IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF PENNSYLVANIA
JAMES P. SCANLAN, et al. : CIVIL ACTION : v. : : AMERICAN AIRLINES GROUP, INC., : NO. 18-4040 et al. :
MEMORANDUM Bartle, J. August 20, 2025 Plaintiffs James P. Scanlan and Carla Riner bring this lawsuit, now certified as a class action, against defendants American Airlines, Inc. (“American”) and American Airlines Group Inc. (“AAG”) under the Uniformed Services Employment & Reemployment Rights Act, 38 U.S.C. § 4301 et seq. (“USERRA”). They aver that defendants violated Section 4316(b)(1) of USERRA in failing to pay or credit American Airline pilots when they are on short-term military leave, which plaintiffs characterize as sixteen days or fewer, but paying or crediting pilots when on jury duty and bereavement leave. Before the court is the motion of defendants to decertify the Rule 12(b)(2) and Rule 12(b)(3) sub-classes certified by the court on May 3, 2022 (Doc. # 228). Defendants argue that plaintiffs cannot satisfy the commonality requirement under Rule 23(a) for any subclass and cannot satisfy the predominance and superiority requirements for the Rule 23(b)(3) subclasses. They also maintain that certification of the Rule could not properly provide relief to each member of the class. After previously not contesting these issues, defendants justify their turnabout “based on subsequent appellate decisions and further consideration of existing decisions.” (Doc. #228-1). I USERRA is the latest of federal statutes designed to protect employees that leave civilian jobs to perform qualifying
military service. Among its purposes is to prohibit discrimination against persons because of their service in the uniformed services. 38 U.S.C. § 4301(3). It provides that those on leave of absence while performing military service are entitled to the same “rights and benefits . . . as are generally provided by the employer of the person to employees having similar seniority, status, and pay who are on . . . leave of absence under a contract, agreement, policy, or plan in effect. . . .” Id. at § 4316(b)(1)(B). Comparability is critical in deciding any claim under this section of USERRA. Scanlan v. Am. Airlines Grp., Inc., 102
F.4th 164, 170 (3d Cir. 2024). Section 1002.150(b) of Title 20 of the Code of Federal Regulations states that an employee “must be given the most favorable treatment accorded to any comparable form of leave when he or she performs service . . . .” It further provides: In order to determine whether any two types of leave are comparable, the duration of the leave may be the most significant factor to compare. For instance, a two-day funeral leave will not be “comparable” to an extended leave for service in the uniformed service. In addition to comparing the duration of the absences, other factors such as the purpose of the leave and the ability of the employee to choose when to take the leave should also be considered. Id. This list of factors is non-exhaustive. Scanlan, 102 F.4th at 171. In Count I of the second amended complaint, plaintiffs allege that AAG violates § 4316(b)(1) with respect to its profit sharing plan. AAG deposits a certain percentage of its profits into the plan and divides the pie among the pilots based on their earnings. AAG does not credit pilots for any time on short-term military leave in calculating earnings for the purposes of the profit sharing plan despite doing so for those on jury duty and bereavement leave. Thus, the pilots who take short-term military leave receive a lesser share than they would otherwise receive.1 Count III asserts that American violates § 4316(b)(1) when it fails to pay pilots while they are on short-term military leave but does so while on jury duty or bereavement
1. In their second amended complaint, plaintiffs also brought a state breach of contract claim against AAG, alleging that it violated the terms of the profit sharing plan (Count II). On summary judgment, the court found for AAG on this claim and our Court of Appeals affirmed. See Scanlan, 102 F.4th at 170. Count II is no longer an issue in this action. leave. Plaintiffs, for purposes of this lawsuit, define short- term military leave as 16 days or fewer. II On October 8, 2021, the court first certified a class of current and former pilots at American Airlines under Rule 23(b)(2). Scanlan, 567 F. Supp. 3d 521 (E.D. Pa. 2021) (Doc. # 124). Initially, plaintiffs sought to certify a class which
included pilots as well as fifteen additional work groups at AAG and American, such as flight attendants. In support of their motion for class certification, plaintiffs asserted that the common question for all class members is “whether short-term military leave is comparable to leave for jury duty or bereavement and thus must be credited . . . or compensated . . . .” Id. at 529. Defendants conceded that “[a]s a general matter, the proof required to resolve the American pilots’ claims will be common” in recognition of their collective bargaining agreement which governs all leave taken by such pilots. (Doc. # 98). They simply opposed class
certification on the basis of intra-class conflicts between pilots and non-pilots. The court agreed with defendants and excluded non-pilots. Scanlan, 567 F. Supp. 3d at 529-30. As to certification under Rule 23(b)(2), defendants argued that injunctive relief was not appropriate as class members seek individualized monetary damages. The court concluded that class treatment of pilots alone was appropriate because American and AAG have standard leave policies that govern this entire work group. Any difference between jury duty and bereavement leave from military leave was a common question and could be litigated on a class- wide basis. The court certified the class under Rule 23(b)(2). It
concluded that the monetary relief sought by the class “incidental to the request for declaratory and injunctive relief and flows from the requested injunction.” See id. at 536. Therefore, certification under Rule 23(b)(3) was not necessary, and the court did not analyze whether the class satisfied its requirements at that time. On January 28, 2022, defendants sought to exclude from the class definition any pilots who retired from either the military or from American, as those pilots would not benefit from any prospective relief and therefore should not be part of a class seeking injunctive relief under Rule 23(b)(2).
Plaintiffs, without objection from defendants, proposed that the court certify a subclass of retired pilots under Rule 23(b)(3). The court amended the class to reflect these definitions on April 6, 2022. See Scanlan, Civ. A. No. 18-4040, 2022 WL 1028038, at *3 (E.D. Pa. Apr. 6, 2022) (Doc. # 155). In doing so, the court found that predominance and superiority existed for the Rule 23(b)(3) subclasses. The subclasses currently certified on May 3, 20222 are: A class of American Airlines, Inc. pilots is certified as follows: (a) For Count I, the court certifies a subclass under Rule 23(b)(2) consisting of current American Airlines, Inc. pilots who presently serve in the military and who: participate at some point in the AAG Global Profit Sharing Plan since its inception on January 1, 2016 through the date of judgment in this action; are employed in the United States or are a citizen or national or permanent resident of the United States and employed in a foreign country while participants in the Plan; took or take short-term military leave in a year during which they were entitled to receive an award under the Plan; and were not credited or imputed earnings for this short-term military leave. (b) Also for Count I, the court certifies a subclass under Rule 23(b)(3) consisting of former pilots of American Airlines, Inc. or pilots of American who are former members of the military who: participated at some point in the AAG Global Profit Sharing Plan since its inception on January 1, 2016 through the date of judgment in this action; were employed in the United States or were a citizen or national or permanent resident of the United States and employed in a foreign country while participants in the Plan; took short- term military leave in a year during which they were entitled to receive an
2. The May 3, 2022 Order was the same as the April 6, 2022 Order except that it amended the definitions of the Count II subclasses. Count II, as noted, is no longer before this court. award under the Plan; and were not credited or imputed earnings for this short-term military leave. . . . . (e) For Count III, the court certifies a subclass under Rule 23(b)(2) consisting of all current American Airlines, Inc. pilots who presently serve in the military and took or take short-term military leave while employed at American at any time from January 1, 2013 through the date of judgment in this action and were not paid for that leave equal to what they would have received had they taken leave for jury duty or bereavement. (f) Also for Count III, the court certifies a subclass under Rule 23(b)(3) consisting of all former American Airlines, Inc. pilots or American pilots formerly in the military who took short-term military leave while employed at American at any time from January 1, 2013 through the date of judgment in this action and were not paid for that leave equal to what they would have received had they taken leave for jury duty or bereavement. Doc. # 166.3 Defendants did not seek permission under Rule 23(f) to appeal the class certification decisions by the court. Instead, following further discovery and the exchange of expert reports, defendants later moved for summary judgment under Rule 56. In
3. Pilots responsible for administering the AAG Global Profit Sharing Plan and pilots who have reached individual settlements with or judgments against AAG regarding inadequate profit sharing or against American for failure to compensate short-term military leave under USERRA are excluded from the class. Id. their motion, defendants argued that short-term military leave was not comparable to jury duty and bereavement leave. To establish this disparity, they contrasted class-wide statistical evidence of the annual average of both duration and frequency of military leave of all pilots in the aggregate with the statistical evidence of the annual average duration and frequency of jury duty and bereavement leave for all pilots.
The defendants also focused on portions of the testimony of Scanlan and Riner (the class representatives), the testimony of a representative from the pilot’s union, and the testimony of an expert witness to establish the differences, not on a pilot-by-pilot basis but categorically, as to purpose and control of military leave on the one hand and jury duty and bereavement leave on the other. Scanlan, 638 F. Supp. 3d 453, 458-59 (E.D. Pa. 2022). Defendants asserted that the undisputed facts demonstrated that pilots collectively “have some control over when they perform [military] service” and for this reason, it was not comparable to jury duty or bereavement leave. (Doc.
# 173-1). Defendants, in support of their motion for summary judgment, did not argue that the court must compare separately each military leave taken by each pilot with the defendants’ policies and the facts related to jury duty and bereavement leave. Scanlan, 638 F. Supp. 3d at 458-59. On November 2, 2022, the court granted the motion for summary judgment of defendants on Count I and Count III. It concluded based on the statistical and other evidence in the record that short-term military leave was not comparable as a matter of law to jury duty and bereavement leave. Id. at 460. The Court of Appeals reversed. It held that there were genuine disputes of material fact for the jury to decide on
the issue of comparability. In reaching its decision, it relied as had this court on the unchallenged statistical evidence referenced by the parties in the record. It also agreed that frequency of leave, in addition to duration, purpose, and control, may be considered by the jury in deciding whether short-term military leave is sufficiently similar to jury duty and bereavement leave. Scanlan, 102 F. 4th at 170-71. III With this background, the court turns to defendants’ motion to decertify the existing subclasses. Rule 23(c)(1)(C) permits a court to alter or amend class certification before the
action reaches a final judgment. Barnes v. Am. Tobacco Co., Ins., 161 F.3d 127, 140 (3d Cir. 1998). The court is obligated to ensure that plaintiffs’ claims continue to meet the requirements of Rule 23. Id. Modification of a class certification order may be warranted if “upon fuller development of the facts, the original determination appears unsound.” Zenith Lab’ys, Inc. v. Carter-Wallace, Inc., 530 F.2d 508, 512 (3d Cir. 1976). A class certification order is “inherently tentative.” General Tel. Co. of the Sw. v. Falcon, 457 U.S. 147, 160 (1992). The court, of course, must always be satisfied “after a rigorous analysis” that class certification is appropriate. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350- 51 (2011).
Rule 23(a)(2) provides that one or more class members may sue as representative parties on behalf of all members only if, among other requirements, “there are questions of law or fact common to the class.” A single question may satisfy commonality. Wal-Mart, 564 U.S. at 359. Plaintiffs must show that all putative class members were “subjected to the same harmful conduct by the defendant.” In re Cmty. Bank of N. Va. Mortg. Lending Pracs. Litig., 795 F.3d 380, 397 (3d Cir. 2015). For questions to be common, they must have the capacity to “generate common answers.” Wal-Mart, 564 U.S. at 350. Defendants now argue that subclasses should be
decertified on the ground that there is no common question or answer under Rule 23(a). Specifically, they assert that comparability under USERRA can only be properly resolved by individually comparing the length of time of each occasion each pilot takes short-term military leave with the average time of a jury duty leave or bereavement leave. There is evidence that some pilots have taken as long as 16 days of military leave at one stretch while others have taken as little as one day at a stretch. Many are somewhere in the middle. In defendants’ view, plaintiffs may not be permitted to prove their case for the subclasses as a whole simply by comparing a statistical average length of military leave of all pilots with the average length of jury duty leave or bereavement leave for all pilots
when there is such disparity in the length of each instance of military leave. Defendants also maintain that use of this statistical evidence of duration in a class action would violate the Rules Enabling Act. 28 U.S.C. § 2072. The defendants cite a number of USERRA cases to support their position. Unfortunately, none of them resolves whether common questions and common answers exist so as to allow a USERRA class action to proceed. At the outset, two of the cases are individual actions where not surprisingly the specific facts related to the plaintiff’s military leave were before the court. No
statistical evidence was involved. The class action questions raised here were not implicated. See Tully v. Dep’t of Just., 481 F.3d 1367 (Fed. Cir. 2007); Waltermeyer v. Aluminum Co. of Am., 804 F.2d 821 (3d Cir. 1986). Defendants cite Clarkson v. Alaska Airlines, Inc., an action under USERRA, in which plaintiff represented a class of “current or former Alaska or Horizon pilots who have taken short-term military leave,” understood in that case to be thirty days or fewer. 59 F.4th 424, 431 (9th Cir. 2023). There, the Court of Appeals reversed the lower court’s decision to grant summary judgment as to defendants Alaska Airlines and Horizon Air Industries. Id. at 429. It concluded that there were genuine disputes of material fact as to the comparability of
military leave to other types of leave offered by defendants such as jury duty, bereavement, and sick leave. Id. at 432. In so holding, the court identified factual disputes with respect to the statistical evidence of the duration and frequency of the various leaves as well as differences in testimony between class and defense witnesses as to purpose and control over timing. Id. at 436-39. It also held that frequency of leave is not a proper comparator. The Court at one point seemed to reject a categorical approach to military leave without providing any extended analysis. Id. at 433-34. Yet, in reversing the district court’s grant of summary judgment, it did not say that
it was improper for plaintiffs to rely on statistical evidence. Regardless, the Court did not have before it the specific question whether the class was properly certified or whether each pilot’s claim must be considered and tried individually. White v. United Airlines, Inc., also cited by the defendants, concerns a USERRA putative class action. Before reaching the issue of class certification, the district court dismissed the complaint for failure to state a claim. The Seventh Circuit reversed the lower court’s decision and remanded for further proceedings. 987 F.3d 616, 619 (7th Cir. 2021). The district court has not yet ruled on the question of class certification.4 Our Court of Appeals in Scanlan likewise did not reach
the issue presented here. On an appeal from this court’s grant of summary judgment, it held that there were genuine disputes of material fact. In doing so, it accepted the statistical evidence presented by the parties in comparing duration and frequency of an instance of military leave to the average duration and frequency of an instance of jury duty leave and bereavement leave. It also concluded that there was conflicting evidence on the issue of control and purpose. In its opinion, the Court of Appeals cited certain undisputed statistics: The parties agree on some of the relevant statistics. During the class period, a single instance of short-term military leave averaged about 3.3 days, while an instance of jury-duty leave averaged 1.8 days and an instance of bereavement leave averaged 2.7 days. The most common duration of both
4. Plaintiff filed a motion for class certification which was denied without prejudice for further discovery and guidance from sister circuits, including our Court of Appeals in this action. Order, White, Civ. A. No. 19-114 (N.D. Ill. Mar. 24, 2024) (Doc. # 127). Discovery is currently ongoing and plaintiff has not yet filed a renewed motion for class certification. short-term military leave and jury-duty leave was one day, while the most common duration of bereavement leave was three days.5
The parties also agree about the frequency with which pilots took each type of leave. On average, pilots took short-term military leave 7 times, jury-duty leave 1.3 times, and bereavement leave 1.2 times annually. But pilots most commonly used each type of leave once per year. And in any given year, pilots took an average of 21.9 total days of short-term military leave, 2.3 total days of jury-duty leave, and 3.1 total days of bereavement leave.
5 A majority of each type of leave – 65.2% of short-term military leaves, 87.2% of jury- duty leaves, and 97% of bereavement leaves – lasted between one and three days.
Scanlan, 102 F.4th at 168-69.
The Court of Appeals further explained that the parties have conflicting evidence on whether pilots have the ability to schedule military leave. Id. at 169. The defendants presented evidence that pilots have a great deal of control while plaintiffs showed contrary evidence. Neither side argued that it was the same for each and every pilot. The Court also noted that bereavement leave serves two purposes, that is to allow pilots to grieve and to protect the public while pilots grieve. Id. On the issue of jury duty, the parties agreed that its purpose is civic duty, albeit compulsory. While the parties do not dispute that one purpose of military leave is civic duty, the defendants assert that it primarily serves as a parallel career for pilots. The court turns to the Supreme Court’s decision in Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442 (2016), which it finds to be instructive. There the district court had certified a collective action under the Fair Labor Standards Act, 29 U.S.C. §§ 201 et seq., and a Rule 23(b)(3) class pursuant to an
analogous state statute. The classes encompassed employees of defendant who claimed that they were not paid for the time spend donning and doffing protective gear at the workplace. The time differential among employees in doing so was a matter of a couple of minutes at most. The action went to trial, and the jury found in favor of the class. Tyson, 577 U.S. at 452. The defendant asserted that it was error to allow plaintiffs to use statistical or representative evidence of an expert as to the average time involved in donning and doffing their gear. It argued that each employee must prove how much time he or she spent at these
tasks. The Supreme Court disagreed. It allowed the use of statistical evidence there and rejected the employer’s position to impose a broad rule that representative or statistical evidence cannot be used in class actions. Id. at 455. In its ruling, the Court stressed that the employer had not kept detailed records of the employees’ hours in violation of statutory mandates. Id. at 456-57 (discussing Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680 (1946)). Under this circumstance, the employer could not be heard to complain about the only evidence available to the employees. Furthermore, representative evidence sometimes is the “only practicable means to collect and present relevant data” to establish liability. Id. at 455. The Court added that if
individual actions had been brought rather than a class action it still would have been proper to introduce statistical evidence to prove the time that the individual had worked. Id. at 455. The Court cautioned courts “to give careful scrutiny to the relation between common and individual questions in a case.” Id. at 453. It went on to explain the difference between those two types of questions: An individual question is one where “members of a proposed class will need to present evidence that varies from member to member,” while a common question is one where “the same evidence will suffice for each member to make a prima facie showing [or] the issue is susceptible to generalized, class-wide proof.”
Significantly, the Court declared: The District Court could have denied class certification ... only if it concluded that no reasonable juror could have believed that the employee spent roughly equal time donning and doffing.
Id. at 459. Unlike Tyson, no one disputes that the records of the individual class members are available in this action. Moreover, the time disparity among employees in Tyson for donning and doffing was characterized by the Supreme Court as “roughly equal.” Id. In contrast, the evidence here establishes without dispute that some pilots have single instances of 16 consecutive days on military leave while others record single instances of only one day on military leave. No reasonable juror could find that sixteen days and one day are roughly equal in duration. Since these disparities are not roughly equal, the statistical evidence, cited by our Court of Appeals in Scanlan, of 3.3 days for the average duration at a
single stretch for all pilots’ military leave cannot support the existence of a critical common question for class action treatment under Rule 23(a). The duration of each occurrence when a pilot takes leave must be evaluated on its own for purposes of comparison with the average annual duration of jury duty of 1.8 days and bereavement leave of 2.7 days. These are individual questions. One size does not fit all. The Supreme Court in Wal-Mart reiterated that commonality requires that each class member must suffer the same injury. 564 U.S. at 349. The Court underscored, however, that for purposes of class certification it is insufficient simply to
raise common questions “even in droves.” The key is “the capacity of a class-wide proceeding to generate common answers apt to drive the resolution of the litigation.” Id. at 350. (quoting Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L. Rev. 97, 132 (2009)). In Ferreras v. American Airlines, 946 F.3d 178 (3d Cir. 2019), the district court certified a class of employees who alleged that defendants had not properly paid them for all
the time worked in violation of the New Jersey Wage and Hour Law, N.J. Stat. §§ 34:11-56a29, et seq. The defendant argued on appeal that class certification was improper because the court would have to engage in individual inquiries as to when they clocked in and when they worked while off the clock. The Court of Appeals agreed with the defendant that both commonality and predominance could not be satisfied. Id. at 186. The Court cited Wal-Mart that merely raising common questions “even in droves” is not sufficient. Id. at 185. There was no capacity in the proceeding to generate common answers. Id. at 185-86. The pilots here may be raising broad common questions
as in Wal-Mart and Ferreras. Nonetheless, as in those cases there are no common answers. Again, there are no common answers on comparability for all class members due to the disparate lengths of military leave. The court also agrees with the defendants that allowing the introduction of statistical evidence on duration to support commonality and thus liability here in this class action would violate the Rules Enabling Act. 28 U.S.C. §§ 2072(a) and (b). The Act provides that “the rules of practice and procedure and rules of evidence . . . shall not abridge, enlarge, or modify any substantive right.” Id. Surely, a pilot in any individual action would not be allowed to introduce the statistical evidence of the average duration of all pilots’
leave to prove comparability with jury duty and bereavement leave when his or her time on leave is not roughly equal with the statistical averages. Instead, the pilot would have to produce the records of his or her own time on military leave. To permit a class of pilots simply to prove their case with average times under such circumstances would improperly enlarge the substantive rights of those pilots under USERRA and abridge the defenses available to opposing parties. It would constitute “Trial by Formula” forbidden by the Rules Enabling Act. See Tyson, 577 U.S. at 458; Wal-Mart, 564 U.S. at 367. Then there is the factor of control, that is the
ability of a pilot to choose when to take military leave. It is a factor along with duration, frequency, and purpose that is to be considered by the jury in deciding whether military leave is comparable to jury duty leave and bereavement leave. Scanlan, 102 F.4th at 170; 20 C.F.R. § 1002.150(b). Unlike duration and frequency, this factor cannot be quantified by the use of averages. Plaintiffs do not explain how an expert or anyone else could testify as to the commonality of control for all pilots in the subclasses. Both the Court of Appeals and this court have recognized that the evidence as to whether a pilot has control over the timing of their military leave varies, depending on rank as well as type and length of military duty to be performed. Scanlan, 102 F.4th at 169; Scanlan, 638 F. Supp.
3d at 458-59. The commonality requirement under Rule 23(a) as to the control factor cannot be met. The issue of control is an individual question as the answer is not susceptible to generalized, claim-wide proof. Tyson, 577 U.S. at 458-59. The plaintiffs submit any issue of individuality can be resolved by having the jury answer a number of separate questions to determine what lengths of military leave are comparable. For example, the jury could be asked whether 16 consecutive days of military leave at a stretch are comparable to the average length of an instance of jury duty leave or bereavement leave. Another question could include three days of
consecutive military leave as the comparator and so on. Regardless as to how convincing this approach may appear at first glance, it does not advance the ball for the plaintiffs. While duration may be the most significant factor, it is by no means the only factor in the comparability analysis. Still lurking is the factor of control, among others. There is no common evidence on this factor as to the group of pilots who took 16 consecutive days of military leave, as to the group who took only 3 consecutive days of military leave or as to any other group for that matter. Control will differ from pilot to pilot, regardless of the group into which a pilot fits. Giving the jury the opportunity to decide what time frame, if any, would properly compare to the average time frame for jury duty
and bereavement leave without dealing with or by ignoring the concomitant factor of control would be a flawed exercise. At the oral argument on the motion to decertify, plaintiffs’ counsel laid bare the weakness of their commonality argument. He suggested that plaintiffs may simply avoid introducing any evidence related to control to avoid the pitfalls raised by defendants even though control is a factor to be considered in comparing military leave with non-military leave. Scanlan, 102 F. 4th at 170-71. The factor of frequency cannot be forgotten. Pilots averaged short-term military leave 7 times per year while they
averaged jury duty 1.3 times and bereavement leave 1.2 times per year. They took an average of 21.9 total days of military leave per year while only 2.3 total days of jury duty and 3.1 total days of bereavement leave. Id. at 169. The jury may consider frequency, even though not specifically listed in 20 C.F.R. § 1002.150(b). See Scanlan, 102 F. 4th at 170-71. Frequency like duration and control is an individual issue not susceptible to a common answer for all pilots. IV For the above reasons, there is no common question (or answer) under Rule 23(a). It is not necessary therefore to pursue the issues of predominance and superiority under Rule 23(b). The court will grant the motion of defendants to
decertify the Rule 23(b)(2) and Rule 23(b)(3) subclasses which were certified on May 3, 2022. The action will proceed as to the named plaintiffs only.