Samuel Duran v. U.S. Bank National Association

In our second encounter with this class action case, plaintiffs Samuel Duran and

Matt Fitzsimmons appeal from the trial court’s order denying class certification. This

case is a wage and hour class action challenging whether defendant U.S. Bank National

Association (Bank) had properly classified its business banking officers (BBOs) as

exempt employees under the outside salesperson exemption. This exemption applies to

employees who spend more than 50 percent of their workday engaged in sales activities

outside their employer’s place of business. The trial court concluded plaintiffs failed to

demonstrate that the case is manageable as a class action. We affirm.



FACTUAL BACKGROUND AND PROCEDURAL HISTORY

I. Background

The factual background behind plaintiffs’ claims and the procedural history of this

case from its inception are well known to the parties and this court, and we will not repeat

it here. In brief, the case centers on plaintiffs’ allegations that they were misclassified as

outside salespersons when they allegedly actually spent the majority of their working

2

hours inside Bank offices.1

Like the court below, we incorporate by reference pages 13

through 24 of the Supreme Court’s opinion in Duran, supra, 59 Cal.4th 1.) That decision

affirmed our opinion reversing the judgment entered in favor of plaintiffs after a bench

trial, leaving open that the trial court could “entertain a new class certification motion.”

(Duran, at p. 50.) That new motion is the subject of the present appeal.

On July 7, 2014, the remittitur from Duran was filed and the case was reassigned

to a new judicial officer on September 8, 2014, following the Bank’s challenge to the

original trial judge.

II. Motions Regarding Certification

On December 18, 2014, the Bank filed a motion to deny class certification.

On January 21, 2015, the trial court denied plaintiffs’ request to reopen discovery,

finding that the prior discovery was “very extensive.”

Plaintiffs retained survey expert Jon A. Krosnick in connection with their

opposition to the Bank’s motion. On February 1, 2015, Krosnick sent an “ ‘advance

letter’ ” to putative class members, alerting them to an impending survey and the fact that

they might be contacted. The letter enclosed $2 as a “thank you” and promised an

additional $25 to $35 for answering a 20-minute phone survey that would be “completely

confidential.”

On February 13, 2015, Ted Biggs, a senior vice president of the Bank, sent a letter

to putative class members informing them of the Bank’s belief that Krosnick’s letter had

been sent in connection with the pending lawsuit. Among other things, Biggs disputed a

statement in Krosnick’s letter promising respondents that their identities would be kept

confidential.



1 An outside salesperson is one who “customarily and regularly works more than

half the working time away from the employer’s place of business” on sales duties. (See

Industrial Welfare Com., Wage Order No. 4-2001, subd. (2)(m).) Here, there is no

dispute that BBOs were primarily engaged in “sales.” (Duran v. U.S. Bank National

Assn. (2014) 59 Cal.4th 1, 26 (Duran).)

3

On February 27, 2015, the trial court granted, in part, plaintiffs’ ex parte

application seeking, among other things, the authority to send a corrective mailing.2 The

court denied the Bank’s corresponding ex parte application seeking to halt the survey and

to bar further “ ‘unapproved’ ” communications with putative class members. The court

found plaintiffs had a right to conduct the survey in order to support their opposition to

the Bank’s motion. At the same time, the court observed it was unclear “how a survey of

putative class members in this unique context could help resolve the manageability

issues” identified by the Supreme Court in Duran.

On April 30, 2015, plaintiffs filed an opposition to the Bank’s motion to deny

certification, characterizing their opposition as a cross-motion for class certification.3

In

support of their motion, plaintiffs included, among other things, an April 2015 survey

report prepared by Krosnick (2015 Survey). In his report, he indicated his task had been

“to conduct a survey of members of the Duran class with which to generate a menu of

margins of errors for various possible numbers of witnesses who could testify at trial and

answer questions to reveal the average number of hours they worked per week, the

proportion of their work hours that were spent performing sales-related activities, and the

proportion of their work hours that were spent performing outside sales-related

activities.” The 2015 Survey is described in more detail below.

On June 5, 2015, the trial court granted, in part, a motion to compel filed by the

Bank, ordering plaintiffs to produce the identities of the 2015 Survey respondents and

their survey responses. The court also ordered plaintiffs to produce the identities of those

respondents to an earlier Krosnick survey (2008 Survey) who had also responded to the



2 Apart from allowing plaintiffs to depose the Bank’s senior vice president “in

order to get to the bottom of what happened,” the trial court denied plaintiffs’ request to

reopen discovery.

3 The parties submitted extensive evidence to the trial court in connection with the

class certification issue, much of which the court found to be “incredibly redundant.”

4

2015 Survey, together with the 2008 Survey responses of those respondents.

4

In its

order, the court observed plaintiffs had failed to provide a “separate detailed trial plan,”

even though this was “one of the most important take-aways from the Supreme Court’s

decision in this case.”

On August 21, 2015, plaintiffs submitted an expanded trial plan. Plaintiffs

proposed a bifurcated trial, with the liability phase focusing on “ ‘the ultimate question’ ”

of whether the Bank had “realistic” expectations that BBOs could perform their job

duties and meet the production goals of their position while spending more than half of

their working time away from the office. As statistical evidence, they proposed using

representative testimony with sample sizes ranging from 15 to 50 BBOs, having

associated respective margins of error ranging from 11 percent to 5.53 percent. This

liability phase would be followed by a restitution phase to determine the amount of

money owed to the class, “[b]ased on the court’s findings of the average hours worked

per week by the randomly chosen BBOs who testify at trial.” Plaintiffs proposed

allowing the Bank to litigate its affirmative defense by calling BBOs outside the random

sample, subject to the trial court’s discretion to cut off the Banks’ presentation of

evidence under Evidence Code section 352.

On October 2, 2015, the Bank filed its reply brief, expanding the evidentiary

record to include 66 additional declarations from BBOs who testified that they regularly

spent the majority of their time outside Bank locations. The Bank also included a

declaration prepared by its expert, Andrew Hildreth, which criticized the 2015 Survey

and its methodology. In brief, Hildreth concluded the survey suffered from self-selection

bias, as well as serious measurement and estimation errors.



4 The 2008 Survey was ruled inadmissible during the first trial.

5

On November 13, 2015, plaintiffs filed their reply in support of their cross-motion

for class certification, along with a third Krosnick report that addressed Hildreth’s

criticisms.

On March 25, 2016, the trial court, pursuant to Evidence Code section 730,

appointed an independent expert, Kent D. Van Liere, to advise it on the survey science

underpinning the parties’ motions.5



5 Evidence Code section 730 provides, in part: “When it appears to the court, at

any time before or during the trial of an action, that expert evidence is or may be required

by the court or by any party to the action, the court on its own motion or on motion of

any party may appoint one or more experts to investigate, to render a report as may be

ordered by the court, and to testify as an expert at the trial of the action relative to the fact

or matter as to which the expert evidence is or may be required. The court may fix the

compensation for these services, if any, rendered by any person appointed under this

section, in addition to any service as a witness, at the amount as seems reasonable to the

court.” In a footnote in its order, the trial court asked that we give “further guidance on

the use of [Evidence Code] Section 730 to appoint an expert as a court technical

consultant,” citing to Diaz-Barba v. Superior Court (2015) 236 Cal.App.4th 1470, 1490.

The use of experts to assist the court as consultants can be beneficial in complex

cases. We note Federal Rules of Evidence, rule 706 also applies to court-appointed

expert witnesses. Cases addressing this rule may prove helpful to our trial courts. As

was observed in Leesona Corp. v. Varta Batteries, Inc. (S.D.N.Y. 1981) 522 F.Supp.

1304, 1312: “[A] court expert serves not only as a witness on whose opinion the Court

can rely for assistance, but also as both a second set of ears for the court and a teacher

who, unaffected by his having been called as a witness by one side or the other, can

explain the technical significance of the evidence presented. [¶] Naturally a court will

rely heavily on and give great weight to the testimony of its own expert . . . .

Nevertheless, although a scientifically difficult patent case such as this is ideally suited to

the appointment of a court expert pursuant to Rule 706, the appointment of such an expert

does not remove from the trial court’s shoulders the burden of understanding and

becoming fully familiar with the technical questions in the case. The court expert serves

to enhance the trial court’s understanding, but it is the court, and not its expert, that

decides the case.”

California courts have acknowledged the value in court-appointed assistance from

experts, especially in complex cases presenting issues beyond a trial judge’s knowledge.

(People ex rel. Brown v. Tri-Union Seafoods, LLC (2009) 171 Cal.App.4th 1549, 1573–

1574; Mercury Casualty Co. v. Superior Court (1986) 179 Cal.App.3d 1027, 1032.)

6

On March 29, 2016, the Bank submitted a second declaration from Hildreth

addressing new material contained in the third Krosnick report.

III. Denial of Class Certification

On May 19, 2016, the trial court filed its order denying class certification. In its

order, the court concluded plaintiffs had failed to carry their burden of showing that

common questions predominated. The court observed, “A primary factor in the Duran

court’s finding that the class may have been improvidently certified the first time around

was its conclusion that the trial court had ‘received no evidence establishing uniformity

in how BBOs spent their time.’ ” The court concluded that, apart from shoring up

plaintiffs’ trial plan, their statistical evidence “continues to fall short of the kind of

showing of uniformity that the Duran court found necessary.” In spite of the voluminous

evidence submitted by both parties, the court found “[t]he only new material of any

importance lies in the disputes between Krosnick and Hildreth” (italics added). The court

determined the certification issue based on its resolution of those disputes.

Relying on Hildreth’s analysis, the trial court concluded that two central flaws in

the 2015 Survey operated to defeat plaintiffs’ certification efforts: (1) “the issues arising

from the differences between the responses to the 2008 Survey and the responses to the

2015 Survey,” and (2) “the related issue of Krosnick’s use of total average hours worked

per week rather than average overtime hours worked per week.” As to the second point,

the trial court surmised that Krosnick had failed to understand plaintiffs’ theory of

liability.

In setting forth its reasoning, the trial court also indicated it was “initially

satisfied” with how plaintiffs’ trial plan proposed addressing the Bank’s affirmative

defenses. However, upon finding the 2015 Survey data to be unreliable “for any

purpose” (italics added), the court determined plaintiffs could not rely on the survey

instrument in establishing a prima facie case. Additionally, the court observed that

representative sampling could not be used to establish an aggregate restitution award

7

because the proposed sample sizes were not scientifically supportable for that purpose.

The court also found it would not be able to adequately manage the Bank’s affirmative

defense—that at least some class members worked most of their time outside the office—

because the 2015 Survey failed to eliminate the need for “a host of ‘mini trials.’ ” In

sum, the court concluded the pertinent factual questions “present severe manageability

problems.” The court thus denied certification on the grounds that plaintiffs had failed to

demonstrate common issues would predominate or that individual issues could be

effectively managed. This appeal followed.

DISCUSSION

I. Standard of Review for Class Certification

The party seeking class treatment must show (1) an ascertainable and sufficiently

numerous class, (2) a well-defined community of interest, and (3) substantial benefits

from certification that render proceeding as a class superior to the alternatives. (Ayala v.

Antelope Valley Newspapers, Inc. (2014) 59 Cal.4th 522, 529–530 (Ayala); Brinker

Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004, 1021 (Brinker).) The

“community of interest” requirement encompasses three factors: predominant common

questions of law or fact; class representatives with claims or defenses typical of the class;

and class representatives who adequately represent the class. (Ayala, at p. 530; Brinker,

at p. 1021.)

The certification question is a procedural one that does not focus on the merits of

the case. (Brinker, supra, 53 Cal.4th at p. 1023; Sav-On Drug Stores, Inc. v. Superior

Court (2004) 34 Cal.4th 319, 327 (Sav-On).) In evaluating predominance, our Supreme

Court explained: “[A court] must determine whether the elements necessary to establish

liability are susceptible of common proof or, if not, whether there are ways to manage

effectively proof of any elements that may require individualized evidence.” (Brinker, at

p. 1024; accord, Ayala, supra, 59 Cal.4th at p. 533.) “The ‘ultimate question’ the element

of predominance presents is whether ‘the issues which may be jointly tried, when

8

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.’ ” (Brinker, at p. 1021; accord, Duran, supra, 59 Cal.4th at p. 28.) In

addition, the trial court must consider whether individual issues, including those arising

from affirmative defenses, can be managed fairly and efficiently. (Duran, at pp. 28–29;

Washington Mutual Bank v. Superior Court (2001) 24 Cal.4th 906, 922–923; Martinez v.

Joe’s Crab Shack Holdings (2014) 231 Cal.App.4th 362, 378 (Martinez).)

We review the class certification ruling for an abuse of discretion. (Ayala, supra,

59 Cal.4th at p. 530; Brinker, supra, 53 Cal.4th at p. 1022.) Trial courts are afforded

great discretion in granting or denying certification because they are ideally situated to

evaluate the efficiencies and practicalities of permitting group action. (Brinker, at

p. 1022; Sav-On, supra, 34 Cal.4th at pp. 326–327.) Generally, the trial court’s ruling

will not be disturbed unless it is unsupported by substantial evidence or rests on improper

criteria or erroneous legal assumptions. (Ibid.; Linder v. Thrifty Oil (2000) 23 Cal.4th

429, 435–436.) In Ayala, our Supreme Court explained: “We review the trial court’s

actual reasons for granting or denying certification; if they are erroneous, we must

reverse, whether or not other reasons not relied upon might have supported the ruling.”

(Ayala, at p. 530; accord, Alberts v. Aurora Behavioral Health Care (2015)

241 Cal.App.4th 388, 399; Martinez, supra, 231 Cal.App.4th at p. 373.)

II. Duran

In Duran, our Supreme Court discussed the use of surveys and statistical sampling

as evidence to prove class action claims. “[I]f sufficient common questions exist to

support class certification, it may be possible to manage individual issues through the use

of surveys and statistical sampling. Statistical methods cannot entirely substitute for

common proof, however. There must be some glue that binds class members together

apart from statistical evidence. While sampling may furnish indications of an employer’s

centralized practices [citation], no court has ‘deemed a mere proposal for statistical

9

sampling to be an adequate evidentiary substitute for demonstrating the requisite

commonality, or suggested that statistical sampling may be used to manufacture

predominate common issues where the factual record indicates none exist.’ ” (Duran,

supra, 59 Cal.4th at p. 31; accord, Martinez, supra, 231 Cal.App.4th at p. 379; see Tyson

Foods, Inc. v. Bouaphakeo (2016) 136 S.Ct. 1036, 1049 [“Whether a representative

sample may be used to establish classwide liability will depend on the purpose for which

the sample is being introduced and on the underlying cause of action.”].)

Our Supreme Court in Duran noted that “the outside salesperson exemption has

the obvious potential to generate individual issues because the primary considerations are

how and where the employee actually spends his or her workday.” (Duran, supra,

59 Cal.4th at p. 27.) The court stressed that “a statistical plan for managing individual

issues must be conducted with sufficient rigor.” (Duran, at p. 31; accord, Mies v.

Sephora U.S.A., Inc. (2015) 234 Cal.App.4th 967, 985 (Mies).) The court noted that “[i]f

the variability [in the class] is too great, individual issues are more likely to swamp

common ones and render the class action unmanageable.” (Duran, at p. 33.)

III. The Trial Court’s Conclusions Regarding the 2015 Survey

A. The 2015 Survey and Hildreth’s Criticism

Plaintiffs’ theory of liability is that the Bank’s uniform classification of the BBO

position as exempt was unlawful because the position allegedly was designed as an inside

sales/telemarketing job, and realistically could only be performed by spending the

majority of time inside the bank. In its order denying certification, the court noted that

plaintiffs’ trial plan reported “[t]he stated purpose of the Krosnick survey was to

determine the appropriate size of a sample group of witnesses needed to testify at trial in

order to prove, by statistical inference [citation], Plaintiffs’ theory of liability and an

aggregate amount of restitution.”

The trial court found plaintiffs satisfied the requirements of ascertainability,

numerosity, and adequacy of representation. But the court ruled plaintiffs failed to show

10

common questions of law or fact predominated over individual issues. In addition, the

court concluded class treatment was not superior to other means of resolving the claims

because individual issues could not be properly managed. We find these conclusions are

supported by substantial evidence, primarily in the form of Hildreth’s declarations.

Krosnick reported that 160 class members were randomly selected for the survey.

A total of 87 interviews were completed6

and the response rate for the survey was 54

percent. He indicated that 95.45 percent of the respondents reported spending 50 percent

or less of their work time performing outside sales-related activities. Ninety point ninetyone

percent of respondents who stated they spent 50 percent or more of their work time

performing sales-related activities also reported that they worked an average of 63.18

hours per week. Using the survey data, confidence intervals and margins of error were

estimated for sample sizes relating to these findings, ranging from 15 to 50 people, with

the results set forth in chart form. For example, as to the issue of time respondents had

spent performing outside sales-related activities, a sample size of 15 people resulted in a

margin of error of 11.01 percent, whereas a sample size of 50 people had a margin of

error of 5.53 percent.

Hildreth summed up his opinion of the 2015 Survey by concluding Krosnick’s

“proposed ‘menu’ of potential sample sizes and accompanying margins of error is

completely inaccurate. In reality, far larger sample sizes than those proposed would be

required to approach the margins of error presented by Dr. Krosnick . . . .” Hildreth

faulted the survey for having various forms of bias and error, as well as for its improper

use of survey data. For example, Hildreth opined that the 2015 Survey was “a textbook

example of self-selection bias” because it is based on a self-selected sample of

respondents. By subtracting 40 hours from Krosnick’s weekly hour figures, Hildreth also



6 As the trial court noted, most of the tables in the technical appendix show the

number of respondents as 66 rather than 87.

11

showed that the 2008 Survey produced estimates alternatively calculated at 13.82 or

14.39 overtime hours per week, whereas the 2015 Survey produced estimates of 23

overtime hours per week. Hildreth observed: “This significant difference between the

two surveys, from the same population, is of an order of magnitude that illustrates the

degree to which the statistics produced by the Krosnick surveys are not reliable.” He also

concluded that Krosnick had not offered any scientific or tested explanation as to why

such a dramatic difference would have been produced by surveying the same population

with the same questions.

7

Hildreth opined that by comparing the total hours worked per week instead of

average overtime hours per week, Krosnick had masked the extent to which the values in

the two surveys differed. Krosnick’s methodology minimized the difference between the

two surveys’ results and the associated estimated margins of error. Hildreth’s analysis

demonstrated that a statistical comparison of the overtime hours only (versus total hours

worked) showed that the responses given by the same respondents “are so different they

might as well be completely unrelated.” He concluded that if Krosnick’s methodology

were used to estimate the average hours of overtime worked per week, “testimony from

the whole putative class would potentially be required to produce an acceptable relative

or absolute margin of error . . . .”

B. The Trial Court’s Reasoning

The trial court started from the premise that in the absence of a viable “trial ready”

trial plan, “certification must be denied.” The court first noted the need for plaintiffs to

show predominance, which was contingent on the 2015 Survey data: “Plaintiffs need the

survey to support selection of a sample size for the liability phase, to provide a basis for



7 Hildreth also reported that 18 individuals responded to both the 2008 Survey and

the 2015 Survey. The average for the 18 matched responses from the same individuals

shows an increase in the purported average hours worked per week by an average of 5.8

hours. All but two individuals had changed their estimates, with six people reporting

fewer hours and 10 others increasing the hours they worked.

12

an aggregate restitution determination and also, in this case, to provide some comfort to

the court that [the Bank’s] affirmative defense will not lead to a proliferation of mini[]

trials in the first phase of the case.”

The trial court then focused on manageability, specifically, on the need for the trial

plan to adequately address both liability with respect to plaintiffs’ prima facie case, as

well as the Bank’s affirmative defense of the outside sales exemption. While initially

satisfied with the manner in which the proposed trial plan had anticipated addressing the

Bank’s affirmative defense, the court concluded plaintiffs could not establish a prima

facie case as to liability because the 2015 Survey data was unreliable for any purpose.

Second, even if the liability phase was allowed to proceed, the trial court

determined that there was no manageable way to determine an aggregate restitution

award using representative sampling because “the sample size is not scientifically

supported for this purpose.” The damages phase of the trial would thus require testimony

from each class member, leading to a “host of ‘mini trials.’ ” Finally, because of the

unreliability suggested in the comparison between the 2008 and 2015 surveys, the court

concluded that, as with the issue of restitution, “the affirmative defense [that at least some

class members in fact worked most of their time outside the office] would [also] appear

to require a host of ‘mini trials.’ ”

The trial court found the differences between the responses to the 2008 Survey and

the responses to the 2015 Survey, along with Krosnick’s use of total average hours

worked per week rather than average overtime hours worked per week, to be dispositive

of the certification issue. The first court noted that the 2008 Survey solicited essentially

the same information as the 2015 Survey with respect to the total hours worked per week,

yet the responses to the 2008 Survey yielded an average of 53.76 hours, whereas the 2015

Survey yielded an average of 63.18 hours. We concur with the court’s observation that

“[e]ven a non-mathematician can readily see that this is a difference of nearly ten hours

per week, which is a large number in this context.”

13

The trial court found this discrepancy to be “tangible evidence” supporting

Hildreth’s arguments regarding self-interest bias, tainting subject matter areas that were

covered by the 2015 Survey only, namely, whether respondents spent most of their time

inside or outside of the office. The bias identified by Hildreth could have skewed all the

survey results upwards, potentially requiring a larger sample size for plaintiffs to prove

their liability theory, as well as increasing the number of class members the Bank would

call to establish its affirmative defense.

The trial court found that a separate problem arose from plaintiffs’ proposal to use

the same randomly selected witness group for the fixing of aggregate damages. The

court noted that Krosnick had used average total hours worked per week rather than

average overtime hours worked per week in his calculations. This choice was

problematic because the difference in total hours worked “yields significantly different

variation calculations than a comparison between far smaller numbers—here, the

overtime hours derived by subtracting 40 from the total hours.” The court agreed with

Hildreth that in the context of restitution, the focus had to be on overtime hours, not total

hours worked. The variation within the class would be much greater in the overtime

context, and “the necessary sample size needed to achieve the relative margin of error

found acceptable by the court in [Bell v. Farmers Ins. Exchange (2004) 115 Cal.App.4th

715 (Bell)] is driven up dramatically.”8



The trial court concluded that the 2015 Survey results “do not support the

conclusion that—given the small class size in this case—a sample group of any size

smaller than the entire class could ever yield an aggregate restitution award ‘as a matter

of just and reasonable inference.’ [Citation.] In other words, the experience of each class

member would need to be tested not only to determine his or her individual share, but



8

In Bell, the parties’ experts estimated average weekly overtime hours at 9.4

hours, with a plus/minus 0.9-hour margin of error. Bell involved a random sample of

nearly 300 individuals. (Bell, supra, 115 Cal.App.4th at pp. 753, 755–756.)

14

also to arrive at an aggregate amount of restitution in the first place. This, of course, has

severe manageability implications.” Although the court found that a survey like the 2015

Survey could theoretically be used to support a finding as to whether the Bank’s policy of

classifying all class members as exempt employees was lawful, it concluded that a

comparison between the overlapping portions of the 2008 and 2015 surveys

“demonstrates that the data from the 2015 Survey is not sufficiently reliable for

Plaintiffs’ intended use.”

We agree with the trial court that, at this point, the viability of plaintiffs’ motion to

proceed as a class action is critically dependent on the evidentiary value of the 2015

Survey. As the Duran court noted, the lower court in the prior trial “received no evidence

establishing uniformity in how BBOs spent their time.” (Duran, supra, 59 Cal.4th at

p. 32.) The disparity between the estimated overtime hours worked per week as reported

in the 2008 and 2015 surveys (approximately 14 hours in 2008 versus 23 hours in 2015)

is more than enough to convince us that substantial evidence supports the court’s decision

to deny certification. The overtime figure nearly doubled in the 2015 Survey, with a

difference of almost 10 hours per person per week on average. Plaintiffs do not

satisfactorily explain this discrepancy, which the lower court found to have “taint[ed] the

use of the 2015 Survey not only on the subject matter where the two surveys overlap but

also on the subject matter covered only by the 2015 Survey—to wit, where respondents

say they spent most of their time, whether it was inside or outside the office.” We

conclude the court did not abuse its discretion in ruling that the 2015 Survey is unreliable

for the purpose of showing that common issues would predominate at trial. We proceed

to specifically address plaintiffs’ arguments on appeal.

IV. Plaintiffs’ Arguments

A. Whether the Trial Plan Adequately Manages the Affirmative Defense

As part of its ruling, the trial court stated that plaintiffs’ trial plan “fails to resolve

the manageability issues posed by the affirmative defenses, and specifically the argument

15

that at least some of the class members in fact worked most of their time outside the

office.” Plaintiffs assert this passage indicates the court erroneously believed that a

liability determination in a misclassification case must identify which class members

were misclassified. They claim their theory of the case, that the Bank did not have a

realistic expectation that its BBOs would primarily engage in outside sales activities, is

fully consistent with Duran, and assert any finding of liability on this point would be

equally applicable to all BBOs.

Plaintiffs rely heavily on Justice Liu’s concurrence in Duran, supra, 59 Cal.4th 1

and on Martinez, supra, 231 Cal.App.4th 362 in asserting that “ ‘the ultimate question is:

what are “the realistic requirements of the job?” ’ ”

9

Plaintiffs argue that liability in a

misclassification case is determined based not on how employees actually spend their

time, but rather on whether the employer realistically expected employees to spend the

majority of their time on exempt duties. In their view, if an employer did not realistically

expect employees to spend the majority of their time on exempt duties, then its

classification of the position as exempt is unlawful and liability is thereby established for

the entire class, even for those class members who actually spent the majority of their

time on exempt duties.

Justice Liu’s concurrence did not alter the Duran court’s conclusion that “the

question is ‘first and foremost’ how the employee’s time is actually spent.” (Duran,

supra, 59 Cal.4th at p. 27.) Instead, the concurrence was drafted “to further elucidate the

proper inquiry at the class certification stage of an employee misclassification case and

the duty of the trial courts to manage individual issues in a class action trial. (Id. at



9

Indeed, in their reply brief plaintiffs assert that their appeal is centered on “how

to interpret Justice Liu’s statement in Duran that an employee who exceeds the 50%

outside-sales threshold should be classified as non-exempt—i.e., eligible for overtime

pay—if devoting that much time to outside sales is not a realistic requirement of the job.”

A concurrence is not the opinion of the court and is not binding. (People v. Amadio

(1971) 22 Cal.App.3d 7, 14.)

16

p. 51.) Justice Liu sought to reemphasize that in some cases, an employer’s “realistic

expectations” or the “realistic requirements of the job” may be shown by common proof.

(Id. at pp. 52–53.) Specifically, he sought to address “[w]hat would it mean to ‘manage

individual issues’ in the context of an employee misclassification case?” (Id. at p. 55.)

He then set forth two ways for a trial court to frame a representative sampling approach

to proving class liability. First, he advised that consideration of individual issues should

inform the design of any sampling or similar statistical approach. Second, he

acknowledged that a defendant is entitled to raise individual issues that challenge the

results of a valid sampling plan as implemented. (Duran, at pp. 55–56.) It does not

appear to us that the concurrence is inconsistent with the majority opinion or that it

suggests a different standard should apply when analyzing certification motions.

In Martinez, the appellate court concluded the trial court abused its discretion by

denying the plaintiffs’ motion for certification of a class of salaried managerial

employees who had allegedly been misclassified. The plaintiffs alleged that the

employer’s operations, hiring, training, and other practices were uniform throughout the

restaurant chain. Because the restaurants were typically understaffed, managers routinely

filled in where needed, working as cooks, servers, bussers, hosts, stockers, bartenders,

and kitchen staff. (Martinez, supra, 231 Cal.App.4th at pp. 368–369.) Each of the

plaintiffs’ declarants estimated he or she had spent the majority of time performing such

hourly tasks, but later admitted they were unable to estimate the amount of time spent on

exempt versus nonexempt tasks. (Id. at pp. 369, 371.) The trial court found all putative

class members performed similar duties and responsibilities pursuant to a finite task list.

(Id. at p. 371, fn. 12.) But it concluded the plaintiffs’ claims were not typical and they

were inadequate class representatives because they were unable to estimate the number of

hours spent on individual exempt and nonexempt tasks and their job duties varied from

day to day. Common questions did not predominate because there were “significant

individual disputed issues of fact relating to the amount of time spent by individual class

17

members on particular tasks,” and such variability would require adjudication of the

affirmative defense of exemption for each class member. (Id. at pp. 371–372.)

The Court of Appeal reversed. Citing Ayala, supra, 59 Cal.4th 522, Martinez

cautioned that courts should “avoid focusing on the inevitable variations inherent in

tracing the actions of individuals and to instead focus on the policies—formal or

informal—in force in the workplace.” (Martinez, supra, 231 Cal.App.4th at p. 380.) The

plaintiffs’ theory of liability—that by classifying all managerial employees as exempt, the

employer had violated overtime laws—was “ ‘by its nature a common question eminently

suited for class treatment.’ ” (Ibid.) Significant common issues pervaded: the employer

operated a chain of restaurants governed by the same policies and procedures; exempt

employees were expected to work at least 50 hours per week; and the parties had

identified a finite task list, suggesting jobs were “ ‘highly standardized.’ ” (Ibid.) The

“gist of plaintiffs’ claim” was that “regardless of the patina of managerial discretion

expressed in their job description, they functioned consistently as utility workers, crosstrained

in all tasks, who could be assigned to fill in where needed without affecting the

labor budget or requiring overtime compensation. The crux of the matter, therefore, lies

in whether a typically nonexempt task becomes exempt when performed by a managerial

employee charged with supervision of other employees.” (Id. at p. 381, fn. omitted.)

Martinez also reasoned that under Sav-On, “courts in overtime exemption cases

must proceed through analysis of the employer’s realistic expectations and classification

of tasks rather than asking the employee to identify in retrospect whether, at a particular

time, he or she was engaged in an exempt or nonexempt task.” (Martinez, supra,

231 Cal.App.4th at p. 382.) “The court here must ask whether [the employer’s] realistic

expectations and the actual requirements of the job resulted in the exercise of

supervisorial discretion by managerial employees or instead relegated employees to [a]

utility role.” (Id. at p. 383.) By focusing on the employees’ inability to identify what

tasks they performed and for what purpose, Martinez concluded the trial court had failed

18

to consider Sav-On’s explicit direction that courts need not assess an employer’s

affirmative exemption defense against every class member’s claim before certification.

(Ibid., citing Sav-On, supra, 34 Cal.4th at pp. 337–338.)

Martinez is distinguishable in that the central dispute here is where BBOs spent

the majority of their time, and not on whether they spent the majority of their time on

exempt tasks. Unlike Martinez, the crux of the matter here is not task classification.

Instead, the issue is how much time each putative class member spent outside the Bank’s

locations. As the trial court stated, “a declaratory judgment that the policy was unlawful

would resolve nothing because to determine restitution each claimant would have to be

called to testify as to where he or she spent most of their time and how much overtime

they worked. Thus a declaratory judgment on the policy, standing alone, serves no

purpose.” The court conceded that the result might have been different “if a bona fide

statistical survey provided reason to believe that only a few class members may have

been properly classified and those could be readily identified; but the survey evidence

here provides no such assurance.”

In our view, the trial court properly focused on manageability issues pertaining to

the affirmative defenses, while fully understanding plaintiffs’ theory of liability.

10

In the



10 The trial court accurately described plaintiffs’ theory of the case as follows:

“Plaintiffs assert that because the BBO position was designed, managed and monitored as

an inside sales job, and as a consequence that is how the BBOs uniformly spent their

time, Defendant violated mandatory overtime wage laws by classifying all BBOs as

exempt. They intend to show, if given the chance, that it was not realistic for Defendant

to expect that putative class members would be able to perform the duties and meet the

production goals of the BBO position by ‘customarily and regularly work[ing] 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’

[citation]. In other words, Plaintiffs intend to show that Defendant ‘had a consistently

applied policy or uniform job requirements and expectations contrary to a Labor Code

exemption.’ [Citation.] Under this theory, the focus of the liability inquiry would be on

the ‘realistic requirements of the [BBOs] job’ [citation], i.e., whether it was realistic for

Defendant to expect that the BBOs would be ‘outside’ more than half of the time, given

19

portion of the court’s decision about which plaintiffs complain, the court observed that if

“the vast majority of the class sample would credibly testify that most of their time was

spent inside the office and the confidence interval was relatively narrow, then one might

infer that the number of witnesses the defense would call on its affirmative defense would

be relatively small. But if the results relied upon by Krosnick are as unreliable as

suggested by a comparison of the 2008 and 2015 surveys, then the court cannot take

much comfort in the notion that Plaintiffs’ trial plan adequately manages the affirmative

defenses. As with the issue of restitution, the affirmative defense would appear to require

a host of ‘mini trials.’ ”11

Even if the trial court did err in its assessment of plaintiffs’ theory, in the same

section of its decision containing the passage quoted above, the court also faulted their

trial plan because, given the overall unreliability of the 2015 Survey, there was no way

for plaintiffs to establish a prima facie case on the issue of liability. Additionally, the

court concluded that even if the liability phase were to proceed, plaintiffs would not be

able use representative sampling to establish an aggregate restitution award. Thus, the

court found multiple flaws in plaintiffs’ trial plan.

12



the nature of their job duties and performance expectations. Such a finding would be

based on evidence of the actual experiences of putative class members who were ‘getting

the job done’ as to where they spent their time doing so.”

11 Trial courts must determine whether individual issues, including those arising

from affirmative defenses, can be managed fairly and efficiently. The Supreme Court in

Duran stated: “If statistical evidence will comprise part of the proof on class action

claims, the court should consider at the certification stage whether a trial plan has been

developed to address its use. A trial plan describing the statistical proof a party

anticipates will weigh in favor of granting class certification if it shows how individual

issues can be managed at trial.” (Duran, supra, 59 Cal.4th at pp. 31–32.) However, the

court cautioned, “a class action trial management plan may not foreclose the litigation of

relevant affirmative defenses, even when these defenses turn on individual questions.”

(Id. at p. 34.)

12 To the extent plaintiffs criticize the trial court for relying on Teamsters v. United

States (1977) 431 U.S. 324, we find no error. The court did not rely on Teamsters in its

20

It is established that even when the party proposing a class asserts “the employer

consistently imposed a uniform policy or de facto practice on class members, the party

must still demonstrate that the illegal effects of this conduct can be proven efficiently and

manageably within a class setting.” (Duran, supra, 59 Cal.4th at p. 29; see Koval v.

Pacific Bell Telephone Co. (2014) 232 Cal.App.4th 1050, 1060 [“existence of a uniform

policy does not limit a trial court’s inquiry into whether class action treatment is

appropriate”] and Arenas v. El Torito Restaurants, Inc. (2010) 183 Cal.App.4th 723, 734

[“trial court concluded plaintiffs’ theory of recovery—that managers, based solely on

their job descriptions, were as a rule misclassified—was not amenable to common proof”

given evidence employees’ “duties and time spent on individual tasks varied widely”].)

“To the contrary, ‘courts have routinely concluded that an individualized inquiry is

necessary even where the alleged misclassification involves application of a uniform

policy. ” (Mies, supra, 234 Cal.App.4th at p. 984.)

Plaintiffs also fault the trial court for rejecting its proposal to address how to

proceed if evidence on the average hours worked produced an unacceptable margin of

error. In that case, they had proposed the parties could continue to submit evidence “until

the margin of error reached an acceptable number.” The trial court rejected this proposal,

stating that “it is too late in this litigation for another ‘provisional’ certification order that

is subject to a further development of a trial plan.” Given that this case has lasted well

over a decade and that one trial has already occurred and been reversed, and that two

appellate court decisions on this matter have already been issued, we fully support the

trial court’s sentiment on this point. Additionally, the Supreme Court in Duran

specifically stated: “The trial court is of course free to entertain a new certification

motion on remand, but if it decides to proceed with a class action it must apply the



decision, but instead was simply quoting a passage in Duran in which the Supreme Court

criticized plaintiffs for relying on Teamsters.

21

guidelines set out here,” including assessing manageability of affirmative defenses.

(Duran, supra, 59 Cal.4th at p. 33, italics added.)

B. Variation in Restitution

Plaintiffs also claim the trial court erred by finding that variation in restitution

defeats certification. They start off on the wrong foot, however, by asserting that the trial

court “held that plaintiffs satisfied the requirements for predominance and manageability

with respect to liability.” Actually, the court found that the 2015 Survey, if it had been

the only survey, might have been used to measure the extent of variation within the class

on the issue of whether BBOs spent more than half of their time in the office. However,

this observation was tempered by the court’s acceptance of Hildreth’s opinion that

self-interest bias caused respondents to overestimate the time spent inside the office.

This bias, the existence of which was supported by the discrepancies between the two

Krosnick surveys, suggested to the court that individual variations would necessitate a

larger sample size than plaintiffs suggested. Thus, even apart from restitution, the court

did not find that the 2015 Survey supported a finding of predominance and manageability

with respect to liability.

On the issue of restitution, plaintiffs assert the trial court erred in concluding the

survey was going to be used as the sole basis for fixing the amount. However, even

plaintiffs admit that the testimony of the randomly drawn sample “would have been a

substantial factor in the court’s ultimate determination of the amount of restitution.”

Plaintiffs proposed establishing aggregate restitution by using the trial court’s finding of

average overtime hours worked by a representative witness group (the size of which

would be based on the flawed 2015 Survey), which would then be extrapolated and

applied in the second phase of trial to the individual BBOs’ payroll records to determine

each individual’s recovery.

We cannot conclude the trial court committed legal error in its assessment of the

impact the 2015 Survey evidence would have in the manageability of the restitution

22

phase of trial. Whether the court was correct in concluding that no sample of any size

less than the entire class could ever yield an aggregate restitution award, the fact remains

that the 2015 Survey evidence offered by plaintiffs was seriously flawed, leaving the

court with no alternative except to anticipate that every class member would need to

testify.13 The court properly rejected Plaintiffs’ plan for determining aggregate restitution

because the plan rested on a survey that was wholly unreliable.

Plaintiffs also contend the trial court erred by holding them to a higher standard

than required by law with respect to the determination of an aggregate restitution award.

Contrary to plaintiffs’ argument, the court was not holding them to a higher standard.

Instead, it was stating its view of the state of the evidence, specifically, that the 2015

Survey was worthless as a sampling tool, necessarily requiring each class member’s

testimony as to restitution.

V. Rejection of 2015 Survey

Plaintiffs use the final arguments in their opening brief to address the central issue

in this appeal, namely, whether the trial court erred in rejecting the 2015 Survey. They

first assert that the survey “satisfied all of Duran’s conditions for the use of statistical

evidence in a class action.” As they note, however, one of these requirements is that the

margin of error cannot be “intolerably high.” (Duran, supra, 59 Cal.4th at p. 46.)

The trial court rejected the 2015 Survey because of the difference between the

average hours worked as reported in the 2008 Survey versus the 2015 Survey. Again, the

2008 Survey reported an average of approximately 54 hours per week, while the 2015

Survey reported an average of approximately 63 hours per week. Because the court

could not reconcile the differences between the two surveys, it concluded the 2015



13 Plaintiffs’ related argument that the burden should have been on the Bank as the

employer to come forward with evidence as to the number of hours worked because it

had kept inaccurate or incomplete records has been addressed previously in Duran,

supra, 59 Cal.4th at p. 41.

23

Survey estimate of the amount of time BBOs were engaged in outside sales work was

also unreliable. Plaintiffs argue that in doing so “the trial court exercised its discretion in

a way contrary to the internal rules governing statistical analysis.” Their arguments are

not persuasive.

Plaintiffs first attribute the discrepancy in the memories of the class members to

“the passage of time,” noting the class period goes back to December 1997. Thus, in

2015, members were being asked to recall events up to 17 years earlier.14

Plaintiffs

appear to urge the differences do not matter here because these same discrepancies will

occur whether the trial proceeds on an individual basis or a class action basis.15

We

disagree. In individual actions, the effects of the passage of time upon each witness’s

memory can be explored by the fact finder. In a class action, it is impossible to do so,

and plaintiffs do not advance any scientific principles suggesting that statistical evidence

can be used to accurately reflect the impact of the passage of time on survey respondents

in a manner that can be extrapolated to the class as a whole. While plaintiffs argue again

that the burden of uncertainty should not be shifted to them because the Bank failed to

keep accurate records, the fundamental inquiry here is whether this matter can proceed as

a class action, not which party is to blame for the fact that plaintiffs’ evidence fails to

support a finding of commonality.

Plaintiffs also argue that the difference between the average hour estimates from

the two surveys is not important because the estimates independently have low relative

margins of error. They note the 2008 Survey had a relative margin of error of 13.50

percent while the corresponding margin of error in the 2015 Survey was 14.96 percent,



14 The trial court was well aware of this point, noting the “gaps in the passage of

time between the underlying events and the surveys and between the surveys themselves

are unusually large . . . . Some of the problems here may be traced to this circumstance.”

15 As an aside, this argument does little to inspire confidence in the 2015 Survey

results.

24

observing that we approved a 9.57 percent relative margin of error in Bell, supra,

115 Cal.App.4th 715. However, the margins of error are artificially low because they are

based on the total hours worked, and not overtime hours. Further, it is the comparison of

the two surveys that reveals the inconsistency, irrespective of whether each survey is

internally consistent within itself.

Plaintiffs’ final argument—that the Bank and its attorneys skewed the results of

the 2015 Survey by sending the Biggs letter to putative class members—is not well taken.

Plaintiffs do not cite to any evidence in the record suggesting that the letter “infuse[d] the

survey with bias.”

In sum, the trial court did not abuse its discretion in concluding that the wide

discrepancy between the 2015 and 2008 survey results demonstrated that the 2015 Survey

was unreliable, and served as tangible evidence that the survey results were tainted by

bias. Accordingly, substantial evidence supports the court’s finding that the survey data

was unreliable as evidence of uniformity in how BBOs spent their time, and unreliable as

statistical support for selecting a representative witness group to testify as to liability or

restitution without causing the inquiry to devolve into a multiplicity of individual mini

trials, especially in light of the Bank’s right to call witnesses outside the sample to

establish its affirmative defense.16



16 After oral argument and the submission of this case, plaintiffs requested this

court to consider an appellate case, now published, on employment class actions. The

case is ABM Industries Overtime Cases (Dec. 11, 2017, A132387, A133077, A133695)

___ Cal.App.5th ___ [2017 WL 6887032] (ABM Industries). We have decided to review

the case and find it clearly distinguishable from our matter.

First of all, ABM Industries involved the denial of class certification where the

employees’ witness, Woolfson, the sole expert on either side concerning the issue of class

certification, was found by the trial court to be unqualified. The employer had offered no

expert to challenge the employees’ witness. Indeed, “ABM had failed to lodge a formal

objection to the evidence [of the employees’ expert], move to strike the [expert’s]

declaration, depose Woolfson on his credentials or conclusions, and/or provide their own

contrary expert opinion.” (ABM Industries, supra, ___ Cal.App.5th ___ [2017 WL

The order denying class certification is affirmed.