*394Opinion

Valerie Alberts and others (hereafter plaintiffs), formerly employed as members of the nursing staff at two acute care psychiatric hospitals owned and operated by Aurora Behavioral Health Care (Aurora), filed the instant wage and hour lawsuit alleging, on behalf of themselves and a class of similarly situated individuals, that Aurora’s uniform practices and de facto policies routinely denied nursing staff employees meal and rest periods required by California law. Plaintiffs allege Aurora intentionally understaffed its hospitals while simultaneously requiring nursing staff to remain at their posts and monitoring patients unless relieved, resulting in class members being denied meal and rest breaks (and failing to receive additional compensation required by California law). Plaintiffs further allege that Aurora required nursing staff members to complete outstanding assignments before leaving at the end of a shift, but actively discouraged or denied requests for overtime compensation and instructed employees to finish outstanding tasks off the clock. Plaintiffs proposed five subclasses: the meal break subclass, the rest break subclass, the overtime subclass, and two derivative subclasses for waiting time penalties owed and inaccurate wage statements.

The trial court denied plaintiffs’ motion for class certification, finding a lack of “commonality” among the subclasses. We conclude the court relied on improper criteria and erroneous legal assumptions in denying certification. However, while we believe that plaintiffs’ theory of liability presents a common question well suited for class treatment, it is unclear what effect, if any, individual issues, such as damages, will have on the manageability of the case. As our Supreme Court has cautioned, “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.” (Duran v. U.S. Bank National Assn. (2014) 59 Cal.4th 1, 29 [172 Cal.Rptr.3d 371, 325 P.3d 916].) Here, the parties and the trial court focused almost exclusively on the existence of common issues, to the exclusion of the issue of manageability. Accordingly, we reverse and remand the matter for further consideration consistent with our holding.

FACTUAL AND PROCEDURAL BACKGROUND

Defendants and respondents Aurora Las Encinas Hospital (Las Encinas) and Aurora Charter Oak Hospital (Charter Oak) are psychiatric hospitals *395owned and operated by defendant and respondent Aurora (collectively, the Hospital). The Hospital provides services at varying levels of care — inpatient, partial hospitalization and intensive outpatient programs — to patients suffering from various psychiatric illnesses, chemical dependency or both (co-occurring disorders).

This action was initiated in August 2009. The operative fourth amended complaint alleges that Valerie Alberts, Rudolph Breilein, Robin Motola, Cyndi Lane, Shelby Edison and Aviance Contreras are members of a putative class of current and former nonexempt employees of the Hospital who, from August 6, 2005, to the present (class period), provided patient care and held the following positions on the Hospital’s nursing staff: registered nurse (RN), licensed vocational nurse (LVN), licensed psychiatric technician (LPT), and mental health worker (also known as behavior health specialist or psychiatric assistant; MHW). The complaint alleges unpaid overtime, failure to provide meal and rest periods, failure to pay waiting time penalties, and failure to provide accurate itemized statements (Lab. Code,1 §§ 1194, 226.7, 203, 226), among other statutory violations, and unfair competition based on these violations (Bus. & Prof. Code, § 17200).

Plaintiffs sought class certification on behalf of approximately 1,053 putative class members. Plaintiffs proposed that the class be split into two primary subclasses, divided between individuals employed by Las Encinas and by Charter Oak. Those subclasses would in turn be divided into six additional subclasses, five of which are at issue here: (1) a rest break subclass, (2) a meal break subclass, (3) an overtime subclass, (4) a waiting time subclass and (5) an itemized statement subclass.2 Plaintiffs argued the central question in establishing classwide liability was whether class members were subjected to common practices and policies which denied them meal and rest breaks and overtime payments.

In support of the class certification motion, plaintiffs submitted declarations by plaintiffs’ counsel, documentation and declarations from 25 (mostly) former employees of Las Encinas and Charter Oak, and excerpts from the depositions of Cheryl Cook, director of nursing (DON) at Las Encinas and the designated most knowledgeable witness regarding the Hospital policies at issue, Evaldo Casas, Las Encinas’s former staffing coordinator, and Brenda Nocon, Charter Oak’s DON. Plaintiffs also submitted two expert witness declarations. The first, from Dr. Brian Kriegler, reflected his (1) development of a sampling design to select representative samples of both potential class members and timekeeping and payroll data for Las Encinas, (2) analysis of Hospital timekeeping and payroll data for sample class members, and (3) *396analysis of data provided as a result of his first two tasks in light of plaintiffs’ allegations regarding class certification. The second expert declaration, by long-term psychiatric nurse Denise Rounds, related to standard practices in acute care psychiatric facilities, and contained Rounds’s analysis of the Hospital’s administration of its nursing staff, with a focus on the impact of staffing levels on the provision of meal and rest breaks to nursing staff employees, off-the-clock work and patient care.

The Hospital opposed certification. The Hospital argued that it maintained lawful meal and rest break policies, that employees were paid for all hours worked, and that plaintiffs had failed to proffer substantial credible evidence that an informal policy of failing to provide employees meal and rest breaks or to perform off-the-clock work could be proved on a classwide basis with common evidence. The Hospital asserted that, absent substantial evidence that its allegedly unlawful practices and procedures were susceptible to classwide proof, liability determinations necessarily hinged on individual determinations as to why meal or rest breaks were missed, and whether the Hospital knew or should have known an employee was working off the clock.

In support of its opposition, the Hospital presented, among other things, documentation and the declarations of 34 (mostly) current nursing staff employees, as well as Charter Oak’s chief operating officer, DON Cook and Hospital’s counsel, with attached excerpts from the depositions of Rounds, DON Cook, Casas, Kriegler and certain members of the putative class. The Hospital also submitted a declaration by its expert statistician, Robert Crandall, containing his assessment of Kriegler’s report, and an analysis of “whether the available data is consistent with the hypothesis that systematic practices resulted in employees systematically not being provided meal breaks.” Finally, the Hospital lodged numerous evidentiary objections to (1) plaintiffs’ and other witnesses’ declarations, (2) the declaration of one of plaintiffs’ attorneys, (3) and the Kriegler and Round declarations.

Plaintiffs, in turn, filed a reply brief, an additional declaration by counsel, with documents, discovery materials and excerpts from numerous depositions attached, and an extensive supplemental declaration by Kriegler. Plaintiffs also responded to the Hospital’s evidentiary objections, and lodged objections of their own against the Hospital’s witness and employee declarations and the Crandall declaration. Both sides submitted briefs addressing then-recent changes in the law regarding class certification. The Hospital also filed responses to plaintiffs’ evidentiary objections, lodged additional objections of its own and moved to strike Kriegler’s supplemental declaration and portions of the reply brief.

On April 10, 2013, following oral argument, the trial court adopted its tentative ruling and denied class certification on the ground that each *397proposed subclass lacked “commonality.” The trial court found the motion to strike moot. It also declined to address the parties’ evidentiary objections, which it deemed irrelevant to the bases for its denial of certification. Plaintiffs filed this timely appeal. (Linder v. Thrifty Oil Co. (2000) 23 Cal.4th 429, 435 [97 Cal.Rptr.2d 179, 2 P.3d 27] (Linder) [order denying motion for class certification is appealable].)

DISCUSSION

I. Applicable class action principles and the standard of review

A. Standards for class certification

“Drawing on the language of Code of Civil Procedure section 382 and federal precedent,” our Supreme Court has articulated three requirements for the certification of a class. (Brinker Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004, 1021 [139 Cal.Rptr.3d 315, 273 P.3d 513] (Brinker).) Specifically, “[t]he party advocating class treatment must demonstrate [(1)] the existence of 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.” (Ibid.)

The second factor, the “community of interest” factor, is comprised of three subfactors: “ ‘ “(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.) In deciding whether the common questions “predominate,” courts must do three things: “identify the common and individual issues”; “consider the manageability of those issues”; and “taking into account the available management tools, weigh the common against the individual issues to determine which of them predominate.” (Dunbar v. Albertson’s, Inc. (2006) 141 Cal.App.4th 1422, 1432 [47 Cal.Rptr.3d 83].)

“A motion to certify a class action is not a trial on the merits, nor does it function as a motion for summary judgment.” (Carabini v. Superior Court (1994) 26 Cal.App.4th 239, 245 [31 Cal.Rptr.2d 520].) “A class certification motion is not a license for a free-floating inquiry into the validity of the complaint’s allegations; rather, resolution of disputes over the merits of a case generally must be postponed until after class certification has been decided . . . .” (Brinker, supra, 53 Cal.4th at p. 1023.) Nor may a court deny certification on the ground that class members must individually prove their damages. (Id. at p. 1022.)

Class certification “ ‘is “essentially a procedural [question] that does not ask whether an action is legally or factually meritorious.” ’ ” (Brinker, supra, *39853 Cal.4th at p. 1023; see Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 326 [17 Cal.Rptr.3d 906, 96 P.3d 194] (Sav-On); Jaimez v. Daiohs USA, Inc. (2010) 181 Cal.App.4th 1286, 1298 [105 Cal.Rptr.3d 443] (Jaimez).) Although a trial court may consider the merits of a proposed class action, it may do so only to determine whether factual or legal questions common to all class members will likely predominate in driving the litigation, thus making the action amenable to class treatment. (Brinker, at pp. 1021, 1025.) The trial court should resolve legal or factual issues only if doing so is “necessary to a determination whether class certification is proper.” (Id. at p. 1023, italics omitted.) “As the focus in a certification dispute is on what type of questions — common or individual — are likely to arise in the action, rather than on the merits of the case [citations], in determining whether there is substantial evidence to support a trial court’s certification order, [the court] considers] whether the theory of recovery advanced by the proponents of certification is, as an analytical matter, likely to prove amenable to class treatment.” (Sav-On, at p. 327; see Ghazaryan v. Diva Limousine, Ltd. (2008) 169 Cal.App.4th 1524, 1531 [87 Cal.Rptr.3d 518].)

“The ‘ultimate question’ the element of predominance presents is whether ‘the issues which may be jointly tried, when compared with those requiring separate adjudication, are so numerous or substantial that the maintenance of a class action would be advantageous to the judicial process and to the litigants.’ [Citations.] ... A court must examine the allegations of the complaint and supporting declarations [citation] and consider whether the legal and factual issues they present are such that their resolution in a single class proceeding would be both desirable and feasible. ‘As a general rule 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, supra, 53 Cal.4th at pp. 1021-1022, fn. omitted.) “[The 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.” (Id. at p. 1024.)

B. Standard of review

A ruling on class certification is reviewed for abuse of discretion. (Brinker, supra, 53 Cal.4th at p. 1022; Sav-On, supra, 34 Cal.4th at p. 326.) “ ‘Because trial courts are ideally situated to evaluate the efficiencies and practicalities of permitting group action, they are afforded great discretion in granting or denying certification.’ ” (Sav-On, at p. 326.) “ ‘A certification order generally will not be disturbed unless (1) it is unsupported by substantial evidence, (2) it rests on improper criteria, or (3) it rests on erroneous legal assumptions.’ ” (Brinker, at p. 1022; see Linder, supra, 23 Cal.4th at pp. 435-436.)

*399An appeal from an order denying class certification presents an exception to customary appellate practice by which we review only the trial court’s ruling, not its rationale. If the trial court failed to conduct the correct legal analysis in deciding not to certify a class action, “ ‘an appellate court is required to reverse an order denying class certification . . . , “even though there may be substantial evidence to support the court’s order.” ’ ” (Bartold v. Glendale Federal Bank (2000) 81 Cal.App.4th 816, 828 [97 Cal.Rptr.2d 226].) In short, we must “ ‘consider only the reasons cited by the trial court for the denial, and ignore other reasons that might support denial.’ [Citation.]” (Jaimez, supra, 181 Cal.App.4th atpp. 1297-1298; accord, Ramirez v. Balboa Thrift & Loan (2013) 215 Cal.App.4th 765, 776 [155 Cal.Rptr.3d 518].)

II. Pertinent wage and hour requirements

“We begin by identifying the principal legal issues and examining the substantive law that will govern. In doing so, we do not seek to resolve those issues. Rather, the question at this stage is whether the operative legal principles, as applied to the facts of the case, render the claims susceptible of resolution on a common basis. [Citations.]” (Ayala v. Antelope Valley Newspapers, Inc. (2014) 59 Cal.4th 522, 530 [173 Cal.Rptr.3d 332, 327 P.3d 165] (Ayala).) Here, as in Brinker, in which our Supreme Court clarified the scope of an employer’s meal and rest break obligations and explained the criteria applied to assess motions to certify wage and hour claims, only a single element of class suitability, the community of interest question- — ■ predominance of common questions — is in dispute. (See Brinker, supra, 53 Cal.4th at pp. 1021-1022.)

Plaintiffs sought certification of three subclasses of claims: violation of meal period provisions, violation of rest period provisions and failure to pay compensation for missed breaks and overtime.3 California’s meal and rest break rules, and the rules governing overtime pay, are contained in wage orders issued by the Industrial Welfare Commission “on an industry-by-industry basis.” (Bradley v. Networkers Internat., LLC (2012) 211 Cal.App.4th 1129, 1149 [150 Cal.Rptr.3d 268] (Bradley); see Brinker, supra, 53 Cal.4th at pp. 1026-1027.) The putative class members in this case are covered by Industrial Welfare Commission wage order No. 5-2001, which applies to health care workers in public facilities, including hospitals. (Cal. Code Regs., tit. 8, § 11050, subd. 2(P)(4).) As set forth below, federal and state regulations require acute psychiatric care facilities to have adequate staff to serve patient needs. At a minimum, an RN must be on duty at all times. The Hospital requires every unit to have at least two staff members on duty, at least one of *400whom is an RN. The central legal issues are whether the Hospital’s alleged practice and policy of (1) purposefully understaffing units while also requiring nursing staff to remain on duty unless relieved resulted in classwide denial of meal and/or rest breaks and (2) altering timekeeping records, requiring staff to perform work off the clock, and denying or discouraging employees from seeking compensation owed resulted in classwide denial of overtime pay.

Pertinent meal period provisions require that “[n]o employer shall employ any person for a work period of more than five (5) hours without a meal period of not less than 30 minutes ....” (Cal. Code Regs., tit. 8, § 11050, subd. 11(A).) “[A]n employer’s obligation is to provide a first meal period after no more than five hours of work and a second meal period after no more than 10 hours of work.” (Brinker, supra, 53 Cal.4th at 1049.) To qualify as a lawful meal break under California law, an employee must be relieved of all duties for an uninterrupted 30 minutes. (Id. at p. 1040; Cal. Code Regs., tit. 8, § 11050, subd. 11(A).) If an employer fails to comply with these requirements it must pay one hour of pay at the employee’s regular rate “for each workday that the meal period is not provided.” (Cal. Code Regs., tit. 8, §11050, subd. 11(B); see Lab. Code, § 226.7, subd. (c).)

The wage order’s rest period provisions require employers to provide “10 minutes’ rest for shifts from three and one-half to six hours in length, 20 minutes for shifts of more than six hours up to 10 hours, 30 minutes for shifts of more than 10 hours up to 14 hours, and so on.” (Brinker, supra, 53 Cal.4th at p. 1029; see Cal. Code Regs., tit. 8, § 11050, subd. 12(A).) As with meal breaks, employers are required to pay one hour of compensation at the regular rate “for each workday that the rest period is not provided.” (Cal. Code Regs., tit. 8, § 11050, subd. 12(B); see Lab. Code, § 226.7, subd. (c).)

Finally, California’s overtime provisions require, in part, that each employee receive one and a half times their regular rate of pay for all hours worked in excess of 40 hours in a workweek, one and a half times their regular pay for all hours worked in excess of eight hours in a single day, and double their regular rate of pay for all hours worked in excess of 12 hours in a single day. (Cal. Code Regs., tit. 8, § 11050, subd. 3(A)(1)(a) & (b).) Section 1194 provides employees a cause of action for unpaid overtime against their employer. (See Martinez v. Combs (2010) 49 Cal.4th 35, 49-50 [109 Cal.Rptr.3d 514, 231 P.3d 259].)

*401III. The trial court erred in denying plaintiffs’ motion for certification of the meal and rest break subclasses

A. Staffing at Charter Oak and Las Encinas: the backdrop

Charter Oak and Las Encinas provide inpatient, partial hospitalization, and intensive outpatient programs to treat adults and adolescents with psychiatric illnesses and chemical dependency problems. Each facility is divided into units according to the type of patient and degree of intensity of treatment required. The Hospital is licensed and subject to inspection by the State Department of Public Health. Residential treatment is licensed and certified by the State Department of Health Care Services. Staffing for acute care facilities is governed by federal and state regulations, which require the Hospital to have enough qualified staff to meet patients’ needs. (See 42 C.F.R. § 482.23(b) (2015) [acute psychiatric care facility must “have adequate numbers of . . . personnel to provide nursing care to all patients as needed”]; Cal. Code Regs., tit. 22, §§71213, subd. (f) [hospital is required to have “staffing requirements based on assessment of patient needs”], 71215, subds. (c)(2) & (d) [there must be “[sufficient registered nursing personnel” in order to provide “direct nursing care based on patient need,” and each unit must have at least one RN on duty at all times].)4

The Hospital maintains a strict policy requiring that every unit have at least two staff members on duty at all times, at least one of whom is a registered nurse. Staffing needs fluctuate, depending on the number of patients in a given unit (the unit’s “census”) and the level of care a patient requires (the patient’s “acuity”). For purposes of staff scheduling, patient acuity is measured on a scale of one to four, expressed as a function of the number of “patient care hours” a patient requires. A level 1 patient demands little staffing attention while, at the extreme end, a level 4 patient — requiring that a care provider be within arm’s reach of the patient at all times to ensure the safety of the patient or others — requires eight patient care hours per eight-hour shift, a one-to-one staffing ratio.

Charter Oak applies a mathematical formula (census multiplied by acuity) to determine the number of staff required to adequately staff each unit for each of three shifts, ranging from the lowest to highest acuity le'vels: (1) 1.1 patient care hours, (2) 1.5 patient care hours, (3) 2 patient care hours and (4) 8 patient care hours. At Las Encinas, staffing needs are determined using a one-to-six ratio of licensed staff to patients. For the first six patients the licensed staff must be an RN, and an additional RN or LVN is added for each *402additional group of up to six patients. Thus, a unit with 12 patients is to be staffed by at least one RN, plus another RN or LVN, a unit with 14 patients is staffed by at least one RN plus another two RN’s or LVN’s. MHW’s are added as required based on unit census and patient acuity.

Staffing at both facilities requires both advance planning and the flexibility to make rapid adjustments Staffing needs can fluctuate rapidly and radically, depending on changes in census (patients admitted or discharged) and/or patient acuity levels (which may shift quickly, especially among patients treated for psychiatric disorders, e.g., a previously low-acuity patient suddenly threatens bodily injury, thus necessitating one-to-one care). Despite advance planning, staffing levels are dynamic and may require immediate adjustment if a unit finds itself with too many staff (due to discharges) or too few (due to a shift in acuity necessitating that staff be pulled from elsewhere to attend to the patient).

Because the Hospital provides treatment to patients suffering from various psychiatric disorders there is an ever-present risk patients will act out or become violent.5 Accordingly, the Hospital has stressed that patient safety is its “number one” priority, and management regularly and repeatedly instructs employees that it is critical that nursing staff remain vigilant and that patients be constantly monitored. To ensure patient safety and maintain adequate staffing levels, no member of the nursing staff is permitted to leave his or her unit for any break unless he or she is relieved. The Hospital has repeatedly stressed to nursing staff that this is a “zero tolerance issue” and employees who violate the policy are subject to discipline. An RN may be relieved only by another RN. An LVN may be relieved by an LVN or an RN; MHW’s may be relieved by any member of the patient care staff. RN’s and LVN’s have independent obligations to provide patient care at a level required to maintain their professional licenses. An RN who abandons patients without being properly relieved of duty, places his or her professional license in jeopardy.6 *403If a conflict arises between patient needs and an employee’s right to take a break, patient needs must always prevail.

Hospital policy requires that members of the nursing staff ensure that their duties are covered while away from their units. The reason for this policy is obvious: staffing levels depend on the number of nursing staff workers required to address patient care needs, so if an assigned staff member takes a break, another appropriate patient care worker (e.g., an RN for an RN) must step in to ensure that patient care needs are met. In November 2005, Las Encinas began scheduling an RN to “float” among units to provide meal break relief for members' of the nursing staff. Hospital policy also provides for 10-minute rest breaks, but neither Las Encinas nor Charter Oak schedules any staff to provide relief for rest breaks. Employees must arrange for any rest breaks on their own.

Plaintiffs claim that both Las Encinas and Charter Oak are chronically understaffed, and that the Hospital schedules additional staff for break relief only sporadically, thereby placing tremendous pressure on employees not to leave their units, even when they are due for and need a break. If no one is scheduled to provide break relief, staff members are generally forced to rely on coworkers to cover for them in order to take a break.7 Often, this means one person is left to cover the job duties of two, while the other staff member runs out to grab food. In violation of staffing ratios, the person covering has twice the acceptable patient load, creating a potentially unsafe condition. Even when relief staff is scheduled, plaintiffs argue that such relief is routinely unavailable because the floater is called away to perform other duties, or has insufficient time in his or her own schedule to enable staff members to take a full 30-minute break.8

The Hospital has a “Timekeeping Adjustment Form” (TAF, or “kronos”) that employees may use to request compensation for missed meal breaks or overtime. Sometime after this lawsuit was filed, Las Encinas modified its TAF to permit employees to seek compensation for missed rest breaks. However, the Hospital conceded in discovery responses that its payroll system does not differentiate between meal and rest breaks, and it cannot confirm that any employee has ever been paid for a missed rest breaks.

*404B. Plaintiffs’ allegations regarding the denial of meal and rest breaks

Against this backdrop, plaintiffs allege that, in order to reduce expenses and boost profits, the Hospital maintains a policy and engages in an intentional business practice to provide only skeletal staffing resulting in units being chronically understaffed, and fails to provide sufficient relief staff to enable employees to take breaks to which they are lawfully entitled. This policy and practice results in class members effectively being forced to remain on duty and routinely being denied legally compliant meal and rest breaks or compensation therefor, and endangers patients and staff. Plaintiffs further allege that, in order to reduce expenses and increase profits while also meeting patient needs, maintaining appropriate staffing ratios and ensuring timely completion of required documents and paperwork, Hospital management regularly directs them to perform duties off the clock, while also actively discouraging employees from seeking or refusing to pay overtime wages.

Plaintiffs argued that the Hospital has common practices and policies which violate the California law requiring that employees be permitted to take a 30-minute uninterrupted meal break, relieved of all duties, for each five hours of work, and be permitted 10-minute rest breaks after the second and sixth hours of work in an eight-hour shift. In plaintiffs’ view, the Hospital was intentionally and chronically understaffed, a practice which routinely denied nursing staff the opportunity to exercise the right to an uninterrupted 30-minute meal break within the first five hours of a shift. For purposes of class certification, the question is whether this theory of recovery can be “proved (or disproved) through common facts and law.” (Bradley, supra, 211 Cal.App.4th at p. 1143.)

C. The trial court’s reasons for denial of class certification

On appeal, we must “ ‘consider only the reasons cited ... for the denial.’ ” (Jaimez, supra, 181 Cal.App.4th at p. 1297.) The trial court provided several reasons for its conclusion that plaintiffs’ meal and rest break claims could not be determined through common proof.9

First, it found that the “facial legality” of the Hospital’s written policies regarding meal and rest breaks in its employee handbook was (virtually) undisputed. Accordingly, it required plaintiffs to demonstrate the existence of a “systematic practice by supervisors and managers of denying employees the benefits afforded to them” by those policies. Plaintiffs failed to satisfy this requirement because there was “too much variance between the declarations *405and deposition responses of putative class members to indicate [the Hospital had] a universal practice of denying employees their meal and rest breaks.”

Second, the court concluded that, even if plaintiffs had shown that employees had missed breaks, the evidence also showed that their reasons for doing so “varied, ranging from understaffing and coercive hospital policy ... to simple choice.” The court found a similar evidentiary variation with regard to whether nursing staff were compensated for missed breaks, or dissuaded by supervisors from seeking such compensation.

Third, although “both Plaintiffs and [the Hospital] presented statistical evidence as to the existence (or lack thereof) of a systematic policy discouraging meal and rest breaks,” the trial court found it “telling that Plaintiffs’ statistical evidence [did] not account for the possibility that some (or most) of the employees voluntarily worked through or delayed their breaks.” “Ultimately,” the crux of the court’s denial of certification was its overarching conclusion that plaintiffs’ motion relied too heavily “on anecdotal evidence to prove the existence of a systematic violation of overtime and break laws.”

As discussed below, each of these reasons is without merit.

1. Plaintiffs disputed the facial legality of the Hospital’s written policy

The trial court found that because the “facial legality” of the Hospital’s written policy regarding meal breaks was undisputed, plaintiffs were required — but had failed — to demonstrate the existence of a “universal practice” by management to deny nursing staff the benefit of that policy. The trial court’s conclusion rests on a flawed premise. Plaintiffs do in fact dispute the “facial” legality of the Hospital’s break policies.

The meal break policy — which applies to all putative class members — states that employees are entitled to “an unpaid thirty-minute break for a meal period, approximately half way between the beginning and ending of the employee’s shift.” California law, however, requires that a meal break be provided during the first five hours of an employee’s shift. (See Brinker, supra, 53 Cal.4th at pp. 1048-1049.) Plaintiffs submitted substantial evidence that, on those occasions when nursing staff were able to take a meal break, the break was almost always not taken during the first five hours of their shift.

California law also requires that an employee be entitled to take a second meal break for shifts that exceed 10 hours. (Brinker, supra, 53 Cal.4th at p. 1049.) Plaintiffs submitted evidence showing that Hospital policy does not provide for a second meal break, and staff are routinely not provided a second *40630-minute meal break in shifts over 10 hours. For example, although the vast majority of the 35 employee declarations submitted by the Hospital contain testimony that the employee sometimes worked overtime, none contains testimony that the employee received a second meal period during shifts that exceeded 10 hours. Further, only one of the Hospital’s 35 employee declarants was paid a meal period premium when the employee was unable to take a meal break within the first five hours of a shift. According to plaintiffs’ statistical expert, approximately one-third of the meal breaks recorded were not provided within the first five hours of an employee’s shift.

The Hospital’s rest break policy is similarly noncompliant; it does not require that rest breaks be provided for each four hours worked, or a “ ‘major fraction’ ” thereof, and does not provide for a third rest break in shifts exceeding 10 hours. (Quoting Brinker, supra, 53 Cal.4th at p. 1032 [finding that plaintiffs’ claim that employer adopted a uniform rest break policy that failed to “give full effect to the ‘major fraction’ language” of the applicable wage order was the sort of claim “routinely, and properly, found suitable for class treatment”].) The witness declarations submitted by the Hospital do not undermine plaintiffs’ central theories of recovery. (See Jaimez, supra, 181 Cal.App.4th at pp. 1300-1301 [concluding trial court focused improperly on evaluating conflicting issues of fact raised by defendant’s declarations, rather than evaluating whether plaintiff’s theory of recovery was likely to prove amenable to class treatment].)

In short, the trial court’s underlying premise that the parties agreed the Hospital’s written break policies were legally compliant was incorrect. Because denial of class certification rested, in part, on that flawed premise, reversal is in order. However, as discussed below, there are more substantive reasons for reversal.

2. Evidence that some employees took breaks is not a basis for denial of class certification

Even if we assume (as the trial court did) that the Hospital’s written meal and rest break policy is “facially legal” and its facial legality is undisputed by plaintiffs, the mere existence of a lawful break policy will not defeat class certification in the face of actual contravening policies and practices that, as a practical matter, undermine the written policy and do not permit breaks. (See, e.g., Brinker, supra, 53 Cal.4th at p. 1040 [an “employer may not undermine a formal policy of providing meal breaks by pressuring employees to perform their duties in ways that omit breaks”]; Jaimez, supra, 181 Cal.App.4th at p. 1303; Boyd v. Bank of America Corp. (C.D.Cal. 2014) 300 F.R.D. 431, 442; cf. Cicairos v. Summit Logistics, Inc. (2005) 133 Cal.App.4th 949, 963 [35 Cal.Rptr.3d 243] [reversing summary judgment for *407employer based, in part, on evidence that truck drivers “felt pressured” not to take rest breaks, and management knew some drivers were not taking breaks].)

Nor was the court correct to require, at the certification stage, that plaintiffs demonstrate a “universal practice” on the part of management to deny nursing staff the benefit of the Hospital’s written break policy.10 The trial court failed to analyze the proper question — whether plaintiffs had articulated a theory susceptible to common resolution. (Benton v. Telecom Network Specialists, Inc. (2013) 220 Cal.App.4th 701, 726 [163 Cal.Rptr.3d 415] (Benton) [“the proper inquiry is ‘whether the theory of recovery advanced by the plaintiff is likely to prove amendable to class treatment’ ”]; Sav-On, supra, 34 Cal.4th at p. 327.) Instead, it asked whether the evidence was sufficient to establish plaintiffs’ ultimate right to recovery. At the certification stage, plaintiffs need only establish that the question of whether the Hospital’s practices or procedures resulted in the denial of lawful breaks can be determined on a classwide basis. Instead of undertaking this analysis, the trial court held that plaintiffs had to prove class members missed all breaks to which they were entitled. This is an incorrect standard for certification that, as other courts have also found, if correct, would prevent certification of virtually any wage and hour class. (See, e.g., Bufil v. Dollar Financial Group, Inc. (2008) 162 Cal.App.4th 1193, 1207 [76 Cal.Rptr.3d 804] (Bufil) [“a class is not inappropriate merely because each member at some point may be required to make an individual showing as to eligibility for recovery”]; Benton, at pp. 725-728 [reversing order denying certification despite evidence that some putative class members received breaks].)

The trial court also improperly denied class certification on the basis that plaintiffs “rel[ied] too much on anecdotal evidence to prove the existence of systemic violation[s]” of wage and hour laws. First, much of the evidence on which plaintiffs rely is not merely anecdotal. Plaintiffs relied on the Hospital’s own practices and policies, schedules and internal correspondence. Plaintiffs also relied on the findings of an independent survey undertaken in September 2008 by the State Department of Public Health, which revealed that, on the day of the inspection, the Hospital failed to provide requisite staff-to-patient ratios, and failed to ensure that staff assigned to monitor patients on a one-to-one basis had no other job duties in order to ensure a safe and secure patient environment. That survey also concluded that staff-to-patient ratios were exceeded when staff members provided break relief for coworkers by *408assuming their job duties. In addition, plaintiffs relied on testimony by DON Cook and Nocon who were aware and instructed staffing coordinators and supervisors to require that staff clocked out and in for meal breaks (whether or not taken) and return immediately to work, and that the Hospital’s staffing coordinator was authorized to alter time cards to avoid paying meal period premiums, and routinely did so. Plaintiffs also presented a declaration from their staffing expert, an experienced psychiatric RN, Denise Rounds. Rounds did not take issue with the manner in which the Hospital made its staffing projections, only its failure to adhere to those projections.11

Furthermore, both sides presented some anecdotal evidence. Plaintiffs presented the declarations of 25 witnesses who claimed they were rarely or never authorized to take rest and meal breaks due to understaffing or the Hospital’s failure to provide break coverage, and were routinely forced to work off the clock or overtime without compensation. In rebuttal, the Hospital submitted 34 declarations by witnesses who said they were routinely authorized to take meal and rest breaks for which break coverage was provided, regularly took breaks (which they had not chosen to waive or delay) on time, understood the process to obtain and received pay for missed breaks and overtime, and were not forced to work off the clock.

The Hospital argued that, to the extent witnesses missed breaks, they did so for a variety of reasons, not all of which depended on workplace demands. The Hospital also argued that, when pressed at depositions, a number of plaintiffs’ witnesses disavowed parts of their declarations, confirming that they had in fact taken meal and rest breaks. The Hospital misstates the record. There is no question that some of plaintiffs’ declarants testified they occasionally had an opportunity to take breaks. However, the fact that some employees may have taken some breaks is an issue that goes to damages. It is not a proper basis on which to deny certification. “ ‘[A] class action,’ ” as clarified by our Supreme Court, “ ‘is not inappropriate simply because each member of the class may at some point be required to make an individual showing as to his or her eligibility for recovery or as to the amount of his or her damages.’ ” (Sav-On, supra, 34 Cal.4th at p. 333.) And, as Jaimez, supra, 181 Cal.App.4th 1286 explained more recently, “The fact that individual [workers] may have different damages does not require denial of the class certification motion.” (Id. at p. 1301, italics omitted.) “ ‘That calculation of individual damages may at some point be required does not *409foreclose the possibility of taking common evidence on the misclassification questions.’ [Citation.] In sum, ‘individualized proof of damages is not per se an obstacle to class treatment. . . .’ [Citation.] It is no bar to certification ‘that individual class members may ultimately need to itemize their damages.’ ” (Id. at p. 1302.)

Plaintiffs do not claim they were universally denied all breaks, nor must they do so to warrant certification. Brinker, supra, 53 Cal.4th 1004 does not require class proponents to establish the universal application of an allegedly illegal policy; rather, a class proponent need only show a “consistent[]” application of the policy. (Id. at p. 1033.) In Brinker’s wake, courts have repeatedly found that a defendant employer’s evidence of an inconsistent application of an illegal policy to be insufficient on its own to defeat class certification. For example, in Benton, the court reversed the denial of class certification, stating, “The mere fact that some technicians may have taken breaks (or declined to take breaks) . . . ‘ “does not show that individual issues will predominate in the litigation.” ’ ” (Benton, supra, 220 Cal.App.4th at p. 730.) Similarly, in Faulkinbury v. Boyd & Associates, Inc. (2013) 216 Cal.App.4th 220 [156 Cal.Rptr.3d 632] (Faulkinbury), the court reversed the denial of class certification, explaining that “[i]n opposition to the motion for class certification, [the defendant employer] submitted declarations from current employees. . . . [I]n light of Brinker, this evidence at most establishes individual issues of damages, which would not preclude class certification.” (Id. at p. 237.) In Hall v. Rite Aid Corp. (2014) 226 CalApp.4th 278 [171 Cal.Rptr.3d 504], the Court of Appeal, in reversing the denial of class certification, stated that with regard to Brinker’s progeny “[t]hose courts have . . . agreed that, where the theory of liability asserts the employer’s uniform policy violates California’s labor laws, factual distinctions concerning whether or how employees were or were not adversely impacted by the allegedly illegal policy do not preclude certification.” (Id. at p. 289.) And in Bradley, supra, 211 Cal.App.4th 1129, the court reversed the trial court’s decision to deny class certification due to variations among putative class members, explaining that Brinker has “expressly rejected . . . [the idea] that evidence showing some employees took rest breaks and other employees were offered rest breaks but declined to take them made . . . certification inappropriate.” (Id. at p. 1143.)

Instead, plaintiffs argue that the Hospital’s system governing rest and meal breaks — which applies to all putative class members — does not comply with California law. This inquiry is amenable to class certification. As our Supreme Court has stated, “The theory of liability — -that [the employer defendant] has a uniform policy, and that that policy, measured against wage order requirements, allegedly violates the law — is by its nature a common question eminently suited for class treatment.” (Brinker, supra, 53 Cal.4th at p. 1033.)

*4103. Disregard of statistical evidence

The trial court chose to ignore Kriegler’s statistical analysis, in part, because he failed “to account for the possibility that some (or most) employees voluntarily worked through or delayed their breaks . . . This was error.

First, as a preliminary matter, it should be noted that the trial court’s choice of language — “voluntary”—does not comport with plaintiffs’ theory of recovery. In determining whether a class should be certified, a trial court must examine all of the evidence presented by the parties and must do so “under the prism of the plaintiff’s theory of recovery.” (Department of Fish & Game v. Superior Court (2011) 197 Cal.App.4th 1323, 1349 [129 Cal.Rptr.3d 719], italics added.) A “voluntary” act is defined by Black’s Law Dictionary as one being “[unconstrained by interference; not impelled by outside influence.” (Black’s Law Diet. (10th ed. 2014) p. 1806, col. 1.) Here, plaintiffs’ theory of recovery is that the missed breaks were not voluntary, but were constrained and compelled by the Hospital’s policies. Specifically, plaintiffs have identified several mutually reinforcing policies that sharply circumscribed the employees’ freedom of action: (1) the Hospital’s policy of chronic understaffing; (2) the Hospital’s policy of making patient safety the “number one priority”; and (3) the Hospital’s policy of having “zero tolerance” for staff who take a break without first being relieved. To these policies must be added the nurses’ code of ethics, which, as discussed above, requires that a nurse’s “primary commitment” be to the patient and that the nurse must take all steps to provide “optimum patient care.” The Hospital’s policies, as understood through the prism of plaintiffs’ theory, effectively and unfairly leverage a reasonable nurse’s ethical obligations, making missed breaks mandatory, not voluntary. A reasonable/ethical nurse under such circumstances would not risk the life or health of his/her patient suffering from a psychiatric disorder in order to take a mandated meal or rest break.

Moreover, the trial court’s focus on the purported “voluntary” nature of the employees’ missed breaks does not comport with the law. California courts routinely consider “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; see Jaimez, supra, 181 Cal.App.4th at p. 1298.) If an employer fails to provide legally compliant meal or rest breaks, the court may not conclude employees voluntarily chose to skip those breaks. (Brinker, supra, 53 Cal.4th at p. 1033 [“No issue of waiver ever arises for a rest break that was required by law but never authorized; if a break is not authorized, an employee has no opportunity to decline to take it.”]; Bradley,

*411supra, 211 Cal.App.4th at p. 1151 [“employer is obligated to provide the rest and meal breaks, and if an employer does not do so, the fact that an employee did not take the break cannot reasonably be considered a waiver” (italics omitted)]; Faulkinbury, supra, 216 Cal.App.4th at p. 236.) Thus, if as plaintiffs claim, notwithstanding its written policy, the Hospital actually operates under policies that render illusory its employees’ ability to take meal and rest breaks in the first place, it cannot be argued that individual issues predominate because some putative class members “chose” to forgo meal and rest breaks. The trial court’s flawed rationale that certification was not warranted because some nursing staff “voluntarily” skipped breaks disregards plaintiffs’ theory of recovery, i.e., that there was no real choice to be made “voluntarily.”

Second, the trial court erred in dismissing plaintiffs’ statistical evidence because Kriegler’s analysis was at odds with Crandall’s.12 The court’s conclusion is based on a misconception regarding the role of statistical evidence in the context of a certification motion. According to the trial court, the fact that the parties’ statistical evidence differed meant that plaintiffs failed to satisfy their burden of proof. This reasoning is faulty in several respects. First, it is predicated on an incorrect assumption that plaintiffs had an obligation to prove the merits of their claims at the certification stage. That is simply not the case. “A motion to certify a class action is not a trial on the merits . . . .” (Carabini v. Superior Court, supra, 26 Cal.App.4th at p. 245; see Sav-On, supra, 34 Cal.4th at p, 327.) “A class certification motion is not a license for a free-floating inquiry into the validity of the complaint’s allegations; rather, resolution of disputes over the merits of a case generally must be postponed until after class certification has been decided . . . .” (Brinker, supra, 53 Cal.4th at p. 1023.)

Third, the trial court erred in entirely disregarding Kriegler’s statistical analysis because Crandall reached a contrary conclusion (particularly where, as plaintiffs pointed out, Crandall’s analysis may have failed to account for some important factors). Crandall opined that the Hospital’s timekeeping records reflected wide variation in the percentage of meal breaks missed per employee, thus suggesting a need for individualized inquiry. But, as noted in Kriegler’s supplemental declaration, Crandall’s analysis does not account for an obvious reason for such variation: employees worked different amounts of time and the more shifts they worked, the more missed meal periods they were likely to have.

*412For example, Crandall opined that a sampling of records indicated that 12 employees had missed over 95 percent of their meal breaks while 22 employees had received all required breaks. However, according to Kriegler, Crandall failed to consider that the records for the 22 employees who he found had received all required breaks, reflected that 21 of them worked fewer than 14 shifts, and 16 had worked fewer than five shifts. Kriegler observed that, once the variation in the difference in the number of days worked by employees was taken into account, 94 percent of the variance to which Crandall referred could be explained by objective information contained in the sampling data, thus negating the need for individualized inquiries. Further, Kriegler observed that for both Charter Oak and Las Encinas there was a strong correlation between the number of missed meal breaks and the number of shifts worked.

The Hospital takes issue with Kriegler’s analysis, in part, because he conceded he made “no assumptions and offer[ed] no conclusions about whether meal breaks were provided,” and thus could not offer an opinion on the ultimate issue — whether the Hospital’s informal policy caused classwide deprivation of meal breaks.13 However, as plaintiffs point out, Kriegler was not tasked with the responsibility to offer a conclusion regarding whether employees actually received lawful meal breaks. Rather, Kriegler’s declaration was offered to show that the Hospital’s timekeeping and payroll data confirmed plaintiffs’ theory that class-wide policies led to the denial of meal breaks for putative class members. The trial court erred in refusing to consider statistical evidence.14

4. There is substantial common evidence of understaffing resulting in denial of breaks

Plaintiffs presented substantial evidence, including the testimony of DON’S Cook and Nocon, and documentary evidence that Hospital policy strictly prohibits staff from leaving their units for breaks without designated relief. State regulations require that the Hospital have sufficient RN’s on duty to provide for patient needs and staff supervision. (Cal. Code Regs., tit. 22, §§ 71225, subd. (c), 71215, subd. (c).) Accordingly, Hospital staffing policy requires that staffing ratios be maintained and that an RN, and at least one additional nursing employee be on duty in each unit at all times. Without *413additional break relief, employees must cover for one another in order to take breaks, causing staffing ratios to increase to an unacceptable level and exposing patients and staff to potential danger.15 Plaintiffs argue that the Hospital’s policy requiring employees to remain on the job unless relieved, combined with its failure to ensure adequate staffing, conflicts with the Hospital’s obligation to provide employees with meal and rest breaks.

Through Kriegler, plaintiffs presented evidence that the Hospital failed to comply with its scheduling projections and that scheduling data reflects it provided only sporadic break relief. In addition, a staffing coordinator from Las Encinas acknowledged that even when break relief staff was scheduled, those staff members were “often . . . assigned] ... to tasks that would prevent them from doing any break relief.” There is similar evidence that only sporadic meal break relief was provided at Charter Oak. Plaintiffs argue that the net effect of the Hospital’s failure to adequately staff units was that employees were unable to take the breaks to which they were entitled, and remained responsible for patients throughout an entire shift. Further, even when the Hospital did provide relief for meal breaks, neither its written policies nor actual practice comply with the requirement that the first meal break be provided within the first five hours of an employee’s shift, or that staff were provided the requisite uninterrupted 30 minutes.

Staffing coordinators from Las Encinas and DON Cook testified that Las Encinas does not provide or schedule any designated relief for rest breaks. Similarly, there is evidence that relief for rest breaks was never or rarely available at Charter Oak. Although the Hospital’s written policy provides for some rest breaks, at both facilities it was left to employees to try to obtain coverage in order to take rest breaks; management took no responsibility to make such arrangements.

Common evidence shows the Hospital’s break policies and practices applied to all putative class members. Plaintiffs’ meal and rest break claims depend on a common contention; whether the Hospital’s policies and practice fulfill its obligation to provide lawful breaks. (See Brinker, supra, 53 Cal.4th at p. 1033; Faulkinbury, supra, 216 Cal.App.4th at p. 237 [employer’s alleged companywide practice requiring employees to remain at their posts throughout their shifts was a common question appropriate for certification]; Bufil, supra, 162 Cal.App.4th at pp. 1205-1206 [allegation that policy prohibited store employees from locking store or ignoring customers without relief constitutes a common question suitable for certification].)

*414The authorities on which the Hospital primarily relies do not advance its cause. Kenny v. Supercuts, Inc. (N.D.Cal. 2008) 252 F.R.D. 641 and Brown v. Federal Express Corp. (C.D.Cal. 2008) 249 F.R.D. 580 were decided according to standards governing Federal Rules of Civil Procedure, rule 23 (28 U.S.C.), and before Brinker, which governs the certification of wage and hour claims under California law (Brinker, supra, 53 Cal.4th at p. 1033). Stiller v. Costco Wholesale Corp. (S.D.Cal. 2014) 298 F.R.D. 611 actually advances plaintiffs’ cause. There the court refused an effort to decertify a class for lack of commonality under Federal Rules of Civil Procedure, rule 23(a) (28 U.S.C.) — even under the more exacting federal standard announced in Wal-Mart Stores, Inc. v. Dukes (2011) 564 U.S. 338 [180 L.Ed.2d 374, 131 S.Ct. 2541] — where a grocery chain employer argued its liability for unpaid detention time could not be answered in the same way for each class member. (Stiller, at pp. 623, 625.) The court concluded that commonality was still satisfied because a single common question existed as to whether the employer had a de facto policy of detaining warehouse employees during closing procedures without pay, and enforced that policy on a class wide basis. (Id. at p. 625.) And, in Dailey v. Sears, Roebuck & Co. (2013) 214 Cal.App.4th 974 [154 Cal.Rptr.3d 480], the court affirmed denial of certification in a case alleging misclassification of auto center employees as exempt on the ground that determining whether employee class members were misclassified (because employer had implemented policies and practices that caused the employees to spend most of their time engaging in nonexempt work) would require an examination of each employee’s specific duties. (Id. at pp. 996-997.) Such individual inquiries are not necessary to establish liability in a case such as this in which all putative class members are required to remain on duty absent relief.

In sum, reversal is required with respect to the meal and rest break claims because the trial court’s order denying class certification rests on erroneous assumptions, improper criteria, and, in some respects, insubstantial evidence. While plaintiffs’ theory of liability with respect to the meal and rest break claims presents a common question suitable for class treatment, it is unclear from the record below whether individual issues, such as damages, or common issues will predominate — that is, whether a class proceeding is not only manageable but superior to the alternatives. Accordingly, we remand for further consideration.

IV. The trial court erred in denying plaintiffs ’ motion for certification of the overtime and off-the-clock compensation claims

“For the same reasons” it denied plaintiffs’ meal and rest break claims, the trial court found that plaintiffs “failed to show a systematic policy of denying employees overtime,” instead “providing] only anecdotal evidence that employees were denied overtime or forced to work off the clock.”

*415A. Background

In seeking certification, plaintiffs argued that Hospital management was not only aware that its practices denied employees lawful breaks, management also refused to compensate employees for missed breaks, covered up its violations by altering time records and insisted employees work off the clock during meal breaks. Further, in accordance with California regulations, the Hospital requires that nursing staff complete all assigned tasks and paperwork before the end of each shift. Nursing staff often are unable to complete paperwork during their regular shifts due to the press of attending to patients’ immediate needs. As a result, staff are effectively forced to continue working beyond the end of their scheduled shifts to complete required paperwork. Plaintiffs assert that the Hospital actively discouraged employees from seeking overtime compensation and instructed employees to complete outstanding tasks off the clock, and that liability on this claim was subject to classwide proof.

B. Class-wide evidence of denied compensation for missed breaks

Plaintiffs asserted that, regardless of the Hospital’s written policies and despite management’s awareness that lawfully compliant breaks were routinely denied, the Hospital had a strict policy of requiring employees to clock out for a 30-minute meal break each shift, regardless of whether the employee actually took that break. Nursing staff was purportedly instructed by supervisors to clock out and continue working through their meal breaks. Supervisors, in turn, were instructed by management, including DON’S Cook and Nocon, to ensure that employees under their supervision engaged in this practice.

Some putative class members said they were threatened with discipline by Casas if they failed to clock out for meal breaks even when there was no one to provide break relief. Others testified to the effect that their supervisors insisted that employees’ time records reflect that meal and rest breaks were taken even when it was virtually impossible to take such breaks. Plaintiffs also presented evidence that nursing staff was directed and pressured to clock out for meal breaks, then immediately to resume caring for patients on their units, and that high-level Hospital management was aware of — indeed directly implemented — this practice.

For example, according to Ramirez, a former staffing coordinator at Las Encinas, DON Cook instructed her that employees must clock out for meal breaks without regard to whether breaks were actually received. Putative class members received similar instructions from Casas or their immediate supervisors. For example, Casas told one class member he had failed to clock *416out for meal periods and would be suspended if that practice continued. The class member declared, “I believe that I told [Casas] that I didn’t punch out because I didn’t take my break. Mr. Casas’ response was simply that I had to clock out and back in to show a meal period.” Another putative class member declared that, when she “told [her] supervisor . . . that [she] had no relief to take a meal break, [the supervisor] would simply instruct [her] to clock out, continue working and clock back in after thirty minutes had lapsed.” Numerous other class members levied similar accusations. Plaintiffs claimed the same practice was in effect at Charter Oak. Plaintiff Robin Motola, formerly a nursing supervisor at Charter Oak, declared that she was instructed by DON Nocon that Motola and her staff “should simply clock out for breaks and then go back to the unit and continue to work so that it would appear that [they] had not missed [their] meal break.”

Further, plaintiffs argued that, if employees failed to clock out for a meal break not taken, the staffing coordinators adjusted time records to “correct this.” According to Kriegler, his statistical analysis of a representative sampling of the Hospital’s scheduling and payroll records demonstrated that “nearly one out of every five recorded meal breaks was either added or edited by a supervisor.” Plaintiffs’ declarations supported this conclusion. One putative class member declared that “[e]ven if we forgot to clock in and out [for meal breaks not taken], it didn’t matter because the Hospital would simply alter the hours on our payslips so that we were not paid for the extra time.” Another testified that Casas “told [her] that he would manually add [the employee’s] meal breaks by hand to [her] time cards even if [she] did not take them.” Ramirez declared that when she was staffing coordinator at Las Encinas, she was instructed to alter time records if they showed an employee had not received a full 30-minute lunch. Based on this evidence plaintiffs argue that common evidence demonstrates that the Hospital was not only aware that the policies and practices employed at its facilities routinely deprived employees of lawful breaks, it actively attempted to cover up and avoid compensating employees for those unlawful practices.

The Hospital argues that TAF’s (or kronos) are used as a mechanism by which employees can fill out a form to request premium pay if they missed a meal break and continued working. The Hospital presented evidence that some employees, including some plaintiffs, were aware and availed themselves of this practice. Plaintiffs assert this evidence is misleading because Hospital policy places the onus on employees to justify, and obtain authorization for, missed meal compensation at the same time management actively discourages employees from requesting such compensation, and routinely denies such requests. Plaintiffs’ declarants stated that nursing supervisors were reluctant to sign off on TAF’s because upper management placed pressure on them not to authorize overtime or missed meal compensation. *417Supervisors were purportedly afraid to authorize missed meal break compensation because they believed employees would suffer discipline, or that they would themselves get in trouble with their supervisors or the DON.

As for rest breaks, neither Las Encinas nor Charter Oak had a mechanism to enable employees to seek compensation for missed rest breaks. DON Cook and staffing coordinator Casas each confirmed that the Hospital lacked “any method for staff to report missed rest breaks and be paid for them.” Charter Oak’s chief operating officer testified she had “never seen a request for a missed rest break premium.” The TAF’s used by each facility (through which employees request compensation for missed meals or overtime) did not permit employees to request compensation for missed rest breaks.16 The record contains no evidence of any instance in which a putative class member was paid for a missed rest break.

Plaintiffs assert that these common policies and practices were implemented directly by Hospital management and applied to all nursing staff on a classwide basis. As a result of these practices, they argue that employees were systematically denied missed break premiums mandated by state law.

Plaintiffs also submitted a statistical analysis performed by Kriegler (based on his analysis of Hospital timekeeping and payroll data) which they argued demonstrates that the Hospital’s classwide practices violate California law. According to plaintiffs, Kriegler’s analysis shows:

(1) A small fraction of missed meal premiums were paid relative to the total number of missed meal breaks (e.g., although Hospital records show 67 late, short or completely missed meal breaks per 100 shifts, only 3.5 missed meal or rest premiums were paid per 100 shifts);

(2) Almost half (44.6 percent) of the meal breaks recorded were less than 30 minutes long;

(3) Over one-third (34.3 percent) of meal breaks recorded were taken more than six hours after the start of a shift;

(4) Most (over 80 percent) of an employee’s first recorded meal breaks were either less than 30 minutes long or started more than six hours after the start of his or her shift;

(5) The majority (87.4 percent) of all work periods longer than 10 hours did not reflect that a second meal break was taken;

*418(6) Among work periods in excess of 10 hours that did have a second meal break recorded, roughly one-third (33.6 percent) of those breaks were under 30 minutes and three-fourths (73 percent) were taken more than 11 hours after a shift started. Fewer than 1 percent of second meal breaks examined were both 30 minutes long and timely provided;17

(7) There was a widespread practice by which management modified timekeeping records: 24.5 percent of the first and 46 percent of the second recorded meal breaks were “round punch meal breaks” (added or edited by supervisors), and most (84 percent) of all sampled class members’ time records showed “round punch work periods”;

(8) On days on which no break relief was reflected in the Hospital’s daily schedules, timekeeping records still reflected an average of 2.2 meal breaks per day (suggesting that the timekeeping data overreports the number of meal breaks).

For purposes of class certification, plaintiffs’ evidence, including Krieger’s analysis, constitutes persuasive common proof of the Hospital’s uniform policies and practices resulting in classwide denial of lawfully compliant breaks to nursing staff. The statistical evidence reflects a common practice by which management modified timekeeping records, and substantiates declarants’ testimony that the Hospital undertook active efforts to hide its wage and hour violations. Further, the fact that only a few missed meal premiums were actually paid relative to the total number of missed meal breaks, combined with evidence that management actively discouraged employees from seeking missed break compensation, reflects a common practice by the Hospital of failing to compensate employees for hours worked and constitutes common proof of liability.

C. Denied compensation for work performed off the clock

Hospital policy requires overtime be approved in advance, and failure to seek approval for overtime may subject an employee to discipline. Plaintiffs assert that the Hospital actively discouraged nursing staff from requesting overtime by criticizing and threatening to discipline employees who worked too much overtime, criticizing and intimidating employees who requested overtime and repeatedly denying legitimate overtime requests. At the same time, employees — especially RN’s, who were required to complete charts and other mandatory paperwork — -were placed under pressure to ensure that all their work was completed each shift. DON Cook testified that “it was critical *419that nursing staff complete documentation on their patients.” As a result, employees were routinely forced to clock out after their shifts, then return to work to complete paperwork. Plaintiffs argue that Hospital management knew employees routinely worked off the clock after their shifts, and actively condoned and encouraged such conduct. One of plaintiffs’ declarants testified that DON Nocon told her “that working off the clock was the normal practice at Las Encinas because the hospital did not want to pay overtime but expected that all work would be completed before the end of the shift.”18 Toward that end, plaintiffs also presented evidence that the Hospital instructed management to adjust time records and eliminate overtime. As a result of these policies and practices, employees were routinely denied compensation for overtime.

The trial court articulated three reasons supporting its conclusion that the overtime subclass lacked commonality: (1) plaintiffs’ evidence failed to address the possibility that employees worked off the clock by choice; (2) anecdotal evidence submitted by both sides demonstrated variation as to whether employees were denied overtime or forced to work off the clock; and (3) under Brinker, supra, 53 Cal.4th at page 1051, the fact that an employee is clocked out creates a presumption that he or she is not working, and plaintiffs failed to rebut this presumption.

With regard to its first reason, the error in the trial court’s factual and legal analysis is readily apparent. First, this basis is pure speculation in light of the fact that the record contains no evidence that any employee did work — or could have worked — overtime by choice. Hospital policy expressly forbids off the clock work. Second, even if we assume there is evidence some members of the nursing staff voluntarily worked uncompensated overtime, such a “choice” is impermissible under California law. A nonexempt employee (such as the putative class members here) may not lawfully volunteer to work off-the-clock without compensation. (See Cal. Code Regs., tit. 8, § 11090, subd. 3(A) [requiring compensation for “all hours worked,” and overtime pay for “all hours worked” in excess of a 40-hour workweek].)

The court’s second stated basis — evidentiary variations as to the number of employees pressured or forced to work uncompensated overtime— goes to the question of damages, not liability. Such evidence does not undermine the propriety of class certification based on plaintiffs’ theory that *420Hospital management routinely instructed and pressured supervisors to refuse to approve employee overtime and required nursing staff to perform uncompensated off the clock work (both by requiring staff to clock out and in for 30-minute meal breaks not taken, and requiring that staff complete assignments before leaving after a shift). (See Jones v. Farmers Ins. Exchange (2013) 221 Cal.App.4th 986, 996-997 [164 Cal.Rptr.3d 633] [reversing certification denial despite variations in extent to which employees were required to perform off the clock work, because trial court applied improper criteria by focusing on individual damages issues, and ignored plaintiffs’ theory of recovery]; see also Bradley, supra, 211 Cal.App.4th at p. 1155.)

Finally, the trial court’s reliance on Brinker to support its conclusion that plaintiffs failed to rebut the presumption that an employee who has clocked out is not working was misplaced. First, this rationale improperly delves into the merits of the case. At the class certification stage the issue is not whether plaintiffs can definitively rebut the presumption. Plaintiffs need show only that the presumption can be addressed with common evidence. (Brinker, supra, 53 Cal.4th at pp. 1051-1052.) Second, Brinker is inapposite on this point. In Brinker, the court found that the plaintiffs failed to identify any common policies or practices that may have resulted in employees being forced to work off the clock. (Id. at p. 1051.) Here, by contrast, plaintiffs presented evidence of common practices and a de facto policy that resulted in systemic forced off-the-clock work. (See Williams v. Superior Court (2013) 221 Cal.App.4th 1353, 1358, 1369-1370 [165 Cal.Rptr.3d 340] [reversing decertification order where dispute centered on whether application of a uniform policy (including, prior approval required for overtime) was lawful. If unlawful, some class members would be damaged and some might not be, but the employer faced liability as to each.].)

In short, reversal is required with respect to the overtime and off-the-clock compensation claims because the trial court’s order denying class certification rests on improper criteria, erroneous legal assumptions and, in some respects, insubstantial evidence. (Brinker, supra, 53 Cal.4th at p. 1022; Linder, supra, 23 Cal.4th at pp. 435-436.) However, as with the meal and rest break claims, it is unclear from the record below whether common issues predominate over individual ones so as to make a class proceeding superior to the alternatives.

V. Remand is required as to derivative claims subclasses

Plaintiffs’ remaining claims regarding certification of subclasses for waiting time penalties (§ 203) and inaccurate itemized wage statements (§ 226, subd. (e)), which are predicated on the claims for rest and meal breaks and *421overtime, are remanded for further consideration regarding predominance and manageability. (See Dilts v. Penske Logistics, LLC (S.D.Cal. 2010) 267 F.R.D. 625, 640,)

DISPOSITION

The order denying the motion for class certification is reversed and remanded for further consideration in accordance with our holding. The parties are to bear their own costs on appeal.

Concurring. — I concur with some of the majority’s reasoning, and I concur in the judgment. Like the majority, I believe that this case presents questions that may be subject to common proof, and that the case must be remanded to the trial court for determination as to whether those common questions predominate and can be managed efficiently in a single trial. I write separately because I believe the majority has not accurately described the nature of the common questions, and as a consequence, they underestimate the difficulties of managing the trial of this case.

The majority opines that plaintiffs called into question the facial legality of the hospital’s policy regarding meal and rest breaks. But this is true only to a limited extent. The policy as written provided for a meal break “approximately half way between the beginning and ending of the employee’s shift.” It also allowed for two 10-minute rest periods “approximately halfway between starting time and an employee’s meal period, and between the meal period and an employee finishing work for the day.” For employees working what the hospital defined as a “normal work schedule” of “eight hours per workday,” the policy on its face plainly provided for a meal break within the first five hours of work, as required by Labor Code section 512, subdivision (a) and associated regulations.1

Plaintiffs have not based their theory of recovery primarily on the facial illegality of the hospital’s policy. Instead, they have proceeded on the theory that the hospital’s practices amounted to a de facto illegal policy. Plaintiffs have alleged that the hospital systematically understaffed its nursing positions, making it impossible for all employees to take all the breaks to which they were entitled. The question of whether such a policy existed presents a common question, subject to common proof. (See Duran v. U.S. Bank *422National Assn. (2014) 59 Cal.4th 1, 37 [172 Cal.Rptr.3d 371, 325 P.3d 916] (Duran) [stating that certification could be appropriate if an employer “knowingly encouraged a uniform de facto practice inconsistent with” labor law]; Brinker Restaurant Corp. v. Superior Court (2012) 53 Cal.4th 1004, 1051 [139 Cal.Rptr.3d 315, 273 P.3d 513] [indicating that certification could be proper if the plaintiffs could provide “substantial evidence of a systematic company policy to pressure or require employees to work off-the-clock”].) As the majority notes, the Charter Oak Hospital applies a mathematical formula to determine the number of staff members needed, depending on the number of patients and severity of their conditions. (Maj. opn., ante, at pp. 401-402.) If the parties in this case introduce evidence establishing the number of nurses required according to the formula and the number of staff actually assigned to work at different times, they will go a long way toward determining whether plaintiffs’ theory, that the hospital was systematically understaffed, is correct.

Yet as the majority acknowledges, the existence of a common question or theory of recovery is not the entire story. For certification to be appropriate, the issues subject to common proof must predominate over individual issues. In addition, the court must be able to manage the litigation of individual issues within a trial. Although a lack of uniformity among class members is not a bar to certification (Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 333 [17 Cal.Rptr.3d 906, 96 P.3d 194]), devising a method for managing variation among class members is not a given. Duran, supra, 59 Cal.4th 1, provided an object lesson in manageability. The case involved a class of 260 employees who worked in sales positions for the defendant bank. The plaintiffs alleged that the employees were not exempt from minimum wage and overtime pay requirements, and the determination of this depended on whether the job reasonably required employees to spend the majority of their days selling outside the bank. The trial court permitted the parties to present evidence from only 20 plaintiffs. (Id. at pp. 16-17.) The court then extrapolated from those 20 to determine liability for all 260. (Id. at p. 24.) The Supreme Court rejected this method in part because the sample was likely skewed, but also because the trial court improperly prevented the defendants from introducing any evidence regarding the class members outside the sample of 20. (Id. at pp. 33-34.) As the Supreme Court noted, by “excluding relevant evidence central to the defense, the court here did not manage individual issues. It ignored them.” (Id. at p. 34.) This was error because “a class action trial management plan may not foreclose the litigation of relevant affirmative defenses, even when these defenses turn on individual questions.” (Ibid.)

This case is even more complex than Duran and may present even greater variability among putative class members. The class is four times as numerous, spread over different positions, with employees who may have had vastly differing experiences because they worked at different times, with *423different lengths of experience and different supervisors. The evidence indicated that different employees had widely differing experiences on the job; some claimed they were frequently denied breaks, while others said this almost never happened. It is not obvious that the level of variability within the class is sufficiently low to allow for sampling (see Duran, supra, 59 Cal.4th at p. 33), nor that the hospital would be capable of defending itself by presenting evidence regarding the work experiences of all plaintiffs as a group. (See id. at pp. 34-35.)

Crucially, the individual evidence in this case would be relevant to defendants’ liability, not merely the extent of damages. As the Duran court noted, “Defenses that raise individual questions about the calculation of damages generally do not defeat certification. [Citation.] However, a defense in which liability itself is predicated on factual questions specific to individual claimants poses a much greater challenge to manageability. This distinction is important. As we observed in City of San Jose v. Superior Court [(1974)] 12 Cal.3d [447,] 463 [115 Cal.Rptr. 797, 525 P.2d 701]: ‘Only in an extraordinary situation would a class action be justified where, subsequent to the class judgment, the members would be required to individually prove not only damages but also liability.’ ” (Duran, supra, 59 Cal.4th at p. 30.) In a case like this one, where plaintiffs’ primary allegations are that the hospital had de facto illegal policies, as opposed to policies illegal on their face, individual evidence is likely to be central both to plaintiffs’ case and to the hospital’s defense. Indeed, the evidence introduced thus far includes not only statistical evidence that may be used to establish or challenge plaintiffs’ common theory of recovery, but also anecdotal reports from individual class members. At trial, the hospital will most likely defend itself by eliciting testimony that employees missed breaks only rarely, or delayed them voluntarily. The majority claims that “[i]f an employer fails to provide legally compliant meal or rest breaks, the court may not conclude employees voluntarily chose to skip those breaks.” (Maj. opn. ante, at p. 410.) But in this case, whether employees chose to skip breaks voluntarily is important to determining whether an illegal policy existed. The individual differences cannot simply be swept under the rug on the ground that they pertain only to damages. This is not a case where liability may be determined solely through common evidence, and the only reason to deal with individual issues is to ask, with respect to each class member, “whether he or she worked for defendant during the relevant period” in a position affected by the alleged illegal practice. (Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at p. 332.) Plaintiffs’ own theory makes the experience of each individual necessary to determining liability.

The trial court is in a far better position than we are to determine whether these obstacles can be overcome, and it is for just this reason that trial courts *424“are afforded great discretion in granting or denying certification.” (Linder v. Thrifty Oil Co. (2000) 23 Cal.4th 429, 435 [97 Cal.Rptr.2d 179, 2 P.3d 27].)