Timothy MCCLEERY et al., Plaintiffs and Appellants, v. ALLSTATE INSURANCE COMPANY et al., Defendants and Respondents.
In this putative class action, property inspectors allege they were engaged by three “service” companies to perform inspections for two major insurers. The inspectors allege they were in fact employees of the insurers and service companies jointly, and were entitled to but deprived of minimum wages, overtime, meal and rest breaks, reimbursement of expenses, and accurate wage statements.
The inspectors moved for class certification, supported by their expert’s declaration that liability could be determined and damages calculated classwide by way of statistical analyses of results obtained from an anonymous, double-blind survey of a sampling of class members.
The trial court summarily rejected the expert’s plan and denied certification on the ground that the inspectors had failed to show that their status as employees (as opposed to independent contractors) could be established on predominately common proof.
We reversed the order and remanded the matter with a direction, as pertinent here, to evaluate plaintiffs’ proposed sampling plan. (McCleery v. Allstate Ins. Co. (Feb. 5, 2016, B256374) 2016 WL 463275 [nonpub. opn.].) On remand, plaintiffs offered a trial plan describing their proposal to establish liability and damages by way of an anonymous survey of all class members. The trial court found common issues existed as to the class members’ employment status. It further found that plaintiffs’ survey method, although flawed in some respects, was carefully crafted for accuracy. However, the court found plaintiffs’ trial plan to be unworkable because it failed to address individualized issues and deprived defendants of the ability to assert defenses. The court therefore again denied certification.
Plaintiffs appeal, contending the trial court applied improper criteria and made incorrect legal assumptions.
We conclude that under the analytic framework promulgated by Brinker Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004, 139 Cal.Rptr.3d 315, 273 P.3d 513 (Brinker ) and Duran v. U.S. Bank National Assn. (2014) 59 Cal.4th 1, 172 Cal.Rptr.3d 371, 325 P.3d 916 (Duran ), the trial court acted within its discretion in denying certification.
A. Procedural Posture Up To the First Appeal
We take the facts and much of the procedural posture from our prior opinion.
Property inspectors Timothy McCleery, Yvonne Beckner, Terry Quimby and April Boyles Jackson filed this action on behalf of themselves and similarly situated persons, alleging defendants Allstate Insurance Company and Farmers Group, insurers for whom the plaintiffs provided property inspection services, and CIS Group LLC/North American Compass Insurance Services Group (CIS), Advanced Field Services, Inc. (AFS), and Capital Personnel Services, Inc. (PMG), service companies contracting to provide inspection services, concocted a scheme to insulate themselves from labor laws by nominally employing plaintiffs as independent contractors while retaining control over all aspects of their work. Plaintiffs purport to represent a putative class of approximately 1,550 property inspectors in California.
Plaintiffs allege the insurers and service companies were in fact their joint employers, and all defendants failed to pay minimum wages and overtime (Lab. Code, § 1194), furnish timely or accurate wage statements (Lab. Code, § 226, subd. (e) ), establish a policy for meal or rest breaks, or reimburse them for employment expenses (Lab. Code, § 2802), and in so doing violated the Unfair Competition Law (Bus. & Prof. Code, § 17200 et seq.; UCL).
In 2013, plaintiffs filed five class certification motions, one for each employer, designating one subclass per employer and a sixth subclass for CIS employees who had suffered retaliation for cooperating with plaintiffs in this litigation. Plaintiffs contended defendants’ liability or lack thereof could be determined on common proof regarding defendants’ status as joint employers and their uniform employment policies, or lack thereof.
Defendants opposed the motions, arguing, as pertinent here, that few if any inspectors provided services for only the defendant insurers, but in fact freelanced for any insurer that would hire them, and wide variation existed in their work practices, as some worked part-time and some full-time, some long hours on any given day and some short, and some with the assistance of subcontractors.
The trial court tentatively concluded that plaintiffs had demonstrated the requisite ascertainability and numerosity for class certification (at least with respect to some subclasses), as well as the suitability of class counsel and diligence and typicality of the class representatives. The court also tentatively concluded that liability issues could be divided into, as pertinent here, two phases, the first to determine whether the defendants were plaintiffs’ joint employers and the second to determine whether plaintiffs had been deprived of legally mandated wage and hour benefits. But the court requested further briefing on the commonality of proof of deprivation of wage and hour benefits.
In response, plaintiffs submitted the supplemental declaration of Dr. John Krosnick, their survey expert, setting forth his plan to use established survey methods and statistical analyses to accurately determine and measure the extent of variations in the inspectors’ work. Dr. Krosnick acknowledged that variations in the inspectors’ work practices might not be amenable to classwide proof—for example, as to overtime, minimum wages, and expenses—but represented that his methodology for the design and implementation of a survey of representative samples of the plaintiffs’ class would result in reliable evidence on issues for which common proof was unavailable, such as wage-statement violations, amounts owed to compensate inspectors for earned and unpaid overtime, differences between earned wages and the minimum wage, compensation for mileage and other earned and unpaid work expenses, and compensation for retaliation. Dr. Krosnick declared that his methodology would account for variations in work practices and would be manageable, scientifically based, and fair, using “well-established methodology of random sampling ․ designed expressly to gauge the amount of variation in an attribute within a population, and well-established statistical procedures for analyzing randomly sampled data,” thereby accurately measuring the degree to which variations in the inspectors’ work affects the plaintiffs’ entitlement to the various employee benefits, “no matter how much variation there is within the population.”
The trial court summarily rejected Dr. Krosnick’s plan, and concluded that defendants’ employment practices as to each class member would necessitate individualized determinations.
We reversed the order denying certification and directed the trial court to evaluate the extent to which Dr. Krosnick’s proposed sampling and statistical methods might render proof of some or all of the liability issues manageable.
B. Post-Appeal Proceedings
On remand, Dr. Krosnick elaborated on his plan. He explained he was in the midst of conducting survey research of all people who performed inspections for Allstate, Farmers, CIS, AFS, and PMG to gather evidence for both liability and damages determinations as to unpaid overtime, missed meal and/or rest breaks, unpaid minimum wage, unreimbursed miles driven, and other business expenses. After gathering this evidence he intended to compute penalties and pre-judgment interest.
Dr. Krosnick explained his research involved “(1) drawing a sample of respondents to represent a population, (2) collecting data from those respondents, and (3) analyzing the data generated to answer the questions of interest.” He and his team of researchers had obtained from plaintiffs’ attorneys “the constructs to be measured in the survey,” which he described as “(1) the amount of overtime worked, (2) the numbers of meal and rest breaks to which inspectors were entitled to take under California law but did not take (assuming that law applied to these individuals), (3) the amount of time inspectors spent performing specific tasks of relevance to the claimed minimum wage violations, (4) the number of miles that inspectors drove to do their work, [and] (5) the amount of money that inspectors spent for other business expenses incurred in connection with their work.”
The researchers designed questions to “gauge the target constructs.” Dr. Krosnick explained that “[t]he questions were drafted according to the principles of optimal questionnaire measurement identified in the published peer reviewed academic literature on survey methodology and related fields, drawing on [his] expertise and years of experience in the field.” Each interview would take approximately 45 minutes, and respondents would be told their answers and participation were confidential.
After ethics approval by IntegReview, an Institutional Review Board providing ethical review services to scientists conducting human subjects research, and some pretesting, Dr. Krosnick retained SSRS, a telephone survey research firm, to conduct the actual interviews. He sent letters to all potential class members explaining the survey, inviting them to participate, and informing them they would be contacted by phone. Each letter contained a “small financial pre-incentive to thank each respondent for reading it,” as well as the offer of $100 for their participation and an added $10 if they initiated telephone contact with the research team.
C. The Interview
Each interview was conducted by telephone by an SSRS employee who read questions from a computer monitor and entered responses by computer into a flowchart program designed by Dr. Krosnick. The flowchart updated both itself and the interviewer’s script according to answers received. For example, if a respondent stated he or she had worked as an inspector for AFS from 2005 to 2008, the program calculated the maximum number of months the respondent could have worked and inserted that number, those years, and the vendor’s name into later questions, obtaining, for example, the following question: “The maximum number of months you could have worked with AFS between 2005 and 2008 is 42 months. For how many of those months did you work with AFS?”
Aside from some initial monitoring, neither Dr. Krosnick nor his research team participated in the interviews.
Respondents were asked preliminary questions to establish whether they had worked for CIS, AFS, PMG, or Allstate, and when and for how long. A respondent giving an answer outside known parameters was coached to give an in-range response. For example, it was known that AFS performed inspections from 2005 to 2008. If a respondent stated he or she had worked for AFS after 2008, the interviewer was instructed to say, “I think I might have misheard you. Let me read the question again.” If the respondent repeated the out-of-range year, the interviewer was instructed to say, “I’ve been told AFS stopped doing inspections in 2008. Let me read this question again.” If the respondent persisted in an invalid answer, the interviewer wrote “9998” as the response, which was treated in Dr. Krosnick’s algorithm as “Don’t know.”
Respondents were asked whether and how often they had received training required for work, attended required meetings, completed inspections that were or were not paid for, worked on inspections that were or were not completed, engaged in work-related planning activities, purchased supplies needed for work, communicated with supervisors about issues unrelated to a particular inspection, kept their own books to ensure accurate payment, traveled for work to places other than where inspections occurred, and took breaks lasting 10 to 29 minutes.
The interviewers advised respondents that activity spent “doing inspections you completed and were paid for” and “doing inspections you completed but were not paid for” included “time spent getting instructions from a company to do those inspections, determining how much you would be paid for each inspection, making appointments for each inspection, mapping your route to each inspection, traveling to the site, inspecting the property, taking photographs, preparing photographs for submission, completing the inspection report, and submitting the completed report.” Respondents were advised that activity spent “attempting to complete inspections that you were not able to complete” included “time spent getting instructions from a company to do those inspections, determining how much you will be paid for each inspection, making appointments for each inspection, mapping your route to each inspection, traveling to the site, and attempting to do the inspection.”
Respondents were asked to recall what time they spent in all of these activities, the number of days per week worked beginning in 2005, and the number of hours worked, divided into the following categories: fewer than three hours; three to five; five to six; six to eight; eight to 10; 10 to 12, and greater than 12 hours. Respondents were asked to recall what percentage of their work fell into each category, when during the workday they took 30-minute breaks, when they took breaks lasting between 10 and 30 minutes, when their breaks were longer than 30 minutes, and what percentage of their breaks fell into each time span.
Respondents were directed to a Web site during the interview, where possible work-related activities were listed, such as performing inspections that either were or were not completed, and for which they had or had not been paid; planning work activities other than inspections; and purchasing supplies; and were asked what percent of their time was spent doing each activity. Before answering, respondents were repeatedly reminded about their prior answers. For example, the interviewer would state, “Before you answer, I’d like to tell you that you already told me that you spent [a specified] percent of your time taking breaks lasting 10 to 29 minutes. So that means that you spent [a computer-calculated] percent of your time doing all of your other things for work. Now, how many of the [calculated] percent did you spend doing the three things ․, combined? You can answer with any number between zero percent and [the calculated] percent.”
As before, if a respondent’s answer fell outside parameters established by prior answers, the interviewer would say, “I think I might have misheard you. Let me read this question again.” If the answer was out of range a second time, the interviewer would say, “I need to type a number between 0 and [the calculated percent] as your answer to this question. Let me read these questions again.” A persistently wrong answer was coded as “998” for “Don’t know” or “999” for “Refused.”
Respondents were asked whether they had incurred expenses beginning in 2005 for necessary work items such as Internet access, cell phones, landlines, fax machines, “printer, computer, ink, map, book, paper, [or] software,” and were asked how sure they were of their responses on a scale of one to five.
Interviewers asked what respondents had read or heard about the lawsuit, and were instructed to “probe repeatedly” with the question, “What else do you remember about the purpose of any lawsuit involving Allstate, Farmers, CIS, AFS, or PMG,” until the respondent finally said, “nothing.”
Finally, respondents were tested on their ability to estimate time. The interviewer stated, “First, please close your eyes and keep them closed until I ask you to open them. ․ Are your eyes closed? [¶] Thank you. Now, with your eyes closed, when I say ‘go,’ please think back, silently to yourself, to the beginning of your day today, when you got up this morning. Then think silently to yourself about everything you did, in the order you did it, who you saw, where you went, and everything that happened. Take your time to go slowly, keep your eyes closed, try to remember as many little details as you can, and don’t stop thinking about those things until I say ‘stop.’ Please don’t say anything out loud about what you’re thinking—just think silently to yourself. OK?” When the respondent was ready, the interviewer would say “go” and start a clock. After a random period ranging from 20 to 40 seconds, the interviewer would say “stop” and ask how many seconds had just elapsed. Respondents were then asked to estimate how long the interview had taken overall.
SSRS personnel recorded the respondents’ answers by computer and forwarded them to Dr. Krosnick. Dr. Krosnick’s research team analyzed the data and mailed $100 incentive checks to each respondent who had elected to receive one.
D. Survey Analysis
In his third trial plan report, dated January 16, 2017, Dr. Krosnick explained that he made computations “to generate estimates of totals of the [tested] quantities for each of the subclasses (AFS inspectors, PMG inspectors, and CIS inspectors),” apportioned separately for inspections done for Farmers and Allstate. For each of the quantities, he would compute a margin of error and “confidence intervals at various different confidence levels” separately for each of the three vendors (AFS, CIS, and PMG) and two insurers “using data from all respondents who reported on their experiences doing inspections for each company,” and would conduct a separate statistical analysis comparing the companies to determine if the results could be pooled.
Once he computed the various quantities, he would apportion each quantity to Farmers and Allstate. To do so, he would rely on defendants’ records “to determine the number of inspections done by each inspector for Allstate and the number of inspections done by each inspector for Farmers during a specific time period. Post-certification, [he would] use this same approach for all survey respondents and apportion the quantities computed with the survey data proportionally for each respondent in proportion to his/her ratio of inspections done for Farmers vs. Allstate.”
Interim estimates of the relative margins of error for the target quantities (average overtime hours worked per respondent per week; missed rest and meal breaks, hours spent on minimum wage tasks, hours owed as penalties, and reimbursement owed per respondent per month for mileage and business expenses) ranged from 10.6 to 25.5 percent if a 95 percent confidence interval was used, and from 7.8 to 18.7 percent using an 85 percent confidence interval.
Dr. Krosnick explained that to minimize bias due to nonresponses, statistical analyses would be conducted “to gauge the match of the participating individuals with the [proposed class]. If discrepancies [were] found, results [would] be obtained after weighting the survey data to maximize resemblance of the participating sample to the population.”
Dr. Krosnick explained that additional analyses would be conducted “to explore systematic non-response.” This would be done “by exploring systematic differences between responses of people who were easy to contact vs. those who were difficult to contact, as indicated by whether the respondent had a telephone number listed in a phone book, how often the respondent had moved from living at one address to another, and how many call attempts were made to reach the respondent.” If systematic differences were observed, he would “explore the robustness of conclusions depending upon whether adjustments are made based on the assumption that hard to reach respondents answered the questions similarly to the ways the uninterviewed individuals would have answered if they had been interviewed.”
He explained he would “also explore unit non-response by comparing the interviewed people to people who were not interviewed but were on the list to be interviewed in terms of a range of characteristics that can be observed from the contact information, such as the gender of the person, the region of California in which they resided, the region of the state in which their telephone area code was located, and other such characteristics.” If systematic differences were observed, he would “assess the robustness of conclusions if adjustments are made to eliminate discrepancies between interviewed individuals and individuals not interviewed.”
In his first trial plan report, dated January 27, 2014, Dr. Krosnick had declared: “It might seem that survey respondents should be viewed as witnesses providing testimony. But in fact, that is not so. Respondents are not testifying witnesses. Instead, survey respondents participate in a scientific measurement procedure overseen by a qualified expert and conducted according to a set of rules designed to assure accuracy to allow the expert to provide testimony. It is the expert who will offer opinions generated using scientific methodology to produce statistical calculations of damages class-wide, and the expert can be cross-examined.”
E. Defendants’ Objections
In opposition to plaintiffs’ trial plan, defendants offered the declaration of Robert W. Crandall, a labor studies expert.
Crandall declared Dr. Krosnick’s survey asked no liability questions related to the employee/independent contractor distinction, and in fact avoided questions about the degree of independence inspectors enjoyed. (Crandall speculated that Dr. Krosnick purposefully avoided these questions because cognitive pre-testing revealed that the inspectors reported high degrees of independence.) This deficit not only left the independent contractor question unanswered but potentially skewed the survey results by artificially narrowing variances. If the range of experience is narrow, Crandall explained, the average experience obtained by Dr. Krosnick through his statistical analyses would more closely represent the experience of the whole. But “if the data were to show that everyone is different because experiences vary widely, then the average may not be representative of the actual experiences of many members of the underlying population.”
Crandall explained that Dr. Krosnick asked no questions pertaining to joint employment by Farmers, and in fact appeared not to understand that a joint employment issue existed in this litigation. In his deposition, Dr. Krosnick testified, “There’s no question in my survey that asks about how many hours a day an individual did work for Farmers.” When asked why his survey included no questions about how often, if ever, an inspector worked five hours or more in a day for Farmers, how often they ever worked at least four hours for Farmers, or what business expenses they incurred while working for Farmers, Dr. Krosnick answered, “I wasn’t asked to do that.”
Dr. Krosnick’s survey similarly asked no questions regarding Farmers’ knowledge or control of any facet of an inspector’s workday, e.g., how many hours the inspector worked, what breaks were or could have been taken, or what meetings were attended or expenses incurred. Dr. Krosnick notably asked no question regarding “why” a given break outcome occurred. He simply asked for counts of shifts of various durations and break lengths, eschewing questions about whether employers provided meal or rest breaks that inspectors declined to take.
Crandall explained that the very precise recall required by Dr. Krosnick’s survey questions about events stretching back 10 years invited significant error. “For example,” he stated, “estimating shifts that were greater than 5 but less than 6 hours in duration with breaks between 10 and 29 minutes and breaks 30 minutes or greater” presented a significant recall burden to survey respondents. Crandall related that when Dr. Krosnick was asked in deposition how accurate he believed the inspectors’ responses were, he testified, “I’m not here to testify to the accuracy of the survey respondents’ memories.” When asked his opinion about whether an inspector could “tell us years after the fact whether they took a 25-minute meal period or a 35-minute meal period,” Dr. Krosnick testified, “I don’t have an opinion of the accuracy of the respondents’ responses.”
Crandall related that when Dr. Krosnick returned from a break during his own deposition he was unable to recall how long the break had lasted. And when asked about a deposition session that had occurred a few weeks earlier, he could not recall how many breaks had been taken or how long they lasted.
Crandall observed that Dr. Krosnick’s survey failed to account for the fact that an inspector could work for multiple entities—including those not party to this case—during a given day or week. So if an inspector worked 4.5 hours for Farmers and 4.5 for Allstate on a given day, for a total of 9 hours, survey answers would indicate one hour of overtime was owed, when in reality none was owed. Dr. Krosnick further failed to ask about whether any respondents subcontracted out any inspections, even though his original survey plan explicitly called for subcontracting to be taken into account.
Finally, although Dr. Krosnick proposed to mesh a summary of responses with data obtained from defendants indicating how many inspections had been performed and on what dates, in reality that data failed to indicate on what dates inspections took place, as the inspection date was always the upload date, not the performance date. The data also failed to indicate in what order multiple inspections on a given date occurred or how long each took. There was therefore no way to know how an inspector had apportioned his or her day between either inspections or employers, especially given that all responses were anonymous.
In the end, Crandall declared, the anonymous nature of Dr. Krosnick’s survey led to inaccurate and unverifiable results.
F. Trial Court Ruling
The trial court found that inspectors fell into several subgroups: those who essentially worked full-time for defendants; those who worked part-time—either because they performed inspections only part-time or sometimes worked for nonparty companies; those who worked with others to perform the assigned inspections; and those who interspersed inspections with other activities, such as school or parenting. The court nevertheless found that common issues predominated over individual ones “as to the employee/independent contractor issues and joint employer issues.”
The court further found that Dr. Krosnick’s survey was “carefully crafted to verify appropriate respondents and accuracy in the responses.” However, the court found that plaintiffs’ statistical sampling alone did not render their claims manageable. It found that Dr. Krosnick’s survey results failed to specify for which insurers inspections were performed, or to explain whether the inspectors’ failure to take meal or rest breaks was due to preference or to the exigencies of the job. Also, the survey’s anonymity foreclosed the defendants from cross-examining witnesses to verify responses or test them for accuracy or bias.
The trial court found that plaintiffs’ trial plan failed to address the wide work-practice variations among inspectors and offered no way to manage individualized issues, but simply ignored them.
In the end, the trial court concluded that plaintiffs’ proposed class action would not be superior to individual actions because their survey failed to address “all of the information needed for an accurate determination of liability,” and the trial plan “deprive[d] defendants of the right of cross-examination and the ability to present their affirmative defenses.”
Finding plaintiffs’ trial plan to be “unworkable,” the court denied certification.
Plaintiffs contend that in denying certification the trial court relied on improper criteria and made incorrect legal assumptions. We disagree.
A. Standard of Review
Code of Civil Procedure section 382 authorizes a suit to be tried as a class action “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.” (Code Civ. Proc., § 382.) Class certification requires demonstration of an ascertainable and sufficiently numerous class, a well-defined community of interest, and the superiority of proceeding as a class. (Brinker, supra, 53 Cal.4th at p. 1021, 139 Cal.Rptr.3d 315, 273 P.3d 513.)
The “community of interest” requirement has 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. (Brinker, supra, 53 Cal.4th at p. 1021, 139 Cal.Rptr.3d 315, 273 P.3d 513; Ayala v. Antelope Valley Newspapers, Inc. (2014) 59 Cal.4th 522, 529-530, 173 Cal.Rptr.3d 332, 327 P.3d 165 (Ayala ).) Generally, “ ‘if the defendant’s liability can be determined by facts common to all members of the class, a class will be certified even if the members must individually prove their damages.’ ” (Brinker, at p. 1022, 139 Cal.Rptr.3d 315, 273 P.3d 513; Cochran v. Schwan’s Home Service, Inc. (2014) 228 Cal.App.4th 1137, 1142, 176 Cal.Rptr.3d 407 (Cochran ).)
In reviewing a trial court’s denial of class certification we examine “whether the operative legal principles, as applied to the facts of the case, render the claims susceptible to resolution on a common basis.” (Ayala, supra, 59 Cal.4th at p. 530, 173 Cal.Rptr.3d 332, 327 P.3d 165; Brinker, supra, 53 Cal.4th at pp. 1023-1025, 139 Cal.Rptr.3d 315, 273 P.3d 513.) “ ‘The certification question is “essentially a procedural one that does not ask whether an action is legally or factually meritorious.” ’ ” (Brinker, at p. 1023, 139 Cal.Rptr.3d 315, 273 P.3d 513.) Courts focus instead on what type of questions—common or individual—are likely to arise, and whether proceeding as a class action, as compared to other forms of action, is a superior method of resolving these questions. (Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 327, 339 & fn. 10, 17 Cal.Rptr.3d 906, 96 P.3d 194 (Sav-On ).) A class action may be certified even if it is unlikely the class will eventually prevail on the merits, as certification in such a situation allows a defendant to obtain a favorable judgment binding all class members. “It is far better from a fairness perspective to determine class certification independent of threshold questions disposing of the merits, and thus permit defendants who prevail on those merits, equally with those who lose on the merits, to obtain the preclusive benefits of such victories against an entire class and not just a named plaintiff.” (Brinker, at p. 1034, 139 Cal.Rptr.3d 315, 273 P.3d 513.)
“Although predominance of common issues is often a major factor in a certification analysis, it is not the only consideration. In certifying a class action, the court must also conclude that litigation of individual issues, including those arising from affirmative defenses, can be managed fairly and efficiently. [Citation.] ‘[W]hether in a given case affirmative defenses should lead a court to approve or reject certification will hinge on the manageability of any individual issues. [Citation.]’ In wage and hour cases where a party seeks class certification based on allegations that 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 pp. 28-29, 172 Cal.Rptr.3d 371, 325 P.3d 916.) “ ‘Individual issues do not render class certification inappropriate so long as such issues may effectively be managed.’ ” (Id. at p. 29, 172 Cal.Rptr.3d 371, 325 P.3d 916.) “Trial courts must pay careful attention to manageability when deciding whether to certify a class action. In considering whether a class action is a superior device for resolving a controversy, the manageability of individual issues is just as important as the existence of common questions uniting the proposed class. If the court makes a reasoned, informed decision about manageability at the certification stage, the litigants can plan accordingly and the court will have less need to intervene later to control the proceedings.” (Ibid.)
“We review the trial court’s ruling for abuse of discretion and generally will not disturb it, ‘ “unless (1) it is unsupported by substantial evidence, (2) it rests on improper criteria, or (3) it rests on erroneous legal assumptions.” ’ ” (Ayala, supra, 59 Cal.4th at p. 530, 173 Cal.Rptr.3d 332, 327 P.3d 165.) If the court’s “reasons for granting or denying certification ․ are erroneous, we must reverse, whether or not other reasons [could have been] relied upon [to] support[ ] the ruling.” (Ibid.; Cochran, supra, 228 Cal.App.4th at p. 1143, 176 Cal.Rptr.3d 407.) In this respect, “ ‘appellate review of orders denying class certification differs from ordinary appellate review. Under ordinary appellate review, we do not address the trial court’s reasoning and consider only whether the result was correct. [Citation.] But when denying class certification, the trial court must state its reasons, and we must review those reasons for correctness. [Citation.] We may only consider the reasons stated by the trial court and must ignore any unexpressed reason that might support the ruling.’ ” (Cochran, at p. 1143, 176 Cal.Rptr.3d 407.) “In other words, we review only the reasons given by the trial court for denial of class certification, and ignore any other grounds that might support denial.” (Bartold v. Glendale Federal Bank (2000) 81 Cal.App.4th 816, 829, 97 Cal.Rptr.2d 226, overturned on other grounds due to Legislative Action in 2001 Cal. Legis. Serv. Ch. 560.)
Because trial courts “ ‘are ideally situated to evaluate the efficiencies and practicalities of permitting group action,’ ” they are “ ‘afforded great discretion’ ” in evaluating the relevant factors and ruling on a class certification motion. (Sav-On, supra, 34 Cal.4th at p. 326, 17 Cal.Rptr.3d 906, 96 P.3d 194.)
B. Plaintiffs’ Trial Plan is Inadequate and Unfair
Here, the trial court’s only statement about predominance of common issues favored plaintiffs. The court stated that defendants’ status as employers could be established by common factors showing the degree of control they reserved over the inspectors’ work. Although the court discussed disparate individual issues at several points in its order, it did so only from the perspective of manageability, and made no finding that individual issues predominated over common ones. The court denied certification notwithstanding its finding (at least impliedly) that a community of interest existed, because it found litigation of individual issues, including those arising from affirmative defenses, could not be managed fairly and efficiently using only Dr. Krosnick’s survey.
Plaintiffs defend Dr. Krosnick’s survey at length, arguing it was methodologically correct and scientifically valid, captured all pertinent variations in hours worked among inspectors, eschewed irrelevant questions, and produced reliable and accurate results. But this misses the point, as the trial court apparently agreed with these propositions, finding the survey “was carefully crafted to verify appropriate respondents and accuracy in the responses.”
The problem is not that Dr. Krosnick’s survey fails as a scientific measurement procedure, but that it fails as a trial plan.
“Class certification is appropriate only if ․ individual questions can be managed with an appropriate trial plan.” (Duran, supra, 59 Cal.4th at p. 27, 172 Cal.Rptr.3d 371, 325 P.3d 916.) Here, the trial court reasonably concluded plaintiffs’ trial plan failed to address how they could fairly establish defendants’ liability on a classwide basis as to any claim.
With respect to overtime and meal and rest breaks, “simply having the status of an employee does not make the employer liable for a claim for overtime compensation or denial of breaks. An individual employee establishes liability by proving actual overtime hours worked without overtime pay, or by proving that he or she was denied rest or meal breaks.” (Sotelo v. Media News Group, Inc. (2012) 207 Cal.App.4th 639, 654, 143 Cal.Rptr.3d 293 (Sotelo ).) Here, Dr. Krosnick admitted in his deposition that his survey asked no question identifying for which insurer class members performed inspections. It is thus unclear how plaintiffs could establish the liability of Farmers, for example, without considering whether any inspector worked for Farmers more than eight hours in a day or 40 in a week. (Lab. Code, § 510.) Nor could the information be extrapolated from Farmers’ records of inspections actually performed, because nothing in those records indicated the dates on which they were performed (only the upload date was known), which inspectors performed them (inspectors sometimes had subcontractors perform them), how long they took or would typically take, or whether the inspector did other work on that day for nonparties.
Plaintiffs’ plan similarly failed with respect to their minimum wage claim (Lab. Code, § 1194), as the inspectors were paid a piece rate for each inspection performed, and plaintiffs offered no explanation how they could establish, by their survey alone, the number of inspections performed for Farmers, how long they took, or what Farmers paid for them.
Regarding meal and rest period claims, inspectors performed inspections for a number of insurance companies, including nonparties, often in the same day, but Dr. Krosnick’s survey failed to ask if anyone ever worked long enough in a day for either Farmers or Allstate to be entitled to a meal or rest period from that insurer or any of its three co-employers.
Plaintiffs argue the mere fact that defendants failed to adopt affirmative meal or rest period policies suffices to establish their liability, with damages to be calculated only as a measure of restitution under the unfair competition law. (See Benton v. Telecom Network Specialists, Inc. (2013) 220 Cal.App.4th 701, 726, 163 Cal.Rptr.3d 415 [theory that defendant violated wage and hour requirements by failing to adopt meal and rest break policies is amenable to class treatment; whether employee was able to take required breaks goes to damages].) Our Supreme Court has not yet ruled on this point of law (see Duran, supra, 59 Cal.4th at p. 31, fn. 28, 172 Cal.Rptr.3d 371, 325 P.3d 916 [“We express no opinion on this question”] ), but even if correct, plaintiffs’ survey offered no way even to guess which of the inspectors’ employers, if any, deprived them of meal or rest breaks—even if the only use of the information would be to calculate the amount of restitution owed. (See Cal. Code Regs., tit. 8, § 11040, subds. 11(A), 12(A).)
Plaintiffs argue damages need not be apportioned separately for any insurer defendant because both Farmers and Allstate are vicariously liable as coconspirators for the misconduct of the service companies, which undeniably employed the inspectors. But this ignores the fact that inspectors also worked for nonparties. No authority of which we are aware would make Farmers liable to an inspector who worked as a joint employee of CIS and a nonparty insurer.
And because plaintiffs made no effort to explain how they could establish through common proof what expenses, if any, inspectors incurred for any particular insurer, or how they were deprived of wage statements, the trial court could reasonably conclude these claims were unmanageable as well under the trial plan. (See Lab. Code, § 226, subd. (e); Cal. Code Regs., tit. 8, § 13520.)
The trial court also reasonably concluded that by anonymizing responses plaintiffs unfairly insulated their survey from any meaningful examination. Even Dr. Krosnick, their only witness regarding the survey, did not know who the survey respondents were or why any class member had chosen not to participate. (Dr. Krosnick’s analysis of nonresponse bias did not consider whether any class members may have declined to participate due to their personal lack of any claim.) He declared respondents should not be thought of as witnesses, and he testified he had no opinion as to their reliability.
In fact, plaintiffs expressly admit they intend to answer the ultimate question in this case based solely on expert testimony—testimony founded on multiple hearsay that defendants could never challenge. As Dr. Krosnick declared, “Respondents are not testifying witnesses. Instead, ․ [i]t is the expert who will offer opinions ․, and the expert can be cross-examined.” But “[a]lthough an expert ‘may rely on inadmissible hearsay in forming his or her opinion [citation], and may state on direct examination the matters on which he or she relied, the expert may not testify as to the details of those matters if they are otherwise inadmissible.’ ” (Korsak v. Atlas Hotels, Inc. (1992) 2 Cal.App.4th 1516, 1525, 3 Cal.Rptr.2d 833.)
Plaintiffs argue defendants need no access to Dr. Krosnick’s data, as they are free to conduct their own survey and present contrary conclusions to the jury. This again misses the point. Defendants have the right to defend against plaintiffs’ claims by impeaching the evidence supporting them. (Goldberg v. Kelly (1970) 397 U.S. 254, 269-270, 90 S.Ct. 1011, 25 L.Ed.2d 287 [“due process requires an opportunity to confront and cross-examine adverse witnesses”].) Plaintiffs’ proposed procedure forestalls defendants’ exercise of this important right.
“A class ․ may establish liability by proving a uniform policy or practice by the employer that has the effect on the group of making it likely that group members will work overtime hours without overtime pay, or to miss rest/meal breaks.” (Sotelo, supra, 207 Cal.App.4th at p. 654, 143 Cal.Rptr.3d 293.) And “California courts and others have in a wide variety of contexts considered pattern and practice evidence, statistical evidence, sampling evidence, expert testimony, and other indicators of a defendant’s centralized practices in order to evaluate whether common behavior towards similarly situated plaintiffs makes class certification appropriate.” (Sav-On, supra, 34 Cal.4th at p. 333, 17 Cal.Rptr.3d 906, 96 P.3d 194.) But no case of which we are aware, and plaintiffs refer us to none, suggests a trial may be conducted solely on the evidence of an expert witness relying on an anonymous double-blind survey, no matter how scientific the survey may be.1
For the foregoing reasons, we conclude the trial court acted within its discretion in denying class certification.
The order denying certification is affirmed.2 Respondents are to receive their costs on appeal.
1. We express no opinion on the scientific validity of Dr. Krosnick’s survey qua survey, and nothing in our opinion is intended to limit the trial court’s ability to examine Dr. Krosnick concerning the matter upon which his opinion is based. (Evid. Code, §§ 801, 802; Sargon Enterprises, Inc. v. University of Southern California (2012) 55 Cal.4th 747, 771, 149 Cal.Rptr.3d 614, 288 P.3d 1237.)
2. The motion of PMG to dismiss plaintiffs’ appeal is denied.
CHANEY, Acting P. J. FN* Judge of the Los Angeles Superior Court, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution.
We concur: BENDIX, J. CURREY, J.*