Aaron SENNE; Michael Liberto; Oliver Odle; Brad McAtee; Craig Bennigson; Matt Lawson; Kyle Woodruff; Ryan Kiel; Kyle Nicholson; Brad Stone ; Matt Daly; Aaron Meade; Justin Murray; Jake Kahaulelio; Ryan Khoury; Dustin Pease; Jeff Nadeau ; Jon Gaston; Brandon Henderson ; Tim Pahuta; Lee Smith; Joseph Newby; Ryan Hutson ; Matt Frevert; Roberto Ortiz ; Witer Jimenez; Kris Watts; Mitch Hilligoss; Daniel Britt ; Yadel Marti; Helder Velaquez; Jorge Jimenez ; Jorge Minyety; Edwin Maysonet; Jose Diaz; Nick Giarraputo; Lauren Gagnier; Leonard Davis; Gaspar Santiago; Grant Duff ; Omar Aguilar; Mark Wagner; David Quinowski; Brandon Pinckney, Individually and on Behalf of All Those Similarly Situated; Jake Opitz; Brett Newsome, Plaintiffs-Appellants, v. KANSAS CITY ROYALS BASEBALL CORP.; Marlins Teamco LLC; San Francisco Baseball Associates, LLC ; Office of the Commissioner of Baseball, dba Major League Baseball, an unincorporated association; Allan Huber Selig, "Bud"; Angels Baseball LP; St. Louis Cardinals, LLC; Colorado Rockies Baseball Club, Ltd.; Cincinnati Reds, LLC ; Houston Baseball Partners LLC; Athletics Investment Group, LLC; Rogers Blue Jays Baseball Partnership ; Padres L.P.; San Diego Padres Baseball Club, L.P. ; Minnesota Twins, LLC ; Detroit Tigers, Inc.; Los Angeles Dodgers LLC; Sterling Mets L.P.; AZPB L.P.; New York Yankees P'Ship; Rangers Baseball Express, LLC ; Milwaukee Brewers Baseball Club, Inc.; Chicago Cubs Baseball Club, LLC ; Pittsburgh Associates, LP; Baseball Club of Seattle, LLP; Los Angeles Dodgers Holding Company LLC; Rangers Baseball, LLC, Defendants-Appellees., 934 F.3d 918

PAEZ, Circuit Judge: It is often said that baseball is America's pastime. In this case, current and former minor league baseball players allege that the American tradition of baseball collides with a tradition far less benign: the exploitation of workers. We are tasked with deciding whether these minor league players may properly bring their wage-and-hour claims on a collective and classwide basis. BACKGROUND I.

Most major professional sports in America have their own "farm system" for developing talent: for the National Basketball Association, it's the G-League; for the National Hockey League, it's the American Hockey League; and for Major League Baseball (MLB), it's Minor League Baseball. MLB and its thirty franchise teams rely heavily on this extensive minor league system, which has nearly 200 affiliates across the country and employs approximately 6,000 minor league players. Nearly all MLB players begin their careers in the minor leagues.

Each minor league club is associated with one of the thirty franchise MLB teams. The minor league system is governed by the Major League Rules (MLRs), which dictate the terms of employment and compensation for both minor and major league players. Under the MLRs, all minor league players are required to sign a seven-year Uniform Player Contract (UPC).

Ostensibly, players are required to sign the UPC for "morale" and "to produce the similarity of conditions necessary for keen competition." The UPC "obligates Player[s] to perform professional services on a calendar year basis, regardless of the fact that salary payments are to be made only during the actual championship playing season." It describes its scope as setting "the terms and conditions of employment during all periods in which Player is employed by Club as a Minor League Player." Players are paid by the MLB franchise affiliated with the minor league team for which they play.

Under the UPC, first-year players are paid a fixed salary of $1,100 per month during the regular ("championship") season that runs from April through September. In addition to their salaries during the championship season, some players receive signing or performance-related bonuses and college scholarships. Beginning in early March each year, the minor league affiliates conduct spring training in Arizona and Florida; every MLB franchise operates a minor league training complex in one of these two states. The parties dispute whether spring training is required, but the UPC strongly indicates that it is mandatory.

Virtually all players are unpaid during spring training. Spring training lasts approximately four weeks, until the championship season begins in April. Some players attest that spring training entails working seven days a week, with no days off. During spring training, teams typically have scheduled activities in the morning prior to playing games in the afternoon. For example, a team spring training

schedule for one of the San Francisco Giants' affiliates describes that at 6:30 AM, there was an "Early Van for Treatment and Early Work" ; at 7:00 AM, the "Regular Van" departed; at 7:45 AM, the "Early Work" began; and then between 9:00 AM and 11:00 AM, the team would perform activities such as "Stretch," "Throwing Program," and "Batting Practice." Lunch was to be at 11:00 AM, before a 12:10 PM bus to a neighboring city for a 1:00 PM away game. At the conclusion of spring training in early April, some players are assigned to minor league affiliates, and begin playing games in the championship season.

During the championship season, minor league teams play games either six or seven days per week. The championship season lasts around five months, beginning in April and ending in September. One of the regular season leagues within minor league baseball is the California League, which-as the name implies-plays games exclusively within California. Players who are not assigned to play for affiliates in the championship season stay at the Arizona or Florida facilities for "extended spring training." Extended spring training continues until June, and involves similar activities to spring training.

Although most players do not get paid during extended spring training, as many as seven MLB clubs do pay for work during extended spring training due to an ambiguity in the MLRs over when players are permitted to be paid. After the championship season ends in September, some players participate in the "instructional leagues," which run from approximately mid-September to mid-October. The parties dispute whether participation in the instructional leagues is mandatory for the players involved, although as with spring training, the UPC strongly implies that participation is required.

Activities and schedules during the instructional league are similar to spring training. And just as with spring training, players are virtually never paid for participation in the instructional league. II. Plaintiffs are forty-five current and former minor league baseball players who bring claims under the federal Fair Labor Standards Act (FLSA) and the wage-and-hour laws of California, Arizona, and Florida against MLB, MLB Commissioner Bud Selig, and a number of MLB franchises.

Plaintiffs allege that defendants do not pay the players at all during spring training, extended spring training, or the instructional leagues. They further allege that because players are "employees" and the activities the players perform during those periods constitute compensable work, defendants have unlawfully failed to pay them at least minimum wage. And according to plaintiffs, while the players are paid-albeit not much-during the championship season, they routinely work overtime, for which they are never compensated as a matter of policy.

In May 2015, plaintiffs filed their Second Amended Consolidated Class Action Complaint, which alleged wage-and-hour claims under the laws of eight states and the FLSA; plaintiffs also sought certification of a FLSA collective action. The district court preliminarily certified the FLSA collective in October 2015. Notice was sent to approximately 15,000 current and former minor league players, of which more than 2,200 opted in. In 2016, defendants moved to decertify the FLSA collective, while plaintiffs moved to certify a Rule 23(b)(2) class as well as Rule 23(b)(3) classes under the laws of eight states.

The district court denied certification for all proposed Rule 23(b)(3) classes, concluding that predominance was not satisfied for two primary reasons. Senne v. Kansas City Royals Baseball Corp. , 315 F.R.D. 523, 572, 577-84 (N.D. Cal. 2016). First, the court concluded that predominance was defeated by the choice-of-law issues presented by the proposed classes, given that (1) the winter off-season training claims entailed work performed in dozens of different states with no common

schedule or situs; and (2) the championship season claims involved frequent travel between state lines for away games. Id. at 580-81. The district court also determined that the inclusion of claims for winter off-season work fatally undermined predominance, as the court would be required to undertake an overwhelming number of individualized inquiries to determine which activities constituted compensable "work" and how much time was spent doing "work." Id. at 577-84. For similar reasons, the court held that plaintiffs were not "similarly situated" and therefore decertified the FLSA collective. Id. at 585-86.

The court also granted the defendants' motion to exclude an expert survey (the "Pilot Survey") submitted by plaintiffs, finding that its methodology and results did not satisfy the requirements of Federal Rule of Evidence 702 and Daubert v. Merrell Dow Pharmaceuticals, Inc. , 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). Id. at 586-90, 113 S.Ct. 2786. The court further refused to certify the proposed Rule 23(b)(2) class, concluding that because the plaintiffs were all former-rather than current-players, they lacked standing to represent a (b)(2) class. Id. at 584-85, 113 S.Ct. 2786.

Plaintiffs moved for reconsideration, narrowing their proposed classes significantly in response to the concerns the district court expressed in its initial certification order. Plaintiffs requested Rule 23(b)(3) certification of an Arizona class and a Florida class for work performed during spring training, extended spring training, and the instructional leagues in those states. Plaintiffs also moved for certification of a 23(b)

(3) California class, covering players who participated in the California League during the championship season. Additionally, plaintiffs sought to certify a reworked FLSA collective of players who participated in the California League or in spring training, extended spring training, and the instructional leagues. In addition to the 23(b)(3) classes and FLSA collective, plaintiffs requested certification of a Rule 23(b)(2) injunctive relief class consisting of current minor league players who participate in spring training, extended spring training, or the instructional leagues in Florida or Arizona.

To cure the court's earlier concerns about standing, four current minor league players moved to intervene to represent the proposed (b)(2) class. On reconsideration, plaintiffs argued that they could meet Rule 23(b)(3) 's predominance requirement and FLSA's "similarly situated" requirement through a combination of the use of representative evidence and application of the so-called "continuous workday" rule. Plaintiffs' representative evidence took a variety of forms, including an expert survey (the "Main Survey"), hundreds of team schedules, payroll data, and testimony from both players and league officials.

The most controversial piece of evidence was the Main Survey, which plaintiffs argued served as representative evidence of hours worked, particularly when used in concert with a continuous workday theory. The Main Survey asked players to report the times they "most often" arrived and departed from the ballpark or training facility during the championship season, spring training, extended spring training, and the instructional leagues, and asked players to estimate how much time they spent eating meals while at the ballpark.

The survey did not, however, ask players about the kinds of activities they performed at the facilities, or how much time they spent performing particular activities. Given these purported shortcomings, defendants moved to exclude the Main Survey, and further argued that even if the survey were admissible under Daubert , it still could not be used to meet the predominance and "similarly situated" requirements due to its alleged flaws.

The district court denied defendants' motion to exclude the Main Survey, finding it admissible under Daubert and concluding that defendants' challenges went "to the weight of the Survey and not its admissibility" and were "better left to a jury to evaluate." The district court further concluded that the Main Survey could be used in combination with other evidence-such as team schedules, testimony, and payroll data-to meet Rule 23(b)(3) 's predominance and FLSA's "similarly situated" requirements, observing that certifying the classes and the FLSA collective "will not preclude Defendants from challenging the sufficiency of the Main Survey and Plaintiffs' damages model on

summary judgment and/or at trial." Because it concluded that the predominance and "similarly situated" requirements could be met with the use of representative evidence and application of the continuous workday rule, the district court recertified the narrowed FLSA collective and certified a California (b)(3) class. However, the district court denied certification for the Arizona, Florida, and (b)(2) classes, holding that choice-of-law concerns defeated predominance for the Arizona and Florida classes and undermined "cohesiveness" for the (b)(2) class.

At defendants' request, the district court certified the FLSA collective certification order for interlocutory review under 28 U.S.C. § 1292. Plaintiffs petitioned us for permission to appeal the denial of certification for the Arizona, Florida, and Rule 23(b)(2) classes, and defendants likewise petitioned to appeal the certification of the California class; we granted both petitions, consolidating those cross-appeals with the FLSA collective appeal.

STANDARD OF REVIEW We review for abuse of discretion the district court's class certification rulings, and review for clear error any findings of fact the district court relied upon in its certification order. Parsons v. Ryan , 754 F.3d 657, 673 (9th Cir. 2014). A district court's choice of law determinations, however, are reviewed de novo. Mazza v. Am. Honda Motor Co. , 666 F.3d 581, 589 (9th Cir. 2012). A district court abuses its discretion where it commits an error of law, relies on an improper factor, omits a substantial factor, or engages in a clear error of judgment in weighing the correct mix of factors.

Stockwell v. City & County of San Francisco , 749 F.3d 1107, 1113 (9th Cir. 2014) (citing Bateman v. Am. Multi-Cinema, Inc. , 623 F.3d 708, 712 (9th Cir. 2010) ). When we review a grant of class certification, "we accord the district court noticeably more deference than when we review a denial." Abdullah v. U.S. Sec. Assocs., Inc. , 731 F.3d 952, 956 (9th Cir. 2013) (quoting Wolin v. Jaguar Land Rover N. Am., LLC , 617 F.3d 1168, 1171 (9th Cir. 2010) ).

ANALYSIS To paraphrase the Chief Justice, these complex appeals require us to call a great number of balls and strikes, as both parties raise numerous challenges to the district court's certification order. For their part, plaintiffs challenge the district court's decision to deny certification for the Arizona and Florida Rule 23(b)(3) classes and the Rule 23(b)(2) class on the grounds that choice-of-law issues defeated the predominance requirement for the Arizona and Florida (b)(3) classes and also thwarted "cohesiveness" for the proposed (b)(2) class.

Defendants, on the other hand, contest the district court's certification of the California (b)(3) class, arguing first that choice-of-law issues defeat both predominance and adequacy, and second, that plaintiffs cannot meet the predominance requirement through the use of their proffered representative evidence: the Main Survey, team schedules, payroll records, deposition testimony, and declarations.

Defendants further charge that the district court erred in certifying the FLSA collective because plaintiffs' representative evidence does not show that the collective members are "similarly situated." Defendants also contend that the district court erred by not "rigorously analyzing" plaintiffs' expert evidence at the class and collective certification stage. We address each argument in turn. I. Class certification is governed by Federal Rule of Civil Procedure 23.

As a threshold matter, a party seeking class certification must satisfy the four requirements of Rule 23(a) : (1) numerosity; (2) commonality; (3) typicality; and (4) adequacy of representation. "Class certification is proper only if the trial court has concluded, after a 'rigorous analysis,' that Rule 23(

a) has been satisfied." Parsons , 754 F.3d at 674 (quoting Wang v. Chinese Daily News, Inc. , 737 F.3d 538, 542-43 (9th Cir. 2013) ). In addition to the requirements of Rule 23(a), a proposed class must also meet the requirements of one or more of the "three different types of classes" set forth in Rule 23(b). Leyva v. Medline Industries, Inc., 716 F.3d 510, 512 (9th Cir. 2013). Here, plaintiffs proposed classes under two of Rule 23(b) 's class types: Rule 23(b)(3) and 23(b)(2).

A class may be certified under Rule 23(b)(3) only if the district court "finds that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy." Fed. R. Civ. P. 23(b)(3). Of these two requirements-predominance and superiority-only predominance is at issue on appeal.

"The predominance inquiry focuses on 'the relationship between the common and individual issues' and 'tests whether proposed classes are sufficiently cohesive to warrant adjudication by representation.' " Vinole v. Countrywide Home Loans, Inc. , 571 F.3d 935, 944 (9th Cir. 2009) (quoting Hanlon v. Chrysler Corp. , 150 F.3d 1011, 1022 (9th Cir. 1998) ). In determining whether the predominance requirement is met, courts have a "duty to take a close look at whether common questions predominate over individual ones" to ensure that individual questions do not "overwhelm questions common to the class." Comcast Corp. v.

Behrend , 569 U.S. 27, 34, 133 S.Ct. 1426, 185 L.Ed.2d 515 (2013) (internal quotation marks and citation omitted). Rule 23(b)(2), on the other hand, requires only that "the party opposing the class ha[ve] acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." Fed. R. Civ. P. 23(b)(2). Although 23(b)(2) classes are most common in the civil rights context, "we have certified many different kinds of Rule 23(b)(2) classes." Parsons , 754 F.3d at 686. II.

We first address whether choice-of-law issues fatally undermine plaintiffs' proposed Rule 23 classes. The district court's decision was split on the impact of choice-of-law questions: as to the proposed Rule 23(b)

(3) California class, the court held that choice-of-law concerns defeated neither Rule 23(b)(3) 's predominance requirement nor Rule 23(a) 's adequacy requirement. Yet as to the proposed 23(b)

(3) Arizona and Florida classes, the district court held the opposite: that choice-of-law issues posed an insurmountable hurdle to meeting both predominance and adequacy. Similarly, the court determined that choice-of-law questions made certification of the proposed Rule 23(b)(2) class inappropriate. Concerns over which state's laws apply to a proposed class "do not necessarily preclude a 23(b)(3) action." Hanlon , 150 F.3d at 1022.

But "[u]nderstanding which law will apply before making a predominance determination is important when there are variations in applicable state law," and potentially varying state laws may defeat predominance in certain circumstances. Zinser v. Accufix Research Inst., Inc. , 253 F.3d 1180, 1189 (9th Cir. 2001) opinion amended on denial of reh'g , 273 F.3d 1266 (9th Cir. 2001). We have been particularly concerned about the impact of choice-of-law inquiries in nationwide consumer class actions and products liability cases. See, e.g. , Mazza , 666 F.3d at 585, 591-94 ; Zinser , 253 F.3d at 1184-90.

A district court considering state law claims brought in federal court must utilize the choice-of-law rules of the forum state-here, California. Klaxon Co. v. Stentor Electric Mfg. Co. , 313 U.S. 487, 496-97, 61 S.Ct. 1020, 85 L.Ed. 1477 (1941). "By default, California courts apply California law unless a party litigant timely invokes the law of a foreign state, in which case it is the foreign law proponent who must shoulder the burden of demonstrating that foreign law, rather than California law, should apply to class claims." In re Hyundai & Kia Fuel Econ.

Litig. , 926 F.3d 539, 561 (9th Cir. 2019) (en banc) (internal quotation marks and citations omitted). To meet their burden, the objectors must satisfy California's three-step governmental interest test, used to resolve choice of law issues. Id. First, the court determines whether the relevant law of each of the potentially affected jurisdictions with regard to the particular issue in question is the same or different.

Second, if there is a difference, the court examines each jurisdiction's interest in the application of its own law under the circumstances of the particular case to determine whether a true conflict exists.

Third, if the court finds that there is a true conflict, it carefully evaluates and compares the nature and strength of the interest of each jurisdiction in the application of its own law to determine which state's interest would be more impaired if its policy were subordinated to the policy of the other state, and then ultimately applies the law of the state whose interest would be the more impaired if its law were not applied. Kearney v. Salomon Smith Barney, Inc. , 39 Cal.4th 95, 45 Cal.Rptr.3d 730, 137 P.3d 914, 922 (2006) (internal quotation marks and citations omitted); see also Hairu Chen v.

Los Angeles Truck Centers, LLC , 7 Cal.5th 862, 866, 249 Cal.Rptr.3d 594, 444 P.3d 727 (2019). In making its choice-of-law determinations, the district court relied heavily on the California Supreme Court's decision in Sullivan v. Oracle Corp. , 51 Cal.4th 1191, 127 Cal.Rptr.3d 185, 254 P.3d 237 (2011), and the parties do not dispute that Sullivan provides the most helpful guidance for the choice-of-law questions before us.

In Sullivan , the California Supreme Court answered a certified question from this court regarding whether California overtime law applied to non-resident employees of a California corporation who worked primarily in their home states of Colorado and Arizona, but also worked in California (and several other states) for "entire days or weeks" at a time. Id., 127 Cal.Rptr.3d 185, 254 P.3d at 239, 243.

Sullivan first concluded that as a matter of statutory construction, California law applied to all work performed for days or weeks at a time within the state's borders, regardless of whether it was performed by residents or non-residents. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 241-43. Next, Sullivan undertook California's three-step governmental interest analysis for choice-of-law questions. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 244-47.

At the first step of the analysis-whether the relevant laws differed-the court noted that California's overtime law "clearly" differed from the laws of the plaintiffs' home states. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 245. At the second step-whether a "true" conflict existed-the court held that the existence of a true conflict was "doubtful, at best." Id.

The court explained that the second step involves examining "each jurisdiction's interest in the application of its own law under the circumstances of the particular case," noting that a court "may make [its] own determination of the relevant policies and interests, without taking 'evidence' as such on the matter." Id. (internal quotation marks, alterations, and citations omitted). Sullivan observed that "California has, and has unambiguously asserted, a strong interest in applying its overtime law to all nonexempt workers, and all work performed, within its borders." Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 245.

The court concluded that "neither Colorado nor Arizona has a legitimate interest in shielding Oracle from the requirements of California wage law as to work performed here." Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 246. In so holding, the court rejected two specific arguments advanced by Oracle. First, Oracle contended that because Arizona and Colorado have workers' compensation statutes with express extraterritorial application, those statutes indicate an interest in extending the protection of their employment laws to their residents working outside the state. Id. Not so, Sullivan held.

While "a state has such an interest, at least in the abstract, when the traveling, resident employee of a domestic employer would otherwise be left without the protection of another state's law," the states had "expressed no interest in disabling their residents from receiving the full protection of California overtime law when working here, or in requiring their residents to work side-by-side with California residents in California for lower pay." Id.

Second, Oracle argued that Arizona and Colorado "have an interest in providing hospitable regulatory environments for their own businesses" and thus "also have an interest in shielding their own businesses from more costly and burdensome regulatory environments in other states." Id. Relying on principles of federalism, Sullivan dismissed this argument.

While "a state can properly choose to create a business-friendly environment within its own boundaries," the federal Constitution does not require a state to substitute " 'the conflicting statute of another state' " for its own laws that are " 'applicable to persons and events' " within that state. Id. (quoting Phillips Petrol. Co. v. Shutts , 472 U.S. 797, 822, 105 S.Ct. 2965, 86 L.Ed.2d 628 (1985) ). Nor does the Constitution "permit one state to project its regulatory regime into the jurisdiction of another state." Id. (citing Healy v.

Beer Insti., Inc. , 491 U.S. 324, 336-37, 109 S.Ct. 2491, 105 L.Ed.2d 275 (1989) ). Finally, although Sullivan held that there was almost certainly no true conflict because neither Arizona nor Colorado had a "legitimate interest" in blocking the application of California law to the work performed in California, the court nonetheless proceeded to the third step of the analysis "for the sake of argument." Id., 127 Cal.Rptr.3d 185, 254 P.3d at 246-47.

Sullivan concluded that the analysis at the third step-determining which state's interest would be more impaired if its policy were subordinated to the policy of the other state-yielded a straightforward answer: [T]o subordinate California's interests to those of Colorado and Arizona unquestionably would bring about the greater impairment. To permit nonresidents to work in California without the protection of our overtime law would completely sacrifice, as to those employees, the state's important public policy goals of protecting health and safety and preventing the evils associated with overwork.

Not to apply California law would also encourage employers to substitute lower paid temporary employees from other states for California employees, thus threatening California's legitimate interest in expanding the job market. By way of comparison, not to apply the overtime laws of Colorado and Arizona would impact those states' interests negligibly, or not at all ...

Alternatively, viewing Colorado's and Arizona's overtime regimens as expressions of a general interest in providing hospitable regulatory environments to businesses within their own boundaries, that interest is not perceptibly impaired by requiring a California employer to comply with California overtime law for work performed here. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 247 (citations omitted) (emphasis added). A. We first conclude that the district court did not err in holding that under Sullivan , California law should apply to the (b)

(3) California class. Although defendants correctly point out that Sullivan is not precisely analogous to the case at hand, the two principal differences on which defendants rely are unpersuasive. Specifically, defendants first rely on the fact that while Sullivan involved a California corporation, "most of the MLB Club Defendants with affiliates in the California League are located outside California." But a close reading of Sullivan indicates that California law should apply to the California class, even though many of the employers are not headquartered in California.

For example, Sullivan expressly contemplated that California's overtime laws may not apply to nonresident employees of an out-of-state business who enter California only "temporarily during the course of a workday," but contrasted such a scenario with employees who work in California for "entire days and weeks," who are covered by California law. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 243 (emphases, internal quotation marks, and citations omitted).

Similarly, Sullivan specifically left open the possibility that other California employment laws, such as pay stub requirements, may not apply to non-resident employees of out-of-state employers-with the clear implication that overtime laws would apply to such employees. See id. , 127 Cal.Rptr.3d 185, 254 P.3d at 243-44. Likewise, far from limiting its holding only to nonresident employees of in-state employers, Sullivan merely emphasized that employees of in-state employers would especially be covered by California law. See id. , 127 Cal.Rptr.3d 185, 254 P.3d at 243.

Second, defendants characterize Sullivan as resting on the court's determination that "neither Arizona nor Colorado ... has asserted an interest in regulating overtime work performed in other states." Defendants argue that here, by contrast, "numerous" states have a competing interest in regulating work performed in California.

But defendants misread Sullivan by erroneously presuming that its conclusion at the third step-that subordinating "California's interests to those of Colorado and Arizona unquestionably would bring about the greater impairment"-hinged entirely on whether Arizona or Colorado law had asserted an interest in extraterritorial application of their wage laws. Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 247. It is certainly accurate to say that Sullivan 's holding was influenced by the fact that neither Arizona nor Colorado law purported to apply extraterritorially.

Yet the court's discussion at step three cannot fairly be read to support the argument that California's "strong interest in applying its overtime law to ... all work performed within its borders," id. , 127 Cal.Rptr.3d 185, 254 P.3d at 245, would suddenly become the lesser -impaired interest in the event another state expressed a clear interest in applying its wage laws to work performed in California.

Rather, Sullivan strongly indicates that California's interest in applying its laws to work performed within its borders for days or weeks at a time would reign supreme regardless of whether another state expressed an interest in applying its own wage laws instead of California's. Although we read Sullivan as clearly mandating the application of California law to the California class, two additional considerations support our conclusion today.

First, because the district court found that plaintiffs had met their burden of showing that California law could constitutionally be applied-a determination defendants do not contest on appeal-the burden shifted to defendants "to demonstrate 'that foreign law, rather than California law, should apply to class claims.' " Mazza , 666 F.3d at 590 (quoting Wash. Mut. Bank v. Superior Court , 24 Cal.4th 906, 103 Cal.Rptr.2d 320, 15 P.3d 1071, 1081 (2001) ).

The district court held that defendants failed to meet this burden, because they had "not gone beyond speculating in a general manner that the claims of some members of the putative California Class might be subject to the law of another state and that the interests of another state might be more impaired by application of California law." Defendants specifically point to one of the named plaintiffs-Mitch Hilligoss-as an example of the alleged "need to conduct choice of law inquiries as to every member of the California class." The district court found this example unpersuasive for several reasons, and we agree.

The defendants argued that Illinois law should apply to Hilligoss' work in California because the time he spent in California was a small proportion of his overall career (around two months out of a six-year career). The district court, however, correctly read Sullivan as indicating that California law should nonetheless apply to Hilligoss' California work.

Indeed, the proportion of time the nonresident employees in Sullivan worked in California was quite small (and in one case, even less than the proportion of Hilligoss' career spent in California): during the relevant three-year period, one worked 20 days, another 74 days, and the third 110 days. 127 Cal.Rptr.3d 185, 254 P.3d at 239. Put differently, the employees in Sullivan worked in California approximately 1.8%, 6.7%, and 10% of the time, respectively. Id . What mattered in Sullivan -and what matters here-is that when the employees worked in California, they did so for "entire days or weeks" at a time.

Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 243. Second, practical considerations strongly support applying California law to work performed in California, at least as a general rule; to hold otherwise "would lead to bizarre and untenable results." See Brief for Professors Peter Hay and Patrick J. Borchers, Dkt. No. 21, at 12-13 (hereinafter "Professors' Amicus Brief"). If the law of the state in which work is performed is not the law that generally applies, employers and employees alike would be subjected to an unworkable scheme.

Employers would be required to properly ascertain the residency status-itself not necessarily an easy task, as any student or seasonal worker could attest-of each of its employees. For every nonresident employee, employers would then have to determine whether the wage laws of that employee's state of residence apply extraterritorially, and then come up with different rules for each of its employees according to their state of residence and any extraterritorial application of their home state's laws.

This would mean that at a single worksite, employees working side-by-side in the same position would not only be owed vastly different minimum wages, but also that an employer would need to set different rules for meal and rest breaks for different employees, and so on and so forth. It cannot be in any state's legitimate "interest" to foist such an administrative nightmare upon both employers and employees. Such a scenario would also result in an enormous competitive advantage-or disadvantage-for prospective employees based solely on their state of residency.

Employers would be incentivized to hire residents of states with low minimum wages and otherwise employer-friendly wage laws, while residents of states with higher minimum wages and more protective employment laws would suddenly be far less appealing. Amici Professors Hays and Borchers persuasively point out that as defendants would have it, a college student still domiciled in Seattle while attending a Nebraska university would have to be paid $15 per hour at a part-time job in Nebraska, "nearly double Nebraska's minimum wage of $8 per hour." Professors' Amicus Brief at 13.

This, of course, would put the student at a crushing disadvantage; what rational employer would hire her? Moreover, given the administrative cost involved in attempting to comply with a patchwork of multiple states' wage laws at a single workplace, some employers might instead choose to stick to hiring only resident employees, or perhaps only non-resident employees from a particular state (presumably one with a low minimum wage and minimally protective employment laws).

We do not foreclose the possibility that there could be some circumstances in which a proper application of California's choice-of-law rules might lead to the application of another state's wage and hour laws to work performed in California. Nor do we create a per se rule or an unrebuttable presumption. We hold only that, given the above considerations, we are more than satisfied that the district court did not err in concluding that under Sullivan , California law applies to the California class. B.

We next address whether the district court erred in determining that choice-of-law considerations defeated predominance and adequacy for the proposed Arizona and Florida Rule 23(b)(3) classes, and conclude that the district court's determination must be reversed. Our conclusion is animated in part by several of the considerations outlined above, which apply with equal force to the Arizona and Florida classes. Moreover, the aforementioned enormous practical implications of a contrary holding would be just as problematic and unworkable in Arizona and Florida as in California. 1.

With those considerations in mind, we apply California's three-step governmental interest analysis, and conclude that Arizona law should apply to the work performed in Arizona, and Florida law to the work performed in Florida. At the first step, we agree with defendants that the differences in state law are "material," meaning that "they make a difference in this litigation." Mazza , 666 F.3d at 590. For example, some states have more expansive

definitions of "work," others have differing available defenses, and we have previously held that the elements for a quantum meruit claim-alleged in both the Arizona and Florida classes-"vary materially from state to state." Id. at 591 (citing Candace S. Kovacic, A Proposal to Simplify Quantum Meruit Litigation , 35 Am. U.L. Rev. 547, 558-60 (1986) ). 2.

"Because the relevant laws differ," we must "next examine each jurisdiction's interest in the application of its own law under the circumstances of the particular case to determine whether a true conflict exists." Sullivan , 127 Cal.Rptr.3d 185, 254 P.3d at 245 (alteration, internal quotation marks, and citation omitted). We are not persuaded, as defendants contend, that a "true" conflict exists.

First, under California's choice-of-law principles, "a jurisdiction ordinarily has the predominant interest in regulating conduct that occurs within its borders." Mazza , 666 F.3d at 592 (internal quotation marks omitted) (quoting McCann v. Foster Wheeler LLC , 48 Cal.4th 68, 105 Cal.Rptr.3d 378, 225 P.3d 516, 534 (2010) ). The dissent contends that "California has long rejected" this approach. Dissent at 953. In noting these principles, we do not ignore the evolution of California's choice of law doctrine.

We recognize that the California Supreme Court "renounced the prior rule, adhered to by courts for many years, that in tort actions the law of the place of the wrong was the applicable law in a California forum regardless of the issues before the court" when it adopted the governmental interest approach. Hurtado v. Superior Court , 11 Cal.3d 574, 114 Cal.Rptr. 106, 522 P.2d 666 (1974).

Yet the California Supreme Court has acknowledged that while it "no longer follows the old choice-of-law rule that generally called for application of the law of the jurisdiction in which a defendant's allegedly tortious conduct occurred without regard to the nature of the issue that was before the court ... California choice-of-law cases nonetheless continue to recognize that a jurisdiction ordinarily has the predominant interest in regulating conduct that occurs within its borders." McCann, 105 Cal.Rptr.3d 378, 225 P.3d at 534 (internal quotation marks and citation omitted) (emphasis in original).

Thus, when conducting the governmental interest analysis, we must also recognize that a state ordinarily has the predominant interest in regulating conduct within its borders. We draw this conclusion not from California's interest in regulating conduct within its own borders, but from California's choice-of-law principles. Thus these principles are not limited to the California class but apply to the Florida and Arizona classes as well.

See Mazza , 666 F.3d at 593 ("The district court did not adequately recognize that each foreign state [not just California] has an interest in applying its law to transactions within its borders."). The district court erred in ignoring these principles as a starting point, instead faulting plaintiffs for not addressing "in detail the interests of either Arizona or Florida in applying their law" and focusing on the absence of Florida or Arizona cases akin to Sullivan -despite the strong indications that Arizona and Florida have the "predominant interest" in applying their laws to work performed within their state.

See Mazza , 666 F.3d at 592. Second, Sullivan relied on several different considerations to arrive at its conclusion that the existence of a true conflict was "doubtful, at best": (1) the states in which the employees resided did not express an intent to apply their laws extraterritorially; (2) the employees' states of residence did not have a "legitimate interest" in shielding an employer from California's wage laws as to work performed in California; and (3) federalism and due process made extraterritorial reach doubtful under the circumstances. See 127 Cal.Rptr.3d 185, 254 P.3d at 245-47.

Although defendants vigorously argue that the first of those rationales is inapplicable here-as discussed in greater detail below-at a minimum, the second and third rationales do apply, and weigh against the existence of a true conflict. As to the first rationale, both defendants and the dissent contend that several states have expressed an interest in applying their wage and hour laws to work performed outside the state.

In support of their position, they cite to a handful of cases where courts (largely district courts or intermediate state courts, with the exceptions of West Virginia and Washington) have applied one state's wage laws to work performed at least partially in another state. For several reasons, we are unpersuaded by defendants' arguments.

For one, we read Sullivan as indicating that under California's choice-of-law principles, a state has a legitimate interest in applying its wage laws extraterritorially only in two limited circumstances, neither of which apply here: one, when a state's resident employee of that state's resident employer leaves the state "temporarily during the course of the normal workday," and two, "when the traveling, resident employee of a domestic employer would otherwise be left without the protection of another state's law." Id. , 127 Cal.Rptr.3d 185, 254 P.3d at 242, 246 (citations, internal quotation marks, and alterations omitted).

Moreover, the cases on which defendants and the dissent rely are, in large part, both factually and procedurally inapposite to the circumstances of this case. For example, defendants rely heavily on Bostain v. Food Exp., Inc. , 159 Wash.2d 700, 153 P.3d 846, 851 (2007) to argue that Washington has an interest in applying its wage laws extraterritorially.

As the California Supreme Court held in Sullivan , however, Bostain "says nothing about a case such as this"-that is, a case which (1) involves work performed entirely in one state, and (2) presents an unavoidable conflict-of-laws issue. 127 Cal.Rptr.3d 185, 254 P.3d at 243. In Bostain , by contrast, either Washington law applied to the work performed in both Washington and other states, or else no state's law applied. Id., 127 Cal.Rptr.3d 185, 254 P.3d at 243, 246.

Significantly, Bostain interpreted an overtime statute that specifically delineated the circumstances under which its provisions would apply to interstate truck drivers; as the Washington Supreme Court noted, interstate truck drivers by definition perform some of their work out of state. 153 P.3d at 848-51. The statute at issue in Bostain did "not limit the requirement for overtime pay to hours worked" within the state's borders. Id. at 851.

Similarly, here, defendants point to no state statutes potentially applicable to the Arizona and Florida class members that limit their application to work performed within the state. 3. Although the existence of a "true" conflict is questionable, we need not decide whether a true conflict exists, as the third step of California's governmental interest test yields a clear answer: the laws of Arizona and Florida should apply to the work performed wholly within their respective boundaries. See Sullivan , 127 Cal.Rptr.3d 185, 254 P.3d at 247.

As the California Supreme Court has explained the step three inquiry: [T]he court does not "weigh" the conflicting governmental interests in the sense of determining which conflicting law manifested the "better" or the "worthier" social policy on the specific issue. An attempted balancing of conflicting state policies in that sense is difficult to justify in the context of a federal system in which, within constitutional limits, states are empowered to mold their policies as they wish.

Instead, the process can accurately be described as a problem of allocating domains of law-making power in multi-state contexts -by determining the appropriate limitations on the reach of state policies-as distinguished from evaluating the wisdom of those policies. Emphasis is placed on the appropriate scope of conflicting state policies rather than on the "quality" of those policies. McCann , 105 Cal.Rptr.3d 378, 225 P.3d at 533-34 (alterations and citations omitted) (emphasis added).

As discussed above, in Mazza , we faithfully applied the principle under California's choice-of-law jurisprudence that "a jurisdiction ordinarily has the predominant interest in regulating conduct that occurs within its borders." Mazza , 666 F.3d at 592 (internal quotation marks omitted) (quoting McCann , 105 Cal.Rptr.3d 378, 225 P.3d at 534 ). We thus had no trouble concluding at step three that "each class member's consumer protection claim should be governed by the consumer protection laws of the jurisdiction in which the transaction took place." Id. at 594.

Notably, we reached this conclusion without specifically inquiring into the interests potentially expressed by any state's statutory language or case law. Rather, our conclusion was dictated by the principle, discussed above, that a jurisdiction ordinarily has the predominant interest in regulating conduct within its own borders. Id. at 591-92 (first citing State Farm Mut. Auto. Ins. Co. v. Campbell , 538 U.S. 408, 422, 123 S.Ct. 1513, 155 L.Ed.2d 585 (2003) ; and then citing McCann , 105 Cal.Rptr.3d 378, 225 P.3d at 534 ).

Moreover, in Sullivan , the court concluded that to subordinate California's ability to apply its own wage laws to work performed within the state would "unquestionably" cause greater impairment to California than to the states that might seek to apply their wage laws to work performed by their residents within California. 127 Cal.Rptr.3d 185, 254 P.3d at 247. As described previously, while this holding was influenced by the absence of an expression of interest by Arizona or Colorado in applying their laws extraterritorially, it did not rise or fall on that ground.

See id., 127 Cal.Rptr.3d 185, 254 P.3d at 244-47. And although defendants point to a handful of cases that have entertained the potential application of one state's wage laws to work performed in another state, they have not pointed to a single state with a potentially-applicable statute that expresses a clear interest in applying to work performed wholly outside the state.

But even if defendants were able to identify any states that had unambiguously expressed an interest in applying their wage laws to work performed entirely in another state, Sullivan strongly militates against concluding that such an expression of interest would be adequate to overcome the principle that the state in which the conduct at issue occurs has the "predominant interest" in applying their own law. See Mazza , 666 F.3d at 592-94 ; Sullivan , 127 Cal.Rptr.3d 185, 254 P.3d at 245-47.

Forcing Arizona or Florida to allow the application of other states' wage laws in this case would be just as destructive to the balance Arizona and Florida have struck between protecting workers and fostering a hospitable business environment within their states as allowing the application of Colorado or Arizona law in Sullivan would have been to the balance California struck between those same interests. See Sullivan , 127 Cal.Rptr.3d 185, 254 P.3d at 246-47.

The district court fundamentally misunderstood the proper application of California's choice-of-law principles-which, when correctly applied, indicate that Arizona law should govern the Arizona class, and Florida law the Florida class. C. We next address whether the district court erred in refusing to certify a Rule 23(b)(2) class for unpaid work at defendants' training facilities in Arizona and Florida on the sole basis that choice-of-law issues undermined "cohesiveness" and therefore made injunctive and declaratory relief inappropriate.

Because the district court's errors in its choice-of-law analysis relating to the proposed Arizona and Florida Rule 23(b)(3) classes apply equally to its refusal to certify the proposed Rule 23(b)(2) class, we also reverse the denial of the (b)(2) class. We further hold that the district court erred in imposing a "cohesiveness" requirement for the proposed Rule 23(b)(2) class.

Although we have never explicitly addressed whether "cohesiveness" is required under Rule 23(b)(2), courts that have imposed such a test treat it similarly to Rule 23(b)(3) 's predominance inquiry -something we have previously rejected in no uncertain terms. See Walters v. Reno , 145 F.3d 1032, 1047 (9th Cir. 1998) ("[W]ith respect to 23(b)(2) in particular, the government's dogged focus on the factual differences among the class members appears to demonstrate a fundamental misunderstanding of the rule.

Although common issues must predominate for class certification under Rule 23(b)(3), no such requirement exists under 23(b)(2)."); see also 2 Newberg on Class Actions § 4:34 (5th ed. 2012) (describing similarity between predominance under Rule 23(b)(3) and "cohesiveness" under Rule 23(b)(2) in courts that have adopted it). We therefore remand for the district court to consider anew whether to certify the proposed Rule 23(b)(2) class. III.

Having addressed the impact of choice-of-law questions, we turn to the issue next up at bat: whether the district court erred in concluding that plaintiffs could meet the predominance requirement for the proposed California, Florida, and Arizona (b)(3) classes through a combination of representative evidence and application of the "continuous workday" rule. Rule 23(b)(3) 's predominance requirement requires courts to ask "whether the common, aggregation-enabling issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues." Tyson Foods, Inc. v.

Bouaphakeo , --- U.S. ----, 136 S. Ct. 1036, 1045, 194 L.Ed.2d 124 (2016) (quoting 2 Newberg on Class Actions § 4:49 (5th ed. 2012) ). A proposed (b)(3) class may be certified as long as "one or more of the central issues in the action are common to the class and can be said to predominate ... even though other important matters will have to be tried separately, such as damages or some affirmative defenses peculiar to some individual class members." Id. (quoting 7AA C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 1778 (3d ed. 2005) ).

"[P]redominance in employment cases is rarely defeated on the grounds of differences among employees so long as liability arises from a common practice or policy of an employer." 7 Newberg on Class Actions § 23:33 (5th ed. 2012). Although the existence of blanket corporate policies is not a guarantee that predominance will be satisfied, such policies "often bear heavily on questions of predominance and superiority." In re Wells Fargo Home Mortg. Overtime Pay Litig ., 571 F.3d 953, 958 (9th Cir. 2009).

Whether the district court was correct in concluding that plaintiffs had satisfied the predominance requirement hinges on the application of two longstanding wage-and-hour doctrines to this case: first, the burden-shifting framework initially set forth in the Supreme Court's seminal decision in Anderson v. Mt. Clemens , 328 U.S. 680, 66 S.Ct. 1187, 90 L.Ed. 1515 (1946) and recently expanded upon in Tyson Foods v. Bouaphakeo , --- U.S. ----, 136 S. Ct. 1036, 194 L.Ed.2d 124 (2016) ; and second, the so-called "continuous workday" rule. We address each of these doctrines and their application to this case in turn.

A. In Mt. Clemens , the Supreme Court acknowledged the difficult bind that employees frequently confronted when seeking to bring wage-and-hour claims against their employers: if their employers had failed to maintain proper timekeeping records, proving the hours of uncompensated work often posed "an impossible hurdle for the employee." 328 U.S. at 687, 66 S.Ct. 1187. Mt. Clemens held that such a catch-22 was not in line with "the remedial nature of [the FLSA] and the great public policy which it embodies." Id.

After all, "[s]uch a result would place a premium on an employer's failure to keep proper records in conformity with his statutory duty; it would allow the employer to keep the benefits of an employee's labors without paying due compensation." Id. To address this problem, Mt.

Clemens established its landmark burden-shifting framework for actions in which the employer has kept inaccurate or inadequate records: if an employee "proves that he has in fact performed work for which he was improperly compensated" and "produces sufficient evidence to show the amount and extent of that work as a matter of just and reasonable inference," then the burden "shifts to the employer to come forward with evidence of the precise amount of work performed or with evidence to negative the reasonableness of the inference to be drawn from the employee's evidence." Id . at 687-88, 66 S.Ct. 1187.

If the employer does not rebut the employee's evidence, damages may then be awarded to the employee, "even though the result be only approximate." Id. at 688, 66 S.Ct. 1187. Mt.

Clemens explicitly rejected the notion that allowing approximate damages in such situations would be unfair due to its speculative and imprecise nature or because employers sometimes make good-faith mistakes over what constitutes compensable "work": The employer cannot be heard to complain that the damages lack the exactness and precision of measurement that would be possible had he kept records in accordance with the [statutory] requirements ...

And even where the lack of accurate records grows out of a bona fide mistake as to whether certain activities or non-activities constitute work, the employer, having received the benefits of such work, cannot object to the payment for the work on the most accurate basis possible under the circumstances ... In such a case it would be a perversion of fundamental principles of justice to deny all relief to the injured person, and thereby relieve the wrongdoer from making any amend for his acts. Id. (internal quotation marks and citation omitted). Seventy years after Mt.

Clemens addressed the use of representative evidence at the trial stage to show damages, Tyson extended Mt. Clemens ' holding to answer two important questions: whether representative evidence may be used at the class certification stage, and whether representative evidence may also be used to establish liability in addition to damages.

In Tyson , employees who worked in more than 400 jobs across three departments at a meat processing plant sued under the FLSA and an Iowa wage law, alleging that Tyson had not paid them overtime for time they spent donning and doffing protective gear; the employees also sought certification of a Rule 23 class and a FLSA collective action. 136 S. Ct. at 1041-42.

The district court certified the class and collective actions, rejecting Tyson's arguments that the claims were inappropriate for resolution on a classwide and collective basis due to the dissimilarity in the types of protective gear worn and the variations in time spent donning and doffing that gear. Id. at 1042-43.

Because Tyson had not kept records of the donning and doffing time, plaintiffs relied on representative evidence to demonstrate both liability and damages: employee testimony, video recordings, and-most significantly-an expert study that computed an estimated amount of time spent donning and doffing for each of the three departments based on hundreds of video observations. Id. at 1043.

Although the expert estimated that the time spent donning and doffing was 18 minutes per day for two of the departments and 21.25 minutes for the other, id. , the survey data showed a great deal of variation in how long it took individual employees to don and doff. Id. at 1055 (Thomas, J., dissenting). Specifically, the time spent donning ranged from around thirty seconds to more than ten minutes, and the time doffing varied from under two minutes to over nine minutes. Id.

After a jury verdict in the employees' favor (albeit one that awarded less than half of the damages recommended by the employees' expert based on the survey data), Tyson moved to decertify the class and set aside the jury verdict, arguing that this variance made class and collective certification inappropriate. Id. at 1044-45. The district court denied the motion, and the Eighth Circuit affirmed. Id.

Tyson sought certiorari on the grounds that using representative evidence "manufactures predominance by assuming away the very differences that make the case inappropriate for classwide resolution," "absolves each employee of the responsibility to prove personal injury," and strips the employer of their ability to "litigate its defenses to individual claims." Id. at 1046. Rejecting these arguments, the Supreme Court affirmed the class and collective certifications. Id. at 1046-47.

Because of Tyson's dereliction of their recordkeeping duties, the employees were entitled to "introduce a representative sample to fill an evidentiary gap created by the employer's failure to keep adequate records." Id. at 1047. The Court held that if the representative sample introduced were admissible and "could have sustained a reasonable jury finding as to hours worked in each employee's individual action, that sample is a permissible means of establishing the employees' hours worked in a class action." Id. at 1046-47.

Stated another way, Tyson concluded that even where "reasonable minds may differ" about whether representative evidence is sufficiently probative of the requirements for liability for a particular cause of action-in Tyson , whether it was probative of the "time actually worked by each employee"-that question is to be resolved by the jury, not at the class certification stage.

Id. at 1049 ("The District Court could have denied class certification on this ground [whether the representative evidence was "probative as to the time actually worked by each employee"] only if it concluded that no reasonable juror could have believed that the employees spent roughly equal time donning and doffing.") (emphasis added). If the proffered representative evidence, however, were "statistically inadequate or based on implausible assumptions," it "could not lead to a fair or accurate estimate of the uncompensated hours an employee has worked." Id. at 1048-49.

But where the evidence is admissible-for expert evidence, using the Daubert standard-then the "no reasonable juror" standard at the class certification stage applies. See id. at 1049. B. Having established the parameters of when representative evidence may be used at the class certification stage, we address the second significant wage-and-hour doctrine relevant to this case: the "continuous workday" rule. The rule was first promulgated by the Department of Labor (DOL) consistent with the Supreme Court's decisions interpreting the FLSA prior to the enactment of the Portal-to-Portal Act in 1947. IBP, Inc. v.

Alvarez , 546 U.S. 21, 27-28, 126 S.Ct. 514, 163 L.Ed.2d 288 (2005). It presumes that once the beginning of the workday is triggered, an employee performs compensable work throughout the rest of the day until the employee completes their last principal activity (or the last activity which is "integral and indispensable" to the employee's principal activities)-whether or not the employee actually engages in work throughout that entire period. See id. at 28, 32-37, 126 S.Ct. 514 ; see also Alvarez v.

IBP, Inc. , 339 F.3d 894, 907 (9th Cir. 2003), aff'd , 546 U.S. 21, 126 S.Ct. 514, 163 L.Ed.2d 288 (2005) (holding that under the continuous workday rule, "work time [is] continuous, not the sum of discrete periods"). Of course, this rule raises inevitable questions: when does the workday begin, and when does it end? The DOL defines the "workday" to generally mean "the period between the commencement and completion on the same workday of an employee's principal activity or activities." 29 C.F.R. § 790.6(b).

The Supreme Court expanded upon this definition, interpreting "principal activity or activities" to also include "all activities which are an integral and indispensable part of the principal activities." IBP , 546 U.S. at 29-30, 126 S.Ct. 514 (internal quotation marks and citation omitted). Thus, any activity which is "integral and indispensable" to principal activities, even if performed outside of a scheduled shift, triggers the beginning of the "workday." Id. at 31-37, 126 S.Ct. 514.

"Among the activities included as an integral part of a principal activity are those closely related activities which are indispensable to its performance," 29 C.F.R. § 790.8(c), such as knife-sharpening performed outside of a scheduled shift by butchers at a meatpacking plant. Mitchell v. King Packing Co. , 350 U.S. 260, 261-63, 76 S.Ct. 337, 100 L.Ed. 282 (1956). C.

With all of that in mind, we turn to how these two doctrines impact this case, and more specifically, whether the district court was correct in concluding that the combination of Tyson and the continuous workday rule enabled plaintiffs to show that they meet Rule 23(b)(3) 's predominance requirement.

Defendants contend that the district court erred in holding that plaintiffs had demonstrated predominance for two main reasons: (1) because the Main Survey asked only about arrival and departure times at the ballpark and not about what activities the players actually performed while at the ballpark, plaintiffs cannot rely on the continuous workday theory because there is no way to determine the beginning or end of the "workday," and (2) the Main Survey revealed significant variations in players' arrival and departure times, even among players employed by the same MLB franchise.

This task requires us to address the proposed Arizona and Florida classes separately from the California class. As an initial matter, however, we note that despite defendants' repeated suggestions to the contrary, the representative evidence offered by plaintiffs was not limited to just the Main Survey, nor are observational studies the only type of evidence permitted to fill in evidentiary gaps under Tyson . We reject defendants' erroneous view of the record and their cramped reading of Tyson . 1. As to the Arizona and Florida classes, we easily affirm the district court's determination.

Recall that these two classes cover time spent participating in spring training, extended spring training, and the instructional leagues-periods during which virtually all players are completely unpaid for their participation. Moreover, these classes do not bring overtime claims, but rather allege minimum wage violations. Therefore-as the district court correctly held-liability can be established simply by showing that the class members performed any compensable work.

That is easily resolved on a classwide basis by answering two questions: (1) are the players employees of defendants, and (2) do the minor league team activities during these periods constitute compensable work under the laws of either Arizona or Florida? We hold that these two "common, aggregation-enabling issues in the case are more prevalent [and] important than the non-common, aggregation-defeating, individual issues," therefore making certification appropriate. Tyson , 136 S. Ct. at 1045 (quoting 2 Newberg on Class Actions § 4:49 (5th ed. 2012) ).

Defendants do not seriously contest that their policy is to deny players compensation during spring training, extended spring training, and the instructional leagues-nor could they credibly do so, given that the MLB's own mandatory contract "obligates Player[s] to perform professional services on a calendar year basis, regardless of the fact that salary payments are to be made only during the actual championship playing season." And as we have long held, such uniform corporate policies "carry great weight for certification purposes." In re Wells Fargo Home Mortg. Overtime Pay Litig ., 571 F.3d at 958.

This is not the "rare[ ]" case where predominance is defeated despite the existence of an employer's "common practice or policy." 7 Newberg on Class Actions § 23:33 (5th ed. 2012). We also agree with the district court that as to these classes, many of defendants' protests go to damages, not liability. Damages may well vary, and may require individualized calculations. But "the rule is clear: the need for individual damages calculations does not, alone, defeat class certification." Vaquero v. Ashley Furniture Indus., Inc. , 824 F.3d 1150, 1155 (9th Cir. 2016) ; see Tyson , 136 S.

Ct. at 1045 (holding that where "one or more of the central issues in the action are common to the class and can be said to predominate," certification may be appropriate "even though other important matters will have to be tried separately, such as damages." (quoting 7AA C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 1778 (3d ed. 2005) )). We do not, however, mean to minimize defendants' criticisms of the Main Survey.

Indeed, we agree that there are a number of legitimate questions about the persuasiveness of the Main Survey, especially if it were the only representative evidence submitted in support of certification. But as we have mentioned, the Main Survey was but one piece of the plaintiffs' representative evidence-evidence that also included hundreds of internal team schedules and public game schedules, payroll data, and the testimony of both players and league officials.

At minimum, if the players are "employees" under either Arizona or Florida law and defendants are unable to prove that any affirmative defenses apply, the team schedules will serve to conclusively demonstrate that the players spent time working for which they were uncompensated. See Ariz. Rev. Stat. Ann. § 23-362 ; Ariz. Admin.

Code. § R20-5-1202(19) (" 'Hours worked' means all hours for which an employee covered under the Act is employed and required to give to the employer, including all time during which an employee is on duty or at a prescribed work place and all time the employee is suffered or permitted to work."); 29 C.F.R. § 778.223 ("As a general rule the term 'hours worked' will include: (

a) All time during which an employee is required to be on duty or to be on the employer's premises or at a prescribed workplace and (

b) all time during which an employee is suffered or permitted to work whether or not he is required to do so."). Moreover, if plaintiffs can persuade a jury that their workday began at a particular time-either because they were required to report at that time, or because they arrived of their own volition but engaged in work activities upon arriving (i.e., were "permitted" to work)-the continuous workday doctrine eliminates the need for plaintiffs to prove which activities they engaged in throughout the day. See IBP , 546 U.S. at 28, 32-37, 126 S.Ct. 514.

Defendants should not "be heard to complain that the damages lack the exactness and precision of measurement that would be possible had [they] kept records in accordance with the [statutory] requirements," even if their "lack of accurate records grows out of a bona fide mistake as to whether certain activities or non-activities constitute work." Mt. Clemens , 328 U.S. at 688, 66 S.Ct. 1187. "Having received the benefits of such work, [defendants] cannot object to the payment for the work on the most accurate basis possible under the circumstances." Id. 2.

We next address whether the district court was correct to hold that predominance had been met for the California class. Given the differences in the types of claims brought by the California class as compared to the Arizona and Florida classes, certification of the California class is more complex and requires additional analysis. Unlike the Arizona and Florida classes, the California class brought claims relating to work performed during the championship season-a time when the players do get paid, albeit not much.

As a result, in order to prove liability on their overtime claims, the California class must show that its members worked more than 8 hours in a day, more than 40 hours in a week, and/or worked 7 days in a workweek. See Cal. Labor Code § 510 ; Mendoza v. Nordstrom, Inc. , 216 Cal.Rptr.3d 889, 393 P.3d 375, 381-82 (2017).

Likewise, to establish liability on their minimum wage claims, the California class must demonstrate that they worked hours for which they were not paid at least minimum wage-but whereas the Arizona and Florida classes can demonstrate liability simply by showing they worked any hours, the California class's burden is made more challenging by the fact that the players receive some pay. See Cal. Labor Code §§ 1182 et seq ; Armenta v. Osmose, Inc. , 135 Cal.App.4th 314, 37 Cal. Rptr. 3d 460, 466-68 (2005). Nonetheless, a number of considerations lead us to affirm the district court's determination.

First, as with defendants' uniform policy of not paying players for participation outside of the championship season, defendants do not credibly dispute that their policy is to never pay overtime and to pay a fixed salary, regardless of the actual number of hours worked. We reiterate that common corporate policies like this "carry great weight for certification purposes," In re Wells Fargo Home Mortg. Overtime Pay Litig ., 571 F.3d at 958, and that predominance is "rarely" defeated in cases where such uniform policies exist. See 7 Newberg on Class Actions § 23:33 (5th ed. 2012).

Second, the team schedules alone-independent of the Main Survey or any other evidence-may suffice to show overtime liability. As the district court noted, plaintiffs' expert testified that approximately 65-85% of California League players had at least one workweek with games on all seven days, and that nearly half of all workweeks included games on all seven days. For those workweeks, the players would be entitled to overtime pay for their work on the seventh day of the workweek. See Cal. Labor Code § 510.

Third, and most significantly, we are persuaded that under Tyson , the representative evidence plaintiffs offered was adequate to meet their burden at this stage. As we observed in the preceding section, defendants do identify multiple legitimate criticisms of the Main Survey, and it is certainly possible that a jury may not find the Main Survey-even in combination with all of plaintiffs' other evidence-adequate proof of liability (or at least not to the extent plaintiffs allege).

In particular, a jury may be persuaded by defendants' arguments that players did not begin compensable work upon arriving at the ballpark or that players stopped engaging in compensable work long before they left the ballpark, such that the Main Survey's estimated arrival and departure times are insufficient to clear the preponderance hurdle.

As we explain below, however, Tyson counsels that such criticisms do not doom certification here unless no reasonable jury could conclude that the combination of the Main Survey and plaintiffs' other representative evidence was probative of the amount of time players actually spent performing compensable work. See Tyson , 136 S. Ct. at 1046-49. And while defendants correctly point out that the Main Survey revealed meaningful variations in players' arrival and departure times, the same was true of the employees' donning and doffing times in Tyson -yet such variation did not preclude certification there.

See id. at 1043 ; id. at 1055 (Thomas, J., dissenting). Because defendants do not challenge the district court's ruling on admissibility under Daubert , the defects they have identified with the Main Survey could only have defeated certification upon a conclusion that all of the representative evidence offered-the Main Survey, schedules, testimony, and the like-could not have "sustained a reasonable jury finding as to hours worked in each employee's individual action." See Tyson , 136 S. Ct. at 1046-47.

As in Tyson , the district court "made no such finding," id. at 1049, and indeed found the opposite: Plaintiffs will be able to use the survey data in combination with other evidence that may be sufficient to allow a jury to draw conclusions based on reasonable inference as to when players were required to be at the ballpark and how long after games they were required to remain at the ballpark. ...

Thus, as in Tyson Foods , it appears that representative evidence can be combined with actual records of time spent engaged in the various activities to derive a reasonable estimate of the amount of time worked by class members. We are then left to ask whether "the record here provides [a] basis for [us] to second-guess that conclusion." Id. After reviewing the record, we are satisfied that we should not disturb the district court's determination, in part due to California's expansive definition of "employ" and "hours worked." Under California law, to "employ" means "to engage, suffer, or permit to work." Cal.

Code Regs. tit. 8, §§ 11040(2)(E), 11100(2)(E) (emphasis added). "Hours worked" means "the time during which an employee is subject to the control of an employer, and includes all the time the employee is suffered or permitted to work, whether or not required to do so." Cal. Code Regs. tit. 8, §§ 11040(2)(K), 11100(2)(H).

Inexplicably, however, defendants claim that under California law, "time spent engaging in activities that are not required by, or under the control of, an employer is not compensable and does not begin or extend a workday." This is a tortured and wholly unsupported reading of the law, and is manifestly contrary to one of the cases defendants themselves cite in support of their argument. See Morillion v.

Royal Packing Co. , 22 Cal.4th 575, 94 Cal.Rptr.2d 3, 995 P.2d 139, 143 (2000) ("[T]he two phrases-'time during which an employee is subject to the control of an employer' and 'time the employee is suffered or permitted to work, whether or not required to do so'-can also be interpreted as independent factors, each of which defines whether certain time spent is compensable as 'hours worked.' ") (citation omitted). Indeed, Morillion counsels that "hours worked" includes all time the employer "permit[s]" an employee to work, even if the work is not required and the employee is not under the employer's control.

See id. Thus, a player who arrives early or stays late at the ballpark of their own volition and performs "work" activities during that time is still owed compensation because the player was "permitted" to work, despite the work not being required. Likewise, under Morillion , if players were expected to arrive or depart at a particular time-whether that requirement was de facto or official-it is immaterial what activities the players actually engaged in while at the ballpark.

Even if the players spent their time at the ballpark doing things like eating or showering, they were still under their employer's control and unable "to use the time effectively for their own purposes," and thus were owed compensation. See id. , 94 Cal.Rptr.2d 3, 995 P.2d at 146. Indeed, Morillion explicitly rejected an analogous argument by the employer in that case: We reject Royal's contention that plaintiffs were not under its control during the required bus ride because they could read on the bus, or perform other personal activities.

Permitting plaintiffs to engage in limited activities such as reading or sleeping on the bus does not allow them to use the time effectively for their own purposes ... Plaintiffs were foreclosed from numerous activities in which they might otherwise engage if they were permitted to travel to the fields by their own transportation. Allowing plaintiffs the circumscribed activities of reading or sleeping does not affect, much less eliminate, the control Royal exercises by requiring them to travel on its buses and by prohibiting them from effectively using their travel time for their own purposes.

Similarly ... listening to music and drinking coffee while working in an office setting can also be characterized as personal activities, which would not otherwise render the time working noncompensable. Id. (alterations, internal quotation marks, and citations omitted). Thus, if plaintiffs use their representative evidence-especially the Main Survey and the testimony of players and league officials-to persuade a jury that they were required to be at the ballpark at particular times, they need not show how the players spent that time.

The fourth and final consideration weighing in favor of affirming the district court's determination is our standard of review. Abuse of discretion is always a relatively deferential standard, but when we review a grant of class certification, "we accord the district court noticeably more deference than when we review a denial." Abdullah , 731 F.3d at 956 (citation omitted). Were we to review de novo, this would likely be a closer call. But as they say, tie goes to the runner-and, under our deferential standard, to the district court. D. Finally, defendants, citing to Wal-Mart Stores, Inc. v.

Dukes , 564 U.S. 338, 352, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), contend that the district court was required to "rigorously analyze" the Main Survey, rather than evaluating its admissibility under Daubert and its appropriateness for meeting class certification requirements under Tyson . Tyson requires that we reject this argument. There, the Court explicitly distinguished the use of representative evidence to establish hours worked in wage and hour claims from the use of representative evidence in cases like Wal-Mart . Tyson , 136 S. Ct. at 1048.

Specifically-as we have explained-for wage and hour cases where the employer has failed to keep proper records, Tyson holds that once a district court has found expert evidence to be admissible, it may only deny its use to meet the requirements of Rule 23 certification if "no reasonable juror" could find it probative of whether an element of liability was met. Id. at 1049. Given the similarities between this case and Tyson , the rule set forward in Tyson controls, and "[defendants'] reliance on Wal-Mart is misplaced." Id. (citation omitted). IV.

We next address whether the district court properly certified the FLSA collective action. FLSA permits employees to bring lawsuits on behalf of "themselves and other employees similarly situated." 29 U.S.C. § 216(b). We recently delineated the appropriate standard for FLSA collective certification in Campbell v. City of Los Angeles , 903 F.3d 1090 (9th Cir. 2018). As we explained in Campbell , "there is no established definition of the FLSA's 'similarly situated' requirement, nor is there an established test for enforcing it." Id. at 1111 (citing Thiessen v. Gen. Elec.

Capital Corp. , 267 F.3d 1095, 1102 (10th Cir. 2001) ). In Campbell , we rejected both the minority approach to FLSA collective certification-which treats a FLSA collective as analogous to a Rule 23(b)(3) class-and the majority "ad hoc" approach. Id. at 1111-1117. The former approach, we observed, is inconsistent with the statute itself, as well as the choice of Congress and the Advisory Committee on Rules to distinguish FLSA collectives from Rule 23 class actions. Id. at 1111-1113.

And while the latter approach-the so-called ad hoc approach-is a "significant improvement" over the minority approach, it has two major flaws that led us to decline to adopt it. Id. at 1113-1116. First, this approach inappropriately "focus[es] on differences rather than similarities among the party plaintiffs," leading district courts to "treat[ ] difference as disqualifying," rather than "treat[ing] the requisite kind of similarity as the basis for allowing partially distinct cases to proceed together." Id. at 1117. Second, because the ad hoc approach allows distr

[…]