Paul Ossmann v. Meredith Corporation

82 F.4th 1007

• Court: Court of Appeals for the Eleventh Circuit
• Decided: September 8, 2023
• Docket: 22-11462
• Status: Published

Opinion

USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 1 of 49

[PUBLISH]

In the

United States Court of Appeals For the Eleventh Circuit

____________________

No. 22-11462 ____________________

PAUL OSSMANN, Plaintiff-Appellant, versus MEREDITH CORPORATION,

Defendant-Appellee.

Appeal from the United States District Court for the Northern District of Georgia D.C. Docket No. 1:19-cv-03200-SDG ____________________ USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 2 of 49

2 Opinion of the Court 22-11462

Before JILL PRYOR and GRANT, Circuit Judges, and MAZE, ∗ District Judge. GRANT, Circuit Judge: Paul Ossmann was the Chief Meteorologist at CBS46, an Atlanta news station. But during his tenure, female colleagues raised repeated complaints that he engaged in inappropriate conduct and sexual harassment—including “compliments” about appearance, sexually charged language, requests for nude photos, and more. After several meetings with Ossmann did not stop the behavior, it became clear to local managers that he could no longer work at CBS46. The managers needed authorization from the station’s parent company to terminate his employment, so the local Human Resources Director moved the Ossmann file up the chain. She sent a termination request form to the corporate office explaining that Ossmann had violated the company’s sexual harassment policies; the form also included Ossmann’s race, the demographics of his colleagues, and identification of potential comparator employees who had engaged in similar conduct. Ossmann, who is white, alleges that he was terminated because of his race in violation of 42 U.S.C. § 1981. The sexual harassment justification, he says, was just a pretext. To survive

∗ The Honorable Corey L. Maze, United States District Judge for the Northern District of Alabama, sitting by designation. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 3 of 49

22-11462 Opinion of the Court 3

summary judgment, Ossmann needed to show that a reasonable jury could conclude that if he were not white, the station would not have terminated him. The district court found that he did not make that showing, and Ossmann asks us to overturn that conclusion. He notes that the station’s new meteorologist is a Hispanic woman, but mostly argues that the existence of race data on the corporate form means that he was fired because he was white. We cannot agree. The presence of race data in the local station manager’s termination request is not enough for any jury to reasonably conclude that Ossmann’s sexual harassment conduct, much of which he admitted, was pretext for the true reason for Ossmann’s firing—his race. We affirm the grant of summary judgment. I. Meredith Corporation hired Paul Ossmann in 2012 as a temporary weekend meteorologist for CBS46. Ossmann became the station’s Chief Meteorologist in mid-2017 and remained in that position until he was terminated less than two years later, in April 2019. During that time, Ossmann’s female co-workers repeatedly complained of his inappropriate behavior and sexual harassment. In April 2017—a few months before his promotion to Chief Meteorologist—a female meteorologist reported that Ossmann had repeatedly told her that she “cockblocked” him over a dispute about vacation scheduling and that he had a dream about them USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 4 of 49

4 Opinion of the Court 22-11462

having sex together. She also reported that Ossmann told another female employee that “his first three-way was with a black woman.” When Ossmann met with Human Resources Director Laurel Berenguer and his then-supervisor Frank Volpicella to discuss the complaint, he admitted to using the term “cockblocked,” but he denied making the other comments. Volpicella issued Ossmann a written warning letter for exercising “poor judgment,” which Ossmann signed to acknowledge receipt. In that letter, Ossmann was “advised that further incidents may result in additional disciplinary action, up to and including termination” and was reminded of Meredith’s “zero tolerance for behavior that could contribute to creating a hostile work environment.” 1 A little more than six months later, a female news producer also complained. She told HR Director Berenguer that Ossmann sent her “highly inappropriate” messages on Facebook. In those messages, Ossmann told her that he masturbated while thinking about her, that he wanted to have sex with her, and that he wanted

1 Meredith’s sexual harassment policy prohibited, among other conduct, “unwelcome sexual advances” and listed examples of inappropriate behavior that could qualify such as “unwelcome sexual jokes or innuendoes, sexual stories, sexual objects, sexual gestures, inappropriate sexual contact, leers, stares, whistles, and blocking a path or exit.” Ossmann adds—without supporting evidence—that these policies only apply to conduct occurring in the workplace. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 5 of 49

22-11462 Opinion of the Court 5

her to send him nude photos. When a meeting was called with Berenguer and Ossmann’s new supervisor, Steve Doerr, Ossmann admitted to sending the messages. He explained that he did so “in an attempt to enter into an off-duty personal relationship.” The meeting concluded with Ossmann apologizing to Berenguer and Doerr for making the female producer feel uncomfortable. The incident was memorialized in a written warning letter titled “Final Written Warning: Exercising Poor Judgment.” But, unlike the first letter, this one was not signed by Ossmann. The parties dispute whether he received it. Ossmann says he was unaware of the letter until it was produced in discovery for this litigation and adds that Doerr told him privately that his conduct did not violate company policy. For his part, Doerr says that he provided Ossmann “with a written warning for violating Meredith’s policy against sexual harassment.” Roughly a year and a half later, yet another female employee raised yet another complaint. She reported that after the news aired Ossmann pulled her aside and said: Not to be like uncle Joe [Biden], I wanted to let you know I look at you all the time. You’re so pretty, put together. I see you walk around and you carry yourself very well. You’re very attractive and that’s attractive to me. You don’t flaunt it. Don’t put it out there. You’re not all a selfie kind of person. You always look nice.

In the moment, she thanked Ossmann for the compliment and told him that she liked working with him. But afterward she reported USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 6 of 49

6 Opinion of the Court 22-11462

to HR Director Berenguer that the comment made her feel uncomfortable—to the point that she immediately called her husband and parents—and that she was disappointed in herself for the way that she handled the situation. She feared that she had set herself up for it to happen again. HR Director Berenguer and Supervisor Doerr again met with Ossmann to discuss the allegations. According to Berenguer’s contemporaneous notes, Ossmann admitted to making the comments, but did not “mean anything by his comments.” In his view, he was just paying his coworker a compliment because he thought that they “had that kind of relationship.” (Ossmann now denies that he admitted to making the comments.) Doerr reminded Ossmann that this was not the first time he had behaved inappropriately with female colleagues. Doerr also suspended him until Lyle Banks, the station’s General Manger, decided on how to proceed. Doerr and Banks reviewed the allegations and decided that based on Ossmann’s “pattern of violating Meredith’s policy against sexual harassment,” they “had no choice but to terminate” him. They believed that termination was necessary “to maintain a safe workplace free from sexual harassment.” Banks directed Berenguer to “put together a recommendation to submit to corporate for approval to proceed with termination.” Berenguer’s standard practice—based on training she received from Meredith’s corporate human resources USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 7 of 49

22-11462 Opinion of the Court 7

director—was to fill out a form called an “EEO Analysis.”2 According to Berenguer, the purpose of the form was to gather all the “relevant information in making a request for review for discharge.” The form was required for any discharge, job elimination, restructuring, or reorganization. The EEO Analysis states that the station requested to terminate Ossmann because he violated the company’s sexual harassment and hostile work environment policies. It described the “cockblocked” and inappropriate messaging incidents, as well as an earlier incident that Ossmann claims was too distant to justify his firing. The form did not describe the final incident because the station moved forward with termination immediately rather than adding it to his file. Banks and Doerr are listed as the decisionmakers. The form includes Ossmann’s race, sex, and age— which Ossmann says tainted the decisionmaking process—and asks whether “other employees [have] been in a similar circumstance and, if so, how was that handled and resolved?” Berenguer listed two employees terminated for violating the company’s sexual harassment policies and another who received a written warning for “conduct unbecoming a manager.” Below that question, the form includes a table titled “Comparables (if applicable).” That table requests the same basic information provided for Ossmann— race, sex, age, job title, salary, and supervisor, among other things.

2 Although there are obvious candidates for what this could stand for, the

record does not offer the answer. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 8 of 49

8 Opinion of the Court 22-11462

But Berenguer did not complete this table with the information for the employees who were also disciplined. Instead, she listed the other five members of the weather team, none of whom had been disciplined, because they had “similar jobs or the same job.” And on the next page, the form includes a statistical analysis on how Ossmann’s termination would affect the demographics of the news station generally and the comparable group—here the weather team—specifically. At the company’s Rule 30(b)(6) deposition, Berenguer repeatedly testified that this information was used to ensure that the company was “being equitable.” Or, putting it another way, she testified that it was used to make sure that the company was not “treating one person in that situation in that comparable group differently than others.” Berenguer emailed the station’s request to Kandis Bock, a Vice President of Human Resources at Meredith. The record is limited on what happened next. The email to Bock is not in the record; nor was she deposed. That said, we know that Bock provided the local station managers with authorization to terminate Ossmann. So four days after suspending Ossmann, Doerr notified him that he was being terminated. Three weeks later, the station replaced Ossmann with Jennifer Valdez, a Hispanic meteorologist. Valdez had been with the station longer than Ossmann, and he admits that she was qualified. Ossmann’s suit alleges race discrimination in violation of 42 U.S.C. § 1981 and breach of his employment contract’s for-cause USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 9 of 49

22-11462 Opinion of the Court 9

provision. 3 After the close of discovery, Meredith moved for summary judgment and the matter was referred to a magistrate judge, who recommended that the motion be granted. The district court adopted a modified version of the magistrate judge’s recommendation and granted summary judgment for Meredith. Ossmann now appeals the district court’s order. II. We review the grant of summary judgment de novo. McAlpin v. Sneads, 61 F.4th 916, 927 (11th Cir. 2023). “A grant of summary judgment will be affirmed if ‘there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.’” Furcron v. Mail Ctrs. Plus, LLC, 843 F.3d 1295, 1303 (11th Cir. 2016) (quoting Fed. R. Civ. P. 56(a)). There is a genuine dispute when “the evidence is such that a reasonable jury could return a verdict for the nonmoving party.” Id. (quotation omitted). The evidence is viewed in the light most favorable to the non- moving party and all reasonable inferences are drawn in that party’s favor. Id. at 1303–04. III. A. Ossmann contends that Meredith terminated him because of his race, in violation of 42 U.S.C. § 1981. “Section 1981 prohibits

3 Ossmann’s complaint also brought claims for disparate discipline and hostile

work environment, but he abandoned those claims below. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 10 of 49

10 Opinion of the Court 22-11462

intentional race discrimination in the making and enforcement of public and private contracts, including employment contracts.” Jenkins v. Nell, 26 F.4th 1243, 1249 (11th Cir. 2022) (quoting Ferrill v. Parker Grp., Inc., 168 F.3d 468, 472 (11th Cir. 1999)). The McDonnell Douglas burden-shifting framework generally applies to discrimination claims based on circumstantial evidence. Id. That framework first requires Ossmann to establish a prima facie case of intentional discrimination. Lewis v. City of Union City, 918 F.3d 1213, 1220–21 (11th Cir. 2019) (en banc). If he does so, he is entitled to a rebuttable presumption of intentional discrimination.4 Id. at 1222. To rebut that presumption, Meredith needs to produce evidence of a valid, nondiscriminatory reason for terminating him. Id. at 1221. If it meets that burden, the presumption of intentional discrimination drops out of the case and Ossmann must demonstrate that Meredith’s “proffered reason was merely a pretext for unlawful discrimination, an obligation that ‘merges with the plaintiff’s ultimate burden of persuading the factfinder that [he] has been the victim of intentional discrimination.’” Id. (quoting Texas Dep’t of Cmty. Affs. v. Burdine, 450 U.S. 248, 256 (1981)) (alterations adopted). Unlike a Title VII discrimination claim—where a lesser “motivating factor” standard sometimes applies—a § 1981 claim

4 The district court concluded that Ossmann made out a prima facie case by

showing that he belongs to a protected class, was qualified for the position, was terminated, and was replaced by someone outside of his protected class. That decision is unchallenged on appeal. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 11 of 49

22-11462 Opinion of the Court 11

requires proof that race was a but-for cause of a termination. Comcast Corp. v. Nat’l Ass’n of Afr. Am.-Owned Media, 140 S. Ct. 1009, 1013, 1017 (2020). This does not require Ossmann to prove that race was the exclusive cause of his termination, but it does require him to prove that but for his race he would not have been terminated. See United States v. Benjamin, 958 F.3d 1124, 1131–32 (11th Cir. 2020); Ziyadat v. Diamondrock Hosp. Co., 3 F.4th 1291, 1297–98 (11th Cir. 2021). So to survive summary judgment, Ossmann needs to show that a reasonable jury could conclude that had he not been white, he would not have been terminated. B. “When evaluating a charge of employment discrimination” we “focus on the actual knowledge and actions of the decision- maker.” Walker v. Prudential Prop. & Cas. Ins. Co., 286 F.3d 1270, 1274 (11th Cir. 2002). Ossmann urges that Bock—the corporate HR Vice President who approved his termination—was the final decisionmaker responsible for terminating him. Meredith disagrees, contending that Banks and Doerr—the local station managers who investigated the allegations and decided that they “had no choice but to terminate”—were the final decisionmakers. It may well be that the better view of this evidence is that the local station managers were the final decisionmakers. But because we must make all reasonable inferences in Ossmann’s favor as the non- moving party, we analyze his employment discrimination claim assuming that Bock was the final decisionmaker. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 12 of 49

12 Opinion of the Court 22-11462

Ossmann offers several arguments. He starts by contending that the EEO Analysis is direct evidence proving Bock’s intentional discrimination. Failing that, he argues that Meredith has failed to satisfy its burden at step two of the McDonnell Douglas framework of producing evidence of its valid, nondiscriminatory reason, and that, in any event, he has met his burden at step three of demonstrating pretext. Alternatively, he says he has presented a convincing mosaic of circumstantial evidence proving intentional discrimination and that he has shown a viable “cat’s paw” theory— under which Bock merely rubberstamped the racial animus of the station managers. We address each of his arguments in turn, but note at the outset that none succeed.5 C. Ossmann’s first set of arguments center on the EEO Analysis form. His opening volley is that the form is direct evidence of illegal discrimination. “Direct evidence of discrimination is evidence that reflects a discriminatory or retaliatory attitude correlating to the discrimination or retaliation complained of by the employee, and, if believed, proves the existence of a fact without inference or presumption.” Fernandez v. Trees, Inc., 961 F.3d 1148, 1156 (11th Cir. 2020) (quotations omitted) (emphasis added). This is a

5 Ossmann also appeals the district court’s order granting summary judgment

on his breach of contract claim. But he concedes that his contract claim is “derivative of the § 1981 termination claim.” Accordingly, because we affirm the grant of summary judgment on Ossmann’s § 1981 claim, we also affirm the grant of summary judgment on his breach of contract claim. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 13 of 49

22-11462 Opinion of the Court 13

“rigorous standard.” Damon v. Fleming Supermarkets of Fla., Inc., 196 F.3d 1354, 1359 (11th Cir. 1999). We accept “only the most blatant remarks,” such as “a management memorandum saying, ‘Fire Earley—he is too old,’” as direct evidence. Id. (quoting Earley v. Champion Int’l Corp., 907 F.2d 1077, 1081 (11th Cir. 1990)). “If the alleged statement suggests, but does not prove, a discriminatory motive, then it is circumstantial evidence.” Fernandez, 961 F.3d at 1156 (quotation omitted). The EEO Analysis does not meet—or even approach—this standard. The form listed several categories of information in a neutral fashion, including Ossmann’s race, as well as the race of the other employees. For this document to prove that Ossmann was terminated because of his race, we would need to infer that it treated his race as a negative factor and that had his race been different Bock would not have approved his termination. Setting aside whether these inferences are even plausible, they are plainly inferences. And where inferences are required, evidence is not direct. The EEO Analysis is circumstantial rather than direct evidence, so we proceed to Ossmann’s remaining arguments. Ossmann next contends that Meredith did not rebut the presumption of intentional discrimination created by his prima facie showing under McDonnell Douglas. To do so, Meredith needed to produce evidence showing that Bock had a legitimate, nondiscriminatory reason for approving Ossmann’s termination. Burdine, 450 U.S. at 254. This justification must be sufficient to “allow the trier of fact rationally to conclude that the employment USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 14 of 49

14 Opinion of the Court 22-11462

decision had not been motivated by discriminatory animus.” Id. at 257. As the district court recognized, the EEO Analysis—which says that the station terminated Ossmann for sexual harassment policy violations—is “evidence of what Bock received and reviewed with respect to Ossmann’s termination.” And the record includes Berenguer’s meeting notes describing underlying incidents of harassment, warning letters, and deposition testimony—all supporting Meredith’s contention that it fired Ossmann because of repeated incidents of sexual harassment and inappropriate comments. This evidence is sufficient to give Ossmann a fair opportunity to demonstrate pretext and tracks rebuttal evidence this court has accepted before. See Vessels v. Atlanta Indep. Sch. Sys., 408 F.3d 763, 770 (11th Cir. 2005); Bogle v. Orange Cnty. Bd. of Cnty. Comm’rs, 162 F.3d 653, 657–58 (11th Cir. 1998). It plainly satisfies step two of the McDonnell Douglas evidentiary framework. We disagree with Ossmann’s argument that our decision in IMPACT v. Firestone compels a different conclusion. 893 F.2d 1189 (11th Cir. 1990). We have described that case as establishing that the employer “must present specific evidence regarding the decision-maker’s actual motivations.” Walker v. Mortham, 158 F.3d 1177, 1181 n.8 (11th Cir. 1998). Here, the only information sent to Bock was the EEO Analysis, and even Ossmann argues that she made her decision based on the contents of that form. And as we explained in Vessels, so long as those “primarily responsible for making the decision[]”—here, the local station managers— USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 15 of 49

22-11462 Opinion of the Court 15

articulate a sufficient race-neutral justification, the employer meets its burden of production even if upper level supervisors who approve the decision have not articulated the exact basis for their approval. 408 F.3d at 770; see also IMPACT, 893 F.2d at 1193–94 (acknowledging that rebuttal evidence need not come from the actual decisionmaker). Like Vessels, this is not a case “where an upper-level manager overruled a subordinate manager’s recommendation or decision without explanation.” 408 F.3d at 770. Meredith has met its burden to rebut Ossmann’s prima facie case, which means the presumption of intentional discrimination drops out of the case. Finally, Ossmann says that the form means he has provided sufficient evidence to show pretext at the third stage of McDonnell Douglas. At this point, Ossmann needs to show that a reasonable jury could disbelieve Meredith’s nondiscriminatory reason— repeated incidents of sexual harassment—and instead conclude that, but for his race, Bock would not have approved the termination. Brooks v. Cnty. Comm’n, 446 F.3d 1160, 1163 (11th Cir. 2006). A “court merely uses the pretext inquiry to guide its determination of the ultimate issue at summary judgment.” Smith v. Lockheed-Martin Corp., 644 F.3d 1321, 1346 n.86 (11th Cir. 2011). The pretext inquiry “merges with the ultimate burden of persuading the court that she has been the victim of intentional discrimination,” so we ask whether there is a genuine dispute of material fact as to whether Meredith intentionally discriminated against Ossmann. Burdine, 450 U.S. at 256. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 16 of 49

16 Opinion of the Court 22-11462

Unlike most discrimination plaintiffs, Ossmann does not offer any evidence that employees of other races were not fired for repeated instances of sexual harassment. Ossmann’s only evidence that Bock intentionally discriminated against him, aside from the fact that he was replaced by a non-white woman, is the inclusion of race on the EEO Analysis form. 6 This form does not remotely approach the amount of evidence necessary for a reasonable jury to conclude that Ossmann was fired because of his race.

6 Ossmann argues that two more pieces of evidence are relevant. First, he says

that some incidents listed on the EEO Analysis were too old to serve as the basis for his termination, and that another was dated in the future. But the listing of previous incidents of sexual harassment does not demonstrate pretext—if anything it bolsters the conclusion that termination was appropriate. And no reasonable jury could conclude that the “future” event is anything but a typographical error given the substantial evidence showing that the incident happened on the same date two years before the date listed in the analysis. Second, Ossmann argues that an inference of pretext can be drawn from Meredith’s shifting explanations for the origin of the EEO Analysis. In its interrogatory responses, Meredith said that Banks directed Berenguer to prepare the EEO Analysis. But when (unsuccessfully) claiming attorney work- product privilege over the document, Meredith told the court that in-house counsel directed Berenguer to prepare it. From this, Ossmann says a jury could infer Meredith’s “consciousness of guilt.” Ossmann does not point to any authority for this inference. The authority that Ossmann does cite— United States v. Wilson—relied on a line of criminal cases recognizing that false exculpatory statements may be used to prove guilt. 788 F.3d 1298, 1311 (11th Cir. 2015) (citing United States v. Holbert, 578 F.2d 128, 129 (5th Cir. 1978)). Meredith’s explanation on the origin of the EEO Analysis was part of a routine assertion of attorney work-product privilege—it is neither a false exculpatory statement nor evidence of pretext. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 17 of 49

22-11462 Opinion of the Court 17

To start, the EEO Analysis form says on its face that Ossmann was terminated for multiple sexual harassment policy violations. In fact, it specifically describes three incidents of harassment. It also includes race data for the weather team and overall station statistics. Because of this, Ossmann contends that his race was “a negative factor” in the decision to terminate him. For its part, the dissent adds that the form required Bock to consider “racial balance” when making her decision and contends that a jury could infer that “Bock would have recommended that the lone black or Hispanic member be suspended rather than terminated to avoid racial imbalance.” Dissent at 4, 14, 17, 19, 23– 24. But “inferences in favor of a plaintiff can be based only on evidence—not on speculation.” Martin v. Fin. Asset Mgmt. Sys., Inc., 959 F.3d 1048, 1058 (11th Cir. 2020); see also Smith, 644 F.3d at 1328 n.25. Speculation is all we have here. We agree that we must infer, at this stage of the case, that Bock considered the data on the last attachment to the form. But considered it how? The evidence does not provide an answer. The language of the form does not tell Bock what she should do with the racial data, and it does not require her to engage in racial balancing—it is completely neutral.7 As was the only

7 The dissent points specifically to FAQ #10 in the EEO form as evidence that

Bock was required to engage in racial balancing. Dissent at 5–7, 11–12. But a closer look at the full text of FAQ #10, rather than the excerpts quoted by the dissent, reveals that the form imposes no racial balancing requirements. FAQ USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 18 of 49

18 Opinion of the Court 22-11462

testimony offered to explain the document. According to Berenguer, the purpose of the data was to make sure that the station was not “treating one person in that situation in that comparable group differently than others.” 8 Indeed, we find puzzling the dissent’s apparent inference that a 20-second pause before Berenguer described the company’s commitment to treating employees equally can be interpreted as substantive evidence of discrimination. See Dissent at 13–14. Particularly when she was being asked about how someone else used the form, and her only source of information was being “walked through” it 14 years earlier when she started her job at the station. No reasonable jury could conclude from the bare fact that this document includes data on the race of all weather employees at the station that Ossmann was fired because of his race. It is just

#10 asks: “In reviewing the existing and proposed organizational charts, is a particular protected category of employee being impacted by the restructure at a higher percentage rate than similarly situated non-protected employees? Conduct a Risk Analysis as appropriate.” Doc. 71-4 at 2. To start, by its terms this inquiry applies only in restructuring actions rather than for termination requests (though the station responded in any event). And one can understand why a reviewing authority would want to know whether a restructuring (read: layoffs) would impact a protected category of employees in a way that presented legal risk to the company. 8 We do not use oral argument to “ask attorneys to provide new evidence with

which to make our own findings of fact.” United States v. Campbell, 26 F.4th 860, 876 n.12 (11th Cir. 2022) (en banc). So unlike the dissent, we do not rely on counsel’s statements at oral argument for additional factual evidence about how the form may have been used. See Dissent at 15. Instead, we confine our review to the certified record. See Id. at 874. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 19 of 49

22-11462 Opinion of the Court 19

as likely (which is to say entirely speculative) that Bock used the race data in favor of Ossmann. We have no evidence either way. And we do not know what Bock would have done had Ossmann not been white because Ossmann does not have evidence that Bock treated employees of other races more favorably. If he did, a jury may have been able to make the inference that race was a but-for cause of his termination. But to infer that without any other evidence is nothing but speculation. On this record, a reasonable jury could not—by a long shot—conclude that Ossmann’s race was a but-for cause of his termination. Our decision in Smith v. Lockheed-Martin Corp. demonstrates the point. 644 F.3d 1321 (11th Cir. 2011). There, the company had fired seven white employees who forwarded a racially offensive email but had not fired black employees for similar conduct. Id. at 1332, 1336. Reversing the district court’s grant of summary judgement in favor of the company, we pointed to three pieces of evidence from which a reasonable jury could infer intentional discrimination. First, there was evidence that the same decisionmaker had discriminated against other white employees in similar (but distinct) investigations into the distribution of racist emails occurring around the same time. Id. at 1341–44. Second, there was evidence that because of a tragic incident at a Lockheed- Martin facility in Texas—a racially based shooting that had received a massive amount of media attention—Lockheed-Martin was under intense pressure to “emphatically prove” that the company was committed to curbing racism against black employees by USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 20 of 49

20 Opinion of the Court 22-11462

ensuring that all white employees engaging in racist conduct were fired. Id. at 1329, 1344–45. Third, there was evidence that a disciplinary review committee used a decision matrix that detailed each investigated employee’s conduct and included a column listing each employee’s race. Id. at 1336. After reviewing the first two categories of evidence, we said that this matrix “strengthen[ed] the reasonableness of the inference” that the decisionmaker “sought to fire all whites who distributed racist emails.” Id. at 1345–46. In light of the other evidence of racial discrimination, “Lockheed’s injection of race into its decision-making process yield[ed] an unavoidable inference that the employee’s race impacted the discipline determination.” Id. at 1346. And bolstering that inference, we pointed out, race was not tracked in an investigation into two black employees’ racist emails, which a jury could infer was “because it was already known that both employees to be disciplined . . . were black and, therefore, would not be terminated for their conduct.” Id. at 1346 n.87. Moreover, the company admitted that it had “no legitimate business purpose” in monitoring the employees’ race. Id. at 1346 n.85 (quotation omitted and alteration adopted). The dissent accuses us of avoiding the “unavoidable” inference that we recognized in Lockheed-Martin. Dissent at 4. But the reason that inference was unavoidable in Lockheed-Martin was the dramatic amount of additional evidence in that case. Here, we have none. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 21 of 49

22-11462 Opinion of the Court 21

The more relevant point from Lockheed-Martin is its warning that an inference “is not a suspicion or a guess. It is a reasoned, logical decision to conclude that a disputed fact exists on the basis of another fact.” 644 F.3d at 1328 n.25. We heed that warning here. What Lockheed-Martin does not establish is that a reasonable inference of intentional discrimination is created any time race is included in a document used to facilitate an employment decision. Other evidence is needed to show that race was used in a discriminatory manner. In a different case, with evidence that it was used for improper reasons in a particular employment decision, a form containing racial data could move the needle. But without any evidence suggesting that the race information was a negative factor against Ossmann, or was used favorably in evaluating a non-white employee, an inference of intentional discrimination is not reasonable. Much less “unavoidable.” The fact that Ossmann was replaced by a non-white employee is not enough. Being replaced by someone outside one’s protected class can help to establish the prima facie case of discrimination for burden-shifting purposes. Maynard v. Bd. of Regents, 342 F.3d 1281, 1289 (11th Cir. 2003). But it is not enough to carry the day on the substantive question of discrimination. See Cuddeback v. Fla. Bd. of Educ., 381 F.3d 1230, 1236 (11th Cir. 2004). Indeed, there is no record evidence suggesting that Bock was even involved in selecting Ossmann’s replacement—which would be necessary to infer that she approved his termination so that she could replace him with someone who is not white. Because Ossmann cannot show that a reasonable jury could conclude that USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 22 of 49

22 Opinion of the Court 22-11462

Meredith intentionally discriminated against him—rather than fired him for his repeated violations of Meredith’s sexual harassment policies—his claim cannot survive summary judgment. D. Notwithstanding the complete lack of evidence of intentional discrimination in this case, the dissent relies on two cases where the defendants affirmatively sought to use race in their decisionmaking. Neither has any application here. The first is Ricci v. DeStefano, where a local fire department sought to intentionally discriminate against white firefighters to avoid disparate impact suits from non-white firefighters. 557 U.S. 557, 561–63 (2009). The Supreme Court held that the fire department needed a “strong basis in evidence to believe it will be subject to disparate-impact liability” before it could engage in intentional discrimination to avoid it. Id. at 585. Meredith has not argued that it engaged in intentional discrimination, so Ricci is not relevant. The second case is the recent decision rejecting the race- based admissions systems of Harvard and the University of North Carolina, Students for Fair Admissions, Inc. v. President & Fellows of Harvard College, 143 S. Ct. 2141 (2023), which the dissenting opinion rightly concedes is not binding on these unrelated facts. Dissent at 7–8. There, both colleges considered race at every stage of their admissions programs, openly admitted to doing so, and insisted that they be permitted to continue. Id. at 2154–2156, 2166. This case could not be more different. It is a run-of-the-mill § 1981 case without any evidence of intentional discrimination. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 23 of 49

22-11462 Opinion of the Court 23

Still, the dissent insists that SFFA helps “rebuff Meredith’s ‘negative factor’ argument,” because the Supreme Court rejected the universities’ argument that race never played a “negative factor” in any applicant’s admission decision. Dissent at 7, 18–19; see 143 S. Ct. at 2169. But in SFFA there was extensive evidence that race was a positive factor in some applicants’ decisions. SFFA, 143 S. Ct. at 2169. In a “zero-sum” game, therefore, race was a negative factor for other applicants. Id. The same conclusion cannot be drawn here; there is neither evidence that race has been used as a positive factor in other employment decisions nor evidence that Meredith treats its employment decisions as zero-sum. In sum, Ossmann lacks direct evidence of discrimination, he lacks evidence that Meredith treated his race as a factor favoring his termination, and he lacks evidence that Meredith treated similarly situated non-white employees more favorably. On the other hand, Meredith has produced extensive evidence of Ossmann’s sexual harassment, which is a valid, nondiscriminatory reason for his termination. On this record, no reasonable jury could infer that Meredith’s justification was pretext for race discrimination. E. Ossmann says that even setting aside the McDonnell Douglas steps, he can prove his claim by presenting “a convincing mosaic of circumstantial evidence that would allow a jury to infer intentional discrimination by the decisionmaker.” Smith, 644 F.3d at 1328 (quotation and footnote omitted). He is correct that the convincing mosaic metaphor offers an alternative to plaintiffs USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 24 of 49

24 Opinion of the Court 22-11462

unable to succeed through the McDonnell Douglas framework. See id.; Vessels, 408 F.3d at 768 n.3. The problem for Ossmann is that his evidence is neither convincing nor a mosaic. He offers no facts besides the EEO form and his replacement’s race, which we have already determined are not enough to show discrimination. And in any event, the mosaic theory’s relevance is highest for employees who cannot make out a prima facie case of discrimination. These employees may, for example, lack adequate comparators but otherwise have circumstantial evidence of discrimination. See Smith, 644 F.3d at 1328; Bailey v. Metro Ambulance Servs., Inc., 992 F.3d 1265, 1273 n.1 (11th Cir. 2021). That is not Ossmann. Plus, the convincing mosaic inquiry is identical to the final stage of the McDonnell Douglas framework: both ask whether there is enough evidence for a reasonable jury to infer intentional discrimination. Smith, 644 F.3d at 1326, 1328. We have already concluded that Ossmann’s circumstantial evidence fails to create a triable question of intentional discrimination. We say so again in rejecting his convincing mosaic argument. Ossmann also raises a “cat’s paw” theory of liability. That theory “provides that causation may be established if the plaintiff shows that the decisionmaker followed the biased recommendation without independently investigating the complaint against the employee.” Stimpson v. City of Tuscaloosa, 186 F.3d 1328, 1332 (11th Cir. 1999); see also Ziyadat, 3 F.4th at 1298. Still, the non-decisionmaker’s racial animus must be a but-for cause of the termination. Ziyadat, 3 F.4th at 1298. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 25 of 49

22-11462 Opinion of the Court 25

For this theory, Ossmann retreats from his claim that Bock acted with discriminatory animus. He instead says that she “accepted the Atlanta managers’ recommendation without investigating” and that the Atlanta managers—Banks and Doerr— recommended his termination for discriminatory reasons. He argues that Doerr’s reasons for recommending Ossmann’s termination—repeated episodes of sexual harassment—were pretext because, Ossmann says, Doerr did not believe that his harassing conduct violated Meredith’s policies. But even if that were true (as unlikely as it may be) Ossmann does not argue the same for Banks, who independently reviewed the facts and directed Berenguer to submit the termination form for corporate approval. Because both acted together, the causal chain between Doerr and the ultimate recommendation to Bock is broken by Banks’s participation in the decision. For Doerr’s alleged racial animus to be a but-for cause of his termination, Ossmann would have had to argue either that Banks also acted with racial animus or that Banks—like Bock—was a mere rubberstamp. He has done neither here. * * * The ultimate question in any discrimination case is whether the defendant intentionally discriminated against the plaintiff based on race. Ossmann failed to show that a reasonable jury could conclude that Meredith terminated his employment because he was white. The district court’s order granting summary judgment for Meredith is therefore AFFIRMED. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 26 of 49

22-11462 Maze, J., dissenting 1

COREY L. MAZE, District Judge, dissenting: I agree with the majority that we must infer Kandis Bock considered Paul Ossmann’s race and the race of his coworkers when Bock approved Ossmann’s termination. Opinion at 19. But I think we must also make the next logical inference: if Bock considered race, then changing race would affect Bock’s decision. The majority says this inference lacks reason and logic. I’ll let you, the reader, decide whether they are right. Below are five facts. Assume the first four are true (I’ll show you the proof later), then ask yourself if it is reasoned and logical to infer that Bock would respond differently depending on the race of the meteorologist in Fact #5: 1. A local station wants to fire a male meteorologist for sexually harassing female coworkers. 2. Meredith corporate policy requires Bock to review and then approve or deny the termination request. 3. As part of her review, Bock must determine whether terminating the meteorologist would impact a minority racial group at a higher percentage rate than white employees. 4. To ensure Bock can conduct this review, corporate counsel orders the local station’s HR director to fill out a form that identifies the meteorologist’s race and the race of his coworkers, then chart each racial group by percentage. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 27 of 49

2 Maze, J., dissenting 22-11462

5. The resulting racial group chart tells Bock that: a. The meteorologist is white, and granting the termination request would increase the percentage of black and Hispanic members of the weather team from 17% to 20% each; or, b. The meteorologist is Black or Hispanic, and granting the termination request would decrease the percentage of his minority racial group from 17% to 0%, meaning his race is no longer represented on the weather team. Based on these facts, I think it is both reasonable and logical to infer that changing the race of the meteorologist would affect Bock’s decision. Otherwise, what’s the point of disclosing his race and the race of his innocent coworkers? The majority avoids answering this question by declaring that the EEO form is “completely neutral” and “does not tell Bock what she should do with the racial data.” Opinion at 19-20. Not so; the EEO form told Bock how to consider race. I’ll show you the form and what Berenguer said about the form below. When we view this evidence in the light most favorable to Ossmann, we must infer that Berenguer sent Bock racial group data so that Bock could conduct a race-based disparate impact review, and under that review, we must infer that Bock “would have responded differently” if the station sought to fire the lone black or Hispanic member of the weather team rather than one of team’s four white members. Comcast Corp. v. Nat’l Ass’n of African USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 28 of 49

22-11462 Maze, J., dissenting 3

American-Owned Media, 140 S. Ct. 1009, 1015 (2020); see also Bostock v. Clayton Cty., __ U.S. __, 140 S. Ct. 1731, 1739 (2020)(“a but-for test directs us to change one thing at a time and see if the outcome changes. If it does, we have found a but-for cause.”). Our precedent supports this inference. When faced with a similar race-conscious document, this Court has said that the inference that race mattered was “unavoidable,” and only a jury could decide what happened: On its face, the ‘matrix’ indicates that race was pertinent to the discipline decisions made, and Lockheed has not explained satisfactorily why this was legitimate. Therefore, although the district court entirely ignored this fact, Lockheed’s injection of race into its decision-making process yields an unavoidable inference that the employee’s race impacted the discipline determination, and it is a jury’s province to decide whether race actually bore on the decision to terminate Mitten. Smith v. Lockheed-Martin Corp., 644 F.3d 1321, 1346 (11th Cir. 2011) (footnotes omitted). If we stuck to Lockheed-Martin, the case would end here. We could cut-and-paste the conclusion and change just two names: [Meredith’s] injection of race into its decision-making process yields an unavoidable inference that the employee’s race impacted the discipline determination, and it is a jury’s province to decide USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 29 of 49

4 Maze, J., dissenting 22-11462

whether race actually bore on the decision to terminate [Ossmann]. Id. But the majority avoids Lockheed-Martin’s unavoidable inference, so we press on. I.

Defendants rarely confess wrongdoing, so plaintiffs like Ossmann often base their inferences on circumstantial evidence. That doesn’t make their inferences unreasonable. Imagine discovering your child reaching into a cookie jar. When you ask why his hand is in the cookie jar, your child says nothing. You didn’t see your child take a cookie. He did not admit that he was taking a cookie. Yet you can reasonably infer: He was taking a cookie!

A. Meredith’s hand is in the cookie jar. “Outright racial balancing is patently unconstitutional.” Students for Fair Admissions, Inc. v. President & Fellows of Harvard Coll., 143 S. Ct. 2141, 2172 (2023) (brackets and quotations omitted). Yet Ossmann offers documentary evidence that suggests Bock considered the impact on racial group balance when deciding whether to approve Ossmann’s termination. Corporate counsel told Laurel Berenguer that, before corporate would sign off on Ossmann’s termination, she needed to complete an EEO Analysis form and send it to Bock. Berenguer testified that the EEO Analysis is a standard form that Meredith requires local stations to send “to corporate to review or request USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 30 of 49

22-11462 Maze, J., dissenting 5

termination,” and Meredith admits that Berenguer sent Ossmann’s EEO Analysis form to Bock “in the usual course of business.” The form’s header confirms that corporate uses the EEO Analysis data to review termination requests:

(circle added). Later on the first page, the form instructs the person requesting a termination to:

FAQ #10 asks the requesting party to disclose whether “a particular protected category of employee” would be impacted by the decision “at a higher percentage rate than similarly situated non- protected employees”:

While FAQ #10 talks about protected groups “being impacted by the restructure,” Berenguer testified that corporate HR trained her to fill out the form for terminations, as well as restructures and reorganizations, and she had followed that policy for 14 years. Consistent with her training and practice, Berenguer responded to FAQ #10 by typing “Reference Risk Analysis,” which refers to this “Risk Analysis” chart that Berenguer created: USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 31 of 49

6 Maze, J., dissenting 22-11462

The green column disclosed the racial makeup of the whole news station. The yellow column disclosed the racial makeup of the weather team, who Berenguer identified by name, race, age, and sex on the previous page:

The gray column disclosed which racial group(s) would be diminished if Bock granted the request. These charts told Bock that terminating Ossmann would not impact minority employees (the African-American and Hispanic reporters who each comprised 1/6 of the weather team), “at a higher percentage rate than similarly situated non-protected employees” (the white employees who filled four of the six spots). They also told Bock that white employees formed the largest racial USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 32 of 49

22-11462 Maze, J., dissenting 7

group in the news department (49%) and the weather team (75%), so terminating a white employee would not negatively impact a protected racial group. The majority acknowledges that Bock considered this racial group data but concludes that no evidence supports how she considered it. Opinion at 19-20. But that conclusion ignores the obvious question: is there any reason Bock would consider the racial group makeup of the news station and the weather team other than to conduct the disparate impact review called for by FAQ #10? We don’t require jurors to check their common sense at the door. If the EEO form tells Bock how to use the racial group data, and Meredith offers no other explanation for sending Bock the racial group data, then a juror can reasonably find that Bock followed the form. B. Meredith says it adds racial group data to termination documents “to ensure equitable treatment of its employees.” And Title VII prohibits employment practices that disparately impact racial groups. 42 U.S.C. § 2000e-2(k). So why then is racial balancing problematic? Ossmann points to the Supreme Court’s recent statement that “[o]utright racial balancing is patently unconstitutional.” Students for Fair Admissions, 143 S. Ct. at 2172 (quotations omitted). That case, however, involved student admissions to college, not personnel decisions. So while Students for Fair Admissions helps Ossmann rebuff Meredith’s “negative factor” argument, see infra USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 33 of 49

8 Maze, J., dissenting 22-11462

Part II(C), its holding is not binding here. But Meredith is not out of the woods yet; there are two other reasons why conducting a disparate impact review was improper. 1. Conducting a group disparate impact analysis when deciding whether to terminate an individual employee flouts the Supreme Court’s decision in Ricci v. DeStefano, 557 U.S. 557, 583–84 (2009). In Ricci, white employees scored higher than minority employees on the New Haven Fire Department’s qualification exam, thus placing them higher in the pecking order for promotions. Citing group statistics, some minority candidates threatened to sue the City under a disparate impact theory unless the City discarded the test results, and some white candidates threatened to sue under a disparate treatment theory if it did. The City threw out the test results to avoid the minority candidates’ disparate impact claim. Despite the City arguing that it had to discard the test results to avoid a disparate impact lawsuit, id. at 579, the Supreme Court held that the City discriminated against high scoring white and Hispanic candidates. The Court recognized the conflict between Title VII’s individual-focused disparate treatment provision and its group-focused disparate impact provision. The Court resolved the conflict in favor of individuals by adopting this rule: “before an employer can engage in intentional discrimination for the asserted purpose of avoiding or remedying an unintentional disparate impact, the employer must have a strong basis in evidence to believe it will be subject to disparate-impact liability if it fails to take USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 34 of 49

22-11462 Maze, J., dissenting 9

the race-conscious, discriminatory action.” Id. at 585. The Court found that the City faced no possibility of disparate impact liability if it accepted the exam results because (a) the exams were job- related and consistent with a business necessity and (b) the City had not failed to adopt an equally valid, less-discriminatory alternative test. Id. at 587 (citing 42 U.S.C § 2000e-2(k)(1)(A), (C)). Like the City, Meredith offers no evidence, much less a strong basis of evidence, that denying the local station’s request to fire Ossmann would have caused minority employees to file a viable disparate impact lawsuit against the application of Meredith’s sexual harassment policy. Id. Nor, I suspect, would anyone argue that Meredith’s sexual harassment policy is not “job- related” and “consistent with business necessity.” Id. So Ricci forbid Meredith from analyzing how Ossmann’s firing would impact racial balance. 2. Meredith’s policy of providing race statistics for corporate HR to review when approving or denying local personnel decisions also contradicts the EEOC’s instruction to keep race-related data away from decisionmakers: § 1602.13 Records as to racial or ethnic identity of employees Employers may acquire the information necessary for completion of items 5 and 6 of Report EEO–1 either by visual surveys of the work force, or at their option, by the maintenance of post-employment records as to the identity of employees where the same is USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 35 of 49

10 Maze, J., dissenting 22-11462

permitted by State law. In the latter case, however, the Commission recommends the maintenance of a permanent record as to the racial or ethnic identity of an individual for purpose of completing the report form only where the employer keeps such records separately f rom the employee’s basic personnel form or other records available to those responsible for personnel decisions, e.g., as part of an automatic data processing system in the payroll department. 29 C.F.R. § 1602.13 (emphasis added); see also United States Equal Employment Opportunity Commission, 2021 EEO-1 Component 1 Data Collection Instruction Booklet, Appendix D (“If an employee declines to self-identify their race and/or ethnicity, employment records or observer identification may be used. Where records are maintained, it is recommended that they be kept separately from the employee’s basic personnel file or other records available to those responsible for personnel decisions.”). The EEOC seems to believe that the inference of disparate treatment of individuals is so strong when decisionmakers consider race that it recommends companies build a wall between those who collect and report race-related data and those who make personnel decisions. The evidence suggests that Meredith blew through that wall. — To sum up, the combination of the EEO Analysis form and Berenguer’s testimony creates a reasonable, logical inference that USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 36 of 49

22-11462 Maze, J., dissenting 11

Bock considered race when deciding whether to terminate Ossmann. That FAQ #10 says that Bock was to use racial data to determine whether granting the requested personnel change would impact protected racial groups “at a higher percentage rate than similarly situated non-protected employees” supports a reasonable, logical inference that Bock would have responded differently if Ossmann was the lone black or Hispanic member of the weather team. That reasonable inference creates a genuine dispute of material fact that only a jury can resolve. See Fed. R. Civ. P. 56(a). II. The majority offers three rebuttals to my reading of the evidence: (1) the EEO form does not tell Bock how to use the racial group data; (2) Berenguer testified that the purpose of the racial group data was to ensure that the station was not “treating one person in that situation in the comparable group differently than others”; and, (3) Ossmann offers no evidence that Bock considered Ossmann’s race as a “negative factor.” I address each rebuttal in turn. A. The majority says that we cannot reasonably infer that Bock considered the potential impact on racial group balance when reviewing Ossmann’s file because “[t]he language of the form does not tell Bock what she should do with the racial data, and it does not require her to engage in racial balancing—it is completely neutral.” Opinion at 20. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 37 of 49

12 Maze, J., dissenting 22-11462

But as I showed in Part I, FAQ #10 told Berenguer to send data that would help Bock determine whether “a particular protected category of employee” would be impacted “at a higher percentage rate than similarly situated non-protected employees.” And the Risk Analysis charts that Berenguer created in response disclosed the percentage of each racial group so that Bock could determine the impact on racial group percentages. This language speaks for itself, and a reasonable juror could read it to require Bock to conduct a disparate impact review of racial groups. B. To support its finding that the EEO form is race neutral, the majority says that Berenguer testified that “the purpose of the [racial] data was to make sure that the station was not ‘treating one person in that situation in the comparable group differently than others.’” Opinion at 20. The majority omits important context. Here’s the full exchange between counsel and Berenguer, with the portion the majority quotes in bold: Q. And you can’t remember from 14 years ago why you fill out that form or the instructions as to why it says age, race, and sex? A. I believe I’ve already answered that question, but, yes, I do remember. So that we can — or not we, but that upon review, it can be determined whether there’s equitable treatment or not. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 38 of 49

22-11462 Maze, J., dissenting 13

Q. Well — but in your case, the comparables were not being disciplined; correct? A. Correct. Q. So then why in the world would you need age, race, and sex when you’re not — when you’re comparing them to people that were not being disciplined? Do you have an answer? Well, that’s — A. I don’t actually. Q. That’s 20 seconds. That’s 20 seconds for the Court — for the record. Go ahead. A. I’m not the one who completes the analysis, so I can speculate that it is making sure that you aren’t treating one person in that situation in that comparable group differently than others. Q. But in the comparables you put in, they weren’t being treated to anything. How can you compare and disprove the negative? Isn’t it true Ms. Berenguer, that this is simply an improper and illegal form used by Meredith? A. No, I don’t believe that to be true. (objections omitted). There are two reasons why Berenguer’s testimony does not show that Bock used race to ensure that USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 39 of 49

14 Maze, J., dissenting 22-11462

Ossmann was treated equally compared to others accused of sexual harassment, as the majority infers, Opinion at 21, rather than to judge the impact termination would have on racial group balance, as I infer. First, when asked why Bock needed to know the race of persons who had not been disciplined for sexual harassment, Berenguer sat stumped for 20 seconds before testifying that she did not know and could only speculate. Having watched Berenguer struggle to answer before admitting that she did not know, I doubt the district court would find Berenguer’s subsequent speculation admissible or that a jury would find it credible. Second, Bock could not perform the disparate treatment analysis that the majority infers from Berenguer’s testimony because, as the majority acknowledges (Opinion at 9), Berenguer did not provide Bock with the race of the other persons who had been accused of sexual harassment. A different question on the form, FAQ #15, requested the data needed to conduct the race- neutral, disparate treatment analysis the majority infers:

The majority’s inference that “the purpose of the [racial group] data was to make sure that the station was not ‘treating one person in that situation in that comparable group differently than others,’” id. at 20-21, may have been true if Berenguer provided the race data USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 40 of 49

22-11462 Maze, J., dissenting 15

necessary to analyze FAQ #15. But again, Berenguer only provided race data for FAQ #10, and Ossmann’s argument stems from that race-conscious part of the form—not FAQ #15. Meredith’s counsel agreed with this distinction between disparate treatment review (FAQ #15) and disparate impact review (FAQ #10) at oral argument. When pressed to explain what Berenguer meant when she testified that she provided Bock with the race of “comparables” who were not accused of sexual harassment, counsel conceded that, “when [Berenguer] was talking about comparables, she was not talking about people who had violated the company’s sexual harassment policy, she was talking about people within the group—people who worked in the group—and whether or not the decision was going to have a disparate impact on the group.” I am not saying that counsel’s statement is “factual evidence.” Opinion at 20 n.8. The EEO form plus Berenguer’s testimony is the evidence. I’m just pointing out that, against its interest, Meredith interprets Berenguer’s testimony to discuss disparate impact review, not disparate treatment review. And if Ossmann, Meredith, and I all view the evidence to show disparate impact review, a reasonable juror can too.1

1 I did not elicit “new evidence” about an attorney’s actions at the panel

stage—the questioning the en banc court found inappropriate in United States v. Campbell, 26 F.4th 860, 875-76 & n.12 (11th Cir. 2022) (en banc). See Opinion at 20 n.8 (citing Campbell). I instead asked counsel to reconcile Berenguer’s USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 41 of 49

16 Maze, J., dissenting 22-11462

C.

Finally, the majority says that we cannot reasonably infer that race mattered to Bock’s decision because Ossmann offers no evidence that race was a “negative factor against Ossmann or was used favorably in evaluating a non-white employee[.]” Opinion at 23-24. The majority’s search for a “negative factor” makes it stray from § 1981 and the Supreme Court’s decision in Students for Fair Admissions.

1. Section 1981(a) gives employees “the same right . . . to make and enforce contracts,” as members of another race. 42 U.S.C. § 1981(a). Section 1981(b) says this right applies to terminations. If a plaintiff alleges that the termination of his contract violated § 1981, a plaintiff must show that race was a ‘but for’ cause of his termination, using this framework:

If the defendant would have responded the same way to the plaintiff even if he had been white, an ordinary speaker of English would say that the plaintiff received the ‘same’ legally protected right as a white person. Conversely, if the defendant would have responded differently but for the plaintiff’s race, it follows that the plaintiff has not received the same right as a white person.

testimony about “comparables” with the EEO form. In other words, I treated “attorneys as attorneys,” not “attorneys as witnesses.” 26 F.4th at 876, n.12. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 42 of 49

22-11462 Maze, J., dissenting 17

Comcast, 140 S. Ct. at 1015. 2 Applying the Comcast standard here, Ossmann must show that Bock “would have responded differently” if the local station requested the termination of a coworker of a different race under similar circumstances. Id.; see also Bostock 140 S. Ct. at 1739 (“a but-for test directs us to change one thing at a time and see if the outcome changes. If it does, we have found a but-for cause.”).

That means Ossmann does not have to prove that Bock considered being white as a “negative factor” in the literal sense— i.e., being white added weight to the local station’s termination request—as long as Ossmann can show that Bock would have responded differently if the request was to terminate either the lone black or Hispanic weather team member. For example, if Bock would have recommended that the lone black or Hispanic member be suspended rather than terminated to avoid racial imbalance, then Ossmann did not receive the “same right” as his coworkers under § 1981; even if Bock did not consider Ossmann’s race as a “negative factor” when considering his termination.

2. The Supreme Court made this point in Students for Fair Admissions. In that case, a group of Asian-American students challenged the admissions policies of Harvard College and the University of North Carolina because the policies allowed the colleges to consider applicants’ race. While both colleges admitted

2 Section 1981 prohibits discrimination against all races, including white. See

McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273, 295 (1976). USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 43 of 49

18 Maze, J., dissenting 22-11462

that they considered race, they argued that “an individual’s race is never a negative factor.” Students for Fair Admissions, 143 S. Ct. at 2169. Rather, race served as a positive factor for some applicants from underrepresented groups. Harvard, for example, likened race to playing a musical instrument: the ability to play was considered a positive factor for some, but the inability to play was not a negative factor for others. Id.

The Court said, “[t]his understanding of the admissions process is hard to take seriously.” Id. “A benefit provided to some applicants but not to others necessarily advantages the former group at the expense of the latter.” Id. In other words, making race a positive factor for some groups necessarily makes race a negative factor for other groups. To deny this, as Justice Thomas put it, “simply defies mathematics.” Id. at 2199 n.9 (Thomas, J., concurring).

The Court explained that Harvard was making decisions that balanced race, as shown by the fact that “[f ]or the admitted classes of 2009 to 2018, black students represented a tight band of 10.0%–11.7% of the admitted pool.” Id. at 2171. To ensure this racial balance, Harvard’s Admissions Committee would begin each meeting “with a discussion of ‘how the breakdown of the class compares to the prior year in terms of racial identities.’ And ‘if at some point in the admissions process it appear[ed] that a group [was] notably underrepresented or ha[d] suffered a dramatic drop off relative to the prior year, the Admissions Committee may USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 44 of 49

22-11462 Maze, J., dissenting 19

decide to give additional attention to applications from students within that group.” Id. at 2170 (citation omitted).

In short, Harvard’s policy was to consider racial group balance when considering applications. To keep numbers in the college’s desired range, Harvard would treat applications differently because of race. So despite Harvard saying that race was not considered a “negative factor” when reviewing an Asian- American’s application, in reality, it was.

3. Meredith’s EEO form worked the same way. Corporate required the local station managers to disclose their employees’ race, including employees not accused of wrongdoing, so that Bock could consider group balance when deciding whether to approve the termination request.

So the majority may be right that Bock didn’t consider Ossmann’s race as a “negative factor” when she opened his file and saw that he was white. But if Bock opened the file and instead saw that the local station was seeking to terminate the lone black or Hispanic member of the weather team, it is reasonable to infer that Bock “would have responded differently” to the termination request. Comcast, 140 S. Ct. at 1015. After all, the goal of FAQ #10 was to ensure that the percentage of a “particular protected group of employees” was not negatively impacted compared to “similarly-situated non-protected employees.” Because this inference is reasonable and logical, we must make it and give the case to a jury—just as we did in Smith v. Lockheed Martin. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 45 of 49

20 Maze, J., dissenting 22-11462

III.

This case would be easy if we followed our reasoning in Lockheed-Martin.

1. Before I explain why, though, I note my agreement with the majority that the Appellant in Lockheed-Martin, Anthony Mitten, had more evidence that Ossmann. As the majority lays out, Opinion at 21-23, Mitten presented three pieces of evidence that created a convincing mosaic of circumstantial evidence: (1) a final decisionmaker previously discriminated against white employees, (2) Lockheed-Martin had motive to fire white employees thanks to a recent racially based shooting, and (3) the disciplinary review committee was given a decision matrix that disclosed the employees’ alleged misconduct and race. While Ossmann’s documentary proof is stronger than Mitten’s (more on that later), he does not present motive or bad act evidence like Mitten.

Having less evidence than Mitten does not, however, negate the reasoned, logical inference that Ossmann creates. Think back to our cookie thief analogy. Imagine that—in addition to your child’s failure to explain why his hand is in the cookie jar—you know that he didn’t eat breakfast (i.e., motive) and that he swiped a cookie last week (i.e., prior bad acts). Adding those facts would strengthen the inference that your child reached into the jar to take a cookie. But removing those facts doesn’t make the original inference that he was taking a cookie unreasonable or illogical. Likewise, the USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 46 of 49

22-11462 Maze, J., dissenting 21

majority’s point that Mitten had more evidence than Ossmann is true but not determinative.

2. The decisive question is whether evidence that Bock considered a document that contained Ossmann’s race and the race of his coworkers (including those not accused of wrongdoing) is enough to create a reasonable inference of discrimination. The Court’s discussion of the decision matrix in Lockheed-Martin is on-point, so I quote it in full:

The discipline ‘matrix,’ on which Mitten’s race was tracked, strengthens the reasonableness of the inference that Heiserman sought to fire all whites who distributed racist emails and, thus, fired Mitten because of his race. The disciplinary review committee and Heiserman relied on the ‘matrix’ to reach their discipline decisions, including Mitten’s. On its face, the ‘matrix’ indicates that race was pertinent to the discipline decisions made, and Lockheed has not explained satisfactorily why this was legitimate. Therefore, although the district court entirely ignored this fact, Lockheed’s injection of race into its decision-making process yields an unavoidable inference that the employee’s race impacted the discipline determination, and it is a jury’s province to decide whether race actually bore on the decision to terminate Mitten. USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 47 of 49

22 Maze, J., dissenting 22-11462

Lockheed-Martin, 644 F.3d at 1345-46 (footnotes omitted). The majority correctly notes that the Court started the paragraph by saying that the matrix “strengthens the reasonableness of the inference” created by the motive and bad act evidence. Id. at 1345.

But the rest of the paragraph stands alone. Most importantly, the Court says that (a) the inclusion of race in the matrix, plus (b) Lockheed-Martin’s inability to explain why race was included in the matrix, “yields an unavoidable inference that the employee’s race impacted the discipline determination, and it is a jury’s province to decide whether race actually bore on the decision to terminate Mitten.” Id. at 1346. In other words, adding race to the decisionmaker’s document created the “unavoidable inference,” not the added motive and bad act evidence.

Not only does Ossmann have similar documentary evidence; he has stronger documentary evidence. Lockheed-Martin’s matrix noted race with one letter. Id. at 1336. Lockheed-Martin did not have a policy that told decisionmakers how to use race, and the man who created the matrix testified that he added race “as merely a decision of personal convenience, intended to aid his putative future reporting of that information to external authorities” and that “it was understood that Lockheed’s principal decision-makers would ‘close one eye to the race entry’ when looking at the ‘matrix.’” Id. (brackets omitted).

Meredith’s EEO Analysis form, in contrast, tells us that Bock was to consider Ossmann’s race and his coworkers’ race to USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 48 of 49

22-11462 Maze, J., dissenting 23

determine whether granting the local station’s request would disparately impact racial group balance. As discussed, the Supreme Court held in Ricci that disparate impact review of personnel decisions violates Title VII, absent circumstances not present here. And the Supreme Court just reaffirmed that “[o]utright racial balancing is patently unconstitutional.” Students for Fair Admissions, 143 S. Ct. at 2172 (quotations omitted).

If Lockheed-Martin’s inclusion of race into the decisionmakers’ documents without instructions on how to consider race created “an unavoidable inference that the employee’s race impacted the discipline determination,” Lockheed-Martin, 644 F.3d at 1346, then Meredith’s insertion of race into the EEO Analysis with instructions to use race for disparate impact review creates the same unavoidable inference. IV. I agree with most of the majority’s opinion. I agree with the majority’s portrait of Ossmann as an unsympathetic plaintiff who likely earned his fate. I agree that the evidence suggests that Ossmann’s station managers wanted to fire him because he would not stop harassing women, not because of his race. And if the decision to fire Ossmann stopped at the local station, I would join the majority’s opinion in full. But the decision to fire Ossmann did not stop at the local level; it ended at corporate HR. And corporate wasn’t satisfied knowing that Ossmann sexually harassed women; corporate needed to know Ossmann’s race, and everyone else’s race, so that USCA11 Case: 22-11462 Document: 50-1 Date Filed: 09/08/2023 Page: 49 of 49

24 Maze, J., dissenting 22-11462

corporate could determine whether firing Ossmann would negatively impact racial group balance. Corporate’s injection of race for this purpose creates a reasonable inference—or, as the Court put it in Lockheed-Martin, an unavoidable inference—that Ossmann did not receive “the same right” that coworkers of another race would have received if Bock opened their file instead. 42 U.S.C. § 1981. Whether Bock would have reached a different decision if the races were changed is a genuine issue of material fact that only a jury can decide. Because the majority will not let the jury decide, I respectfully dissent.