PAUL PARKER v. BNSF RAILWAY COMPANY (2024)

OPINION

Curtis Rookaird, through his estate representative Paul Parker, challenges his termination from BNSF Railway Company (BNSF) under the anti-retaliation provision of the Federal Railroad Safety Act (FRSA), 49 U.S.C. § 20109(d). After a jury found in Rookairds favor, the Ninth Circuit vacated the verdict and remanded the case to the district court to reconsider its partial summary judgment for Rookaird on the issue of whether his performing an air-brake test had contributed to BNSFs decision to discharge him. Rookaird v. BNSF Ry. Co., 908 F.3d 451, 463 (9th Cir. 2018) (Rookaird I). On remand, the district court conducted a bench trial on the issue and decided in BNSFs favor. The district court found that BNSF had conceded that Rookairds refusal to stop performing the air-brake test contributed to its decision to discharge Rookaird, but the court nonetheless concluded that BNSF was entitled to an affirmative defense by showing that the air-brake test “contributed very little” to BNSFs decision. Rookaird appeals, contending that the district court erred in its analysis of BNSFs affirmative defense and in certain evidentiary rulings.

We conclude that the district courts application of the FRSA does not comply with the text of the statute, which prohibits the discriminatory discharge of an employee due even “in part” to the employees refusal to violate or assist in violating a railroad safety law, rule, or regulation. 49 U.S.C. § 20109(a)(2). Nor is the district courts conclusion consistent with relevant case law, including our reasoning in a prior case that, under the burden-shifting framework required for FRSA cases, “[a plaintiff] would be entitled to relief even if [the protected activity] played only a very small role in [the employers] decision-making process.” Frost v. BNSF Ry. Co., 914 F.3d 1189, 1197 (9th Cir. 2019) (citing Rookaird I, 908 F.3d at 461). BNSF needed to demonstrate by clear and convincing evidence not merely that it could have fired Rookaird absent his engaging in the protected activity, but rather that BNSF would have fired Rookaird. 49 U.S.C. § 42121(b)(2)(B)(ii); 29 C.F.R. § 1982.104(e)(4); see Brousil v. U.S. Dept of Labor, Admin. Review Board, 43 F.4th 808, 812 (7th Cir. 2022) (citing Speegle v. Stone & Webster Constr., Inc., ARB No. 13-074, 2014 WL 1870933, at *7 (Dept of Labor Admin. Review Bd. Apr. 25, 2014)). An FRSA affirmative defense is a “steep burden,” see Araujo v. New Jersey Transit Rail Operations, Inc., 708 F.3d 152, 162 (3d Cir. 2013), particularly when a district court finds that an employer concedes that the protected activity contributed to the decision to terminate the employee who engaged in it.

We have jurisdiction under 28 U.S.C. § 1291. We affirm the district courts evidentiary rulings, and we vacate and remand the affirmative defense issue for further proceedings consistent with this opinion.

I. FACTS AND PROCEDURAL HISTORY

A. Rookairds Tenure at BNSF

1. On and Before February 23, 2010

Curtis Rookaird began working for BNSF, a national freight train operator, in 2004. Before February 23, 2010, Rookaird did not have a disciplinary record. He began his shift at 2:30 p.m. on February 23, 2010, working on a three-person “switcher” crew at BNSFs Swift depot with engineer Peter Belanger and brakeman Matthew Webb. Rookaird, as a conductor, was in charge of the crew. The crew was supposed to take a van from Swift to Ferndale, and then take train cars back to Custer, where his crew would move 42 railway cars onto storage tracks. The crew was then expected to take a van to Cherry Point to service BNSF customers. The crew initially encountered an approximately two-hour delay while waiting for the paperwork they needed to start the day. At Custer, before moving the cars to storage, Rookaird and his crew performed an air-brake safety test. This test took around twenty to forty minutes to complete.

Rookaird and his crew performed this air-brake test in accordance with the companys standard operating procedures and recent changes to local processes. BNSF train crews nationwide perform routine air-brake tests—sometimes called “air tests”—on a daily basis. Air-brake tests typically take twenty to twenty-five minutes to complete. BNSF had started conducting remote audits to ensure that employees were performing air-brake tests. On February 11, weeks before Rookaird was at Custer with his crew, BNSF implemented a new plan to manage the railway in the Cherry Point area more efficiently by moving the starting point for its switcher crews from Bellingham to Swift. The new plan decreased employees hours, but did not propose reducing or eliminating air-brake tests to gain efficiency.

BNSF, however, encountered less efficient operations that month while implementing the new plan, with train cars often sitting idle for longer than usual. BNSF Assistant Superintendent Stuart Gordon testified during the bench trial that complications were happening at that time, resulting in delays. Trainmaster Dan Fortt, who reported to Gordon, testified that BNSF was then “days behind” on their service commitments, calling the place an “operational nightmare.” Nonetheless, the BNSF company rules stated that, if an employee was in doubt about whether to perform an air-brake test, he should “take the safe course.”

Fortt, headquartered in Canada and then serving BNSF as a front-line supervisor, radioed Rookairds crew on February 23 while they were conducting the air-brake test. He asked them why they were doing it. Fortt commented: “Im not from around here, and I dont know how you guys do anything. But where Im from, we dont have to air test the cars.” Fortt had the authority to instruct the crew more explicitly to stop the air-brake test, but he did not do so. The crew completed the test.

Gordon, who supervised Fortt, had told Fortt to question Rookaird about why the crew was performing an air-brake test. Gordon objected that Rookairds crews performance of the test was unnecessary. Gordon concluded that the crew was inefficient that day because performing the air-brake test had delayed the operation. When later asked during his testimony which part of the job the crew was not doing efficiently, Gordon urged: “An air test, Im telling you.” When asked for any other basis for the crews inefficiency, Gordon did not mention any other specifics.

Rookairds crew started moving the cars into storage. The crew got another call from Fortt, this time asking how much longer they would take to finish storing the cars. It was around 7:30 p.m., and Rookaird estimated they would take another hour or two. At Gordons direction, Fortt told Rookairds crew to pack up and report back to the Swift depot because another crew was going to finish the job. Gordon called Rookairds crew back in part to question them as to why they performed the air-brake test. Fortt agreed that Rookairds crew was taking too long to do their job. Fortt also thought that the air-brake test was unnecessary and “contributing” to slowing the process down. Fortt was not aware of any other reasons for the crews delay.

Back at the Swift depot, Gordon told Rookaird and the crew to “tie up.” Gordon claimed that he did not then plan to pursue disciplinary action against Rookaird or the crew. Rookaird completed his tie-up timeslip at 8:02 p.m. but did not sign it. Rookaird went to the breakroom, where he had a heated exchange with another employee, Ron Krich. Gordon believed he overheard Rookaird encouraging Krich to “slow down,” suggesting an intentional effort to log more hours and earn overtime pay, though ultimately no one testified to having heard Rookaird use those exact words. Gordon came over and told Rookaird to leave. When Rookaird did not leave, Gordon again told Rookaird to go home. Rookaird then left. Rookaird reported that his final off-duty time for that day was 8:30 p.m., within the thirty-minute grace period permitted by BNSFs policies and practices.

Gordon immediately conferred with Fortt and then referred Rookairds crew for discipline in an e-mail to Gordons immediate supervisor, Jeff Beck, as well as to the General Manager of the Northwest Division, Doug Jones, and investigating officer Robert Johnson. Gordon testified that he told these individuals that Rookairds crew did an air-brake test that day “that wasnt necessary.”

2. The Investigation & Rookairds Termination

On February 26, 2010, BNSF sent Rookaird a letter alerting him that BNSF was investigating his actions on February 23, 2010. The letter informed Rookaird that BNSF was investigating him for: (1) failure to work efficiently, (2) dishonesty in reporting his off-duty time, (3) failure to provide a signed tie-up timeslip, and (4) failure to comply with instructions when specifically told to leave the property. BNSF justified its discipline based on its Policy for Employee Performance and Accountability (PEPA). The PEPA established the consequences BNSF could impose for each type of rule violation. The list included aggravated offenses for which BNSF had the option to terminate an employee based on a single violation. These aggravated offenses included gross dishonesty and insubordination.

Johnson conducted a twelve-hour investigation hearing on March 12, 2010. Rookaird had union representation present at the hearing. Johnson did not call Krich as a witness to testify about whether Rookaird talked to him in the breakroom about a “slow down.”

Johnson sent the transcript to the director of employee performance James Hurlburt, as well as to Doug Jones, who had the decision-making authority to discharge Rookaird. In an e-mail dated March 17, 2010, Johnson, as the investigating officer, wrote to Hurlburt and Jones: “I am not going to tell you this crew was the only crew that played the slowdown game. However, I will tell you that this crew made a fatal mistake and we have to make an example out of them.” Jones later testified that Johnsons email had “passion associated with it” and “a lot of emotion put in there.”

Hurlburt recommended that Jones dismiss Rookaird. Jones agreed, concluding that Rookaird had committed rule violations that merited immediate dismissal. Jones testified at the bench trial that Rookaird could not have been dismissed for failure to work inefficiently, because he did not have any prior record of discipline for that. Jones did testify, however, that Rookairds encouraging Krich to slow down contributed to Joness decision.

BNSF terminated Rookaird on March 19, 2010, effective immediately. BNSFs stated reasons for terminating Rookaird echoed those given in the investigation notice letter, all based solely on what happened on February 23: (1) “failure to work efficiently,” (2) “dishonesty when reporting [his] off duty time,” (3) “failure to provide a signed [ ] [t]ie-up timeslip,” and (4) “failure to comply with instructions when told to leave the property․” BNSF also disciplined the two other crew members who were with Rookaird that day, Webb and Belanger, for their failure to work efficiently. Webb and Belanger received a thirty-day suspension and probation.

B. Procedural History

Rookaird sued BNSF in 2014 under the anti-retaliation provision of the FRSA, 49 U.S.C. § 20109(a)(2), (d), alleging that BNSF discharged him in part for his refusal to stop the air-brake test. BNSF claimed that it would have fired Rookaird even if he had not engaged in the air-brake test. The district court entered a partial summary judgment in favor of Rookaird, concluding that the protected activity was a contributing factor in his termination. The district court found that disputed issues of material fact remained as to whether Rookaird had engaged in an activity protected by the FRSA and whether BNSF could prove its affirmative defense. Rookaird I, 908 F.3d at 455. The case proceeded to a jury trial in 2016 on these issues. The jury returned a verdict for Rookaird and awarded him more than $1 million in damages. Id.

On BNSFs first appeal, we affirmed in part that Rookaird had engaged in a protected activity when he refused to stop the air-brake test. Id. at 455–59. 1 We reversed in part, however, after concluding that, under the relevant burden-shifting framework, the district court had improperly conflated Rookairds prima facie and substantive showings. Id. at 459. We concluded instead that Rookaird was not entitled to summary judgment on the contributing-factor element at the substantive stage of the proceedings. Id. at 459–62. We vacated the jury verdict and remanded. Id. at 463. We expressed no view on how the district court should handle the issue of BNSFs affirmative defense on remand. Id.

On remand, the district court determined that three triable issues remained: (1) the contributing-factor element of the substantive stage of Rookairds case, (2) BNSFs affirmative defense, and (3) damages. The court set a jury trial for October 12, 2021.

On September 11, 2021, Rookaird passed away. Paul Parker, as the personal representative of Rookairds estate, substituted for Rookaird as the plaintiff. The parties then stipulated to a bench trial.

The district court held a four-day bench trial in October 2021. The district court confronted two critical evidentiary issues relating to the designation of certain witnesses prior testimony and comparator evidence. We discuss these evidentiary issues in turn, and then turn to the district courts decision.

1. Testimony Designations

Plaintiffs counsel sought to designate the previous trial or deposition testimony of seventeen witnesses, including Rookairds testimony. At a pretrial conference, the district court said that it would “permit the testimony to come in as [plaintiffs counsel] presents, if he believes that thats the way he wants to present that portion of his case.” At the start of the bench trial, the district court indicated that the designated testimony “will be reviewed and considered,” but not read into the record. The district court ordered plaintiffs counsel to file a declaration about the specific efforts that he had taken to make witnesses available for live testimony during the bench trial. Plaintiffs counsel described challenging circumstances with respect to witnesses who had died, retired, lived out of state, or had developed mental health problems that inhibited them from testifying. Ultimately, the district court accepted the designated testimony from many prior witnesses including Rookaird, Hurlburt, Johnson, Webb, Belanger, and Beck. The district court denied the designations for other remaining witnesses, including Krich and Rookairds wife.

2. Comparator Evidence

BNSF had sent Rookairds counsel a spreadsheet from BNSFs internal database with limited descriptions of other employees disciplinary records. The district court granted Rookairds request for more detailed information and directed BNSF to produce relevant specific information about twelve comparators that were selected by Rookaird. BNSF provided Rookaird with information about three employees who had no prior disciplinary record and who were dismissed for “willful dishonesty,” including falsification of their tie-up time records, among other violations. BNSFs records revealed that other employees received only a reminder, warning, suspension, or other minor discipline for similar infractions.

3. The District Courts Decision

The district court issued its Findings of Fact and Conclusions of Law on March 28, 2022. In addition to the facts stated above, the district court made a factual finding of particular note: “BNSF concedes that Mr. Rookairds conducting of the air test contributed to the crews supposed inefficiency and delay.” The district court based this finding on multiple grounds: Joness admission that he had fired Rookaird in part because of the delays that occurred that day, which included a twenty-five-minute delay for the air test; Gordons admission that Rookairds crew was being inefficient in part due to the air-brake test; and Fortts admission that the air-brake test contributed to slowing down the crews process.

In its legal conclusions, the district court relied on this factual finding to conclude: “BNSF concedes that the crews inefficiency was partly caused by Mr. Rookairds decision to conduct an air test—a test that BNSF managers thought was unnecessary to conduct in the first place.” The court continued: “Because Mr. Rookaird was fired for his inefficiency and because the inefficiency was partly caused by the protected activity of refusing to stop the air test, the Court concludes that the air test ‘tend[ed] to affect in [some] way the outcome of [BNSFs] decision’ to fire Mr. Rookaird.” And because refusing to stop the air-brake test “affected” BNSFs decision to terminate Rookaird, it was a “contributing factor” in his termination. The district court further concluded that Rookaird proved his substantive burden to show that refusal to stop conducting the air-brake test was a contributing factor in his dismissal.

The district court nonetheless concluded that BNSF met its burden to prove its affirmative defense—that is, to show by clear and convincing evidence that, absent Rookairds refusal to stop the air-brake test, BNSF would have still fired him for “many reasons unrelated to his inefficiency.” The district court reasoned: “[T]hough the air test was a contributing factor in Mr. Rookairds termination, the Court concludes that the test contributed very little.”

The district court stressed that the air-brake test only took twenty to forty minutes of the crews five-and-a-half hours of work, and no BNSF officer explicitly commanded Rookaird to stop the test. The district courts conclusion rested on Rookairds inaccurate and unsigned February 23 tie-up timeslip, Rookairds repeated refusal to leave work despite Gordons two orders telling him to stop work, and Rookairds exchange with Krich. The district court further pointed to the routine nature of the air-brake test, combined with the lesser discipline given to Rookairds two crewmates Webb and Belanger. From this, the district court concluded that BNSF, having satisfied its affirmative defense, was not liable for unlawful retaliation under the FRSA. Plaintiffs counsel moved to alter or amend the judgment, or in the alternative for a new trial. The district court denied the motion on August 5, 2022.

Rookaird timely appealed. On appeal, he contends that the district court erred in concluding that BNSF established its affirmative defense in light of FRSA law and the record. 2 Rookaird also argues that the district courts evidentiary rulings, which excluded certain testimony offered by plaintiffs and admitted evidence offered by BNSF concerning BNSFs new comparator evidence, constitute reversible error.

II. STANDARD OF REVIEW

After a bench trial, we review a district courts conclusions of law de novo, and we review its findings of fact for clear error. See Yu v. Idaho State Univ., 15 F.4th 1236, 1241–42 (9th Cir. 2021). We review de novo the legal inferences that a district court made while applying the law to the facts. See Suzys Zoo v. Commr, 273 F.3d 875, 878 (9th Cir. 2001). We review evidentiary rulings for abuse of discretion and only reverse an evidentiary ruling for prejudicial error. Wagner v. Cnty. of Maricopa, 747 F.3d 1048, 1052 (9th Cir. 2013).

III. DISCUSSION

A. The Text and History of the FRSA

Construction began on the nations first commercially chartered railway in 1828 when Charles Carroll, the last surviving signer of the Declaration of Independence, laid the first stone in Baltimore Harbor. 3 In 1970, more than 140 years later, Congress enacted the FRSA as its first comprehensive railroad safety law. 4 The FRSA aims “to promote safety in every area of railroad operations and reduce railroad-related accidents and incidents,” affecting railroad workers, passengers, and the general public. See 49 U.S.C. § 20101. Important safety measures include the air-brake tests that railway employees must perform. See Rookaird I, 908 F.3d at 456. A problem with a trains air-brake system can contribute to a crash or derailment, resulting in hazardous material spills, fatalities, or other issues that can gravely impact the lives of persons on the train or living nearby. 5 A train that cannot stop safely threatens disaster for railroad crews and the general public.

To encourage accurate reports about railroad safety, Congress amended the FRSA, first in 1980 and again in 2007, to add anti-retaliation measures ensuring that employees could report their safety concerns without fear. See Araujo, 708 F.3d at 156–57, n.3 (citing H.R. Rep. No. 110-259, at 348 (2007) (Conf.Rep.), as reprinted in 2007 U.S.C.C.A.N. 119, 181); Lee v. Norfolk S. Ry. Co., 802 F.3d 626, 629–30 (4th Cir. 2015) (citing the Federal Railroad Safety Authorization Act of 1980, Pub.L. No. 96-423, § 10, 94 Stat. 1811 (1980); H.R. Conf. Rep. No. 110-259, at 348 (2007), as reprinted in 2007 U.S.C.C.A.N. 119, 180–81); see also Christopher W. Bowman, Whistleblower Protections of the Federal Rail Safety Act: An Overview, 8 Wm. Mitchell J. L. & Prac. 1, 1 (June 2015).

Under the FRSA as amended:

A railroad carrier ․ may not discharge, ․ reprimand, or in any other way discriminate against an employee if such discrimination is due, in whole or in part, to the employees lawful, good faith ․ refus[al] to violate or assist in the violation of any Federal law, rule, or regulation relating to railroad safety ․

49 U.S.C. § 20109(a)(2).

To prevail on an FRSA claim, the plaintiff must prove four elements: (1) the plaintiff engaged in an FRSA-protected activity; (2) the employer knew that the plaintiff engaged in that protected activity; (3) the plaintiff suffered an unfavorable personnel action; and (4) the protected activity “was a contributing factor in the unfavorable personnel action.” Rookaird I, 908 F.3d at 455 (citation omitted). The “contributing factor” element is the only element at issue in this appeal. To prove these four elements, a plaintiff must meet the burdens of proof set forth in 49 U.S.C. § 42121(b)(2)(B), which provides the rules and procedures governing whistleblower cases under the Wendell H. Ford Aviation Investment and Reform Act for the 21st Century (AIR-21). 49 U.S.C. § 20109(d)(2)(A)(i) (incorporating section 42121(b)(2)(B) by reference); see also Araujo, 708 F.3d at 156–57. As clarified by our earlier opinion, the AIR-21 standard sets forth a burden-shifting framework that contains two distinct stages: the prima facie showing and the substantive showing. 49 U.S.C. § 42121(b)(2)(B); Rookaird I, 908 F.3d at 460. In each stage, if a plaintiff makes the required showing, the burden shifts to the respondent, such that each of the two stages contains two steps.

At the prima facie stage, the plaintiff must first establish the existence of facts sufficient to raise an inference that the protected activity was a “contributing factor” in the unfavorable personnel action. 49 U.S.C. § 42121(b)(2)(B)(i); 29 C.F.R. § 1982.104(e)(1)–(3). Second, the employer then has the burden to demonstrate, by clear and convincing evidence, that it “would have taken the same unfavorable personnel action in the absence of that behavior.” 49 U.S.C. § 42121(b)(2)(B)(ii); 29 C.F.R. § 1982.104(e)(4).

If the analysis proceeds to the substantive stage, the first step of the substantive analysis requires the plaintiff to demonstrate by a preponderance of the evidence that the protected activity was a “contributing factor” in the unfavorable personnel action. 49 U.S.C. § 42121(b)(2)(B)(iii); 29 C.F.R. § 1982.109(a). If the plaintiff makes this showing, the second step of the substantive analysis again shifts the burden to the employer, allowing the employer to present an affirmative defense. “[I]f the respondent-employer demonstrates by clear and convincing evidence that the employer would have taken the same unfavorable personnel action in the absence of that behavior,” the plaintiff cannot prevail. 49 U.S.C. § 42121(b)(2)(B)(iv); 29 C.F.R. § 1982.109(b); see also Rookaird I, 908 F.3d at 454, 459–60.

While McDonnell Douglas has provided courts with a default burden-shifting scheme to analyze discrimination claims, the FRSA burden-shifting framework is “much more protective of plaintiff-employees than the McDonnell Douglas framework,” because a claimant need only show that his protected activity was “a contributing factor” in any disciplinary action or termination, “not the sole or even predominant cause.” Araujo, 708 F.3d at 158 (citing 49 U.S.C. § 42121(b)(2)(B)(ii)); see also Greatwide Dedicated Transp. II, LLC v. U.S. Dept of Labor, 72 F.4th 544, 554 (4th Cir. 2023) (quoting Formella v. U.S. Dept of Labor, 628 F.3d 381, 389 (7th Cir. 2010) (internal citation omitted)) (the AIR-21 standard is “more favorable to the complaining employee”); see also McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973).

As the Supreme Court recently stated, “the contributing-factor burden-shifting framework is meant to be more lenient than most” employment discrimination burden-shifting frameworks in that it “is not as protective of employers as a motivating-factor framework.” Murray v. UBS Securities, LLC, 601 U.S. 23, 35, 39 (2024). “That is by design.” Id. at 39. “Congress has employed the contributing-factor framework in contexts where the health, safety, or well-being of the public may well depend on whistleblowers feeling empowered to come forward.” Id. We “cannot override that policy choice by giving employers more protection than the statute itself provides.” Id.

We turn to the district courts analysis of the substantive stage of Rookairds claim followed by BNSFs affirmative defense.

B. An FRSA Affirmative Defense at the Substantive Stage

1. Rookairds Protected Activity Was a Contributing Factor

At the outset, it is undisputed that Rookaird proved his substantive case by showing, by a preponderance of the evidence, that his refusal to stop conducting the air-brake test was a contributing factor to BNSFs decision to discharge him. See 49 U.S.C. § 42121(b)(2)(B)(iii); 29 C.F.R. § 1982.109(a).

The incorporation of a contributing factor standard into the FRSA “reflects a judgment that ‘personnel actions against employees should quite simply not be based on protected [whistleblowing] activities’—not even a little bit.” Murray, 601 U.S. at 36–37 (quoting Marano v. Dept of Justice, 2 F.3d 1137, 1141 (Fed.Cir. 1993) (internal citation omitted)) (cleaned up). The FRSA prohibits discharge or reprimand due even “in part” to an employees lawful, good faith refusal to violate a railroad safety law, rule, or regulation. 49 U.S.C. § 20109(a)(2). It is well-established under the FRSA that “[a] ‘contributing factor’ includes ‘any factor, which alone or in connection with other factors, tends to affect in any way the outcome of the decision.’ ” Rookaird I, 908 F.3d at 461 (quoting Gunderson v. BNSF Ry. Co., 850 F.3d 962, 969 (8th Cir. 2017)) (internal citation omitted).

The contributing factor standard was intended to overrule existing case law, which required a whistleblower to prove that his protected conduct was a “ ‘significant,’ ‘motivating,’ ‘substantial,’ or ‘predominant’ factor in the adverse personnel action ․” Murray, 601 U.S. at 28 (citation omitted). An FRSA contributing factor “may be quite modest.” Frost, 914 F.3d at 1197 (citing Rookaird I, 908 F.3d at 461). “The Supreme Court has interpreted similar language as creating liability if the employees action causes ‘even the slightest’ influence on the employers decision.” Yowell v. Admin. Review Board, U.S. Dept of Labor, 993 F.3d 418, 424 (5th Cir. 2021) (quoting CSX Transp., Inc. v. McBride, 564 U.S. 685, 692, 695–99 (2011)).

The record plainly supports the district courts conclusion that Rookairds refusal to stop the air-brake test contributed to his termination. The district court concluded, at the substantive stage, that Rookairds decision to proceed with the air-brake test “tend[ed] to affect in [some] way the outcome of [BNSFs] decision” to fire Rookaird, because Rookaird was fired “in part” for his inefficiency on February 23. The district court based its conclusion on the testimony of Jones, Gordon, and Fortt. Given the testimony evidence, as noted above, the district court concluded: “BNSF concedes that the crews inefficiency was partly caused by Mr. Rookairds decision to conduct an air test.” “Because Mr. Rookaird was fired for his inefficiency, ․ the air test ‘tend[ed] to affect in [some] way the outcome of [BNSFs] decision to fire Mr. Rookaird,” making the air-brake test a “contributing factor” under the FRSA. Cf. Rookaird I, 908 F.3d at 461.

Under the FRSA, the district courts contributing factor findings and conclusions affect BNSFs ability to prove at the affirmative defense stage that it would have terminated Rookaird absent the air-brake test.

2. BNSFs Affirmative Defense

Our review proceeds to the next and final step of the burden-shifting framework, which is our primary question in this appeal: whether BNSF proved, by clear and convincing evidence, that it would have taken the same unfavorable action against Rookaird absent his refusal to stop the air-brake test. The district court concluded that, “though the air test was a contributing factor in Mr. Rookairds termination, ․ the test contributed very little.” Despite acknowledging that Rookairds refusal to stop the test contributed in part to the decision to terminate him, the district court nonetheless determined “that BNSF is not liable for unlawful retaliation under the [FRSA].” That conclusion can stand only if BNSF was entitled to its affirmative defense.

In the FRSA context, “[u]nder the AIR-21 standard, [a plaintiff] would be entitled to relief even if [the protected activity] played only a very small role in [the employers] decision-making process.” Frost, 914 F.3d at 1197. The final stage of the AIR-21 burden-shifting framework allows an employer to defeat a claim for unlawful retaliation under the FRSA if the employer proves, by clear and convincing evidence, that it would have taken the same unfavorable personnel action absent the protected activity. See 49 U.S.C. § 42121(b)(2)(B)(iv); 29 C.F.R. § 1982.109(b); Rookaird I, 908 F.3d at 454, 459–60. The clear and convincing standard is an “intermediate burden of proof” between a “preponderance of the evidence” and “proof beyond a reasonable doubt.” Araujo, 708 F.3d at 159 (citation omitted); see also OTR Wheel Engg, Inc. v. West Worldwide Servs., Inc., 897 F.3d 1008, 1020 (9th Cir. 2019) (internal citations omitted). The Supreme Court has described the “clear and convincing” standard as “highly probable,” such that the material offered “instantly tilt[s] the evidentiary scales” in the direction of the party providing the proof. Colorado v. New Mexico, 467 U.S. 310, 316 (1984) (cleaned up). In other words, an employer must meet a “steep burden” to prove an FRSA affirmative defense under the AIR-21 framework. See Araujo, 708 F.3d at 162. The burden is steep for FRSA defendants because, unlike defendants under other whistleblowing statutes that utilize the AIR-21 framework, an FRSA employer cannot discharge an employee for his protected activity even “in part.” See 49 U.S.C. § 20109(a).

We rely on the plain text of the FRSA and give the statutes words their “ordinary meaning.” See Araujo, 708 F.3d at 158 (cleaned up). “In part” ordinarily means relating to a portion or division of a whole, to some extent. See Part & In Part, Oxford English Dictionary (2d ed. 1989). Put another way, “in part” means any amount greater than zero. We will not adopt a reading of the FRSA that renders any of its words as surplusage. See Rookaird, 908 F.3d at 466 (cleaned up). The text of 49 U.S.C. § 20109(a) does not permit an employer to discharge an employee if that decision is based, to any extent, on the employees engaging in the protected activity. See Araujo, 708 F.3d at 158 (cleaned up).

Because the FRSA is the controlling statute, we hold that, in an FRSA action, the affirmative defense step in the substantive stage of the AIR-21 framework cannot disregard the FRSAs original inquiry about whether unfavorable action was due “in part” to the employees engaging in a protected activity. 49 U.S.C. § 20109(a)(2). Because we rely on the language of the FRSA, we limit our holding to FRSA actions. While the existence of an affirmative defense means that an employer can defeat FRSA liability in some circumstances, an employer faces a “steep burden,” see Araujo, 708 F.3d at 162, particularly when the protected activity plays a part—however small—within the contours of the employers adverse personnel decision. An FRSA employer may still be able to prove an affirmative defense where a plaintiff shows only a correlation between the protected activity and adverse action, but this is less likely where a factfinder determines that the protected activity caused an employers adverse action, even in part. This is true because an employers grounds for an unfavorable personnel action must be “independently significan[t]”—necessary and sufficient—apart from the employees engagement in the protected activity. See Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7). The employer cannot prevail unless it proves by clear and convincing evidence that the discriminatory action still would have occurred absent the protected activity. See id.; 49 U.S.C. § 42121(b)(2)(B)(iv); 29 C.F.R. § 1982.109(b).

While applying the FRSA affirmative defense standard, the district court reasoned that BNSF could still prove its affirmative defense because Rookairds refusal to stop the air-brake test “contributed very little” to BNSFs decision to terminate him. The proper inquiry, however, is not whether the protected activity “contributed very little” to the firing; the proper inquiry is whether BNSF would have fired Rookaird regardless of whether he had conducted an air-brake test. See 49 U.S.C. § 42121(b)(2)(B)(iv); 29 C.F.R. § 1982.109(b). Under the FRSA, the protected activity cannot contribute even “in part” to the employers termination decision, so the FRSA affirmative defense standard needs to proceed with an analysis about whether and how the termination decision would have occurred absent the protected activity, given that the protected activity cannot contribute to the employers adverse action decision even in part. See 49 U.S.C. § 20109(a)(2); cf. Frost, 914 F.3d at 1197; Murray, 601 U.S. at 28, 35–37, 39 (citations omitted).

Because we have not had much occasion to interpret what clear and convincing evidence an employer would need to provide to meet its affirmative defense burden under the FRSA, we look to our precedent and that of our sister circuits to discern what this district court and others should consider. As recognized by the Seventh Circuit in Brousil, the Administrative Review Board has provided “guiding factors” that factfinders can consider when deciding whether an employer would have taken the same adverse action against an employee absent any protected activity. Those factors include “how ‘clear’ and ‘convincing’ the independent significance is of the non-protected activity” and “the facts that would change in the ‘absence of’ the protected activity.” Cf. Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7). Put another way, the FRSA requires the employer to prove what it “would have done,” not merely what it could have done. Cf. id. As such, the FRSA tasks an employer with proving a counterfactual scenario, such that the factual findings “instantly tilt the evidentiary scales” in their direction. Colorado, 467 U.S. at 316.

When an employer presents circumstantial evidence to prove its FRSA affirmative defense, such evidence may include a lack of temporal proximity between the non-protected activity and the unfavorable action; a poor employee work history and disciplinary record; the consistent application of the employers policies; similarly situated comparators; and the appropriate proportionality of the unfavorable action to the basis for it. Cf. Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7); Van Asdale v. Intl Game Tech., 577 F.3d. 989, 1003 (9th Cir. 2009) (holding, without reaching the burden-shifting stage, that the employees “positive record” could lead a reasonable factfinder to decide that the protected activity contributed to the termination).

To this end, an employers claim that the plaintiff was “technically in violation of written rules” does not meet the affirmative defense burden without proof that the employer consistently applied the same or similar policies to similarly situated comparator employees. See Araujo, 708 F.3d at 162–63; see also Greatwide, 72 F.4th at 558 (rejecting FRSA affirmative defense without proof that employer consistently terminated employees for similar violations under a range of possible disciplines); BNSF Ry. Co. v. U.S. Dept of Labor, 816 F.3d 628, 640–41 (10th Cir. 2016) (rejecting FRSA affirmative defense when employer did not prove that it fired employees for similar violations); Pan Am Ry. v. U.S. Dept of Labor, 855 F.3d 29, 37 (1st Cir. 2017) (rejecting FRSA affirmative defense when employer did not prove that dishonesty in other instances was “of a similar character”); Fresquez v. BNSF Ry. Co., 52 F.4th 1280, 1307–08 (10th Cir. 2022) (rejecting FRSA affirmative defense when employer did not show consistent application of stand-alone dismissible policy); cf. Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7); Kuduk v. BNSF Ry. Co., 768 F.3d 786, 793 (8th Cir. 2014) (agreeing that BNSF proved its affirmative defense after presenting “uncontroverted evidence that it consistently enforced this policy”).

Whats more, an employers “shifting explanations” for its discipline call into question whether it would have administered the same discipline absent the protected activity. See Vieques Air Link, Inc. v. U.S. Dept of Labor, 437 F.3d 102, 110 (1st Cir. 2006); see also Greatwide, 72 F.4th at 559; BNSF Ry. Co., 816 F.3d at 641; Weatherford U.S., L.P. v. U.S. Dept of Labor, 68 F.4th 1030, 1041 (6th Cir. 2023).

An FRSA plaintiff can further show an “overwhelming” case of FRSA retaliation by presenting evidence that the employer tried to dissuade him from engaging in the protected activity. See BNSF Ry. Co., 816 F.3d at 640 (rejecting FRSA affirmative defense when supervisors tried to dissuade plaintiff from engaging in protected activity); cf. Araujo, 708 F.3d at 163.

Rookairds lack of disciplinary record, a tenuous offer of proof from BNSF that it applied its policies consistently to relevant comparators, and BNSFs efforts to dissuade Rookaird from performing the air-brake test raise questions about how BNSF could have proved that it would have terminated Rookaird absent the air-brake test in light of the district courts finding that BNSF conceded that the air-brake test contributed to its decision to discharge Rookaird. 6 Rookaird had no disciplinary history at BNSF before February 23, 2010, and BNSF provided no evidence that it had planned to terminate, investigate, or otherwise reprimand Rookaird before the events on February 23. BNSF showed that gross dishonesty and insubordination gave adequate stand-alone bases for dismissal after a single violation, but dismissal was discretionary and not a mandatory consequence for either violation. To show the consistent application of its policies, BNSF relied on one case with three comparators who were dismissed for willful dishonesty in their tie-up slips, but BNSF only notified Rookaird that the company was investigating and terminating him for “dishonesty,” not “willful” or “gross” dishonesty.

In any event, the district court did not base its conclusions on these comparators; it only considered that BNSF suspended and placed on probation the other two members of Rookairds crew for their failure to work efficiently on February 23. But Rookairds fellow crewmen make for poor comparators when they may very well have been impermissibly reprimanded for the same protected activity as Rookaird. The district court did not base its decision on BNSF showing comparators who were fired for insubordination. Rookairds supervisors, Fortt and Gordon, both tried to dissuade Rookaird from performing the air-brake test. Gordon concluded that the air-brake test delayed the crew and recalled the crew from their shift based on asserted inefficiency, and he did not state any other reason to believe that the crew worked inefficiently.

When we consider the propriety of an FRSA affirmative defense, we ask whether a district court considered such factors with the understanding that the protected activity cannot contribute to the employers adverse action decision even in part. See 49 U.S.C. § 20109(a)(2); 49 U.S.C. § 42121(b)(2)(B)(iv); 29 C.F.R. § 1982.109(b). See, e.g., Araujo, 708 F.3d at 162–63; Greatwide, 72 F.4th at 558–60; BNSF Ry. Co., 816 F.3d at 640–41; Pan Am Ry., 855 F.3d at 37; Fresquez, 52 F.4th at 1307–08; Weatherford U.S., L.P., 68 F.4th at 1041; cf. Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7); Kuduk, 768 F.3d at 793.

We vacate and remand the affirmative defense issue for the district court to consider whether BNSF met its burden to prove that the company would have terminated Rookaird absent his refusal to stop performing the air-brake test, given that the air-brake test cannot contribute even in part to a termination decision.

C. Evidentiary Rulings

Finally, Rookaird appeals two categories of evidentiary rulings: (1) the exclusion of certain testimony designations; and (2) the admission of BNSFs comparator evidence.

Under our abuse of discretion standard, we will not reverse an evidentiary ruling unless the decision is “beyond the pale of reasonable justification under the circumstances.” Boyd v. City and Cnty. of San Francisco, 576 F.3d 938, 943 (9th Cir. 2009) (quoting Harman v. Apfel, 211 F.3d 1172, 1175 (9th Cir. 2000)). In addition, “[a] party seeking reversal for evidentiary error must show that the error was prejudicial, and that the verdict was ‘more probably than not’ affected as a result.” Id. (quoting McEuin v. Crown Equip. Corp., 328 F.3d 1028, 1032 (9th Cir. 2003)) (cleaned up); see also Wagner, 747 F.3d at 1052.

Regarding the testimony designations, plaintiffs counsel argues that he relied upon prior assurances from the district court that such evidence would be permitted. The district court, however, had said that it would only consider admitting such evidence whenever plaintiffs counsel presented it, without previously ruling that any designation would be admitted. Plaintiffs counsel also delayed the presentation of his deposition designations and his explanation of witness unavailability that might have justified admitting the deposition testimony. The district court appropriately weighed the probative value of these designations and did so on a timeline that was reasonable given the timing of the proposed designations. Finally, the district court did admit the designated testimony of several of Rookairds key witnesses. The decision to exclude certain designated testimony by the district court is not “beyond the pale of reasonable justification under the circumstances.” See Boyd, 576 F.3d at 943 (quotation omitted). Moreover, plaintiffs counsel does not identify how these rulings unfairly prejudiced his case, as he must. See id. (quotation omitted). We review evidentiary rulings only for abuse of discretion, and we conclude that the district court did not abuse its discretion in these rulings. See Wagner, 747 F.3d at 1052.

Similarly, the district courts admission of the comparator evidence late in the bench trial was reasonable based on when plaintiffs counsel initially received this evidence and raised potential issues with it. As such, the district court did not abuse its discretion by admitting the comparator evidence. See Boyd, 576 F.3d at 943 (quotation omitted); Wagner, 747 F.3d at 1052.

We affirm the district courts evidentiary rulings.

IV. CONCLUSION

The FRSA prohibits an employer from taking an adverse action against an employee even “in part” based on their engagement with a protected activity. See 49 U.S.C. § 20109(a); 49 U.S.C. § 42121(b)(2)(B); Frost, 914 F.3d at 1197; Araujo, 708 F.3d at 162; Murray, 601 U.S. at 28, 35–37, 39. Because of this language, specific to the FRSA, the district court erred in its analysis of BNSFs proffered affirmative defense when it reasoned that Rookairds air-brake test could contribute to BNSFs decision to fire him so long as it was “very little.”

We AFFIRM the district courts evidentiary rulings, and we VACATE and REMAND the judgment for further proceedings consistent with this opinion based on the record that the district court developed at the bench trial.

Each party shall bear its own costs.

This is an easy appeal that warrants a short memorandum disposition affirming the district court. The only question before us is whether the courts factual findings are clearly erroneous; they are not. The majority opinion has distorted the case beyond recognition by misreading both the relevant statute and the district courts decision and by telling the district court to redo exactly what it already did right. Moreover, the majority opinions mangling of the law will affect a wide range of retaliation cases. I respectfully but emphatically dissent.

A. The District Courts Analysis Was Sound.

The Wendell H. Ford Aviation Investment and Reform Act for the 21st Century (“AIR-21”) describes a straightforward, well-understood set of substantive provisions for proving claims of retaliation. The plaintiff bears an initial burden to prove that protected activity was one reason that contributed, even in small part, to dismissal. 49 U.S.C. § 42121(b)(2)(B)(iii); Rookaird v. BNSF Ry. Co., 908 F.3d 451, 454 (9th Cir. 2018). If the plaintiff meets that burden, the statute provides that the defendant has an affirmative defense: proving by clear and convincing evidence that “the employer would have taken the same unfavorable personnel action in the absence of that behavior.” Id. § 42121(b)(2)(B)(iv); Rookaird, 908 F.3d at 460. In other words, even when protected activity contributed to a firing decision, the employer can “defeat the claim by demonstrating ‘by clear and convincing evidence that the employer would have taken the same unfavorable personnel action in the absence of [the protected activity].’ ” Frost v. BNSF Ry. Co., 914 F.3d 1189, 1197 (9th Cir. 2019) (alteration in original) (emphasis added) (quoting 49 U.S.C. § 42121(b)(2)(B)(iv)). The employer may not retaliate, even a little; but Congress chose to exempt the employer from liability if the firing would have happened anyway. 1

The reason why the AIR-21 standards are so well understood is that Congress chose to apply those same standards in a wide range of statutes, including Sarbanes-Oxley Act of 2002, 18 U.S.C. § 1514A(b)(2); Motor Vehicle and Highway Safety Improvement Act of 2012, 49 U.S.C. § 30171(b)(2)(B); Surface Transportation Assistance Act, 49 U.S.C. § 31105(b); FDA Food Safety Modernization Act, 21 U.S.C. § 399d(b)(2)(C); Consumer Product Safety Improvement Act of 2008, 15 U.S.C. § 2087(b)(2)(B); Energy Policy Act of 1992, 42 U.S.C. § 5851(b)(3); American Recovery and Reinvestment Act of 2009, Pub. L. No. 111-5, § 1553(c)(1)(B), 123 Stat. 115, 299; Criminal Antitrust Anti-Retaliation Act of 2019, 15 U.S.C. § 7a-3(b)(2); William M. (Mac) Thornberry National Defense Authorization Act for Fiscal Year 2021, Pub. L. No. 116-283, § 6314, 134 Stat. 3388, 4601 (amending 31 U.S.C. § 5323(g)(3)(A)); Taxpayer First Act, 26 U.S.C. § 7623(d)(2)(B). Most pertinent here is the Federal Railway Safety Act (“FRSA”).

The FRSA provides that a railroad carrier may not fire or discipline an employee because the employee refused to violate a railroad safety rule. 49 U.S.C. § 20109(a). The FRSA expressly incorporates the AIR-21 substantive provisions, commanding that any enforcement action “shall be governed under the rules and procedures set forth in section 42121(b), including ․ the legal burdens of proof set forth in section 42121(b).” Id. § 20109(d)(2)(A).

The district court conducted a trial in order to determine two questions: the initial substantive burden and the affirmative defense. The district court found that Curtis Rookaird met the initial burden of proving that air-brake testing—a protected activity—was one factor that contributed, albeit “very little,” to his dismissal. Parker v. BNSF Ry. Co., No. 2:14-cv-00176-RAJ, 2022 WL 897604, at *6–*7 (W.D. Wash. March 28, 2022). Because the “contributing factor” bar is so low, Rookaird established the elements of the claim. See Frost, 914 F.3d at 1197 (holding that “ ‘contributing factors’ may be quite modest” and may “play[ ] only a very small role” in the employers decision); Rookaird, 908 F.3d at 461 (“A ‘contributing factor’ includes any factor, which alone or in connection with other factors, tends to affect in any way the outcome of the decision.” (citations and some internal quotation marks omitted)). On appeal, neither party challenges the district courts first determination.

The district court next found that BNSF Railway Company proved its affirmative defense by presenting clear and convincing evidence that it would have fired Rookaird anyway, even if Rookaird had not engaged in the protected activity of testing the brakes. Parker, 2022 WL 897604, at *6–*7. That second finding is the only substantive issue on appeal. 2

The record amply supports the district courts finding. On February 23, 2010, Rookaird and two other employees performed work, including about half an hour of conducting air-brake tests on railcars. Id. at *1–*2. In the judgment of the supervisor, the crew took far too long, and the supervisor told the crew to come back to the depot. Id. at *2. Once the crew returned, after about five-and-a-half hours of work during which little was accomplished, the supervisor told the crew to clock out and go home. Id. The other two employees did as they were told. Id. at *4. Rookaird, however, dishonestly recorded his time of departure as 8:30 p.m., approximately half an hour after the actual time, and he did not sign his timesheet. Id. at *2. Rookaird also declined to follow the supervisors instruction to go home; instead, he started an argument with another employee, refused a second command to go home, and left only after receiving a third command to go home. Id. at *2–*3.

After considering all the evidence, the district court first explained that the air-brake testing “contributed very little” to BNSFs decision to fire Rookaird. Id. at *7. The air-brake testing accounted for only twenty to forty minutes of the crews five-and-a-half hours of inefficient work, no one told the crew to stop the air-brake testing, and air-brake tests were routine. Id. Unlike other violations, discussed below, inefficient work is not an independently dismissible violation.

In addition to the air-brake testings being only a small part of Rookairds inefficiency that day, “Rookaird was fired for many reasons unrelated to his inefficiency.” Id. at *6 (emphasis added). Specifically, BNSF fired Rookaird for lying on his timesheet and failing to sign his timesheet, violations of work rules that independently warranted dismissal, and BNSF fired Rookaird for twice disobeying orders to leave the premises and for causing a heated argument with a co-worker while he remained on site, also an independently dismissible violation. Id. at *6–*7. Indeed, both the general manager who decided to fire Rookaird and the Human Resources employee who reviewed the record and concurred in the firing decision testified that the dishonesty and insubordination justified Rookairds dismissal. 3 In sum, the testing comprised only about ten percent of the time that Rookaird and his crewmates worked inefficiently (which is not an independently dismissible offense anyway), and the testing had nothing at all to do with Rookairds dishonesty and insubordination (either of which is an independently dismissible offense).

The court additionally observed that BNSF imposed a much lesser sanction on the other two members of Rookairds crew. Id. at *7. Although those crew members, too, had worked inefficiently, they had not committed gross dishonesty or insubordination. The lesser punishment for the other crew members supports the conclusion that—consistent with BNSFs written policies—BNSF viewed Rookairds dishonesty and insubordination as the most egregious misconduct.

The record thus strongly supports—if not compels—the district courts finding that BNSF met its burden of proving the affirmative defense. I would affirm.

B. The Majority Opinion Significantly Errs.

The majority opinion errs in many significant ways, which I group as follows: (1) the opinion ignores the “clear error” standard of review; (2) the opinion crafts a new, confusing, nonsensical, and unsupported legal standard and pointlessly remands for the district court to apply that bizarre standard; and (3) the opinion commits a laundry list of additional errors.

1. The Opinion Ignores the “Clear Error” Standard of Review.

The only substantive issue on appeal is the district courts finding that BNSF would have fired Rookaird anyway, even if he had not tested the air brakes. In a range of contexts, we and other circuits consistently have held that whether an employer would have fired an employee anyway is a factual determination that is reviewed for clear error. See Clairmont v. Sound Mental Health, 632 F.3d 1091, 1108 (9th Cir. 2011) (holding that whether an employer “would have reached the same adverse employment decision even in the absence of the employees protected conduct” is “purely a question of fact” (brackets and internal quotation marks omitted)); see, e.g., Baloga v. Pittston Area Sch. Dist., 927 F.3d 742, 752 n.7 (3d Cir. 2019) (citing an earlier precedent for the rule that “whether the employer would have taken [an] action regardless” is a “question[ ] for the jury”); Koszola v. FDIC, 393 F.3d 1294, 1300 (D.C. Cir. 2005) (holding that the appellate court reviews “for clear error” “the district courts finding by clear and convincing evidence that the [employer] would have fired [the employee] regardless of any alleged protected activity”); Johnson v. Univ. of Cincinnati, 215 F.3d 561, 584 (6th Cir. 2000) (holding that whether the employer “would have terminated [the employee] in the absence of his protected conduct ․ is a question of fact for the jury to decide”); Bellaver v. Quanex Corp., 200 F.3d 485, 495 (7th Cir. 2000) (holding that whether the employer “would have fired [the employee] in the absence of discrimination” is a determination “best left in the hands of a jury”); Hall v. Marion Sch. Dist. No. 2, 31 F.3d 183, 193 (4th Cir. 1994) (holding that the determination “whether [the employee] would have been fired ‘but for’ her protected speech ․ is a factual one, and therefore, is not to be reversed absent clear error” (internal citation omitted)); Daniels v. Quinn, 801 F.2d 687, 689 (4th Cir. 1986) (holding that “whether the employee would have been discharged ‘but for’ [protected] speech” is a “classic motivational question [that] is one of fact”).

The majority opinion states at the outset that we review the district courts findings of fact for clear error, Op. at 15, but the opinion nowhere applies that standard to the district courts finding. “We review a district courts findings of fact following a bench trial for clear error, ․ and [we] will reverse only if the district courts findings are ․ illogical, implausible, or without support in inferences from the record.” Chaudhry v. Aragón, 68 F.4th 1161, 1171 (9th Cir. 2023) (citations and internal quotation marks omitted). The majority opinion never mentions the “illogical, implausible, or without support” standard and never explains how the district courts amply supported factual finding contravenes that standard. Instead, the opinion simply ignores the applicable standard of review.

2. The Opinion Manufactures Legal Error by Announcing a Nonsensical New Legal Standard.

Apparently recognizing that no clear error exists, the majority opinion manufactures legal error, suggesting that the district court misunderstood the applicable legal standard. The majority opinion is plainly mistaken. The district court fully comprehended and faithfully applied the correct legal rule.

The statutory scheme is not complicated. Both as a matter of logic and of statutory text, the affirmative defense arises only if the plaintiff first proves that protected activity contributed, even a little, to the employers firing decision. The statute refers in earlier sub-parts to behavior that is protected and then queries, in the affirmative defense, whether the employer would have fired the employee “in the absence of that behavior.” 49 U.S.C. § 42121(b)(2)(B)(ii), (iv) (emphasis added). In other words, we reach the affirmative defense only after a plaintiff has proved that the employer impermissibly considered protected conduct. The finding of a contributing factor is the necessary predicate for the affirmative defense, not some smoking gun that disproves or discredits the affirmative defense (especially where, as here, the district court found that the protected conduct contributed only very little to the firing decision).

The employer then has the burden of proving by clear and convincing evidence that it would have fired the employee anyway, if the employee had not engaged in protected activity. As the majority opinion correctly recognizes, this is a hypothetical, “counterfactual” inquiry that asks what the employer would have done had the protected activity not occurred. Op. at 25. Here is how we have described the rule in more formal terms: Even when protected activity contributes to a firing decision, the employer can “defeat the claim by demonstrating by clear and convincing evidence that the employer would have taken the same unfavorable personnel action in the absence of the protected activity.” Frost, 914 F.3d at 1195 (brackets omitted) (citation and internal quotation marks omitted).

That is precisely the question that the district court asked, and answered, here. The court could not have been clearer. It stated the correct legal standard three separate times. “An employer ․ can defeat the plaintiffs claim if the employer demonstrates, by clear and convincing evidence, that the employer would have taken the same unfavorable personnel action in the absence of the protected activity.” Parker, 2022 WL 897604, at *5 (internal quotation marks omitted). “[A]n employer may defeat the retaliation claim if it can demonstrate by clear and convincing evidence that it would have taken the same unfavorable action absent the protected activity.” Id. “An employer can defeat a claim for unlawful retaliation if it can prove, by clear and convincing evidence, that the employer would have taken the same unfavorable personnel action in the absence of the protected activity.” Id. at *6 (internal quotation marks omitted). It twice summarized the pertinent question here in light of that legal standard: “whether BNSF could prove, by clear and convincing evidence, that it would have fired Mr. Rookaird absent the air test.” Id. at *1, *5. And it twice applied that standard in reaching its conclusion: “The Court concludes, by clear and convincing evidence, that absent the air test BNSF would have still fired Mr. Rookaird.” Id. at *6. “In all, the Court forms the ‘abiding conviction’ that even if Mr. Rookaird did not engage in the protected activity of refusing to stop the air test, BNSF would have still fired him for his gross dishonesty and insubordination. Thus, the Court concludes that BNSF has successfully proved its defense by clear and convincing evidence.” Id. at *7 (internal citation omitted).

The majority opinion disregards the obvious fact that the district court fully appreciated and applied the correct legal rule. Instead, the majority opinion announces a new, confusing, nonsensical, and unsupported legal standard. The majority opinion then vacates the district courts decision and remands for reconsideration under the baffling new standard. I strongly disagree.

The majority opinion reasons as follows. Unlike some statutes, the wording of the FRSA includes the provision that an employer may not retaliate “in whole or in part” against an employee. 49 U.S.C. § 20109(a). According to the majority opinion, the “in part” wording affects not only the employees substantive case, but also the employers affirmative defense. According to the majority opinion, “the FRSA affirmative defense standard needs to proceed with an analysis about whether and how the termination decision would have occurred absent the protected activity, given that the protected activity cannot contribute to the employers adverse action decision even in part.” Op. at 24 (emphasis added). We must consider the affirmative defense “with the understanding that the protected activity cannot contribute to the employers adverse action decision even in part.” Op. at 28. This nebulous standard is wrong in so many ways, it is hard to know where to begin. 4

Fundamentally, the majority opinion conflates the employees substantive burden with the employers affirmative defense. As we and other courts consistently have held, the affirmative defense asks simply what the employer would have done in the absence of the protected activity. That counterfactual inquiry assumes that the protected activity did not occur and asks what decision the employer would have made. In considering the affirmative defense, the legal standard that the employee must meet to prove the substantive case passes out of the picture and is simply irrelevant. The majority opinion cites a long list of cases that purportedly support its novel rule. But not a single case supports its new rule. As discussed in more detail below, no case gives the “in part” wording any special import at all. And no case comes anywhere near suggesting that the “in part” wording has any effect whatsoever on the affirmative defense. Instead, as we and other circuits have repeatedly held, the affirmative defense asks simply whether the employer would have fired the employee in the absence of the protected activity. See supra note 4 (listing cases).

The reason why the “in part” wording has no meaningful effect is that the AIR-21 standard already encompasses the fact that an employee has only a minimal burden of proving retaliation. As we put it succinctly in Frost, “[u]nder the AIR-21 standard,” an employee can meet the contributing-factor standard “even if [the protected activity] played only a very small role in BNSFs decision-making process.” 914 F.3d at 1197 (emphasis added). The substantive standards here are no different than in any other case applying the AIR-21 standards. Not a single case has held that the “in part” wording has any effect whatsoever on either the employees substantive case or the affirmative defense.

Finally, the majority opinions new standard is nonsensical. The affirmative defense asks simply what the employer would have done if the factfinder assumes that the protected activity never had occurred. Because the factfinder must assume that the protected activity does not exist at all, it makes no sense for the factfinder also to keep in mind the “understanding” that the employer cannot fire the employee for protected activity, even in part. Op. at 28. In considering the hypothetical world in which the employee did not engage in protected activity, the answer is either “the employer would have fired the employee anyway” or “the employer would not have fired the employee.” Whatever answer the factfinder reaches, the employees burden at the initial substantive step cannot possibly affect that answer.

I do not envy the district court in this case in reconsidering its decision. I see no reason why the district court could not simply reinstate its original order, adding the notation that it has done so while bearing in mind that an employer cannot retaliate even in part. The law ordinarily does not suffer such irrelevant formalities. Nor do I envy juries in future cases, or our court in future appeals, in applying this nonsensical legal standard, which is premised on a snippet of statutory text that relates only to the substantive case, not to the affirmative defense.

The majority opinions reason for relying on that particular snippet of statutory text is a transparent attempt to limit the damage that its opinion does to our caselaw. The opinion specifically limits its holding to FRSA cases. Op. at 23. But its rule clearly applies more broadly, to all cases involving the AIR-21 standards, because those standards already encompass the minimal level of retaliation that an employee must show.

3. The Majority Opinion Commits a Laundry List of Additional Errors.

The majority opinions flailing analysis also commits a laundry list of additional errors. I address a few of them here.

a. The majority opinion mentions, no fewer than six times, BNSFs purported concession as to the contributing-factor determination. Op. at 4–5, 13 (twice), 21, 27. The opinion describes that purported concession as “particularly” important and a “finding of particular note,” and the opinion implies that the concession makes it harder for BNSF to prove the affirmative defense. Id. at 5, 13. Those statements are wholly illogical and unprecedented. The affirmative defense kicks in only if the plaintiff establishes an improper contributing factor. The majority opinion cites no legal support, and none exists, for the proposition that it matters how the plaintiff establishes that fact—whether by “concession,” by a ruling at summary judgment, by stipulation, by a finding by the trier-of-fact, or otherwise. 5

b. As another example, BNSF concluded that Rookaird was dishonest in recording his sign-out time as 8:30 p.m., because he actually signed out at 8:02 p.m. The opinion implies that Rookaird was not in fact dishonest because the sign-out time was “within the thirty-minute grace period permitted by BNSFs policies and practices.” Op. at 8. The district court found to the contrary, Parker, 2022 WL 897604, at *6–*7, and the majority opinion never explains why that finding was clearly erroneous. Moreover, the majority opinion misunderstands the “grace period.” Nothing in the record suggests that an employee could add 30 minutes to their workday for no reason at all. The policy appears to have set a goal for employees to log out for the day within 30 minutes of completing actual work on the railcars, presumably because, as was true with the work that Rookairds crew performed on the day in question, the actual labor sometimes was not near the computer system. A similar policy applied to arrival times: an employee was expected to be doing actual work on the railcars, or elsewhere, within 30 minutes of signing in. But nothing in the record, nor in common sense, suggests that employees could add up to 60 minutes of work time just because a computer system allowed it.

To the contrary, every BNSF employee who testified on the subject testified that Rookaird had no justification for adding 28 minutes to his sign-out time and that, accordingly, he violated company policy. Rookaird himself testified that he had no justification for adding the time, other than the fact that the computer system allowed it. The other two members of Rookairds crew, who received the same command to sign out, accurately logged their sign-out times. Evidence in the record shows that BNSF had disciplined other employees for adding time to their workdays without justification. The majority opinions statement that Rookairds adding of 28 minutes to his workday was “within the 30-minute grace period permitted by BNSFs policies and practices,” Op. at 8, is wholly unsupported by the record. There is no question that Rookaird was dishonest in completing his timesheet.

c. The majority opinion implies that the district court misunderstood the legal standard by deciding only whether BNSF “could have” fired Rookaird. Op. at 5, 25. The majority opinion implies that the court did not determine whether BNSF “would have” fired Rookaird. Id. I am baffled by the suggestion. As described in detail above, the district court stated the legal standard eight times and, in all eight instances, the court used the correct “would have” formulation. Parker, 2022 WL 897604, at *1, *5 (thrice), *6 (twice), *7 (twice). Nothing suggests that the court misunderstood the legal standard.

d. The majority opinions discussion of an employers efforts to dissuade an employee from engaging in protected activity is wrong on both the facts and the law. The majority opinion states—as fact—that BNSF attempted to dissuade Rookaird from engaging in protected activity. Op. at 27–29. But the district court made no such factual finding, and the majority opinion relies on its own inferences from ambiguous testimony to reach that conclusion. The majority opinion plainly steps outside its role as reviewing court and impermissibly acts as finder of fact.

The majority opinion compounds the factual error by committing legal error. The opinion quotes BNSF Railway Co. v. United States Department of Labor, 816 F.3d 628, 641 (10th Cir. 2016), for the proposition that an employee “can ․ show an ‘overwhelming’ case of FRSA retaliation by presenting evidence that the employer tried to dissuade him from engaging in the protected activity.” Op. at 27. But BNSF Railway used the word “overwhelming” only in a parenthetical; to a “cf.” citation; describing another case as having “less than overwhelming” facts. BNSF Railway, 816 F.3d at 641 (emphasis added). Retaliation cases are inherently fact-specific, and no one factor necessarily controls. Contrary to the majority opinions implication, there is no rule that any time an employer attempts to dissuade an employee, the employee necessarily has an “overwhelming” case. And this entire topic is irrelevant here because the district court did not find that BNSF engaged in dissuasion.

e. The majority opinion next appears to fault the district court for reasoning, in part, that the air-brake testing contributed very little to BNSFs firing decision. As described above, the district court found that BNSF established the affirmative defense for a combination of reasons: (1) the protected activity contributed very little to the firing decision; (2) wholly apart from the protected activity, Rookaird engaged in egregious conduct constituting several independently dismissible violations; and (3) BNSF imposed lesser sanctions on employees who did not engage in that same unprotected conduct. The majority opinion focuses on the first reason only and states: “The proper inquiry, however, is not whether the protected activity ‘contributed very little’ to the firing; the proper inquiry is whether BNSF would have fired Rookaird regardless of whether he had conducted an air-brake test.” Op. at 24. The district court plainly asked the proper question whether BNSF would have fired Rookaird anyway; it did not inquire solely into whether the protected activity contributed very little. To the extent that the majority opinion reads the district courts opinion as relying solely on that one factor, the majority opinion is clearly wrong.

Nor was there anything improper about the district courts considering whether the protected activity played a large or a small role in the decision to fire the employee. As a matter of common sense, the role that the protected activity played in the firing decision bears directly on the credibility of an employers explanation that it would have fired the employee in the absence of the protected activity. For example, if the protected activity was the centerpiece of a firing decision, an employer will have a much harder time convincing a finder of fact that it would have fired the employee anyway. Or, as here, if the protected activity played only a small role and the nonprotected conduct was egregious, then the employers “we would have fired him anyway” explanation has more credibility.

Nothing in the law suggests that a factfinder must disregard the logically salient factor of the role that the protected activity played in the firing decision. The majority opinion cites no support for that unprecedented, irrational rule, and the rule runs counter to the same decisions that the majority opinion cites favorably. See, e.g., Brousil, 43 F.4th at 812 (“[T]he [Administrative Review Board] has admonished factfinders to ‘holistically consider any and all relevant, admissible evidence’ ․ [and] no circuit court or later ARB decision has limited the factfinders ability to look at all relevant evidence.”).

f. The majority opinion misunderstands the employees initial substantive burden of proving that protected activity must have contributed to the firing decision. The opinion states, without citation or other support: “An FRSA employer may still be able to prove an affirmative defense where a plaintiff shows only a correlation between the protected activity and adverse action, but this is less likely where a factfinder determines that the protected activity caused an employers adverse action, even in part.” Op. at 23–24. An employee must show that protected activity actually contributed to the firing decision; merely showing that some activity correlated with the firing decision will not suffice. Rookaird, 908 F.3d at 462–63. The majority opinions mischaracterization of the initial substantive burden fails to respect the law of the case and will cause yet more confusion in this—until now—settled area of law.

I dissent.

FOOTNOTES

. In our analysis, we noted that sufficient evidence supported the jurys finding that Rookaird refused “in good faith” to violate a railroad safety rule or regulation by completing the air-brake test. Rookaird I, 908 F.3d at 456.

. Rookaird also appeals the district courts denial of his Fed. R. Civ. P. 59 motion to alter or amend the judgment, or in the alternative for a new trial. In view of our decision, stated hereafter, we need not reach Rookairds challenge to the denial of his Rule 59 motion.

. Library of Congress, Today in History – February 28, LOC Digital Collections, at https://perma.cc/Z6SR-6TFG.

. See Frank J. Mastro, Preemption is Not Dead: The Continued Vitality of Preemption Under the Federal Railroad Safety Act Following the 2007 Amendment to 49 U.S.C. § 20106, 37 Transp. L. J. 1, 2 (2010).

. See Daniel Gilbert and Tom McGinty, Brake-Related Failures Dog Freight Railroads, Wall St. J., July 16, 2013, at https://perma.cc/27EL-EM6L; see also CSX Train Derailment with Hazardous Materials Release, Hyndman, Pennsylvania, August 2, 2017 vii, 12 (N.T.S.B. Nov. 23, 2020), at https://perma.cc/5U7Q-T2EP; Katherine Shaver, Bad Brakes Caused Derailment, Wash. Post, Mar. 6, 2002, at https://perma.cc/XMD3-3Z3L; Ian Austen, A Decade After a Deadly Derailment, Some Wonder if Canadas Railroads Are Safe, N.Y.Times, Feb. 24, 2023, at A12.

. The non-protected activity and the adverse action occurred in close temporal proximity, as did the protected activity. See Brousil, 43 F.4th at 812 (citing Speegle, ARB No. 13-074, 2014 WL 1870933, at *7) (cleaned up); Van Asdale, 577 F.3d. at 1003 (quotation omitted); Vieques Air Link, 437 F.3d at 108–09 (concluding that the employer failed to prove by clear and convincing evidence that the employees protected activity was unconnected to the adverse personnel action where the action “followed almost immediately on the heels of reports [plaintiff] made about [ ] safety violations”); cf. BNSF Ry. Co., 816 F.3d at 639–41 (rejecting FRSA affirmative defense where investigation and termination commenced within a month of the protected activity). Because Rookairds protected and non-protected activity both occurred proximate to his termination, this factor may not counsel strongly one way or the other toward BNSFs affirmative defense.

. The distinction between the substantive case and the affirmative defense can be explained using a simple analogy. Suppose that I am hosting a dinner party, and I plan to serve lasagna as the main dish. On the morning of the party, I realize that I am out of lasagna noodles and, although I have many after-dinner teas, I do not have one of my favorite teas to serve. I am pressed for time, but I decide to go to the grocery store. Both facts—lack of pasta and lack of a favorite tea—may have contributed to the decision to go shopping. But if I am asked after the fact, I would report that, even if I had stocked plenty of my favorite tea, I would have gone to the store anyway. After all, the most critical need for the dinner party was the main dish. The lack of tea contributed to my decision to go to the store, but I would have gone to the store regardless of my tea supply. Similarly, even when protected activity contributed to a firing decision, the employer escapes liability if the firing would have happened anyway. Nothing required Congress to provide this affirmative defense, of course, but we must apply the statute that Congress enacted.

. I agree with the majority opinion that the evidentiary challenges fail. I also follow the majority opinions lead in referring to the plaintiff as “Rookaird.”

. The Human Resources manager testified at the first trial, and the transcript of his testimony was admitted at the second trial. The general manager testified, in person, at the second trial, and the district judge observed that testimony firsthand. The district judge, as trier of fact and assessor of credibility, was entitled to believe him. See, e.g., United States v. Ginn, 87 F.3d 367, 369 (9th Cir. 1996) (holding that “we must respect the exclusive province of the trier of fact to determine the credibility of witnesses” (brackets and internal quotation marks omitted)).

. The majority opinion cannot mean that, any time protected activity contributed a little to the firing decision, the affirmative defense is unavailable. That ruling would nullify the affirmative defense and contravene our caselaw and the law of all other circuits. E.g., Frost, 914 F.3d at 1197; Pan Am Rys., Inc. v. U.S. Dept of Labor, 855 F.3d 29, 36 (1st Cir. 2017); Metro-North Commuter R.R. Co. v. U.S. Dept of Labor, 886 F.3d 97, 106 (2d Cir. 2018); Wiest v. Tyco Elecs. Corp., 812 F.3d 319, 329 (3d Cir. 2016); Greatwide Dedicated Transp. II, LLC v. U.S. Dept of Labor, 72 F.4th 544, 553–54 (4th Cir. 2023); Yowell v. Admin. Review Bd., U.S. Dept of Labor, 993 F.3d 418, 422 (5th Cir. 2021); Weatherford U.S., L.P. v. U.S. Dept of Labor, Admin. Bd., 68 F.4th 1030, 1040 (6th Cir. 2023); Brousil v. U.S. Dept of Labor, Admin. Review Bd., 43 F.4th 808, 812 (7th Cir. 2022); BNSF Ry. Co. v. U.S. Dept of Labor Admin. Review Bd., 867 F.3d 942, 945 (8th Cir. 2017); Fresquez v. BNSF Ry. Co., 52 F.4th 1280, 1296 (10th Cir. 2022). That ruling also would make no sense in light of the majority opinions decision to remand, because it is undisputed that the air-brake testing contributed to the firing decision.

. The purported “concession” at issue here is not a formal concession by a party. To the contrary, the court conducted a trial to rule on this disputed issue of fact. Parker, 2022 WL 897604, at *1. In its post-trial decision, the district court cited testimony by BNSF employees that they considered the air-brake testing, and the court wrote that BNSF “concedes” the issue. Id. at *6. Read in context, the court meant only that BNSFs own employees testified that they considered the testing; the court clearly did not intend to give its statement any greater weight than that. Regardless, as explained in text, whether BNSF formally conceded the issue or simply lost at trial on the issue is irrelevant to assessing the affirmative defense. The key point is that the contributing-factor finding is merely a necessary predicate for the affirmative defense, not some critical fact that by itself discredits the defense.

GOULD, Circuit Judge:

PAUL PARKER v. BNSF RAILWAY COMPANY (2024)

Authorities cited

Opinion · public domain