matter."50 (20 Cal.4th at p. 790.)

Arguably, the present case is distinguishable from Ramirez in that there is little dispute as to the sales-driven nature of the BBO's job duties. In other words, the issue at trial was not whether BBO's spent the majority of their time engaged in sales-related activities. Rather, the focus was on whether BBO's spent the majority of their time physically outside USB property. Nevertheless, this distinction does not detract from the central issue on appeal, which is whether the trial court erred in prohibiting USB from inquiring into the exempt status of the absent class members.

D. Bell v. Farmers Ins. Exchange, supra, 115 Cal.App.4th 715

Plaintiffs cite our opinion in Bell III for the proposition that ample authority exists to support the use of statistical sampling and representative testimony to determine classwide liability. They claim our endorsement in Bell III of the use of statistical methods to determine damages in wage and hour class actions "would apply equally to determining liability." While the trial court also appeared at various times to rely on our prior opinion as a justification for using the RWG as the basis for determining USB's liability to the entire class, Bell III is manifestly inapposite.

In Bell III, a class of approximately 2,500 claims representatives sued their insurance company employer for unpaid overtime compensation. The employer asserted the representatives were exempt from overtime under the administrative exemption. (Bell III, supra, 115 Cal.App.4th 715, 720.) This defense was rejected by the trial court in an order granting the employees' motion for summary judgment. (Ibid.) On appeal, we upheld the court's finding that the plaintiff employees were nonexempt. (Bell v. Farmers Ins. Exchange (2001) 87 Cal.App.4th 805 [105 Cal.Rptr.2d 59].) Thereafter, the employer filed two unsuccessful motions to decertify the class. (See Bell III, supra, at p. 721.)

Each side then retained expert statisticians to address the damages phase of the trial. As noted previously, the plaintiffs retained Drogin, the same expert used by plaintiffs in the present case. (Bell III, supra, 115 Cal.App.4th 715, 722.) At a trial management hearing, the parties agreed to take depositions of a randomly chosen sample of 50 individuals as a first step towards determining classwide aggregate damages based on a representative sampling of class members. (Ibid.) The experts worked together to identify the sample sizes associated with preset absolute margins of error, ranging from 30 minutes to two hours of overtime per week. (Id. at p. 723.) After the trial court suggested a one-hour margin of error would be satisfactory, the parties conducted more depositions and the experts concluded the one-hour margin could be achieved using a sample size of 286 plaintiffs. The parties ultimately deposed a total of 295 individuals and, after agreeing on the employees' work patterns, the two experts calculated an average weekly overtime figure of 9.42 hours with a margin of error of 0.9 hours per week, or approximately 9.6 percent as the relative margin of error.51 (115 Cal.App.4th at p. 723.)

A brief jury trial followed in which Drogin and Paul Regan (also retained by plaintiffs in the present case) testified for the plaintiffs and the employer's statistician testified for the employer. (Bell III, supra, 115 Cal.App.4th 715, 724.) The employers' expert presented a case for using a median52 weekly overtime figure of 7.27 hours as opposed to the mean (average) figure of 9.42 hours advocated by the plaintiffs. The jury returned a special verdict largely along the calculations arrived at by Drogin and Regan, including approximately $1.2 million in damages for unpaid double-time overtime compensation. (Id. at pp. 724-725.) The employer appealed.

In Bell III, we noted our task had been made easier by the fact that the employer did not challenge Drogin's scientific methodology or his qualifications. (Bell III, supra, 115 Cal.App.4th 715, 747.) Instead, the employer argued that the use of "statistical inference" violated its due process rights by relieving class members of the burden of proving they had worked overtime. (Id. at p. 749.) We rejected this argument: "[S]tatistical sampling does not dispense with proof of damages but rather offers a different method of proof, substituting inference from membership in a class for an individual employee's testimony of hours worked for inadequate compensation. It calls for a particular form of expert testimony to carry the initial burden of proof, not a change in substantive law." (Id. at p. 750.)

We performed a balancing analysis under Doehr and found the defendant's interest was in the proper determination of its overall liability only, and not the amount of damages awarded to any particular class member. (Bell III, supra, 115 Cal.App.4th 715, 751-752.) We found the plaintiffs had a strong interest in using statistical inference to resolve their claims efficiently. (Id. at p. 752.) We also noted the defendant appeared to have conceded during trial management hearings that a one-hour margin of error satisfied its due process concerns. (Id. at p. 755.) We held the award for time-and-a-half overtime "reflected a level of accuracy consistent with due process under the Doehr balancing test." (Ibid.)

We came to the opposite conclusion with respect to the inclusion of double-time hours within the 9.4 average weekly hours of unpaid overtime work. (Bell III, supra, 115 Cal.App.4th 715, 756-757.) The sample group's average of 0.37 hours of unpaid double time was extrapolated to the entire class, resulting in an absolute margin of error of 0.12 hours, or about 32 percent. The distribution was highly skewed, as only 16 employees in the sample group accounted for half of the double-time hours. (Id. at p. 756.) While margin of error alone does not afford a "bright-line constitutional distinction,"53 we observed the parties' experts had not offered "foundational calculations for the determination of double-time or propose[d] an appropriate class size, margin of error, or sampling methodology." (115 Cal.App.4th at p. 756.) Accordingly, we invalidated the double-overtime portion of the damages award. (Id. at p. 757.)

E. The Present Case Does Not Comport with Bell III

The procedures we approved in Bell III are only superficially similar to the procedures utilized in the present case.54 Again, in Bell III we did not have occasion to consider the use of a representative sample to determine classwide liability, since liability was not an issue on appeal. Accordingly, the only issue we addressed was the damages calculation itself, and not whether the plaintiff employees had a right to recover damages in the first place. And our assessment was based on a record evidencing cooperation and agreement among the parties and their counsel.

Second, we agree with USB that the trial court here did not follow established statistical procedures in adopting its RWG-based trial methodology. In Bell III, the size of the representative sample used as the basis for calculating straight overtime damages was determined after extensive calculations had been made by expert statisticians based on surveys of class members and pilot studies. Indeed, we rejected the double-overtime award, in part, because these surveys and pilot studies had failed to consider how to properly calculate such damages. (Bell III, supra, 115 Cal.App.4th 715, 756.) Similar concerns are present here, as the trial court chose the size of the representative group without any consideration as to probable margin of error and without the benefit of any surveys or pilot studies.

Third, the expert witnesses in Bell III cooperated in many aspects relating to the methodology for calculating damages, including agreeing to use the data from the employees' work schedules and collaborating to arrive at the appropriate sample size necessary to achieve the one-hour-per-week margin of error. No such cooperation occurred in this case as the expert witnesses never worked together.

Fourth, it appears that the sampling done in the Bell III case was at all times random, unlike here where a comparatively high number of RWG members opted out before trial, and the trial court allowed evidence from the two named plaintiffs not randomly chosen to be extrapolated to the entire class.

Fifth, the restitution award here was affected by a 43.3 percent margin of error, more than 10 percentage points above the margin of error for the double-overtime award we invalidated in Bell III. In absolute terms, the average weekly overtime hour figure could conceivably be as low as 6.72 hours per week, as opposed to the 11.86 hour figure arrived at here. While we again will not set a bright line for when a margin of error becomes so excessive as to be deemed unconstitutional, we are troubled by this result.55

Finally, another factor present in this case that was not present in Bell III involves the repeated restrictions the trial court placed on USB's ability to present arguably relevant evidence in its defense. In Bell III, we found only a single pretrial ruling that had significantly restricted the employer's right to contest the plaintiffs' proof of damages, a ruling the employer did not challenge on appeal. (Bell III, supra, 115 Cal.App.4th 715, 759.) We also concluded the employer was not prejudiced by the procedures followed in that case. (Ibid.) Unlike Bell III, in which the employer had acquiesced to statistical proof of damages and had waived the right to impeach the employees' testimony at trial (id. at pp. 757-759), USB steadfastly and repeatedly objected to all phases of the trial management plan. Contrary to plaintiffs' contention, we do not agree USB waived its objection to specific aspects of the plan by its "total opposition to statistical methodology." Nor does it follow, even if USB did resist efforts to cooperate, that the trial court was compelled to use the methodology it ultimately selected. In sum, plaintiffs' reliance on Bell III is misplaced.

Here, as we explain further below, the trial court exceeded acceptable due process parameters by limiting the presentation of evidence of liability to the testifying BBO's only. Fundamental due process issues are implicated not only by the unprecedented and inconsistent use of statistical procedures in the liability and damages phases, but also by the manner in which USB was hobbled in its ability to prove its affirmative defense. Under the court's plan, as reinforced by its rulings on motions in limine and related evidentiary matters, USB was barred from introducing manifestly relevant evidence.56 This evidence potentially could have greatly mitigated the damages awarded and possibly could have defeated plaintiffs' class action claim entirely.

F. State and Federal Case Law Does Not Support This Use of Representative Sampling

USB claims California law precludes classwide liability determinations based on evidence obtained from a representative sample in employment cases alleging misclassification. USB relies on several state and federal wage and hour class action cases for the proposition that surveying, sampling, and statistics are not valid methods of determining liability because representative findings can never be reasonably extrapolated to absent class members in misclassification claims given that time spent performing exempt tasks may differ between employees.57 While all the cases cited by USB involve rulings on motions to certify or decertify class actions, they support the conclusion that improper procedures were followed in this case.

(4) Jimenez v. Domino's Pizza, Inc. (C.D.Cal. 2006) 238 F.R.D. 241, involved certification under rule 23 of the Federal Rules of Civil Procedure (28 U.S.C.) (Rule 23)58 on behalf of a class of general managers alleging failure to pay overtime wages and provide rest or meal periods in violation of California law. (Jimenez, supra, at pp. 245-246.) The plaintiffs claimed the employer improperly classified them as executive, administrative, and professional employees though they regularly spent the majority of their time performing nonexempt functions, such as preparing food and cleaning the stores. (Id. at p. 246.) The district court denied certification, explaining that the plaintiffs' proposed trial plan of using surveys and representative testimony to prove misclassification would not avoid the need to conduct individualized inquiries, in part because the employer defendant would have had the right "to cross-examine each general manager to determine whether there is liability as to that specific person." (Id. at p. 253.) This case supports the premise that when liability for unpaid overtime depends on an employee's individual circumstances, employer defendants retain the right to assert the exemption defense as to every potential class member.

In Walsh two account managers filed a class action lawsuit against their former employer. Similar to the present case, the defendant in Walsh had allegedly wrongly classified these employees as exempt from overtime wage laws under the outside salesperson exemption. (Walsh, supra, 148 Cal.App.4th 1440, 1445-1446.) After the trial court certified a class action, the employer moved to decertify an account manager subclass, contending common questions of law and fact did not predominate over individual issues. The employer presented evidence that performance of the managers' primary functions varied significantly, depending upon territory, number of customers and job orders, support from customer service representatives, and the personal approach of each manager. (Id. at p. 1455.) The trial court granted the motion to decertify. (Id. at p. 1452.)

The Court of Appeal concluded that the Supreme Court's decision in Sav-On, supra, 34 Cal.4th 319 [(reversing a Court of Appeal decision that had overturned a trial court's class certification order)] did not strip the trial court of its discretion to consider whether factual variations among individual employees rendered the class action device an inferior method of adjudicating the claims. (Walsh, supra, 148 Cal.App.4th 1440, 1458.) Again, this case clearly supports the premise that due process principles require individualized inquiries where the applicability of an exemption turns on the specific circumstances of each employee, even in cases where the employer's misclassification may be willful.

In Dunbar this court affirmed a trial court's order denying a motion for certification in an overtime wage and hour case involving the executive exemption. The trial court in Dunbar had acknowledged there were issues common to the putative class of store managers, such as whether stocking shelves and operating cash registers are managerial tasks. However, the trial court indicated that the tasks performed across the class were so dissimilar that it could not " `reasonably extrapolate findings from the named plaintiff to the absent class members.' " (Dunbar v. Albertson's, Inc., supra, 141 Cal.App.4th 1422, 1430.) In concluding the trial court properly considered the manageability of individualized issues, we observed: "It is not sufficient, in any event, simply to mention a procedural tool; the party seeking class certification must explain how the procedure will effectively manage the issues in question, and [the] plaintiff has failed to do so here." (Id. at p. 1432.) Again, this case supports the view that the use of sampling to extrapolate liability in an exemption context can be problematic.59

In In re Wells Fargo Home Mortgage Overtime Pay Litigation (9th Cir. 2009) 571 F.3d 953 (Wells Fargo I), the Ninth Circuit reversed a class certification order, finding the district court had placed undue weight on the defendant's uniform policy of classifying mortgage consultants as exempt. (Id. at p. 959.) The lower court had found numerous individualized inquiries would be necessary to resolve the matter but ultimately granted certification based on the defendant's uniform exemption policies. (Id. at p. 956.) On appeal, the Ninth Circuit held the lower court had erred in relying so heavily on the internal exemption policy in the face of admittedly serious issues regarding individual variations among the plaintiffs' job duties and experiences. (Id. at p. 959.) The reviewing court emphasized that regardless of whether such a policy is in place, "courts must still ask where the individual employees actually spent their time" (ibid.) in determining whether overtime laws have been violated.

On remand after the decision in Wells Fargo I, Judge Patel determined class certification was not warranted. (In re Wells Fargo Home Mortgage Overtime Pay Litigation (N.D.Cal. 2010) 268 F.R.D. 604 (Wells Fargo II). Significantly, for our purposes, the court rejected the plaintiff's suggestion that time-consuming individualized inquiries could be avoided by using random sampling of class members to determine whether the class, as a whole, qualified for any of the asserted exemptions, including the outside salesperson exemption. (Id. at p. 612.) The court observed: "In order to adjudicate Wells Fargo's exemption defenses, especially the outside sales person exemption, a substantial quantity of individual inquiries will be necessary. Although there are some issues in this case amenable to common proof, individual inquiries will predominate over common questions."60 (268 F.R.D. at p. 613, italics added.)

We find the opinion in Wells Fargo II to be particularly instructive. In Wells Fargo II, the plaintiff had hypothesized a trial scenario in which, out of 30 testifying employees, 27 were found to be nonexempt with an average of eight hours of overtime per week. Citing this hypothetical, the court concluded the proposed plan was inadequate: "Assume that the court permitted proof through random sampling of class members, and that the data, in fact, indicated that one out of every ten [class members] is exempt. How would the finder of fact accurately separate the one exempt [class member] from the nine nonexempt [class members] without resorting to individual mini-trials? Plaintiff has not identified a single case in which a court certified an overbroad class that included both injured and uninjured parties.... In fact, the court has been unable to locate any case in which a court permitted a plaintiff to establish the nonexempt status of class members, especially with respect to the outside sales exemption, through statistical evidence or representative testimony.... Although `the decision to use [statistical sampling] tools is within the discretion of the district court,' in this case, they would be of extremely limited help to resolving the key issues. [Citation.]" (Wells Fargo II, supra, 268 F.R.D. 604, 612, italics added.)61

G. The Sampling Plan Used Here Was Flawed

The procedural flaws anticipated by the state and federal courts in the opinions summarized above appear to have been fully realized in the present case. In the first place, there was no statistical foundation for the trial court's initial assumption that 20 out of 260 is a sufficient size for a representative sample by which to extrapolate either liability or damages. Neither party proposed a trial plan based solely on the selection of a representative group of plaintiffs, let alone a group of 20. The court appears to have arrived at this procedure on its own, without reliance on legal precedent or the advice of expert witnesses. In their brief on appeal, plaintiffs state that the trial court "ultimately adopted a trial management plan modeled on Dr. Drogin's proposal." While Drogin did propose the use of representative testimony from a randomly selected group of plaintiffs, Drogin did not offer any advice as to the size of the group. Further, Drogin indicated that the group was to be selected only after a survey of the BBO's duties and hours had been conducted.62 As we have shown above, courts are generally skeptical of the use of representative sampling to determine liability, even in cases in which plaintiffs have proposed using expert testimony and statistical calculations as the foundations for setting the sample size. Here, the trial management plan was lacking in any expert input or principled statistical foundation.63

We note while this appeal was pending, the United States Supreme Court issued its opinion in Wal-Mart Stores, Inc. v. Dukes (2011) 564 U.S. ___, ___ [180 L.Ed.2d 374, 131 S.Ct. 2541] (Wal-Mart). The court reversed a Ninth Circuit order that had affirmed the certification of a class consisting of some 1.5 million female employees who claimed they had suffered discriminatory treatment.64 The court found representative sampling studies did not justify certification. Quoting from the dissenting opinion below, the court stated " `[i]nformation about disparities at the regional and national level does not establish the existence of disparities at individual stores, let alone raise the inference that a company-wide policy of discrimination is implemented by discretionary decisions at the store and district level.' [Citation.] A regional pay disparity, for example, may be attributable to only a small set of Wal-Mart stores, and cannot by itself establish the uniform, store-by-store disparity upon which the plaintiffs' theory of commonality depends." (Wal-Mart, supra, 564 U.S. ___, ___ [180 L.Ed.2d 374, 394].)

After discussing the trial procedures that apply to pattern-or-practice cases, including the right of a defendant to raise any individual affirmative defenses (Wal-Mart, supra, 564 U.S. ___, ___ [180 L.Ed.2d 374, 399-400]), the court took issue with procedures the Ninth Circuit had authorized: "The Court of Appeals believed that it was possible to replace such proceedings with Trial by Formula. A sample set of the class members would be selected, as to whom liability for sex discrimination and the backpay owing as a result would be determined in depositions supervised by a master. The percentage of claims determined to be valid would then be applied to the entire remaining class, and the number of (presumptively) valid claims thus derived would be multiplied by the average backpay award in the sample set to arrive at the entire class recovery—without further individualized proceedings. [Citation.] We disapprove that novel project.... [A] class cannot be certified on the premise that Wal-Mart will not be entitled to litigate its statutory defenses to individual claims." (Wal-Mart, supra, 564 U.S. ___, ___ [180 L.Ed.2d 374, 400], italics added.) The same type of "Trial by Formula" that the United States Supreme Court disapproved of in Wal-Mart is essentially what occurred in this case.65 It is important to appreciate this portion of the Wal-Mart opinion was the expression of a unanimous court. For reasons Judge Patel wisely foresaw, we find this approach to be untenable.

II. The Trial Procedure Denied USB Its Right to Due Process

USB claims the trial court's refusal to allow USB to introduce evidence to challenge the claims of the other 239 class members violated its due process rights. We agree.

A. The Evidence USB Sought to Introduce Was Relevant

As outlined above, USB repeatedly attempted to introduce evidence pertaining to non-RWG class members, including those for whom USB had offered specific evidence refuting their potential claims for recovery, such as sworn declarations and/or deposition testimony in which they averred to having spent the majority of their time outside the office. USB notes the judgment awards an average of over $50,000 to each absent class member, notwithstanding that USB offered evidence that potentially could have prevented, at a minimum, approximately one-third of these individuals from receiving any recovery. The four former named plaintiffs who testified at their depositions that they spent more than half their time outside the office (strongly suggesting they were properly classified as exempt), were together awarded over $160,000 in overtime compensation.66 Yet the trial court excluded their testimony at trial on the ground that it was "irrelevant" because it did not comport with the court's trial plan. Unfortunately, relevancy was dictated by the court's trial plan rather than by the trial itself as it unfolded in the courtroom.

We agree with USB that the evidence it sought to introduce is highly relevant in this misclassification case. Under Evidence Code section 351 "Except as otherwise provided by statute, all relevant evidence is admissible." " `Relevant evidence' means evidence, including evidence relevant to the credibility of a witness or hearsay declarant, having any tendency in reason to prove or disprove any disputed fact that is of consequence to the determination of the action." (Evid. Code, § 210.) "Although proffered evidence may have some relevance, `[t]he [trial] court in its discretion may exclude evidence if its probative value is substantially outweighed by the probability that its admission will (a) necessitate undue consumption of time or (b) create substantial danger of undue prejudice, of confusing the issues, or of misleading the jury.' (§ 352.) We review a trial court's evidentiary rulings for an abuse of discretion." (Winfred D. v. Michelin North America, Inc. (2008) 165 Cal.App.4th 1011, 1026 [81 Cal.Rptr.3d 756].)

(5) The evidence USB sought to introduce, if deemed persuasive, would have established that at least one-third of the class was properly classified. Thus, this evidence USB sought to introduce is unquestionably relevant and therefore admissible.67 "Describing a party's fundamental right to present evidence at trial in a civil case, Witkin observes: `One of the elements of a fair trial is the right to offer relevant and competent evidence on a material issue. Subject to such obvious qualifications as the court's power to restrict cumulative and rebuttal evidence ..., and to exclude unduly prejudicial matter [citation], denial of this fundamental right is almost always considered reversible error. [Citations.]' [Citation.] ... [A] party's opportunity to call witnesses to testify and to proffer admissible evidence is central to having his or her day in court." (Elkins v. Superior Court (2007) 41 Cal.4th 1337, 1357 [63 Cal.Rptr.3d 483, 163 P.3d 160].)

Plaintiffs claim the trial court properly refused to admit the declarations because they lacked credibility in light of " `their actual authorship, the circumstances of preparation and internal inconsistencies and ambiguities.' " They also claim USB offered the declarations after the close of evidence, and that they had already been excluded in motions in limine "because of incompatibility with the trial plan." USB cannot be faulted for offering the declarations after the close of evidence, since they were excluded by the pretrial rulings. Further, the weight accorded to the declarations could have been addressed after the evidence was introduced, rather than preempting its introduction entirely.68 Further, to the extent deemed inconsistent with the trial plan, we note our Supreme Court has observed "That a procedure is efficient and moves cases through the system is admirable, but even more important is for the courts to provide fair and accessible justice." (Elkins v. Superior Court, supra, 41 Cal.4th 1337, 1366 [marital dissolution proceeding].) We conclude the trial court erred in foreclosing USB the opportunity to raise individualized challenges to the absent class members' claims.

B. The Error Is Prejudicial

"Trial court rulings on the admissibility of evidence, whether in limine or during trial, are generally reviewed for abuse of discretion. [Citations.] `The trial court's error in excluding evidence is grounds for reversing a judgment only if the party appealing demonstrates a "miscarriage of justice"—that is, that a different result would have been probable if the error had not occurred.' [Citations.]" (Pannu v. Land Rover North America, Inc. (2011) 191 Cal.App.4th 1298, 1317 [120 Cal.Rptr.3d 605].)69 Here, USB was not entirely prevented from presenting evidence in support of its defense, but the deprivation was profound. "The erroneous denial of some but not all evidence relating to a claim [citations] differs from the erroneous denial of all evidence relating to a claim .... In the former situation, the appellant must show actual prejudice; in the latter situation, the error is reversible per se." (Gordon v. Nissan Motor Co., Ltd. (2009) 170 Cal.App.4th 1103, 1115 [88 Cal.Rptr.3d 778].)

In granting plaintiffs' motions in limine restricting the evidence USB would be allowed to present, the trial court effectively prevented USB from establishing its affirmative defense as to classwide liability. The record on appeal supports the inference that a large percentage of the absent class member plaintiffs were properly classified and that USB did not owe them any overtime pay. Thus, there is a reasonable probability that in the absence of the error, USB would have received a more favorable result.70

In large part, the trial court appears to have relied on the lack of a uniform policy requiring BBO's to spend the majority of their time outside the office. In Vinole v. Countrywide Home Loans, Inc. (9th Cir. 2009) 571 F.3d 935, issued on the same day as its opinion in Wells Fargo I, the Ninth Circuit discussed the types of common proof that could suffice to establish the predominance of common issues. While an employer's "uniform application of an exemption to employees" is one factor, district courts should also consider "whether the employer exercised some level of centralized control in the form of standardized hierarchy, standardized corporate policies and procedures governing employees, uniform training programs, and other factors susceptible to common proof." (Vinole, supra, at p. 946.) In discounting USB's claims that BBO's were directed to spend the majority of their time outside their branch offices, the trial court emphasized "the amount of work time BBO's are in or out of bank locations is not monitored or tracked in any way."71 If this is the case, however, it follows that the only way to determine with certainty if an individual BBO spent more time inside or outside the office would be to question him or her individually.

Fundamentally, the issue here is not just that USB was prevented from defending each individual claim but also that USB was unfairly restricted in presenting its defense to classwide liability. With that in mind, the cases relied on by plaintiffs are inapposite. Both Long v. Trans World Airlines, Inc. (N.D.Ill. 1991) 761 F.Supp. 1320 [protective order limited discovery of information from plaintiff flight attendants to a representative sample of class members], and In re Antibiotic Antitrust Actions (S.D.N.Y. 1971) 333 F.Supp. 278 [states sought recovery for alleged overcharges in the sale of certain antibiotics], concerned the damages phase of a trial, not the liability phase.

Plaintiffs also claim the declarations were properly excluded because many of the declarations were repudiated by the BBO's who signed them. According to plaintiffs, there was evidence that five of the 10 plaintiff declarants referenced by USB in its appellate brief were, in fact, misclassified. That some of these declarations were later repudiated by their authors does not mean they are not relevant to USB's defense. Further, plaintiffs do not account for the prior four named plaintiffs who all testified in their depositions that they spent more than half their time outside the office. This evidence, in our view, is both relevant and highly probative as to USB's affirmative defense.

Plaintiffs declare: "At bottom, this appeal is about whether the class action will survive as an effective method to try wage and hour misclassification cases." We doubt the situation is quite this dire. Bell III itself was a class action involving wage and hour misclassification, suggesting that not all such cases are doomed to failure under current law. Plaintiffs also claim it would take 520 days to complete a trial of all 260 class members' claims, and argue the trial court here "did exactly what this court and the California Supreme Court have been urging for years" in attempting to efficiently dispose of these claims. To the contrary, we have never advocated that the expediency afforded by class action litigation should take precedence over a defendant's right to substantive and procedural due process.

C. Application of the Doehr Test to the Present Case

The denial of due process that occurred here is sufficient to satisfy the three-part Doehr test. The private interest at issue here is approximately $15 million in back wages and prejudgment interest that USB is compelled to pay under the judgment. This is clearly a significant interest and thus this factor weighs in favor of USB.

With respect to the risk of erroneous deprivation through the procedures under attack and the probable value of additional or alternative safeguards, we conclude the trial court's management of this case created a high risk that USB will be compelled to pay money to absent plaintiffs who may not be entitled to recovery, either because they generally came within the outside salesperson exemption during their tenure with USB, or because they never worked overtime. The trial court forbade USB from introducing evidence as to any non-RWG class member's right to recover, notwithstanding the 70-plus declarations that had been signed, under penalty of perjury, by BBO's attesting they were properly classified. A fair procedure would have allowed USB the opportunity to inquire into the specific circumstances of these absent class members. Further, even assuming the entire class is entitled to recover unpaid overtime, the method of determining the restitution owed failed to comport with established statistical principles and resulted in such a high margin of error as to render the judgment constitutionally infirm.

With respect to the probable value of additional or alternative safeguards, we can envision procedures that would have lessened the danger of an erroneous result. In striking the double-overtime award in Bell III, we noted that under Doehr "the absence of evidence regarding possible alternatives is relevant not only to the risk of error but also to the ancillary government interest in the procedure. The government cannot have an interest in a procedure if superior alternative procedures are available." (Bell III, supra, 115 Cal.App.4th 715, 756-757.)

As to the final prong of the Doehr test, plaintiffs in this case have a strong interest in obtaining redress from USB. We also do not gainsay the substantial governmental interest in conserving scarce judicial resources. Class action lawsuits are intended to conserve judicial resources and to avoid unnecessarily repetitive litigation. Efficiencies must be maintained, sometimes resulting in imperfect results. A certain amount of variability can be tolerated. However, the trial management plan followed here prevented USB from submitting any relevant evidence in its defense as to 239 class members out of a total class of 260 plaintiffs. Whether the trial court would have given credence to such evidence is beside the point. A trial in which one side is almost completely prevented from making its case does not comport with standards of due process.

(6) In sum, the court erred when, in the interest of expediency, it constructed a set of ground rules that unfairly prevented USB from defending itself. These ground rules were the product of the trial court. We do not suggest that the implementation of any particular additional procedural tool would have satisfied due process. We simply hold that the court, having agreed to try this matter as a class action, denied USB the opportunity to defend itself by flatly foreclosing the admission of potentially relevant evidence.

III. The Authority Relied on by Plaintiffs Is Distinguishable

Plaintiffs refer us to Dilts v. Penske Logistics, LLC (S.D.Cal. 2010) 267 F.R.D. 625, 638 (Dilts), in which the district court found the use of statistical sampling would be an "acceptable method" to prove liability in a class action. The court in Dilts first noted that "California and Federal courts have not discouraged the use of statistical sampling in determining class member damages." (Ibid.) As to liability, the court stated "the use of statistical sampling, at least when paired with persuasive direct evidence, is an acceptable method of proof in a class action. [Citation.] Thus, certification will not be denied simply because Plaintiffs anticipate using representative evidence at trial." (Ibid., italics added.) While we do not disagree with the proposition that statistical sampling is a tool that may be utilized in appropriate cases, it does not follow that it was proper for the trial court in this case to limit presentation of USB's affirmative defense solely to the 21 members of the representative group.72

Dilts also concerned a case at the certification stage. The district court in that case had not yet developed or approved of a particular trial plan, observing that at this stage of the proceedings, "the exact details of [the employees'] plan are unnecessary to determine whether common issues predominate in this matter." (Dilts, supra, 267 F.R.D. 625, 638.) The court merely allowed for the possibility that the plaintiffs, with help from their expert witness, would be able to come up with an acceptable trial plan based on representative testimony, and found that the use of representative testimony was not per se a violation of the defendant's due process rights. (Id. at p. 639.)73 In sum, Dilts does not persuade us that the procedures followed by the trial court after the initial certification in the present case are satisfactory.

Plaintiffs also point to several federal cases that we cited to with approval in Bell III: In re Chevron U.S.A., Inc. (5th Cir. 1997) 109 F.3d 1016 (Chevron), In re Simon II Litigation (E.D.N.Y. 2002) 211 F.R.D. 86 (Simon II), and Hilao, supra, 103 F.3d 767. Those cases also fail to validate the procedures followed by the trial court in this case.

The class in Hilao was composed of roughly 10,000 Philippine nationals who had been tortured, summarily executed or "disappeared" by the regime of Ferdinand E. Marcos. (Hilao, supra, 103 F.3d 767, 771-772.) After liability and exemplary damages had been determined by a jury in the first two phases of the trifurcated proceeding, the court permitted compensatory damages to be extrapolated to the class as a whole on the basis of 137 claims that were randomly selected and tried. (Id. at pp. 782-784.) The number of claims was chosen after an expert in statistics testified that the examination of a random sample of 137 claims would achieve " `a 95 percent statistical probability that the same percentage determined to be valid among the examined claims would be applicable to the totality of claims filed'." (Id. at p. 782.)

Unlike Hilao, here the trial court used a random sample group to determine liability as well as to determine the foundational figures to be used in extrapolating restitution to the entire class. The number of sample members was chosen without the input of any statistical expert. Moreover, USB did not waive any challenge to the computation of restitution. The trial court's use of this sampling procedure to determine both liability and monetary recovery appears to be entirely unprecedented.74 We also note that the class here is comprised of 260 members, not the 10,000-plus members in Hilao, suggesting that it would not have been impossible, even within the framework of this class action lawsuit, to conduct some type of individualized inquiries as to each plaintiff's entitlement to damages.

Chevron is also not favorable to plaintiffs. That case concerned a tort claim for industrial pollution of a residential subdivision. The trial plan "provided for a unitary trial on the issues of `general liability or causation' on behalf of the remaining plaintiffs, as well as the individual causation and damage issues of the selected plaintiffs, and ordered the selection of a bellwether group of thirty (30) claimants, fifteen (15) to be chosen by the plaintiffs and fifteen (15) to be chosen by Chevron." (Chevron, supra, 109 F.3d 1016, 1017.) The goal of this trial "was to determine its liability, or lack thereof, in a single trial and to establish bellwether verdicts to which the remaining claims could be matched for settlement purposes." (Ibid.)

The Fifth Circuit found the district court's plan to be invalid. (Chevron, supra, 109 F.3d 1016, 1021.) The court was particularly concerned that the district court's trial plan was "devoid of safeguards designed to ensure that the claims against Chevron of the non-represented plaintiffs as they relate to liability or causation are determined in a proceeding that is reasonably calculated to reflect the results that would be obtained if those claims were actually tried." (Id. at p. 1020.) Instead, the court found the procedure created potential liability to 3,000 plaintiffs "by a procedure that is completely lacking in the minimal level of reliability necessary for the imposition of such liability." (Ibid.) The court focused its concern on the nonrepresentative nature of the bellwether sample group. (Id. at pp. 1020-1021.)

While the trial court here did not follow the trial plan faulted in Chevron, it still failed to comport with the reliability standards annunciated in that decision. Here, the sample used as the basis for the RWG was not a true random sample because it included the two named plaintiffs, who were not selected as part of the initial sample. Nor can we say with confidence that the sample was a "statistically significant" one, in that the 20-person figure was selected by the trial court with no input from any expert witness as to its representativeness for extrapolation purposes.75 Accordingly, we must disagree with plaintiffs' assertion that "the trial plan in the instant case properly applied well-accepted principles of statistical inference and representative testimony."

Simon II is also distinguishable, in part because it involved hundreds of thousands of potential plaintiffs. (Simon II, supra, 211 F.R.D. 86, 153.) Further the defendant in that case was not restricted to the sample group members in presenting its defense: "In addition to statistical evidence, parties will be permitted to present to the jury relevant lay testimony, expert testimony, and documentary evidence—subject to the constraints of the Federal Rules of Evidence and the practical considerations of trial management." (Id. at p. 154.) The district court recognized that "Experts have developed appropriate modeling techniques for reaching statistically significant and reliable conclusions." (Id. at p. 153.) In Bell III, we recited this passage in support of the general proposition that there is "little basis in the decisional law for a skepticism regarding the appropriateness of the scientific methodology of inferential statistics as a technique for determining damages in an appropriate case." (Bell III, supra, 115 Cal.App.4th 715, 755, italics added.) We did not cite to Simon II in support of the proposition that liability determinations in class actions may be made by extrapolating from a random sample, particularly where the sampling methodology was derived without the benefit of expert statistical advice.

IV. The 43.3 Percent Margin of Error

(7) USB claims the results obtained in the present case are so unreliable as to render the judgment unconstitutional, emphasizing the 43.3 percent margin of error. We note the spread between Drogin's high-end and low-end estimates for the number of weekly overtime hours worked per employee is 10.28 hours.76 Setting aside issues of whether the sampling method was invalid as a means of proving liability, a due process violation is clearly implicated where the method for determining restitution has the potential to increase a defendant's aggregate liability by close to double that which would be warranted if the low end of the margin were applied. (See Bell III, supra, 115 Cal.App.4th 715, 751-753; Hilao, supra, 103 F.3d 767, 786.)

USB claims the trial court's reliance on the factors we mentioned in dicta in Bell III, when we declined to create a bright-line rule for when a margin of error becomes unconstitutional, are not present here and could not salvage the result obtained here even if they were. As noted above, in Bell III we stated: "The reliability of an estimate subject to a large margin of error might conceivably be bolstered by evidence of a high response rate, probable distribution within the margin of error, absence of measurement error, or other matters." (Bell III, supra, 115 Cal.App.4th 715, 756.) The trial court found a high response rate in that 19 out of the 20 RWG members testified, there was no measurement error because the court's findings were "facts" that Drogin extrapolated from, and because the inadmissible Krosnick survey served as the basis for determining probable distribution.

First, USB correctly notes that the passage in Bell III quoted above is dicta. Second, the passage presupposes the use of accepted statistical principles in arriving at the initial result. Here, acceptable statistical principles were not followed in Phase I of this trial and it is undisputed that the findings obtained therein were used as the basis for the restitution calculations in Phase II. Regardless of whether Drogin accurately extrapolated his results from the data contained in the statement of decision, it is the underlying data itself that is constitutionally suspect.77

Additionally, USB claims the response rate was not "extremely high" because six of the original randomly selected RWG members did not respond, namely, the four members who opted out after the second opt-out notice was sent, Smith (who was removed by the court), and Borsay Bryant (who did not testify). USB also claims there was measurement error because several RWG witnesses testified to having worked a range of hours without specifying how often they worked at either the high or the low end of that range, and there was no admissible evidence as to probable distribution of hours worked by non-RWG members. We find these arguments compelling.

"Underlying the contemporary reliance on the methodology of inferential statistics is a recognition that `[e]xperts have developed appropriate modeling techniques for reaching statistically significant and reliable conclusions.' [Citations.]" (Bell III, supra, 115 Cal.App.4th 715, 754.) Here, the plan used by the trial court was not based on appropriate modeling techniques as developed by experts. As in the flawed double-time award in Bell III, the "parties' experts did not offer foundational calculations for the determination of double-time or propose an appropriate class size, margin of error, or sampling methodology." (Id. at p. 756.) The results achieved here thus do not reflect a statistically significant and reliable methodology. Accordingly, the judgment must be reversed.

The trial court's error was prejudicial as there was evidence in the form of deposition testimony, as well as the pretrial declarations obtained by USB, that a substantial portion of the class was properly classified as exempt. For all the reasons stated above we conclude the judgment must therefore be reversed. In light of our decision, we need not address USB's other arguments in favor of reversing judgment. We must, however, address the issue of whether the trial court erred in certifying this case as a class action initially or in maintaining it as such by denying USB's two motions to decertify. We conclude the trial court abused its discretion in denying USB's second motion to decertify.

V. The Trial Court Erred In Denying USB's Second Motion to Decertify

USB contends that the trial court abused its discretion in certifying the class of BBO's in 2005, and in later denying the motions for decertification USB filed in 2007 and 2008. USB claims the court's rulings were in error because the law and the evidence in this case rendered class treatment inappropriate as "individual issues predominated and were unmanageable." It claims it was harmed "because the court improperly proceeded on a class basis and subjected USB to its statistically invalid and unconstitutional trial plan that erroneously and improperly resulted in an adverse $15 million judgment."

A. Class Certification and the Standard of Review

Class actions are authorized when there is a question of "common or general interest, of many persons, or when the parties are numerous, and it is impracticable to bring them all before the court, one or more may sue or defend for the benefit of all." (Code Civ. Proc., § 382.) While class actions "offer a means of avoiding `repetitious litigation' [citation] and `a multiplicity of legal actions dealing with identical basic issues ...' [citation]" (Bell III, supra, 115 Cal.App.4th 715, 741) by greatly expediting the resolution of claims, class actions can create possibilities for injustice in an individual case. (City of San Jose v. Superior Court (1974) 12 Cal.3d 447, 458 [115 Cal.Rptr. 797, 525 P.2d 701] (City of San Jose).)

(8) Class certification is appropriate when there is "(1) ... a sufficiently numerous, ascertainable class, (2) ... a well-defined community of interest, and (3) [proof] that certification will provide substantial benefits to litigants and the courts, i.e., that proceeding as a class is superior to other methods. [Citations.] In turn, the `community of interest requirement embodies three factors: (1) predominant common questions of law or fact; (2) class representatives with claims or defenses typical of the class; and (3) class representatives who can adequately represent the class.' [Citation.]" (Fireside Bank v. Superior Court (2007) 40 Cal.4th 1069, 1089 [56 Cal.Rptr.3d 861, 155 P.3d 268]; see Sav-On, supra, 34 Cal.4th 319, 326.)

Common issues predominate when they would be "the principal issues in any individual action, both in terms of time to be expended in their proof and of their importance ...." (Vasquez v. Superior Court (1971) 4 Cal.3d 800, 810 [94 Cal.Rptr. 796, 484 P.2d 964].) "[T]he community of interest requirement is not satisfied if every member of the alleged class would be required to litigate numerous and substantial questions determining his individual right to recover ...." (City of San Jose, supra, 12 Cal.3d 447, 459.) Our Supreme Court has stated: " `The ultimate question in every case of this type is whether ... the issues which may be jointly tried, when compared with those requiring separate adjudication, are so numerous or substantial that the maintenance of a class action would be advantageous to the judicial process and to the litigants.' [Citations.]" (Lockheed Martin Corp. v. Superior Court (2003) 29 Cal.4th 1096, 1104-1105 [131 Cal.Rptr.2d 1, 63 P.3d 913].)

A trial court's rulings on class certification are reviewed for abuse of discretion. (Sav-On, supra, 34 Cal.4th 319, 326.) " `Because trial courts are ideally situated to evaluate the efficiencies and practicalities of permitting group action, they are afforded great discretion in granting or denying certification.... [Accordingly,] a trial court ruling supported by substantial evidence generally will not be disturbed "unless (1) improper criteria were used [citation]; or (2) erroneous legal assumptions were made [citation]" [citation].... "Any valid pertinent reason stated will be sufficient to uphold the order."' [Citations.]" (Id. at pp. 326-327.)

B. Decertification

While USB contests all the certification decisions made by the trial court, we focus on the denial of USB's second motion to decertify, which was filed after the completion of Phase I of the trial.

(9) A trial court always retains the option of decertification. (Lazar v. Hertz Corp. (1983) 143 Cal.App.3d 128, 144 [191 Cal.Rptr. 849].) Decertification may become appropriate in cases where variances in admissible evidence cast serious doubts as to the prevalence of common issues affecting liability. (See Ali v. U.S.A. Cab Ltd. (2009) 176 Cal.App.4th 1333, 1350 [98 Cal.Rptr.3d 568] ["When variations in proof of harm require individualized evidence, the requisite community of interest is missing and class certification is improper."]; compare with Bufil v. Dollar Financial Group, Inc. (2008) 162 Cal.App.4th 1193, 1207, 1208 [76 Cal.Rptr.3d 804] [certification appropriate where employer had policy of prohibiting certain employees from taking breaks].)

The three-factor community of interest analysis applies equally to an order decertifying a class as well as an order granting certification. (See Walsh, supra, 148 Cal.App.4th 1440, 1450-1451 [order decertifying subclass].) If individualized issues arise out of a defendant's affirmative defense, the predominance factor can be defeated: "In examining whether common issues of law or fact predominate, the court must consider the plaintiff's legal theory of liability. [Citation.] The affirmative defenses of the defendant must also be considered, because a defendant may defeat class certification by showing that an affirmative defense would raise issues specific to each potential class member and that the issues presented by that defense predominate over common issues." (Id. at p. 1450.)

C. Examples of Decertification Rulings

In Keller v. Tuesday Morning, Inc. (2009) 179 Cal.App.4th 1389, 1391 [102 Cal.Rptr.3d 498], a trial court had certified a class action lawsuit filed by store managers alleging the defendant had wrongly failed to pay overtime wages. Two years later, the court granted a motion to decertify the class, finding that individual issues predominated over common issues. (Ibid.) In support of its motion to decertify, the defendant had introduced a statistical analysis based on the declarations of 45 managers indicating they spent anywhere from 10 to 100 percent of their worktime on managerial tasks. (Id. at p. 1394.) In granting the motion to decertify, the trial court observed "the question of mandated management policies was subject to classwide proof, yet the amount of time a manager spent performing these acts and his or her exercise of discretion are matters of individual inquiry." (Id. at p. 1396.) The appellate court sustained the decision to decertify the class, finding the lower court's conclusions were supported by substantial evidence. (Id. at p. 1399.)

Other appellate courts have found decertification appropriate in similar circumstances. In Marlo v. United Parcel Service, Inc. (9th Cir. 2011) 639 F.3d 942 (Marlo), the district court decertified a class action lawsuit after it determined the plaintiff had not provided " `common proof to support a class-wide judgment as to liability....' [Citation.]" (Id. at p. 948.) The putative class was comprised of full-time supervisors, who had allegedly been misclassified under the executive and administrative employee exemptions. (Id. at p. 944.) Four years after certification, the district court decertified the class on the ground that the plaintiff had failed to establish predominance under Rule 23(b)(3). (Marlo, supra, at p. 944.) The plaintiff had relied heavily on an annual employee survey conducted by the employer, which the court found "was neither reliable nor representative of the class." (Id. at p. 945.) The court also found individual testimony and evidence of the exemption policy was insufficient to allow a fact finder to make a classwide judgment. (Ibid.)

On appeal, the Ninth Circuit concluded the district court had not erred. (Marlo, supra, 639 F.3d 942, 949.) The appellate court noted the existence of a common classification policy did not necessarily establish the entire class was misclassified "because the policy may have accurately classified some employees and misclassified others." (Id. at p. 948.) The court also cited to our decision in Dunbar in concluding the district court did not err in requiring a week-by-week determination of exempt status. (Marlo, supra, at p. 948.) An additional survey of 160 class members was also deemed " `unrepresentative, unreliable, and [having] essentially no probative value' " because the survey's designer could not state whether the sample was representative of the class. (Id. at p. 949; see also Walsh, supra, 148 Cal.App.4th 1440, upholding decertification of the account manager subclass.)

D. Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319 [17 Cal.Rptr.3d 906, 96 P.3d 194]

In response to USB's argument at the initial certification stage that commonality did not exist because Ramirez requires an individualized analysis to determine whether an employee is properly classified as exempt, the trial court stated that the Supreme Court's opinion in Sav-On "dissecting the merits" of Ramirez was the "complete answer to that argument." In deciding to certify the class, the court relied heavily on USB's company policy of classifying BBO's as exempt, and also on evidence that it failed to train and monitor its employees regarding their exempt status.

The Supreme Court did consider Ramirez in its subsequent opinion in Sav-On. In Sav-On, two salaried drugstore managers filed a class action against their employer, arguing that the defendant had misclassified them under the managerial exemption based on their job title and description, without reference to their actual work. In seeking to certify the class, the plaintiffs offered evidence suggesting putative class members were obligated to perform nonexempt work during more than half of their workdays and were thus entitled to overtime pay. (Sav-On, supra, 34 Cal.4th 319, 324-325.)

On review, the Supreme Court found substantial evidence supported the trial court's conclusions that "deliberate misclassification was defendant's policy and practice" and that "classification based on job descriptions alone resulted in widespread de facto misclassification." (Sav-On, supra, 34 Cal.4th 319, 329.) Emphasizing the great deference given to a trial court's certification order, the opinion concluded the court had not abused its discretion in certifying the class. (Id. at pp. 329, 331.) In so ruling, the court found Ramirez inapposite not just because that case did not involve the managerial exemption, but also because Ramirez did not involve a class action claim. As the court in Sav-On observed, "we did not even discuss certification standards [in Ramirez], let alone change them." (Sav-On, supra, at p. 336.) The court concluded: "Accordingly, Ramirez is no authority for constraining trial courts' `great discretion in granting or denying certification' [citation] or ... for applying a particular set of `factors' whenever plaintiffs in an overtime case seek class certification." (Ibid.)

(10) By the time USB filed its second decertification motion, the RWG procedures employed by the trial court had already severely impinged on USB's right to prove its exemption defense as to the 239 absent class members. "A class action may be maintained even if each member must individually show eligibility for recovery or the amount of damages. But a class action will not be permitted if each member is required to `litigate substantial and numerous factually unique questions' before a recovery may be allowed. [Citations.] ... `[I]f a class action "will splinter into individual trials," common questions do not predominate and litigation of the action in the class format is inappropriate. [Citation.]' [Citations]" (Arenas v. El Torito Restaurants, Inc. (2010) 183 Cal.App.4th 723, 732 [108 Cal.Rptr.3d 15].)

As the Supreme Court stated in Sav-On, "if unanticipated or unmanageable individual issues do arise, the trial court retains the option of decertification." (Sav-On, supra, 34 Cal.4th 319, 335.) When USB filed its second motion to decertify, it had already made repeated efforts to introduce evidence pertaining to its affirmative defense. Thus, this case is further distinguishable from Sav-On, wherein the court stated that because the burden of proof is on the employer, it was not equitable to require plaintiffs to demonstrate that the entire class was nonexempt at the initial certification stage. (Id. at p. 338.) The trial court's denial of the second motion to decertify was based on the erroneous legal assumption that a finding of liability due to misclassification could be determined by extrapolating the findings based on the RWG to the entire class. As we have demonstrated, this conclusion was legally unsound because the methodology employed by the trial court was legally unsound; it violated USB's right to present relevant evidence in its defense.

E. The Trial Court Erred in Relying on USB's Uniform Classification of BBO's

USB also claims that the trial court gave excessive weight to its finding that USB classified all BBO's as exempt and did so without any inquiry as to "any particular employee's job duties, hours worked, performance or any other factor."78 We agree that the court erred in focusing on USB's policies.

In Spainhower v. U.S. Bank National Assn. (C.D.Cal., Mar. 25, 2010, No. 2:08-CV-00137-JHN-PJWx) 2010 U.S.Dist. Lexis 46316, the district court denied a motion to certify a class action of in-store bank managers. The defendant (the same defendant as in the present case) contended the putative class members were exempt from overtime pay and meal and rest requirements under the executive, administrative, and outside sales exemptions. (Id. at pp. *2-*3.) After discussing Vinole, Wells Fargo I, and Wells Fargo II, the court found the plaintiffs had failed to establish questions of fact common to the class predominated over questions affecting only individual members, as required under Rule 23(b)(3) of the Federal Rules of Civil Procedure (28 U.S.C.). (Spainhower, supra, at p. *11.) The plaintiffs, as in this case, asserted the defendant's expectation was solely that the employees would make their branch goals, and the bank "had no expectation as to how those goals were to be met." (Ibid.) As the district court recognized, "It is Plaintiffs' own argument that weighs against class certification. With substantial discretion as to how to operate one's branch comes the likelihood of substantial differences in how each member of the proposed class spent his or her workday. These likely variances weigh against the notion that common proof would establish how each employee actually spent his or her time or what percentage of each employee's work was spent in exempt versus nonexempt activities." (Id. at pp. *11-*12.)

At this juncture, we need not speculate as to whether a workable trial plan could have been devised to account for these individual inquiries. In view of the many courts that have considered this problem at the classification stage, it is doubtful that such a plan could be successfully implemented. Here, the trial court attempted to manage the individual issues in the first phase of this trial by resorting to an unproven statistical sampling methodology that denied USB the right to properly defend the claims against it. As we have demonstrated, the plan fell short. Accordingly, we conclude the failure to grant USB's second motion to decertify was an abuse of discretion.

"Under the abuse of discretion standard, `a reviewing court should not disturb the exercise of a trial court's discretion unless it appears that there has been a miscarriage of justice.' [Citation.]" (Cristler v. Express Messenger Systems, Inc. (2009) 171 Cal.App.4th 72, 80 [89 Cal.Rptr.3d 34].) We have already concluded that the trial court's trial management plan constituted a miscarriage of justice. Accordingly, the class is ordered decertified.


The judgment is reversed. We further order the case decertified as a class action. The order awarding expert witness fees for Miles Locker to plaintiffs is reversed.79 The matter is remanded to the trial court to reconsider the two named plaintiffs' meal and rest break period violation claims in light of the Supreme Court's soon-to-be-issued ruling in Brinker Restaurant Corp. v. Superior Court, supra, 165 Cal.App.4th 25, review granted October 22, 2008, S166350.

Costs on appeal are awarded to USB.

Marchiano, P. J., and Margulies, J., concurred.


1. The two consolidated appeals are from the judgment after court trial (A125557) and from an award of expert witness fees with respect to plaintiffs' expert witness Miles Locker (A126827).
2. The Supreme Court heard oral argument on November 8, 2011. A decision should issue shortly after this opinion is filed.
3. California employees are generally entitled to overtime pay for work of more than eight hours a day and 40 hours per week. (See, e.g., Lab. Code, § 510; Cal. Code Regs., tit. 8, § 11040 et seq.) There are, however, several classes of employees who are exempt from overtime laws, such as outside salespersons.
4. Labor Code section 1194, subdivision (a), authorizes an employee who received less than the applicable legal minimum wage or overtime compensation to seek recovery from his or her employer.
5. While Labor Code section 1171 uses the term "outside salesman," the regulation that is the primary focus of this opinion uses the term "outside salesperson." For the sake of consistency, we will use the term "outside salesperson."
6. For convenience, this opinion will use the BBO abbreviation to refer to both SBB's and BBO's, with exceptions made in reference to individual plaintiffs.
7. USB does not maintain any documents showing the daily or weekly hours worked by BBO's.
8. For example, BBO's are expected to consult with customers and prospects at their place of business "in order to understand their current business practices and identify their financial goals and needs." BBO's are also expected to represent the bank at various civic and community functions, and to conduct joint outside sales calls with other branch employees at potential customers' places of business.
9. While USB claimed BBO's were properly classified under three different exemptions, the primary issue at trial was the application of the outside salesperson exemption from overtime wages.
10. Code of Civil Procedure section 382 authorizes class action lawsuits "when the question is one of a common or general interest, of many persons, or when the parties are numerous, and it is impracticable to bring them all before the court . . . ."
11. When the case was certified, the class contained 282 class members. By the time of judgment, the class consisted of 260 members.
12. Subsequently, the latter date was extended to September 26, 2005.
13. On June 22, 2005, this court denied USB's petition for writ of mandate challenging the trial court's decision to certify the class action. (U.S. Bank National Assn. v. Superior Court, A110059.)
14. Employees are administratively exempt if they "(1) perform `office or non-manual work directly related to management policies or general business operations' of the employer or its customers, (2) `customarily and regularly exercise[] discretion and independent judgment,' (3) `perform[] under only general supervision work along specialized or technical lines requiring special training' or `execute[] under only general supervision special assignments and tasks,' (4) [are] engaged in the activities meeting the test for the exemption at least 50 percent of the time, and (5) earn twice the state's minimum wage." (Eicher v. Advanced Business Integrators, Inc. (2007) 151 Cal.App.4th 1363, 1371 [61 Cal.Rptr.3d 114], quoting Cal. Code Regs., tit. 8, § 11040, subd. 1(A)(2).) California's overtime pay requirements also do not apply to employees whose earnings exceed one and one-half times the minimum wage if more than one-half of their compensation consists of commissions. (Cal. Code Regs., tit. 8, § 11040, subd. 3(D).)
15. It appears plaintiffs were referring to the following passage in Sav-On: "Contrary to defendant's implication, our observation in [Ramirez v. Yosemite Water Co. (1999) 20 Cal.4th 785 [85 Cal.Rptr.2d 844, 978 P.2d 2]] that whether the employee is an outside salesperson depends `first and foremost, [on] how the employee actually spends his or her time' [citation] did not create or imply a requirement that courts assess an employer's affirmative exemption defense against every class member's claim before certifying an overtime class action." (Sav-On, supra, 34 Cal.4th 319, 337, italics added.) We do not read this passage as applying to the trial phase of a class action lawsuit.
16. USB also filed a supplemental brief detailing its due process objections to plaintiffs' proposed methodology. In large part, USB relied on our then recent opinion in Dunbar v. Albertson's, Inc. (2006) 141 Cal.App.4th 1422 [47 Cal.Rptr.3d 83] for the proposition that surveys, sampling, and other statistical data could not be used to infer that all class members were improperly classified because the critical inquiry in exemption cases turns on how each individual employee actually spent his or her worktime.
17. During the hearing, USB's counsel acknowledged there is no right to a jury in a trial of section 17200 claims.
18. In the statement, USB reminded the trial court that it had "repeatedly objected to the use of `representative' testimony and extrapolation in the trial of this matter because it is unconstitutional and violates due process."
19. Section 17200 prohibits three types of wrongful business practices: any (1) unlawful, (2) unfair, or (3) fraudulent business practice or act. (People ex rel. Bill Lockyer v. Fremont Life Ins. Co. (2002) 104 Cal.App.4th 508, 515 [128 Cal.Rptr.2d 463].) "With respect to the unlawful prong, `[v]irtually any state, federal or local law can serve as the predicate for an action' under section 17200. [Citations.] `"[I]n essence, an action based on [section 17200] to redress an unlawful business practice `borrows' violations of other laws and treats these violations, when committed pursuant to [a] business activity, as unlawful practices independently actionable under section 17200 et seq. and subject to the distinct remedies provided thereunder."' [Citation.]" (Ibid.)
20. In their declarations, two of these former plaintiffs, Sean MacClelland and Michael Lewis, claimed they were unaware they had been selected for the RWG when they opted out. Lewis stated plaintiffs' counsel called him several times to encourage him to opt out of the case. MacClelland reported plaintiffs' counsel left intimidating messages on his answering machine. In opposition, plaintiffs submitted a letter dated November 2, 2006, in which both witnesses were informed they had been randomly selected to "have your case tried in this case."
21. USB argued that Proposition 64 (adding Bus. & Prof. Code, § 17203), which passed in November 2004, should be interpreted to mean that in order to have standing all class members must be shown to have lost money or property as a result of a defendant's alleged unlawful activity, not just the named plaintiffs. The trial court noted that this issue was then under review by our Supreme Court, and concluded USB had not presented any binding authority in support of its argument. The Supreme Court subsequently held the new standing requirement applies to the class representatives only, and not to absent class members. (In re Tobacco II Cases (2009) 46 Cal.4th 298, 306 [93 Cal.Rptr.3d 559, 207 P.3d 20].)
22. We note the pleadings and attachments generated by this motion are voluminous.
23. Fitzsimmons acknowledged having received a class action settlement of $8,000 as a result of a lawsuit involving BBO's who had worked at Bank of America, his employer prior to working for USB.
24. Karen Racusin, a regional sales manager, testified for USB and stated that she participated in meetings in which Penza discussed his sales strategies. He never stated that he spent the majority of his time inside his office making telemarketing calls. Based on the number of loan applications that he turned in, and his statements about his sales strategies, she understood that he was out of the office the majority of the time.
25. Nancy McCarthy, a former SBB, also testified as an RWG member. At deposition, she stated that she spent half her time inside the office and half her time outside the office. She also estimated she was outside the office more than half the time during 40 of the 75 weeks she was an SBB. At trial, she testified on direct that she consistently spent 80 percent of her time inside and 20 percent outside. She explained the discrepancy between her deposition testimony and her trial testimony by saying she initially had guessed as to how she spent her time, and only after the deposition did she start "breaking down" what she did on a typical day.
26. Petty's supervisor David Greiner testified that Petty's position was initially designated "business banking officer" and was later changed to "relationship manager." The SBB position was not part of Greiner's group at USB, and Petty was not an SBB. The primary focus of Petty's job was on managing and growing his existing portfolio, and not on seeking outside sales.
27. Pat Collins, one of Duran's supervisors, testified that she informed her BBO's that they would be interviewed by USB attorneys for this lawsuit. None of them indicated to her that they felt pressured into signing declarations.
28. In the declaration, Bradley confirmed he spent the majority of his time outside the bank engaged in sales activities.
29. Another USB executive testified that around 1997 and 1998, key executives at USB decided to compete with major banks by creating the BBO position. The position would be dedicated to selling products, rather than leaving that task as the third or fourth option of branch employees. At the time, USB had a 1.5 percent share of the California market.
30. According to a USB executive, this strategy was important because USB did not advertise, so BBO's were relied on to spread the bank's name through community meetings, groups, and activities.
31. The trial court issued similar rulings on this point during the testimony of other USB managers.
32. When asked whether, based on her review of the weekly activity reports submitted by her BBO's, she had an understanding as to the number of sales calls they conducted on a weekly basis, the court restricted her response to members of the RWG.
33. The weekly pipeline reports prepared by BBO's show the number of telemarketing calls and appointments they make, and their prospective loans. A report will identify the type of prospective loan, the dollar size of the loan, and the location where the loan will be booked. The purpose of the pipeline report is to forecast when a loan will book, and for BBO's to show progress towards their loan goals.
34. On November 14, 2007, the trial court filed its order after hearing, denying USB's motion for judgment. In recounting the trial, the court stated: "For this purpose 25 names were randomly drawn from the class of whom 20 were to serve as trial witnesses . . . on behalf of the class an approach authorized by [Bell III]." The court noted plaintiffs' case was "problematic" due to the impaired credibility of certain RWG members who had signed prior inconsistent declarations, or who had made willful misrepresentations about their employment or education history, sometimes while under penalty of perjury.
35. USB asserted the deposition testimony of the four prior named plaintiffs was admissible as party admissions under Evidence Code section 1220. USB also argued that the deposition testimony and the 70-plus class member declarations were admissible under Evidence Code section 1230 as prior statements of unavailable witnesses, and as statements against interest because the declarations rendered their claims invalid.
36. In its order denying the motion, the trial court observed: "Defendant's lack of enthusiasm for the trial management plan is already preserved for appellate review."
37. "Confidence intervals are a technical refinement, and `confidence' is a term of art. For a given confidence level, a narrower interval indicates a more precise estimate. For a given sample size, increased confidence can be attained only by widening the interval. A high confidence level alone means very little, but a high confidence level for a small interval is impressive, indicating that the random error in the sample estimate is low." (1 Kaye & Freedman, Modern Scientific Evidence: The Law and Science of Expert Testimony (2010-2011) § 6:34, pp. 361-363, fns. omitted.)
38. At a hearing on the statement of decision, the trial court stated its view, "perhaps invoking the vernacular somewhat," that USB treated its BBO's as "cannon fodder." The court also stated: "[I]f you don't grasp the thrust of the Court's finding that it was completely irrelevant to the bank where these folks spent their time as long as the [3] percent market share was increased, then you're missing the basic thrust of the Court's findings at the end of Phase I. That's the key to the case, in the Court's view; grossly over-simplified."
39. Troy Petty's claim for overtime was deemed released by a separation agreement he had previously entered into with USB. Nevertheless, the trial court included the data pertaining to him in the evidence to be used in Phase II. Borsay Bryant, an RWG member, did not appear for trial. The court found Bryant had not waived his claim for compensation and that he would be treated like any other nontestifying class member for purposes of Phase II.
40. USB's motion in limine No. 14 sought to exclude or require an offer of proof in an Evidence Code section 402 hearing as to the testimony of Miles Locker, designated by plaintiffs to testify on the interpretation of wage and hour laws by the Division of Labor Standards Enforcement. The court granted the motion. Locker did not testify.
41. "In a random sample, each member of the population must be equally likely to be selected for the sample, and the chances of being selected must not depend on the selections." (Wazzan & Sulzer, Statistical Analysis and Interpretation of Data Commonly Used in Employment Law Litigation (Spring 2006) 8 Duq. Bus. L.J. 79, 83.)
42. Drogin did not consider individual days of vacation as shown in payroll data if those days were not included in the statement of decision. If such days had been included, this would possibly have led to a different result. However, with respect to the information in the statement of decision, he deemed his results accurate.
43. A pilot study was used in Bell III, which is a tool that is often used in statistical sampling to learn about variations in a population in order to accurately determine the appropriate sample size needed to obtain a certain predefined level of accuracy. Drogin agreed that a pilot study was not conducted here. On cross-examination, he stated he did not see any problem with the fact that no such study was done in this case, noting these studies can take a long time and "sometimes it's not practical to do a pilot study."
44. In Bell III, his pilot study was used to arrive at a reliable estimate regarding the appropriate sample size. The sample size in Bell III was 295 class members out of a population of approximately 2,400 individuals.
45. USB's request for judicial notice (filed in A125557 on June 18, 2010) of plaintiffs' reply to opposition to motion to remand and plaintiffs' request for judicial notice (filed in A125557 on July 2, 2010) of the order granting plaintiff's [motion] to remand, both filed in the United States District Court for Northern California, are denied as the materials were not before the superior court and are not necessary for our determination of the issues raised on appeal.
46. Additionally, Hildreth noted Brian Smith was removed from the sample because he was not a true BBO, suggesting that 5 percent of the remaining class members were also not true BBO's. Similarly, Borsay Bryant's refusal to testify at trial suggests that 5 percent of the remaining class members would also have elected not to testify. Also, Petty could not recover because he had signed a release, but his data was still used to extrapolate to the class. Hildreth stated that for purposes of providing representative data one cannot favor certain characteristics of sample members over other characteristics.
47. Hildreth did not agree with Drogin that the midpoint assumption was reasonable because there was no way to know if the BBO's worked the same number of hours at the lower end of the range as they did at the upper end of the range. Hildreth opined that in order to obtain an estimate with a margin of error at plus or minus one hour the entire class of 260 would need to be sampled. This conclusion was based on the findings associated with the 21 testifying BBO's.
48. Also on this day, USB filed its motion to tax costs. Among the challenges raised, USB contended that the $25,125 in expert fees paid to Miles Locker should not be shifted to it because the court had excluded Locker as an expert witness. USB subsequently filed notices of appeal from the judgment after court trial and from the award of Locker's expert witness fees.
49. In Doehr, a statute permitted plaintiffs in civil suits to obtain ex parte attachments against defendants with a showing so minimal—an averment by the plaintiff that the defendant is liable—that it resulted in a "significant risk of erroneous deprivation." (Doehr, supra, 501 U.S. 1, 21.) The court noted the statute enabled one of the private parties to "`make use of state procedures with the overt, significant assistance of state officials,'" that involve state action "`substantial enough to implicate the Due Process Clause.'" (Id. at p. 11, quoting Tulsa Professional Collection Services v. Pope (1988) 485 U.S. 478, 486 [99 L.Ed.2d 565, 108 S.Ct. 1340].) The court concluded that, absent exigent circumstances, the private party's interest in attaching the property did not justify the burdening of the private property owner's rights without a hearing to determine the likelihood of recovery. (Doehr, supra, at p. 18.)
50. The court in Ramirez also recognized that because an employee might attempt to evade the exemption through substandard performance, courts must also examine "whether the employee's practice diverges from the employer's realistic expectations, whether there was any concrete expression of employer displeasure over an employee's substandard performance, and whether these expressions were themselves realistic given the actual overall requirements of the job." (Ramirez, supra, 20 Cal.4th 785, 802.)
51. The relative margin of error was determined by dividing the absolute margin of error by the estimated weekly hours as follows: 0.9/9.4=0.096, or 9.6 percent. (See Bell III, supra, 115 Cal.App.4th 715, 723-724.)
52. The median is the middle value of a set of numbers arranged in order of size.
53. We noted that a large margin of error "might conceivably be bolstered by evidence of a high response rate, probable distribution within the margin of error, absence of measurement error, or other matters." (Bell III, supra, 115 Cal.App.4th 715, 756.)
54. While it has become acceptable to use statistical inference in determining aggregate damages in a class action suit (e.g., 3 Conte & Newberg, Newberg on Class Actions (4th ed. 2002) § 10:3, pp. 479-482), it also is understood that the possibility of error involved in such an approach may exceed constitutional bounds.
55. It also appears that the claims procedures implemented in Bell III were not used in this case. Rather than requiring absent class members to submit estimates of their average weekly overtime hours, each member was simply awarded the equivalent of 11.86 hours in overtime pay for each week they were employed as BBO's.
56. Plaintiffs acknowledge that USB "consistent with the trial plan, could not introduce evidence about class members outside the RWG group."
57. "The law governing California class actions is comprised of a mixture of federal and state law: California law controls if it exists. Otherwise, `"[i]n the absence of California authority, California courts may look to the Federal Rules of Civil Procedure (FRCP) and to the federal cases interpreting them [citation]." [Citation.]' [Citation.]" (In re BCBG Overtime Cases (2008) 163 Cal.App.4th 1293, 1298 [78 Cal.Rptr.3d 257].)
58. Class actions in federal court brought under rule 23(a) of the Federal Rules of Civil Procedure (28 U.S.C.) must meet the following four prerequisites for class certification: "(1) the class is so numerous that joinder of all members is impracticable; [¶] (2) there are questions of law or fact common to the class; [¶] (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and [¶] (4) the representative parties will fairly and adequately protect the interests of the class." (Fed. Rules Civ.Proc., rule 23(a), 28 U.S.C.)
59. Importantly, both Walsh and Dunbar were decided before the instant case was litigated in the trial court, yet neither case was discussed in the rulings by the trial court below.
60. Rule 23(b)(3) of the Federal Rules of Civil Procedure (28 U.S.C.) allows a class action to be certified provided a court finds "that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy."
61. The court noted that its analysis might have been different had the plaintiff included a statistical study in support of her motion. (Wells Fargo II, supra, 268 F.R.D. 604, 612, fn. 2.)
62. In a footnote, Drogin posed the following scenario: "For example, suppose that results of the questionnaire show 90% of class members are misclassified as exempt. Further, suppose that the random sample included 50 class members, and 80% of the random sample were misclassified as exempt according to the questionnaire responses, but the court ruled that only 72% of the sample were misclassified. Then, the adjustment factor would be 72%/80% = 0.90 and the resulting estimate of percentage misclassified in the class as a whole would be 90%*(0.90) = 81%." This is exactly the kind of calculation Judge Patel faulted in Wells Fargo II, supra, 268 F.R.D. 604, 612.
63. See 1 Saks, Modern Scientific Evidence: The Law and Science of Expert Testimony (2010-2011) section 5:17, stating that in order to yield valid results the size of a sample should be based on three factors: (1) "The homogeneity of the variable to be measured in the population"; (2) "[h]ow narrowly the researcher needs to zero in on the answer" (the desired margin of error); and (3) "[h]ow confident the researcher needs to be that the obtained range around the population parameters is correct" (the desired confidence ratio). Here, none of these factors was taken into account in connection with the initial selection of the 20-person RWG.
64. Interestingly, Drogin was also involved in the Wal-Mart case: "Drogin conducted his analysis region-by-region, comparing the number of women promoted into management positions with the percentage of women in the available pool of hourly workers. After considering regional and national data, Drogin concluded that `there are statistically significant disparities between men and women at Wal-Mart ... [and] these disparities ... can be explained only by gender discrimination.' [Citation.]" (Wal-Mart, supra, 564 U.S. ___, ___ [180 L.Ed.2d 374, 393].)
65. While Wal-Mart is not dispositive of our case, we agree with the reasoning that underlies the court's view that representative sampling may not be used to prevent employers from asserting individualized affirmative defenses in cases where they are entitled to do so.
66. USB points out it unsuccessfully attempted to introduce the sworn statements it had obtained from 78 class members refuting misclassification, members who are now slated to recover at least $6 million from USB, including prejudgment interest.
67. The hearsay rule does not prevent the admission of statements made by a party opponent. (See Evid. Code, § 1220.)
68. "`The term "due process of law" asserts a fundamental principle of justice which is not subject to any precise definition but deals essentially with the denial of fundamental fairness, shocking to the universal sense of justice.' [Citation.] `"The trial of a case should not only be fair in fact, but it should also appear to be fair." [Citations.] A prime corollary of the foregoing rule is that "A trial judge should not prejudge the issues but should keep an open mind until all the evidence is presented to him."' [Citation.]" (In re Marriage of Carlsson (2008) 163 Cal.App.4th 281, 290-291 [77 Cal.Rptr.3d 305].)
69. "No judgment shall be set aside, or new trial granted, in any cause, on the ground of misdirection of the jury, or of the improper admission or rejection of evidence, or for any error as to any matter of pleading, or for any error as to any matter of procedure, unless, after an examination of the entire cause, including the evidence, the court shall be of the opinion that the error complained of has resulted in a miscarriage of justice." (Cal. Const., art. VI, § 13.)
70. USB draws our attention to non-RWG member Nicholas Sternad, who testified at a deposition that he performed exempt administrative and outside sales duties. At the hearing held on May 23, 2005, the trial court granted plaintiffs' summary adjudication motion to disallow the administrative exemption. The trial court refused to give weight to Sternad's testimony, apparently believing that his was an isolated case: "[T]he supplemental evidence shows nothing more than that one employee performed duties and activities not listed on Defendant's BBO job description nor remotely similar to Defendant's own expectations of duties and activities for the BBO position ...." In its supplemental brief in support of the motion for summary adjudication, plaintiffs claimed Sternad ceased performing BBO duties in October 2002. Ultimately, Sternad was awarded about $450,000 (including prejudgment interest) as "restitution," covering the entire period of his employment, from 1997 to the present, including for the period after October 2002.
71. While we recognize the trial court believed the overall requirements of the job could not be met if a BBO spent more than half of his or her worktime outside the office, USB did offer to produce evidence in its defense that could have shown at least some portion of the class was properly classified.
72. "[U]nder current law sampling is a practical option only at the damages stage. There is no conceptual obstacle to using sampling to measure liability, but it would require a major change in tort law. Tort liability is binary: a defendant is either liable or not, and if liable, the defendant must compensate the plaintiff in full. At best, sampling applied to liability can only provide an estimate of the probability that defendant is liable to any plaintiff in an arbitrarily chosen case. This estimate equals the number of liability verdicts divided by the total number of sample cases. Thus, sampling could be used to determine liability only if the tort law recognized probabilistic liability measures." (Bone, Statistical Adjudication: Rights, Justice, and Utility in a World of Process Scarcity (1993) 46 Vand. L.Rev. 561, 597.)
73. In another case from the same district, the court granted a motion to decertify an employment class action that, in part, sought overtime compensation for a subclass of managers. (Weigele v. FedEx Ground Package System, Inc. (S.D.Cal. 2010) 267 F.R.D. 614, 617.) The plaintiffs proposed to determine liability by use of random sampling of subclass members whose testimony would be extrapolated to the entire subclass of managers. (Id. at p. 624.) The court noted: "The major problem with this proposal is that it would likely require a large number of testifying witnesses and no clear or easy manner of determining liability." (Ibid.) The court also noted that the plaintiffs could call fewer witnesses using a reduced level of confidence; however, "as the level of confidence decreases, it becomes increasingly problematic to rely on the results of sampling to extrapolate to the whole class." (Ibid.) The court concluded that logistical difficulties and uncertain class treatment was not a superior method of resolving the litigation. (Id. at p. 625.)
74. We note the Supreme Court in Wal-Mart acknowledged Hilao, but ultimately did not find it persuasive: "Finally, the Court of Appeals determined that the action could be manageably tried as a class action because the District Court could adopt the approach the Ninth Circuit approved in [Hilao]. ... The Court of Appeals [in the opinion below] `s[aw] no reason why a similar procedure to that used in Hilao could not be employed in this case.' [Citation.] It would allow Wal-Mart `to present individual defenses in the randomly selected "sample cases," thus revealing the approximate percentage of class members whose unequal pay or nonpromotion was due to something other than gender discrimination.' [Citation.]" (Wal-Mart, supra, 564 U.S. ___, ___ [180 L.Ed.2d 374, 388].)
75. "How many cases need to be sampled? This depends in large part on the variability of the population. The more diverse the population, the larger the sample must be in order to reflect the population accurately. The more homogeneous the population, the fewer cases that need to be sampled." (Saks & Blanck, Justice Improved: The Unrecognized Benefits of Aggregation and Sampling in the Trial of Mass Torts (1992) 44 Stan. L.Rev. 815, 842.)
76. The absolute margin of error here (11.86 minus 5.14) equals 6.72 (low range); 11.86 plus 5.14 equals 17.00 (high range); 17.00 minus 6.72 equals 10.28. If the 6.72 low range figure is doubled, the result is 13.44 hours, or only 1.58 hours higher than the hourly figure used to compute the award made here. We also note the spread between the high and low range is almost as large as the 11.86 hour figure used to compute plaintiffs' recovery.
77. See Chevron, supra, 109 F.3d 1016, 1020-1021 ("Our substantive due process concerns are based on the lack of fundamental fairness contained in a system that permits the extinguishment of claims or the imposition of liability in nearly 3,000 cases based upon results of a trial of a non-representative sample of plaintiffs. Such a procedure is inherently unfair when the substantive rights of both plaintiffs and the defendant are resolved in a manner that lacks the requisite level of confidence in the reliability of its result.")
78. USB also claims the trial court abused its discretion in extending certification to a class that included numerous BBO's it claims are not entitled to restitution under section 17200. USB contends individuals lack standing to sue unless they have suffered an injury in fact and lost money or property as a result of the alleged unlawful business practice. USB contends individualized issues exist as to whether each class member suffered injury, and as to whether some class members were barred from seeking relief regardless of alleged injury. We need not address this argument.
79. In light of our ruling, USB's request for judicial notice (filed in A126827 on Aug. 2, 2010) of plaintiffs' refusal to withdraw Locker as an expert witness is denied.


1000 Characters Remaining

Leagle.com reserves the right to edit or remove comments but is under no obligation to do so, or to explain individual moderation decisions.

User Comments

Listed below are the cases that are cited in this Featured Case. Click the citation to see the full text of the cited case. Citations are also linked in the body of the Featured Case.

Cited Cases

  • No Cases Found

Listed below are those cases in which this Featured Case is cited. Click on the case name to see the full text of the citing case.

Citing Cases