Norman Wang v. University of Pittsburgh

• Court: Court of Appeals for the Third Circuit
• Decided: July 7, 2026
• Docket: 25-1816
• Status: Published

Opinion

UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT No. 25-1816

NORMAN WANG, Appellant

v.

UNIVERSITY OF PITTSBURGH; UNIVERSITY OF PITTSBURGH MEDICAL CENTER; UNIVERSITY OF PITTSBURGH PHYSICIANS; SAMIR SABA; MARK GLADWIN, et al. _____________________________ On Appeal from the U.S. District Court, W.D. Pa. Judge Marilyn J. Horan, No. 2:20-cv-01952

Before: SHWARTZ, BIBAS, and PHIPPS, Circuit Judges Argued: Mar. 2, 2026; Filed: July 7, 2026 _____________________________

OPINION OF THE COURT BIBAS, Circuit Judge. Academic debate is a full-contact sport. Debating ideas and questioning arguments can bruise people’s egos. But like boxing, academic debate bars low blows. Academics may not harm one another by recklessly dis- regarding the truth. Administrators may not demote professors for calling out discrimination. And managers may not create or tolerate hostile work environments. A culture that cancels instead of counsels sacrifices persuasion on the altar of power. In doing so, it forgets that the remedy for disfavored speech is more speech, not coerced silence. Dr. Norman Wang is an academic cardiologist. He worried that race-based affirmative action in medical schools and hos- pital residencies discriminates against some minorities in favor of others, could violate the law, and harms the very doctors whom it tries to help. Though most commentators opposed his views, some research supported it. After he published an article expressing those views, his bosses (also mostly doctors) alleg- edly denounced it as false and racist, fanned the flames on social media, demoted him at the hospital system, and got his article retracted. Wang sued the university, hospital system, doctors, journal, and publisher for defamation and retaliation. The District Court dismissed some claims on the pleadings and granted summary judgment against Wang on the rest. That was partly wrong. True, Wang has not plausibly alleged state action. But he has plausibly alleged defamation against five defendants, and there are disputed facts about whether his employers retaliated against him for opposing their allegedly illegal discrimination. Thus, we will let many of Wang’s claims go forward. I. AFTER WANG CRITIQUES RACIAL PREFERENCES, HIS BOSSES STRIKE BACK A. Wang’s article questions race preferences in cardiology Pittsburgh enjoys a public-private partnership to train doc- tors. The University of Pittsburgh (Pitt) is a public university with a medical school. It partners with a private hospital sys- tem, the University of Pittsburgh Medical Center (UPMC), and its subsidiary, University of Pittsburgh Physicians (which employs doctors in the UPMC hospital system). Medical

2 students study at the university and join residents and fellows in some training at the hospital system. Wang, like many of his colleagues, is both a professor at the medical school and a cardiologist at the hospital system. He has worked at both for almost two decades, serving patients and training residents and fellows. He no longer teaches regu- larly at the medical school. Nine years ago, he worked his way up to become director of a cardiology fellowship program at the hospital system. As program director, he helped select and train fellows. Admission to medical schools, residencies, and fellowships is a competitive process based on grades, standardized-test scores, and “holistic” factors. App. 2642. Around 2018, Wang noticed that the American Heart Association, American Col- lege of Cardiology, and Accreditation Council for Graduate Medical Education (the hospital’s accreditor for residency and fellowship programs) were pushing another factor: race. Wang was concerned. Though he is “very much for diver- sity,” he feared that these diversity programs “look[ed] overtly discriminatory and quota-based” and failed to “treat people as individuals.” App. 2467, 2469, 2490. So in March 2020, he cri- tiqued these practices in the peer-reviewed Journal of the American Heart Association. His seventeen-page article included four tables, four figures, lots of data, and 108 endnotes full of scholarly sources. After explaining how these programs may violate the law and harm members of minority groups, he con- cluded: “Long-term academic solutions and excellence should not be sacrificed for short-term demographic optics.” App. 2642.

3 B. Wang’s bosses demote and ostracize him When Wang first published his article in March, it did not make a stir. All that changed in the summer of 2020. On July 29, a doctor emailed the president of the hospital system’s Phy- sicians Services Division to flag and complain about Wang’s conclusions (not his data). The president looped in the medical school’s dean, who agreed, adding that the article was “pseudo- scholarly … with handpicked data,” as well as “incredibly offen- sive and racist.” App. 2657, 2662. At the same time, another doctor emailed Dr. Kathryn Ber- lacher, who pinged Dr. Samir Saba. Like Wang, Berlacher and Saba work at both the hospital system and the medical school. Saba is Wang’s superior in both workplaces; Berlacher is his colleague at the medical school but at the hospital gives guid- ance to him as director of the cardiology fellowship program, an umbrella that oversees Wang’s subspecialty program. The doctor who contacted Berlacher said she “can’t argue with facts” but objected to the “quotations he chose to use, the biased conclusions he made from the data, etc.” App. 2655. Berlacher agreed that, while “we can’t argue w[ith] facts,” “[i]t’s v[ery] clearly offensive and racist.” App. 2654. Saba concurred, expressing his “overwhelming feeling … that [Wang] cannot continue in [his leadership] role.” App. 2673. He looped in Dr. Mark Gladwin, chair of the medical school’s Department of Medicine and head of training at the hospital system, on the email chain. Gladwin lamented that, though Wang’s article “was written before the George Floyd events, [the] timing of the print publication could not be worse.” App. 2667. Later, Saba replied that “[e]motions are very high, understandably.” Id. So on July 31, two days after getting those complaints,

4 Berlacher and Saba met with Wang and fired him from his pro- gram directorship at the hospital system. At that meeting, Wang raised his concerns about racial and ethnic preferences in “our [graduate-medical-education] con- tracts with the fellows” at the hospital system and said he “just wanted us to follow the law.” App. 961-63. Wang did not explic- itly state that the hospital was acting illegally. At some point, though, Saba told him: “you know what we’re trying to do here, Norm,” App. 2490–91. Saba also said “that the views that [Wang had] expressed [we]re not in line with the views and policies that [we] were trying to create,” App. 2498. Once Wang had been fired, the next question was whether to rebut Wang’s article or just smear it. Gladwin and the medi- cal-school dean urged university and hospital-system leaders to “take[ ] the high road” by publishing an opposing paper iden- tifying specific errors. App. 2657. They did not. Instead, Berlacher, Saba, and other doctors launched a social- media campaign to discredit Wang’s article. Two days after meeting with him, Berlacher tweeted that it “misinterprets data and misquotes people” and “is scientifically invalid and racist.” App. 145. Next, Berlacher authored a tweet on the hospital sys- tem’s Cardiology Department account (which she controlled) accusing Wang of using “misquotes, false interpretations, and racist thinking.” App. 145, 233. Saba reposted that tweet. That fusillade fanned “a firestorm on [T]witter, [F]acebook[,] and in all national cardiology circles.” App. 3013. Doctors called for Wang to be cut off from training to “#RetractRacists” and pro- tect students. App. 2783. One fellow emailed Berlacher, describ- ing an “uncomfortable” experience that he had had with an

5 “intense grilling” by Wang as evidence that he disliked minor- ities. App. 2701. Berlacher took this email as confirmation that it was “unsafe” for fellows to work for him and that he needed to be removed from his directorship and all educational roles “immediately.” App. 2700. So they did. On August 4, two days after their tweets, Ber- lacher and Saba emailed Wang that “any educational environ- ment in which you partake is inherently unsafe, increasing our learners’ risk for undue bias and harm.” App. 2706. They banned him from any role teaching fellows, residents, or med- ical students. Berlacher had already tweeted to #medtwitter, reassuring other doctors that the students were safe. Over the next three days, the medical school’s dean and Gladwin sent hospital- and school-wide emails, denouncing Wang’s article as hurtful and offensive. As a result, Wang suffered. He was uncomfortable going to work because long-time colleagues stopped talking to him. He cleaned out his office. Other doctors worried that he might kill himself, and the police even called to see if he was being har- assed. C. Hospital and university doctors also get the journal to retract Wang’s article On August 3, Berlacher, Saba, the medical-school dean, and four other hospital-system doctors emailed the Journal of the American Heart Association’s editor-in-chief, urging the jour- nal to retract Wang’s article. They decried it as “deeply offen- sive and disturbing,” with “several misquotes” and “proven scientific falsehoods.” App. 2708. They attached to the email an eight-page “point-by-point rebuttal” that they “[we]re

6 planning to submit[ ] for publication.” Id. Starting July 31, hos- pital-system employees also sent back-channel messages to the American Heart Association, the journal’s proprietor. Two days after Berlacher, Saba, and others emailed the editor- in-chief, the Association and the journal’s editors retracted the article. The retraction notice said: “[T]he University of Pitts- burgh Medical Center (UPMC)[ ] has notified the Editor-in- Chief that the article contains many misconceptions and mis- quotes and that together those inaccuracies, misstatements, and selective misreading of source materials strip the paper of its scientific validity.” App. 237. The Association also put out a press release that day. The next day, an updated notice added two alleged misquotations as a “sample of the misconceptions and misquotes identified in correspondence from UPMC.” App. 163. “Based on errors such as these, the editors” were retracting the article. App. 164. Later that month, an editor’s note on behalf of the journal, lead-authored by Dr. Marc Simon of the hospital system and medical school, accused “the retracted article … [of] misrepresent[ing] facts” and “evidence.” App. 195. Yet the following months brewed a backlash to the back- lash. Multiple institutions objected that Wang had been disci- plined without due process for exercising his academic free- dom. The university’s Tenure and Academic Freedom Commit- tee asked the medical-school dean to reinstate Wang. The American Association of University Professors thought the same. The U.S. Department of Education started investigating whether the university’s “campaign of denunciation” violated Title VI and other laws. App. 3223. The medical-school dean said that “there was a lot of consternation” over potentially

7 “limiting academic freedom” and “this was clearly a big issue for the faculty senate.” App. 2147. So in late October, under pressure, he told Wang that he had never lost any faculty status or privileges at the university and notified Berlacher, Saba, and Gladwin that they had no authority to bar Wang from university activities. D. The District Court rejects Wang’s defamation and retaliation claims Within a week of the first complaint about his article, Wang was demoted, his article was retracted, and he went viral on Twitter for being a dishonest racist. So Wang sued a range of defendants on three theories. First, he claimed defamation by the university, the hospital system, its physician practice, Ber- lacher, Saba, Simon, the American Heart Association, and the journal’s publisher (Wiley Periodicals). But the District Court dismissed that claim. Even though most of the statements were defamatory statements of fact or opinions based on undisclosed facts, it reasoned, the Association’s and publisher’s statements were true and Wang (a limited-purpose public figure) had failed to plead that the rest of the defendants acted with actual malice. Second, Wang brought various civil-rights retaliation claims against the university, the hospital system, and its phy- sician practice, as well as Berlacher, Saba, and Gladwin. The District Court dismissed his Title VI claim against the hospital for failing to plead that a primary purpose of the federal funds was to fund employment, and the Title VI and 42 U.S.C. § 1981 claims against the university because any adverse action had come from the hospital system, not the university. Wang also

8 brought claims under Title VII and the Pennsylvania Human Relations Act (PHRA). The District Court dismissed these claims against the university as to Wang’s statements in his arti- cle but not as to his statements in his July 31 meeting with Ber- lacher and Saba. Later, the court granted summary judgment against Wang on all remaining civil-rights claims, reasoning that Wang had not engaged in any protected activity at the July 31 meeting. Third, Wang brought a First Amendment retaliation claim under 42 U.S.C. § 1983 against those same defendants. Apply- ing Monell’s municipal-liability standard, the District Court dismissed the claims against the university as failing to allege action under a policy or by policymakers. See generally Monell v. Dep’t of Soc. Servs., 436 U.S. 658 (1978). It also dismissed the claims against the hospital system and its physician prac- tice, and later granted summary judgment to the individual defendants, reasoning that the hospital and the physician prac- tice were not state actors, that Berlacher and Saba were not act- ing as employees of the university, and that Gladwin was not involved in Wang’s removal. Wang now appeals all these adverse rulings. II. THE DEFAMATION CLAIM SURVIVES AGAINST FIVE DEFENDANTS To make out a defamation claim, Wang must plausibly plead and prove that a defendant’s published communication was defamatory and sufficiently linked to him. 42 Pa. Cons. Stat. § 8343(a). We review the District Court’s dismissal de novo, taking the complaint’s factual allegations in the light most favorable to Wang. Fowler v. UPMC Shadyside, 578 F.3d

9 203, 206, 210–11 (3d Cir. 2009) (discussing Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009)). The District Court erred in dismissing Wang’s whole defa- mation claim. He has plausibly pleaded that the challenged statements were defamatory. Defendants either did not argue or cannot show that their statements were substantially true. Wang appears to be a limited-purpose public figure who must plead that the defendants spoke with actual malice, but he pleads enough to clear that hurdle as to Berlacher, Saba, the Association, the university, and the hospital system. So we will let much of Wang’s defamation claim proceed. A. The challenged statements asserted facts or were based on undisclosed facts A statement can be defamatory if “it tends so to harm the reputation of another as to lower him in the estimation of the community or to deter third persons from associating or deal- ing with him.” Tucker v. Fischbein, 237 F.3d 275, 282 (3d Cir. 2001) (Alito, J.) (citation omitted). Statements of fact can be defamatory. So too can statements of opinion that “give[ ] rise to the inference that there are undisclosed facts that justify the forming of the opinion.” Meyers v. Certified Guar. Co., 221 A.3d 662, 670 (Pa. Super. Ct. 2019) (quoting Restatement (Second) of Torts § 566 cmt. b (1977)). To see if a statement implies plausibly defamatory facts, we look at it in context. Graboff v. Colleran Firm, 744 F.3d 128, 136 (3d Cir. 2014). The District Court rightly found that Berlacher’s and Saba’s tweets were opinions based on undisclosed facts, so they could amount to defamation. They launched a campaign to ostracize him, making his article go viral and generating a backlash.

10 Berlacher tweeted from her personal account: “@PittCardiol- ogy I’m looking at you. What do we stand for? What do you think of this OPINION piece that misinterprets data and mis- quotes people? @JAHA_AHA this is scientifically invalid and racist.” App. 145, 235. Next, the hospital-system’s official Twitter account @PittCardiology tweeted that Wang’s article used “misquotes, false interpretations and racist thinking to defend a single person’s conclusion.” App. 145, 231. Saba then retweeted that on behalf of one of the hospital system’s institutes. We cannot dismiss these statements as mere academic jousting. Of course, science is a process of revision, so “tenta- tive scientific conclusions” based on disclosed data and meth- ods are not actionable. Pacira Biosciences, Inc. v. Am. Soc’y of Anesthesiologists, Inc., 63 F.4th 240, 247 (3d Cir. 2023). But these were not unfalsifiable “disagreements about the reliabil- ity of the methodology and data underlying the statements.” Id. Rather, these tweets accused Wang of misquoting and misread- ing data and sources—both of which are falsifiable claims. So they were more hit job than academic argument. The Association made defamatory statements, too. It issued two retraction notices (through the publisher) plus a press release. The first notice alleged that the paper contained “many mis- conceptions and misquotes and that together those inaccura- cies, misstatements, and selective misreading of source mate- rials strip the paper of its scientific validity.” App. 143. It gave no examples. The companion press release called Wang’s article “a misrepresentation of the facts,” containing “deliberate mis- information [and] misrepresentation.” App. 144. Later that month, the journal’s editorial board published an “Editor’s

11 Note,” lead-authored by Simon, accusing Wang of “misrepre- sent[ing] facts” and “evidence.” App. 145. The “sting” of the tweets was that Wang’s article was sloppy, riddled with errors, scientifically invalid, and moti- vated by racism. See Oweida v. Trib.-Rev. Publ’g Co., 599 A.2d 230, 239 (Pa. Super. Ct. 1991); accord Schiavone Constr. Co. v. Time, Inc., 847 F.2d 1069, 1083 (3d Cir. 1988) (applying analogous New Jersey law). Though “a simple accusation of racism” is not enough in Pennsylvania, these tweets went fur- ther and “could be construed to mean that [Wang] was acting in a racist manner” and that his racism led him to distort facts and evidence in connection with his work as an academic and a doctor. MacElree v. Phila. Newspapers, Inc., 674 A.2d 1050, 1055 (Pa. 1996). In context, they implied that the article was full of errors driven by motivated reasoning. The Association’s notices, press release, and article likewise accuse him of spreading half-truths and lies. Wang plausibly alleges “the kind of harm which has grievously fractured his standing in the community of respectable society.” Tucker v. Phila. Daily News, 848 A.2d 113, 124 (Pa. 2004) (internal quotation marks omitted). It would have been fine just to say that Wang’s article was wrong, harmful, even dangerous. Those are normal academic opinions. See Pacira, 63 F.4th at 245–47 (reasoning that a sci- entific article’s claim that a drug was “inferior” or “not supe- rior” was an opinion, a tentative scientific conclusion that could not be defamatory). But these statements went much fur- ther, alleging misquotations, misstatements, and outright false- hoods. Wang’s academic integrity is central to his career as a professor, researcher, and doctor. And in Pennsylvania, “any

12 language which unequivocally, maliciously and falsely imputes to an individual or corporation [a] want of integrity in the con- duct of his or its business is actionable.” Cosgrove Studio & Camera Shop, Inc. v. Pane, 182 A.2d 751, 754 (Pa. 1962). When a journal retracts a paper for its “many” misstatements and inaccuracies, a reasonable reader infers that the journal found many falsehoods and errors because someone can objec- tively disprove the article’s statements and claims. See Pacira, 63 F.4th at 246. So such statements can be actionable. B. Defendants cannot win on truth yet Defamation requires falsity. Defendants may raise truth as an affirmative defense and prove that their statements were true. 42 Pa. Cons. Stat. § 8343(b)(1). Courts may not dismiss on this ground, though, unless a statement’s truth is apparent on the complaint’s face. See Bethel v. Jendoco Constr. Corp., 570 F.2d 1168, 1174 (3d Cir. 1978). Berlacher, Saba, and Simon forfeited the truth defense at this stage by not raising it below. The District Court found that the Association’s and publisher’s statements were true, but in doing so mistakenly looked to the second retraction notice, which defendants had attached as an exhibit to their motion to dismiss. But the complaint was ambig- uous; though the language it quoted appeared in both the first and second retraction notices, the date it gave was that of the first retraction notice. Wang disputed that he was referring to the second retraction notice. Given that ambiguity, the District Court should not have relied on that exhibit, as it was “material beyond the complaint.” Dunn v. Castro, 621 F.3d 1196, 1204 n.6 (9th Cir. 2010).

13 Yet even if we read the complaint’s fuzzy language as incor- porating the second retraction, making it fair game, the District Court erred by accepting truth as a defense. The defendants identified only two of Wang’s statements as even potentially misleading (and thus truthfully criticized as misquotations), yet neither bears the weight put on it. The first of the “misquotes” unravels under scrutiny. App. 163. Wang’s article quoted a source stating that Ohio State Uni- versity “simply made it a priority to rank [underrepresented in medicine] applicants more aggressively than in previous years, thus achieving success in matching them regardless of recruit- ing efforts, with the implication being that [the school] accepted less competitive applicants in an effort to increase diversity.” Id. (first alteration in original). That quotation was accurate. Yet the Association objected that it was “taken out of context and falsely implies that the Ohio State University (OSU) pro- gram ranked racial and ethnic groups underrepresented in med- icines with lower academic credentials in order to achieve diver- sity.” Id. Even so, “[t]he full statement” identified by the Asso- ciation still supports Wang’s reading: Underrepresented minor- ities, on average, scored slightly lower on the U.S. Medical Licensing Exam. App. 163. And though the source cautioned that the Licensing Exam scores are not perfect predictors of success and that the observed difference was not statistically significant, it admitted that the exam “scores are considered the most objective variable for comparison.” App. 205. Wang’s quotation was fair, accurate, and supported his conclusion that medical-education programs were giving explicit racial prefer- ences that arguably violated civil-rights law.

14 The second allegation of misquotation is closer to the mark. Wang hypothesized that doctors from underrepresented minor- ity groups do not improve medical access for underserved pop- ulations. He wrote that “a 1995 California study … demon- strated [that] primary care physicians who were not board cer- tified were 1.6 times more likely to work in rural underserved areas when compared with board-certified counterparts.” App. 420. That was not a quotation. But the substance of the para- phrase was off. The study actually found: “Non-board-certified family and general physicians were 1.6 times more likely than other non-board-certified primary care physicians to locate in rural underserved areas.” App. 210. No party pressed any other misquotations, misstatements, or misrepresentations. Yet these two “misquotes” were pre- sented as “[a] sample of the misconceptions and misquotes,” falsely implying that there were many more. App. 163. There were not “many” other misquotations—just one modest error. App. 237. Defendants exaggerated greatly, and the District Court should not have considered documents beyond those in the complaint, so the truth defense cannot defeat Wang’s defa- mation claim at the dismissal stage. C. Wang, a limited-purpose public figure, pleads actual malice for some Wang appears to be a limited-purpose public figure. He is not famous. But he “voluntarily inject[ed] himself … into a particular public controversy and thereby bec[a]me[ ] a public

15 figure for” discussing affirmative action in medicine. Gertz v. Robert Welch, Inc., 418 U.S. 323, 351 (1974). As a public figure, Wang must have plausibly pleaded that defendants spoke “with ‘actual malice’—that is, with knowledge that [each statement] was false or with reckless dis- regard of whether it was false or not.” N.Y. Times Co. v. Sulli- van, 376 U.S. 254, 280 (1964). That does not require malice in the ordinary sense of ill will. Because defendants rarely admit lying or recklessness, “objective circumstantial evidence can suffice to demonstrate actual malice.” Schiavone, 847 F.2d at 1090. 1. Wang pleads Berlacher’s and Saba’s actual malice. As a threshold matter, Saba never raised his actual-malice argu- ment before the District Court. But since the court ruled on that issue anyway, we address it and conclude that Wang properly pleaded enough objective circumstances against both Ber- lacher and Saba. Recall the complaint’s allegations, which we take as true, along with the documents integral to or explicitly relied on by the complaint. See In re Burlington Coat Factory Sec. Litig., 114 F.3d 1410, 1426 (3d Cir. 1997). Wang published his peer-reviewed article in March 2020. Four months later, doctors alerted hospital-system and university staff to Wang’s article and views. Panicked, the doctors and staff “criticized [Wang’s] conclusions.” App. 140. Berlacher and Saba quickly reacted to their criticisms by tweeting and retweeting social- media posts accusing him of misinterpreting data, misquoting sources, and racist thinking. App. 140, 141, 145. Yet Wang’s seventeen-page article was supported by 108 endnotes. And Wang alleges that the “citations and quotations in [his] article were clearly attributed, easily verified, and correct.” App. 150.

16 Mental states, including malice and intent, may be alleged generally, as long as the plaintiff pleads enough supporting facts to make inferring the mental state plausible. See Fed. R. Civ. P. 9(b); Biro v. Conde Nast, 807 F.3d 541, 544–45 (2d Cir. 2015). From Wang’s allegations, we could draw several plau- sible inferences. One is that Berlacher and Saba did not ade- quately investigate their claims that Wang had misquoted his sources. That inference supports dismissal, for a mere “failure to investigate” is not enough to show actual malice. McCafferty v. Newsweek Media Grp., Ltd., 955 F.3d 352, 359 (3d Cir. 2020). But it is just as plausible to infer that Berlacher and Saba knew, or at least recklessly disregarded the possibility, that their tweets were flat wrong. Several facts support that infer- ence. First, the things that they said Wang got wrong were, as Wang alleges, easy to verify. Berlacher and Saba could easily have checked Wang’s sources, all readily accessible and properly cited, and found that his citations and quotations were almost uniformly accurate. Second, the tweets themselves imply that Berlacher and Saba did in fact check Wang’s sources. As we explained above, some statements of opinion “give rise to the inference that there are undisclosed facts that justify the … opinion.” Meyers, 221 A.3d at 670 (cleaned up). Alleging that someone repeatedly “misquote[d]” his sources implies that the speaker checked them. App. 145. These facts distinguish Wang’s claims from most actual- malice cases, which usually turn on allegations that journalists did a shoddy job. See, e.g., McCafferty, 955 F.3d at 359; Biro, 807 F.3d at 546. Wang does not just allege that Berlacher and Saba relied on bad sources (as in Biro) or failed to interview

17 people familiar with the subject (as in McCafferty). Nor does he allege that Berlacher and Saba failed to “read every” rele- vant source on the question. Church of Scientology Int’l v. Dan- iels, 992 F.2d 1329, 1334 (4th Cir. 1993). Rather, Wang’s def- amation claim fits squarely into the space left open by McCaf- ferty. McCafferty had not alleged that the defendant “made up” the allegedly defamatory material or “published facts contrary to information it otherwise knew.” 955 F.3d at 360. Wang, by contrast, alleges facts—especially Berlacher’s and Saba’s own tweets—supporting an inference that they had checked Wang’s citations and quotations and found them (with one minor excep- tion) substantially accurate. Nonetheless, they trashed his arti- cle on Twitter, telling the public that Wang had repeatedly mis- quoted his sources, knowing that was “contrary” to the infor- mation they had. Id. A complaint alleging that a speaker directly contradicted information that he had can survive a motion to dismiss. See id.; Tholen v. Assist Am., Inc., 970 F.3d 979, 984– 85 (8th Cir. 2020). We could also infer that Berlacher and Saba intentionally avoided learning the truth. “[P]urposeful avoidance of the truth” shows actual malice. Harte-Hanks Commc’ns, Inc. v. Connaughton, 491 U.S. 657, 692 (1989). When a speaker chooses “not to listen” to an objective source that could verify a story, that choice supports finding actual malice. Id. at 684. Actual malice is even more apparent where facts show that a speaker or publisher decided—before investigating—to run with a particular narrative. See id. Wang’s article was peer-reviewed. That means it “was vet- ted pursuant to the usual procedures for review of all articles in JAHA.” App. 143. That peer-review vetting process is designed

18 to weed out the egregious errors that Berlacher and Saba accused Wang of making. And the allegations plausibly suggest that they “doubted the veracity of [their] story.” Tucker, 237 F.3d at 286–87. Given that, a cite check was the only thing that could have confirmed what they tweeted. If they failed to check the citations in the face of those doubts, then they “purposefully avoided further investigation with the intent to avoid the truth.” Michel v. NYP Holdings, Inc., 816 F.3d 686, 703 (11th Cir. 2016). Particularly given those concerns, the speed of Berlacher and Saba’s stampede strongly suggests a reckless rush. Within a couple of days of the first complaints, they demoted Wang. Within two days of that, they lambasted him on social media. No one seems to have paused to do a proper cite check; neither of the alleged errors was a misquotation, and one was not even wrong. Berlacher and Saba’s coordinated campaign of denun- ciation was thus a stark departure from normal academic prac- tice. The allegations in the complaint suggest that the doctors were reacting to Twitter outrage, not reasoned scholarly inquiry backed by facts. This circumstantial evidence “permit[s] the conclusion that the defendant[s] in fact entertained serious doubts as to the truth of [their] publication[s].” St. Amant v. Thompson, 390 U.S. 727, 731 (1968). To be sure, each inference is plausible. And two of them suffice to let the defamation claims against Berlacher and Saba go forward. Because we must draw “all reasonable inferences” in Wang’s favor at this stage, even if we thought he would struggle to “ultimately prevail on the merits,” we would let his claim proceed. Connelly v. Lane Const. Corp., 809 F.3d 780, 790–91 (3d Cir. 2016).

19 2. Wang fails to plead Simon’s actual malice. The case against Simon is inadequate. True, Simon, as lead author for the journal’s editorial note, denounced Wang’s article as “a misrepresentation of evidence” with “misrepresented facts.” App. 195. He “deeply regret[ted]” publishing it and was “grate- ful to the medical community for assisting in correcting this error.” Id. But he largely relied on the allegations of error that they had gotten from the hospital system, Wang’s employer. Plus, Simon (unlike Berlacher and Saba) never accused Wang of misquoting his sources—an accusation that would have been verifiable with a simple cite check. The most he said was that Wang had misrepresented the facts and evidence. Though that statement (an opinion premised on undisclosed facts) can still be defamatory, it is harder to disprove, making actual mal- ice harder to allege. What is more, the complaint contains little suggesting that Simon rushed to retaliate against Wang without verifying his suspicions, at least as compared to Berlacher and Saba. So Wang has failed to plausibly allege that Simon wrote with actual malice. 3. The allegations against the publisher are likewise inad- equate. Alleging that a publisher of another’s statement acted with actual malice is hard. Wang must plausibly allege more than that it radically departed from its usual standards. McCaf- ferty, 955 F.3d at 359–60. He does not. The publisher just passed on information from the Association, the hospital sys- tem, and its doctors, which it reasonably thought reliable. Again, mere failure to investigate, on its own, is not enough. Tucker, 237 F.3d at 286. 4. Wang pleads the Association’s actual malice. The clos- est call is about the Association. On the one hand, in places it

20 simply relayed the hospital’s and doctors’ statements, which it reasonably thought reliable. But the Association went further, and it did so hastily. Within two days of being asked to retract the article, it issued not only a retraction, but also a press release in its own name and voice. It “denounce[d] the views expressed in the article” as “contrary to our organization’s core values and historic commitment to promoting diversity and inclusion.” App. 166. It called the criticism for “misrepresentation” “right.” Id. And it went much further, accusing him of “delib- erate misinformation or misrepresentation”—intent that it could hardly know. App. 167. In its updated retraction state- ment, the Association did not merely pass along information as a publisher; it expressly accused Wang of lying. Nor did the Association notify Wang of any errors or give him a chance to rebut those charges before publishing the accu- sations. That is questionable in the academy. Cf. Bd. of Regents v. Roth, 408 U.S. 564, 573 (1972) (explaining that an academic might be entitled to “notice and an opportunity to be heard” and “to refute the charge” when officials challenge his “repu- tation, honor, or integrity” (internal quotation marks omitted)). Although a departure from professional standards, standing alone, does not prove actual malice, it may contribute to that inference when coupled with inflammatory accusations and a deficient investigation. Thus, the Association’s allegedly false charges of intentional deception—made hastily, without a thor- ough review, and without giving Wang a chance to rebut them—are enough to plead actual malice. Cf. McCafferty, 955 F.3d at 359–60.

21 D. The university and hospital system may be vicariously liable for defamation Once the District Court found that Wang had failed to state a defamation claim against the individual defendants, it rea- soned that their employers could not be vicariously liable either. But with the defamation claim against two individuals back in the case, vicarious liability is back on the table. Employers can be liable for defamatory statements that their employees make within the scope of their employment. To be within the scope of their employment, conduct must be the kind of action that the worker is hired to do, must happen substantially within the authorized work time and space, and must be at least partly motivated to serve the employer. Restate- ment (Second) of Agency § 228 (1958); Fitzgerald v. McCutch- eon, 410 A.2d 1270, 1272 (Pa. Super. Ct. 1979). Wang claims that Berlacher and Saba made their statements as part of their employment by the university or hospital sys- tem. He plausibly alleges that, as professors and leaders of the cardiology department, it was within the scope of their employ- ment to “accuse[ ] [Wang] of professional incompetence, and academic dishonesty.” App. 149. But none of Wang’s factual allegations names the physician practice or statements made on its behalf, let alone plausibly alleges its actual malice or vicar- ious liability. So the District Court properly dismissed the def- amation claim against the physician practice. * * * In sum, Wang plausibly alleged that Berlacher and Saba, speaking for the university and hospital system, defamed him,

22 as did the Association. It is plausible that they either knew they were overclaiming the whole time, or at least buried their heads in the sand to avoid checking his citations and investigating their doubts. That is enough for now. The New York Times v. Sullivan actual-malice standard is supposed to be a shield for free expression, not a sword to silence disfavored speech. III. WANG’S CIVIL-RIGHTS RETALIATION CLAIMS SURVIVE Three different civil-rights laws converge to protect Wang’s calling out what he reasonably believed was illegal activity. At multiple stages, for multiple reasons, the District Court rejected this claim. But text and precedent support it. No one may retaliate against an employee “because he has opposed any practice made an unlawful employment practice by” Title VII. 42 U.S.C. § 2000e-3(a). Section 1981 creates a cause of action for “retaliation against a person who has com- plained about a violation of another person’s contract-related ‘right.’ ” CBOCS W., Inc. v. Humphries, 553 U.S. 442, 445, 452 (2008); 42 U.S.C. § 1981(a). Pennsylvania also bars “discrimi- nat[ion] … against any individual because [he] has opposed any practice forbidden by” the PHRA. 43 Pa. Cons. Stat. § 955(d). Despite some differences, Title VII, the PHRA, and § 1981 all require the same three elements for a prima facie case of retaliation: protected activity by Wang, adverse employment action against him, and a causal connection between the two. Est. of Oliva ex rel. McHugh v. New Jersey, 604 F.3d 788, 798 & n.14 (3d Cir. 2010); Wilkerson v. New Media Tech. Charter Sch. Inc., 522 F.3d 315, 318, 320 (3d Cir. 2008). Once a plaintiff makes out a prima facie case, defendants may give legitimate, non-retaliatory reasons why they took

23 their adverse actions. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 (1973). The plaintiff may then show that those reasons are pretextual. Id. at 804. Reviewing de novo the Dis- trict Court’s dismissal of Wang’s claims relating to his article and its grant of summary judgment on his claims relating to comments made during the July 31 meeting, we hold that these claims should go forward. Burtch v. Milberg Factors, Inc., 662 F.3d 212, 220 (3d Cir. 2011); Moore v. City of Phila., 461 F.3d 331, 340 (3d Cir. 2006), as amended (Sept. 13, 2006). A. Wang suffered for criticizing his employers both in his article and at the July 31 meeting 1. Wang opposed policies that he reasonably thought were unlawful. Wang criticized what he believed was the hospital’s and university’s illegal practice of giving racial preferences. He made arguments like the ones that the Supreme Court later accepted in Students for Fair Admissions v. President & Fel- lows of Harvard College, 600 U.S. 181 (2023). That criticism was protected. Moore, 461 F.3d at 341. Yet the District Court disagreed because he had never called out the university or hos- pital system by name, and “advocacy, unconnected to employ- ment practices,” is not enough. Curay-Cramer v. Ursuline Acad., 450 F.3d 130, 136 (3d Cir. 2006). That was mistaken. To start, it is not clear that an employee must oppose his own employer’s illegal actions. All the statute seems to require is “oppos[ing] any practice made … unlawful.” 42 U.S.C. § 2000e-3(a) (emphasis added). We have not passed on that issue, nodding toward it only in dictum. Curay-Cramer, 450 F.3d at 135. And we need not decide it here because, in context, Wang’s article plausibly opposed the university’s and hospital

24 system’s admissions and hiring practices. He worked for those institutions. His article opposed “racial and ethnic preferences within graduate medical education training programs” indus- try-wide and nationwide. App. 415. He wrote that the accredit- ing bodies’ diversity requirements, which the university and hospital system had to follow, may be illegal. Discovery bore out Wang’s complaint. Gladwin read his arti- cle as criticizing affirmative-action practices, “[i]ncluding in graduate medical education” and the “selection of fellows in cardiology” at the hospital system and physician practice. App. 1339–40. In a school-wide email, Gladwin decried it as “anti- thetical to our values.” App. 2764. Because these comments “identif[ied] [Wang’s] employer and the practice—if not spe- cifically, at least by context,” the District Court erred in finding that Wang’s claims in the article were not protected activity. Curay-Cramer, 450 F.3d at 135. At the July 31 meeting, Wang testified that he raised his concerns about racial and ethnic preferences in “our [graduate medical education] contracts with the fellows” at the hospital system, and that he “just wanted us to follow the law.” App. 961–63. Saba told him, “you know what we’re trying to do here, Norm.” App. 2490–91. Saba also told him “that the views that [he had] expressed [we]re not in line with the views and policies that they were trying to create.” App. 2498. They under- stood his criticism perfectly and took it personally. Because there is a genuine dispute of material fact over whether Wang’s comments were protected conduct, the District Court erred in granting summary judgment on Wang’s Title VII/PHRA/ § 1981 retaliation claim.

25 2. The university and hospital system took non-trivial ad- verse actions against Wang. An adverse action is one that “well might have dissuaded a reasonable worker from making or sup- porting a charge of discrimination.” Burlington N. & Santa Fe Ry. Co. v. White, 548 U.S. 53, 68 (2006) (cleaned up). As the District Court recognized, the demotion and ban on contacting fellows, residents, and students are enough for the hospital sys- tem. See id. at 68; Gilbert v. Des Moines Area Cmty. Coll., 495 F.3d 906, 917 (8th Cir. 2007). Wang also alleged that his bosses and colleagues at the uni- versity and hospital had retaliated against his speech by harass- ing him to “isolate[ ] [him] in his work environment in hopes that he would resign.” App. 104. Viral tweets called him racist. School-wide emails from leaders acting as both hospital and university employees denounced him and his views as “anti- thetical to our values and deeply hurtful to many of our URM faculty.” App. 2764. His superiors at both the university and hospital were aware of the environment. Saba “predict[ed] … that the social envi- ronment here will become untenable for Norm [Wang] and this will lead to him leaving UPMC or to his reallocation to a periph- eral site.” App. 597. The blowback was so intense that a col- league feared that Wang might kill himself, telling Berlacher: “One thing that crossed my mind … [i]s the issue of his overall suicide risk .… it must not be easy to get all this push back and come back to work.” App. 3051–52 (second elision in origi- nal). With all that Wang endured, “a reasonable jury might well conclude that this pattern of harassment might dissuade a rea- sonable worker from bringing or supporting a charge of dis- crimination.” Moore, 461 F.3d at 348.

26 The multi-front reputation-bashing that impugned Wang’s integrity as an academic is the sort of conduct that “might [have] impair[ed]” Wang in current or future “employment sit- uations,” and “well might have dissuaded a reasonable worker from making or supporting a charge of” retaliation. Nelson v. Upsala Coll., 51 F.3d 383, 387 (3d Cir. 1995) (first two quota- tions); Burlington, 548 U.S. at 68 (third one) (internal quotation marks omitted). Advocating for a paper’s retraction for academic misconduct can harm Wang’s “reputation in the academic com- munity” by “diminish[ing] his academic reputation, thus deny- ing him opportunities” for future employment and publications. Wu v. Thomas, 863 F.2d 1543, 1545 (11th Cir. 1989) (upholding professors’ claims of Title VII retaliatory harassment). An aca- demic’s reputation is his livelihood, so the harm to Wang’s rep- utation is non-trivial. The retraction itself could likewise dissuade a reasonable academic from reporting discrimination. In academia, publica- tions and citations are the coin of the realm. They affect salary, promotions, and job opportunities. It is “publish or perish,” so losing a publication is no small matter—not to mention the stigma of a retraction. Theidon v. Harvard Univ., 948 F.3d 477, 483 n.4 (1st Cir. 2020). In response, the university (though not the hospital system) says Wang cannot impute its employees’ actions to it. But that confuses § 1983’s state-action inquiry with Title VII’s. Under Title VII, we routinely impute employees’ harassment to their employers when senior employees know about it. See Huston v. Procter & Gamble Paper Prods. Corp., 568 F.3d 100, 107 (3d Cir. 2009). And even with dual employers, under Title VII we do not limit liability to the employer who had more control.

27 Faush v. Tuesday Morning, Inc., 808 F.3d 208, 215 (3d Cir. 2015). Wang raised a genuine dispute about whether Pitt “man- agement knew or should have known about the harassment, but” instead of stopping it, fanned the flames. Jensen v. Potter, 435 F.3d 444, 453 (3d Cir. 2006) (Alito, J.), overruled in part and on other grounds by Burlington, 548 U.S. at 60, 64. At the very least, his superiors at the university (Saba and Gladwin) knew of the retaliation and “failed to undertake prompt correc- tive measures within [their] control.” EEOC v. Glob. Horizons, Inc., 915 F.3d 631, 641 (9th Cir. 2019) (quoting EEOC guid- ance). Because employees “sufficiently senior in [the univer- sity’s] governing hierarchy, or otherwise in a position of admin- istrative responsibility” over Wang, knew of his harassment and publicly accused him of academic misconduct, there is a genuine dispute of material fact over whether the university is responsible for the adverse actions. Huston, 568 F.3d at 107. 3. There is enough for causation. As the District Court rightly found, there are genuine disputes about causation and possible legitimate, nondiscriminatory reasons for the adverse actions. Title VII (and thus the PHRA and § 1981) requires only but-for causation. Univ. of Tex. Sw. Med. Ctr. v. Nassar, 570 U.S. 338, 360 (2013). That is easily met here, as Wang’s oppo- sition to race-based discrimination led to everything that hap- pened to him in the summer and fall of 2020. Thus, Wang’s Title VII/PHRA claims against the hospital and University pro- ceed. Because Wang’s § 1981 claim against Pitt was dismissed before discovery even though it was sufficiently pleaded, we will vacate the District Court’s dismissal of that claim and remand for further proceedings.

28 B. The Title VI claim survives against the hospital, but not the university Title VI outlaws discrimination (and by extension, retalia- tion) by recipients of federal funds. 42 U.S.C. § 2000d; 28 C.F.R. § 42.107(e) (protecting complainants from retaliation); 34 C.F.R. § 100.7(e) (same). Title VI retaliation claims are gov- erned by the same framework as Title VII, PHRA, and § 1981 claims. See Woodson v. Scott Paper Co., 109 F.3d 913, 920 (3d Cir. 1997); Hankins v. Temple Univ. (Health Scis. Ctr.), 829 F.2d 437, 441 (3d Cir. 1987). The university and hospital sys- tem both take federal money. To state a claim under Title VI, though, Wang must also plead that “a primary objective of the Federal financial assistance is to provide employment.” 42 U.S.C. § 2000d-3. The District Court dismissed the Title VI claim against the university for lack of retaliation; the reasoning above applies equally to revive this part of the claim. And it dismissed the Title VI claim against the hospital system for failing to allege the required primary purpose. Yet Wang did allege that the hos- pital system “receives federal funds to employ residents and fel- lows in its residency, G[raduate] M[edical] E[ducation], and fellowship programs.” App. 137. Medical residency straddles the line between education and employment. Cf. Doe v. Mercy Cath. Med. Ctr., 850 F.3d 545, 556, 559 (3d Cir. 2017) (finding it plausible that residency programs are education and residents are employees). At this stage, all Wang need do is allege that the hospital system gets federal money to fund residents. He barely clears that bar.

29 But his allegations against the university come up short. He alleges only that it “receives federal funds that, among other things, assist it in providing a medical education for students at” the school. App. 137 ¶ 6. That statement does not allege a primary purpose of funding employment. 42 U.S.C. § 2000d-3. The District Court dismissed Wang’s Title VI claim on other grounds, so Wang will have an opportunity to amend. IV. WANG’S FIRST AMENDMENT RETALIATION CLAIM FAILS Wang also sued the university, hospital system, physicians practice, Berlacher, Saba, and Gladwin under 42 U.S.C. § 1983 for violating his First Amendment rights. The District Court dismissed the claims against the institutional defendants for lack of state action and later granted summary judgment to Ber- lacher and Saba for the same reason and to Gladwin because he was not involved in demoting Wang. Even if Wang could make out a First Amendment retaliation claim, those rulings were right. A. Most defendants did not act on Pennsylvania’s behalf Because the First Amendment does not restrict private par- ties, only “state actors” can be liable. Leshko v. Servis, 423 F.3d 337, 339 (3d Cir. 2005). People and business count as state actors when the state encourages or joins in the activity, or is “inte- grally related” to the actor, so that the “source fairly can be said to be the state.” Id. at 340. What is more, § 1983 has a tighter standard for vicarious liability than Title VII. Title VII follows agency law, often holding employers vicariously liable and imputing supervisors’ actions to them. Moody v. Atl. City Bd. of Educ., 870 F.3d 206, 215–16 (3d Cir. 2017). Yet for his § 1983 claim, Wang must show that “the state has exercised

30 control over the particular conduct that gave rise to the plain- tiff’s alleged constitutional deprivation.” Kach v. Hose, 589 F.3d 626, 649 (3d Cir. 2009). Wang has not made that showing. Though affiliated with the public university, the hospital system and physician practice are private corporations. They, not the university, are responsible for their employees’ actions. See id. at 646. Their academic affil- iation alone is not enough to make them liable; “[a]ction taken by private entities with the mere approval or acquiescence of the State is not state action.” Am. Mfrs. Mut. Ins. Co. v. Sulli- van, 526 U.S. 40, 52 (1999). Wang’s allegations trying to link the hospital system’s and physician practice’s actions to the university are conclusory. True, the people who disciplined him and retaliated against him were employed by both the hospital system and university. Yet dual employment is not enough for state action. See Borrell v. Bloomsburg Univ., 870 F.3d 154, 160 (3d Cir. 2017). The key question is whether each employee “was wearing his [hospital- system] hat or his [university] hat when he decided to” take adverse action. Id. Berlacher, Saba, and Gladwin were all wear- ing their hospital-system hats. In dismissing Wang from his hospital-system directorship and restricting his access to hos- pital-system residents and fellows, Berlacher and Saba acted as hospital-system employees. They testified that the decision to remove Wang as director and bar him from contacting residents and fellows was made without consultation with anyone from the university. Berlacher’s and Saba’s tweets, as well as the tweet from the hospital system’s account, neither stated the university’s views nor were made on its behalf. See Lindke v. Freed, 601 U.S. 187, 191 (2024) (treating social-media posts

31 as state action only when the official who posted had actual authority to speak for the state and purported to do so on social media). And Berlacher did not even have supervisory authority over Wang at the university. Saba’s purported bar on Wang’s interacting with medical students unaffiliated with the hospital system is a closer case. True, university officials knew about the supposed bar yet did not lift it for two months. Even so, that is not enough for state action. To start, it was effectively a restriction in name only, as Wang never taught classes at the medical school, mostly teach- ing residents and fellows at the hospital. And Saba never pur- ported to bar him on behalf of the university. Three days later, Berlacher and Saba sent another email from Saba’s hospital- email address, telling Wang: “[W]e can no longer have you serve in any medical education role in the institution, specifi- cally pertaining to the general cardiology and electrophysiol- ogy fellows, while also including medical students and resi- dents.” App. 861. In context, they were wearing their hospital- system hats. Saba had no authority to take away his university privileges. He did not consult with the medical-school dean. The dean later reversed Saba’s decision, telling Berlacher, Saba, and Gladwin that “personnel decisions regarding Dr. Wang at [the hospital] are not applicable to the Univer- sity.” App. 912. Nor is Gladwin responsible for that supposed bar. Though he emailed the medical school that “[w]e have taken immediate action and removed [Wang] from [his] leadership position,” he said nothing about interacting with medical students and did not in fact remove Wang from any positions. App. 2764. He, Berlacher, and Saba all testified without contradiction that the

32 decision was made without Gladwin. Gladwin’s after-the-fact defense of Berlacher’s and Saba’s actions is not enough to cre- ate state action, as he “played no affirmative part in depriving” Wang of his rights. Rizzo v. Goode, 423 U.S. 362, 377 (1976). B. Wang’s claims against the university fail That leaves the public university itself. We have called Pitt an “instrumentalit[y] of the state, both in name and in fact.” Krynicky v. Univ. of Pittsburgh, 742 F.2d 94, 103 (3d Cir. 1984). Thus, we have previously held, “the University of Pitts- burgh [is a] state actor[ ] when [it] make[s] employment deci- sions.” Molthan v. Temple Univ. of Commw. Sys. of Higher Educ., 778 F.2d 955, 961 (3d Cir. 1985). Yet the District Court and the parties treat it as a “municipal subdivision” subject to Monell liability. Frazer v. Temple Univ., 25 F. Supp. 3d 598, 607 (E.D. Pa. 2014); Monell, 436 U.S. at 691. We need not decide whether that was right. Even if it was, on the face of Wang’s complaint, the university is not plausibly liable under § 1983. That provision does not allow vicarious lia- bility. Parkell v. Danberg, 833 F.3d 313, 330 (3d Cir. 2016). To hold the university liable, Wang must show either that its pol- icy or custom directly caused his constitutional harm or that it knowingly acquiesced in constitutional violations. Id.; Monell, 436 U.S. at 694. He alleges neither. On policy or custom, Wang does not argue that the university’s policies harmed him, but implies just the opposite—that the employees violated the uni- versity’s policies by infringing on his academic freedom. Plus, the crucial decisions were made by Berlacher and Saba; Gladwin and the medical-school dean just relayed and com- mended others’ actions. And though Wang argues that the either

33 the university “implemented” or the dean “ratified” Berlacher and Saba’s bar on interacting with medical-school students, he never says how. App. 324–25; City of St. Louis v. Praprotnik, 485 U.S. 112, 124 (1988). Although the medical-school dean knew of the bar, it apparently applied only at the hospital-sys- tem and was decided by Wang’s superiors there. Neither Gladwin nor the dean controlled Wang’s hospital-system employ- ment, and Wang did not interact with medical students at the university. In short, Wang’s conclusory allegations do not plau- sibly allege the university’s “participation in or actual knowledge of and acquiescence in the wrongful conduct.” Chavarriaga v. N.J. Dep’t of Corr., 806 F.3d 210, 222 (3d Cir. 2015). So the District Court properly rejected Wang’s First Amendment retaliation claim. * * * * * The First Amendment is part “of a profound national com- mitment to the principle that debate on public issues should be uninhibited, robust, and wide-open.” N.Y. Times v. Sullivan, 376 U.S. at 270. And civil-rights protections “serve as a spur or catalyst to cause employers to self-examine and to self- evaluate their employment practices and to endeavor to elimi- nate, so far as possible, the last vestiges of discrimination.” McKennon v. Nashville Banner Publ’g Co., 513 U.S. 352, 358 (1995) (cleaned up). Taking Wang’s allegations as true, Berlacher, Saba, the Asso- ciation, the university, and the hospital system did not just exer- cise their own academic freedom; they defamed and retaliated against Wang for calling out what he reasonably believed was their own discrimination. His defamation claim may proceed

34 against those five, as may his civil-rights retaliation claims against all appellees. And the District Court should grant him leave to amend his Title VI claim against the university. But his First Amendment claim fails for lack of state action. We will thus AFFIRM IN PART, VACATE IN PART, and REMAND to the Dis- trict Court for further proceedings consistent with this opinion.

35 Counsel for Appellant Shawn M. Rodgers GOLDSTEIN LAW PARTNERS John R. Renner [Argued] Michael E. Rosman Michelle A. Scott CENTER FOR INDIVIDUAL RIGHTS Counsel for Appellee Wiley Periodicals Inc. Elizabeth Seidlin-Bernstein Michael L. Berry [Argued] BALLARD SPAHR Counsel for Appellee American Heart Ass’n, Inc. Howard A. Chajson Steven L. Ettinger [Argued] John C. Conti Jeffrey J. Wetzel DICKIE MCCAMEY & CHILCOTE Counsel for Appellee Marc Simon Bryon M. Chowka Robert J. Ridge [Argued] CLARK HILL Counsel for Appellee University of Pittsburgh Jeremy D. Engle MARCUS & SHAPIRA Jeremy W. Brinster Brian Boynton [Argued] Bruce M. Berman Molly M. Jennings WILMER CUTLER PICKERING HALE & DORR Counsel for Appellees University of Pittsburgh Medical Center, University of Pittsburgh Physicians, Samir Saba, Mark Gladwin, and Kathryn Berlacher Emilie R. Hammerstein Taylor Brailey Terrence H. Murphy Theodore A. Schroeder [Argued] LITTLER MENDELSON

37 SHWARTZ, Circuit Judge concurring in part, dissenting in part.

I agree with the disposition my colleagues reached on all claims except Wang’s defamation claims. I conclude the defamation claims fail for two reasons. First, the challenged statements are non-actionable opinions of academic research and scientific methodology based on disclosed facts. Second, even if those statements were actionable, Wang is a limited purpose public figure, but his complaint does not allege facts from which to infer that the challenged statements were made with actual malice. Accordingly, I would affirm the District Court’s order dismissing Wang’s defamation claims.

I1

Wang is a cardiologist affiliated with the University of Pittsburgh Medical Center (“UPMC”) and the University of

1 The Majority’s fact section merges allegations from the operative complaint (the “FAC”) with facts revealed during discovery and characterizes some facts differently from my view. Because the defamation claims were dismissed under Federal Rule of Civil Procedure 12(b)(6), this fact section relies only on the factual allegations and other materials that may be considered when evaluating such a ruling. “In reviewing a motion to dismiss under Rule 12(b)(6), we treat as true all well-pleaded facts in the complaint, which we construe in the ‘light most favorable to the plaintiff.’” Santomenno ex rel. John Hancock Tr. v. John Hancock Life Ins. Co. (U.S.A), 768 F.3d 284, 290 (3d Cir. 2014) (citations omitted).

1 Pittsburgh School of Medicine (“Pitt”). 2 At UPMC, Wang serves as a doctor and previously directed the fellowship program in clinical cardiac electrophysiology. At Pitt, Wang is a faculty member.

Wang wrote an article entitled “Diversity, Inclusion, and Equity: Evolution of Race and Ethnicity Considerations for the Cardiology Workforce in the United States of America from 1969 to 2019” (the “Article”). App. 139 (FAC ¶ 20), 178- 94. He submitted the Article to the Journal of the American Heart Association (“JAHA”). JAHA is published by the American Heart Association (“AHA”) and Wiley Periodicals, Inc. 3 (“Wiley”). In March 2020, AHA and Wiley published the Article in JAHA after the “normal review and vetting process.” App. 139-40 (FAC ¶ 21).

The Article “trac[ed] the history of the use of race and ethnicity as factors in determining admission into medical schools, residency programs, and fellowships.” App. 139 (FAC ¶ 20). It “asserted that the medical profession had not been successful in reaching its goals of increasing the percentages of underrepresented races and ethnicities in the medical profession generally, and cardiology in particular.” App. 139 (FAC ¶ 20). The Article argued that, “to achieve those goals,” “programs . . . applied different standards to applications by members of underrepresented races and ethnicities and raised questions about the legality,

2 UPMC is the sole owner of Defendant University of Pittsburgh Physicians (“UPP”), which employs physicians on behalf of UPMC, including Wang. I refer to both entities collectively as “UPMC.” 3 AHA and Wiley are private corporations.

2 effectiveness, and wisdom of doing so.” App. 139 (FAC ¶ 20). The Article contained 108 endnotes with embedded hyperlinks.

In late July 2020, “various individuals on social media” criticized the Article, App. 140 (FAC ¶ 22), including Dr. Samir Saba and Dr. Katheryn Berlacher, who are both affiliated with UPMC and Pitt. On August 2, 2020, Saba retweeted a Twitter post about the Article and wrote the Article “uses misquotes, false interpretations, and racist thinking to defend a single person’s conclusion[s]. We are outraged that @JAHA_AHA published this shameful and infuriating piece.” App. 145 (FAC ¶ 43) (quoting App. 233). Later that day, Berlacher retweeted a Twitter post about the Article and added that it “misinterprets data and misquotes people” and is “scientifically invalid and racist.” App. 145 (FAC ¶ 44) (quoting App. 231, 235). 4

On August 5, 2020, AHA and Wiley retracted the Article “[d]espite the fact that [it] was vetted pursuant to the usual procedures for review of all articles in JAHA.” App. 143 (FAC ¶ 39). The Article remains available on JAHA’s website but contains a label noting its retracted status. In connection with the retraction, Wang contends that AHA and Wiley made three defamatory statements against him. First, AHA and

4 Later on August 6, Saba emailed UPMC employees and “asserted that the Dean of UPSOM, Dean Anantha Shekhar, sought the retraction of [the Article].” App. 145 (FAC ¶ 43). Wang initially claimed this statement was defamatory, see App. 150 (FAC ¶ 74), but omitted it from his list of defamatory claims in his briefing, thereby abandoning it, see Appellant’s Br. at 35-37.

3 Wiley issued a statement on August 5 that accompanied the Article’s retraction (the “Retraction Statement”). It stated, in relevant part:

The author’s institution, the University of Pittsburgh Medical Center (UPMC), has notified the Editor-in-Chief that the article contains many misconceptions and misquotes and that together those inaccuracies, misstatements, and selective misreading of source materials strip the paper of its scientific validity. . . . The Editor-in-Chief deeply regrets publishing the article and offers his apologies.

App. 143 (FAC ¶ 39), 237. The Retraction Statement did not provide examples of misconceptions or misquotes from the Article but noted JAHA would “be publishing a detailed rebuttal.” App. 237. The next day, AHA and Wiley posted an updated Retraction Statement (the “Updated Retraction

4 Statement”) 5 with virtually the same paragraph 6 as well as two specific “misconceptions and misquotes identified in correspondence from UPMC” to serve as a “sample” of the Article’s errors. 7 App. 163-64.

5 Even though the Updated Retraction Statement was not explicitly mentioned in the complaint, we may consider it because it: (1) was explicitly referenced in a document referenced in the complaint, (2) was attached to the motion to dismiss, and (3) is undisputedly authentic and integral to the complaint. See Buck v. Hampton Twp. Sch. Dist., 452 F.3d 256, 260 (3d Cir. 2006) (recognizing that courts can consult “documents that are attached to or submitted with the complaint, and any matters incorporated by reference or integral to the claim, items subject to judicial notice, matters of public record, orders, [and] items appearing in the record of the case” (alteration in original) (citations and quotation marks omitted)); Schmidt v. Skolas, 770 F.3d 241, 249 (3d Cir. 2014) (“[A] document integral to or explicitly relied upon in the complaint may be considered without converting the motion to dismiss into one for summary judgment.” (citations and internal quotation marks omitted)); see also In re Burlington Coat Factory Sec. Litig., 114 F.3d 1410, 1426 (3d Cir. 1997) (“Plaintiffs cannot prevent a court from looking at the texts of the documents on which its claim is based by failing to attach or explicitly cite them.”). 6 The Updated Retraction Statement uses “void” rather than “strip.” App. 163, 237. 7 One example claimed that Wang “falsely implie[d] that the Ohio State University (OSU) program ranked racial and ethnic groups underrepresented [classified as underrepresented minorities or “URM”] in medicines with

5 Second, AHA issued a statement (the “AHA Statement”), providing, in relevant part:

The [Article] has rightfully drawn criticism for its misrepresentations and conclusions. As an organization focused on the relentless pursuit of longer, healthier lives for everyone everywhere, the [AHA] denounces the views expressed in the article and regrets its role in enabling those views to be promoted. Those views are a misrepresentation of the facts and are contrary to our organization’s core values and historic commitment to promoting diversity and inclusion in medicine and science. . . . While [JAHA] and the other AHA scientific journals are editorially independent of the Association,

lower academic credentials in order to achieve diversity,” App. 163, but JAHA noted that a complete recitation of the source Wang relied on for the statement said that “USMLE score comparison shows no significant difference between URM and non-URM fellows matched to the OSU program,” App. 163. The second example claimed Wang “falsely implie[d] that working in an underserved area may be attributable to [an] inability to secure a job in other areas because of low professional qualifications” based on a study comparing board- certified doctors to non-board-certified doctors. App. 163-64. The Updated Retraction Statement explained that the source showed that the study compared two different groups of non- board-certified doctors. The Updated Retraction Statement included the sources it relied upon in endnotes with embedded hyperlinks.

6 we take our responsibility to ensure factual accuracy seriously. The independent editors of JAHA and the [AHA] are reviewing the journal’s peer-review and publication processes to ensure future submissions containing deliberate misinformation or misrepresentation are never published. The journal can and will do better.

App. 166-67; see also App. 144 (FAC ¶ 39).

Third, JAHA published another article entitled “A Path Forward,” which “stated that [Wang’s] institution had requested the article be retracted ‘[based on] specific scientific errors as well as misleading and incomplete quotations.’” App. 144 (FAC ¶ 39) (quoting App. 170-71).

Separately, Wang also identified as defamatory an article published by JAHA and authored by Dr. Marc Simon, a UPMC and Pitt colleague. This piece, entitled “Equity, Diversity, and Inclusiveness in Cardiovascular Healthcare and Medicine,” (the “Simon Article”), App. 145 (FAC ¶¶ 45), 195- 96, claimed that the Article “misrepresented facts” and “evidence” “[in its efforts] to argue against affirmative action in the field of cardiology,” App. 145 (FAC ¶ 45) (quoting App. 195). The Simon Article’s sources were set forth in endnotes with hyperlinks.

Wang filed defamation claims against Defendants, alleging that: (1) Saba and Berlacher defamed him in their August 2 tweets, (2) Simon defamed him in the Simon Article, (3) UPMC and Pitt defamed him via their employees’ statements, (4) AHA defamed him in the Retraction Statement, the AHA Statement, and A Path Forward, and (5) Wiley

7 defamed him in the Retraction Statement and A Path Forward.

II 8

In assessing a defamation claim, a “court must determine: (1) whether the defendants have harmed the plaintiff’s reputation within the meaning of state law; and (2) if so, whether the First Amendment nevertheless precludes recovery.” Marcone v. Penthouse Int’l Mag. for Men, 754 F.2d 1072, 1077 (3d Cir. 1985) (citations and quotations omitted).

A

To succeed on a defamation claim under Pennsylvania law, a plaintiff must show, among other things, “[t]he defamatory character of the communication” at issue. 42 Pa.

8 We exercise plenary review of a district court’s order granting a motion to dismiss for failure to state a claim. Burtch v. Milberg Factors, Inc., 662 F.3d 212, 220 (3d Cir. 2011). To survive a motion to dismiss, a complaint must contain sufficient factual matter, accepted as true, to state a claim to relief that is plausible on its face. A claim has facial plausibility when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged. Sheridan v. NGK Metals Corp., 609 F.3d 239, 262 n.27 (3d Cir. 2010) (citations and quotations omitted); see also James v. City of Wilkes-Barre, 700 F.3d 675, 679 (3d Cir. 2012) (“[W]e disregard rote recitals of the elements of a cause of action, legal conclusions, and mere conclusory statements.”).

8 Cons. Stat. § 8343(a)(1). 9 “A communication is defamatory if it tends so to harm the reputation of another as to lower him in the estimation of the community or to deter third persons from associating or dealing with him.” Birl v. Philadelphia Elec. Co., 167 A.2d 472, 475 (Pa. 1960) (citations omitted); Graboff v. Colleran Firm, 744 F.3d 128, 136 (3d Cir. 2014). Statements of opinion, as opposed to statements of fact, may be viewed as having such an effect if they are based on undisclosed facts. Remick v. Manfredy, 238 F.3d 248, 261 (3d Cir. 2001); Parano v. O’Connor, 641 A.2d 607, 609 (Pa. Super. Ct. 1994) (holding that “expressions of . . . opinion” were “not actionable unless they imply undisclosed, false and defamatory facts”).

In the course of academic debate, criticisms and characterizations of scientific methodology and scholarly analysis are better understood as statements of opinion rather than fact. See ONY, Inc. v. Cornerstone Therapeutics, Inc., 720 F.3d 490, 497 (2d Cir. 2013) (“[W]hile statements about contested and contestable scientific hypotheses constitute assertions about the world that are in principle matters of verifiable ‘fact,’ for purposes of the First Amendment and the laws relating to fair competition and defamation, they are more closely akin to matters of opinion, and are so understood by the relevant scientific communities.”); Pacira BioSciences, Inc. v.

9 A Pennsylvania defamation plaintiff must also show: (1) the statement’s publication by the defendant, (2) the statement’s application to the plaintiff, (3) the understanding by the recipient of its defamatory meaning, (4) the understanding by the recipient of it as intended to be applied to the plaintiff, (5) special harm resulting to the plaintiff from its publication, and (6) the lack of a conditionally privileged occasion. 42 Pa. Cons. Stat. § 8343(a).

9 Am. Soc’y of Anesthesiologists, Inc., 63 F.4th 240, 246 (3d Cir. 2023) (determining that “tentative scientific conclusions” were opinions in libel claim); Saad v. Am. Diabetes Ass’n, 123 F. Supp. 3d 175, 178 (D. Mass. 2015) (holding that medical journal’s expression of concern about “the reliability of the data” in professor’s articles was protected opinion).

The Majority holds that accusations of “misquoting and misreading data and sources” are “falsifiable,” thereby rendering them “more hit job than academic argument.” Majority at 11. Setting aside the Majority’s characterization of these statements as a “hit job,” one academic accusing another of “misreading” data is not actionable because it reflects an opinion about the other’s work based on disclosed facts. See Pacira, 63 F.4th at 247 (“[M]ere disputes about the reliability of a scientific study’s disclosed methodology cannot create an actionable falsehood . . . as such disputes do not address whether the statements themselves are verifiable.”); ONY, 720 F.3d at 497 (declining to recognize a cause of action where “plaintiff allege[d] that the inferences drawn from . . . data were the wrong ones”).

That is what happened here. First, AHA and Wiley criticized the Article for “contain[ing] many misconceptions and misquotes and that together those inaccuracies, misstatements, and selective misreading of source materials strip the paper of its scientific validity.” App. 143, 163-65, 237. 10 Likewise, the Simon Article stated that the Article “misrepresented facts in its effort[] to argue against affirmative

10 The AHA Statement and A Path Forward made similar claims.

10 action in the field of cardiology” and included the “misrepresentation of evidence.” App. 145, 195-96.

In making these claims, each author explicitly referred to the Article—the subject of their criticism—and discussed the bases for their opinions, namely the sources listed in the Article’s endnotes. Indeed, the Updated Retraction Statement even identified specific errors as a “sample of [Wang’s] misconceptions and misquotes.” App. 163.

Thus, JAHA’s readers—medical professionals and academics equipped to evaluate the challenged statements and the sources on which they were based—had the opportunity and information to assess for themselves the disputed conclusions. See Pacira, 63 F.4th at 249 (determining statements were based on disclosed facts where “[academic] journal’s readers were provided the basis for the statements, have the expertise to assess their merits based on the disclosed data and methodology, and thus are equipped to evaluate the opinions the authors reached”); Cheng v. Neumann, 51 F.4th 438, 447 (1st Cir. 2022) (concluding an opinion was based on disclosed facts and thus “not actionable because [it] provide[d] a link to the source material, which, in the context of the [controversy], enable[d] readers to draw their own conclusions ‘based on facts accessible to everyone’” (citations omitted)).

The Majority insists “[i]t would have been just fine to say that Wang’s article was wrong, harmful, or even dangerous” because “[t]hose are normal academic opinions.” Majority at 12. That is exactly what the speakers did here. In any event, whether “normal” or not, courts should be cautious when weighing in on academic debates because doing so chills scientific progress and the rigorous discourse required to bring

11 it about. See Joseph v. Springer Nature, No. 20 civ. 4672 (JPC), 2021 WL 1372952, at *7 (S.D.N.Y. Apr. 12, 2021) (declining to consider a retraction statement as actionable where the underlying dispute concerned “a critique of a scientific work” in a contested area of research “that the Court has no business solving”), aff’d sub nom. Joseph v. Springer Nature Am. Inc., No. 21-959-cv, 2021 WL 6105369 (2d Cir. Dec. 21, 2021) (summary order). As the Court of Appeals for the Seventh Circuit held, “scientific controversies must be settled by the methods of science rather than by the methods of litigation. . . . More papers, more discussion, better data, and more satisfactory models—not larger awards of damages— mark the path toward superior understanding of the world around us.” Underwager v. Salter, 22 F.3d 730, 736 (7th Cir. 1994); see also Pacira, 63 F.4th at 247-48 (holding that a defendant’s characterization of academic work as employing a “flawed method” was not actionable because “data and methodology may be the basis of future scholarly debate, but they do not form the basis for trade libel under New Jersey law” as concluding “otherwise would risk ‘chilling’ the natural development of scientific research and discourse”). 11

11 Of course, academic disputes are not always beyond the scope of defamation claims. Indeed, there is a marked distinction between criticizing an academic’s work and the academic themselves. As the Court of Appeals for the Seventh Circuit observed, “[i]f a professor is falsely accused of plagiarism or sexual harassment or selling high grades or other serious misconduct, rather than of having unsound ideas, he has the same right to damages as any other victim of defamation.” Dilworth v. Dudley, 75 F.3d 307, 310 (7th Cir. 1996).

12 Second, Saba and Berlacher’s tweets are also non- actionable opinions about an academic work based on disclosed facts because they explicitly refer to the Article they critiqued and the reasons for their views. See Veno v. Meredith, 515 A.2d 571, 573-75 (Pa. Super. Ct. 1986) (concluding that an article constituted non-actionable opinion because it “merely stated an opinion” on another article and the defendant author did not “imply that he was privy to some additional facts, beyond [the article] which supported his opinion”). Furthermore, although their characterization of the Article as “racist” may be inflammatory, it is not actionable under Pennsylvania law. See McCafferty v. Newsweek Media Grp., Ltd., 955 F.3d 352, 358 (3d Cir. 2020) (“In Pennsylvania, ‘a simple accusation of racism’ is not enough . . . Rather, the accusation must imply more, by for instance suggesting that the accused has personally broken the law to ‘act[ ] in a racist manner.’” (second alteration in original) (quoting MacElree v. Philadelphia Newspapers, Inc., 674 A.2d 1050, 1055 (Pa. 1996)); 12 see also McAndrew v. Scranton Republican Publ’g

12 The Majority relies on MacElree v. Philadelphia Newspapers, Inc., 674 A.2d 1050, 1055 (Pa. 1996). There, the court determined that calling a local district attorney “the David Duke of Chester County” was actionable because it “could be construed to mean that [he] was acting in a racist manner in his official capacity as district attorney” by “abusing his power . . . to further racism and his own political aspirations.” Id. at 1054-55. In reaching this holding, however, the court made clear “[t]his is not to say that every accusation of racism is defamatory.” Id. at 1055. Interpreting MacElree, this Court recognized that an accusation of racism absent “specific, unlawful wrongdoing”

13 Co., 72 A.2d 780, 783 (Pa. 1950) (“It would doubtless be very annoying for a man to be charged with bigotry of any kind, yet it has been held that for a man to be charged with such bigotry is not defamation.”); Rybas v. Wapner, 457 A.2d 108, 110 (Pa. Super. Ct. 1983) (“A publication which charges that an individual is actuated by unpleasant or undesirable prejudice may offend his sensitivities, but is not thereby libelous.”).

Because Defendant’s statements are nonactionable opinions, I would affirm the order dismissing Wang’s defamation claims. 13

was not actionable. McCafferty v. Newsweek Media Grp., Ltd., 955 F.3d 352, 358 (3d Cir. 2020). I apply this same logic to conclude that the accusations that Wang’s article were “racist” are insufficient to support a cause of action given the lack of any assertion that he engaged in unlawful conduct. 13 Although the allegedly defamatory statements are best understood as opinions, even if interpreted as factual, they are truthful statements about the accuracy of certain matters Wang asserted, and are thus not actionable for three reasons. First, Wang cited a paper discussing admission standards for URM applicants at OSU. He quoted the paper as stating: [W]e simply made it a priority to rank [URM] applicants more aggressively than in previous years, thus achieving success in matching them regardless of recruiting efforts, with the implication being that we accepted less competitive applicants in an effort to increase diversity. App. 414 (second alteration in original). After this quote, Wang added: “[e]ncouraging the explicit use of race and

14 ethnicity for employment reveals a lack of knowledge regarding legal permissibility and fellow status.” App. 414. The Retraction Statement highlighted that Wang omitted the sentence immediately following the quoted excerpt, which provided:

Although the measurement of an applicant’s achievement reflected in his or her application and, by extrapolation, the applicant’s potential for being a successful fellow, cannot be fairly quantified, USMLE score comparison shows no significant difference between URM and non- URM fellows matched to the OSU program.

App. 163. In other words, the Article suggested that the original work stood for the proposition that less qualified candidates were accepted into the OSU program in the name of diversity, but the original work stated that there was no significant difference in test scores (i.e., qualification) between candidates from different racial groups.

Second, in a section of the Article relating to the qualification of URM, Wang suggested another researcher’s hypothesis that “working in an underserved area may be attributable to inability to secure a job in other areas because of low professional qualifications” was bolstered by:

a 1995 California study that demonstrated primary care physicians who were not board certified were 1.6 times more likely to work in

15 rural underserved areas when compared with board-certified counterparts. App. 163-64, 420. However, as the Updated Retraction Statement noted, this 1995 study in fact stated: Among physicians completing postgraduate medical education after 1974, board-certified family physicians were [three] times more likely to locate in medically underserved rural communities than were other primary care physicians. Non-board-certified family and general physicians were 1.6 times more likely than other non-board-certified primary care physicians to locate in rural underserved areas. App. 164. In other words, the original source compared rural employment rates for non-board-certified family and general physicians for non-board-certified primary care physicians, but Wang described it as comparing rural employment rates between board-certified and non-board-certified primary care physicians. While Wang did not include quotation marks, he (incorrectly) described the results of a study and, at the very least, misrepresented what it demonstrated. Third, the Retraction Statement’s assertion of “many” errors does not change the truth-as-a-defense analysis. The use of the word “many” could change the meaning of a statement such that it is defamatory—for example, one alleged act of misconduct may be less damaging than fifty allegations, depending on the severity of that single act. However, “defamation case law affords leeway in the reporting of numbers if the discrepancy does not alter the gist of what is being reported and does not produce a different effect in the mind of the reader.” Croce v. N.Y. Times Co., 345 F. Supp.

16 B

Even if any of Defendants’ statements were actionable under state law, Wang’s claims are still prohibited under the First Amendment’s public figure doctrine. Like my colleagues, I conclude that Wang is a limited purpose public figure and failed to allege actual malice as to Defendants Wiley

3d 961, 987 n.5 (S.D. Ohio 2018), aff’d, 930 F.3d 787 (6th Cir. 2019). Indeed, Pennsylvania courts recognize that “[t]he law does not require perfect truth, so long as any inaccuracies do not render the substance and ‘gist’ of the statements untrue. . . . The ‘gist’ of a statement is true if the effect upon a reader is the same regardless of the inaccuracy.” ToDay’s Hous. v. Times Shamrock Commc’n, Inc., 21 A.3d 1209, 1215 (Pa. Super. Ct. 2011); Rubin v. CBS Broad. Inc., 170 A.3d 560, 565 (Pa. Super. Ct. 2017) (“Substantial truth ‘absolve[s] a defendant even if she cannot justify every word of the alleged defamatory matter; it is sufficient if the substance of the charge be proved true, irrespective of slight inaccuracy in the details.’” (alteration in original) (quoting Masson v. New Yorker Mag., Inc., 501 U.S. 496, 516-17 (1991))). The gist of the statement remains accurate because it conveyed that more than one statement mischaracterized its source. Thus, if the alleged defamatory statements are viewed as factual rather than opinions, then truth as a defense would apply to all of Wang’s defamation claims. See Budhun v. Reading Hosp. & Med. Ctr., 765 F.3d 245, 259 (3d Cir. 2014) (“While ordinarily a party may not raise affirmative defenses at the motion to dismiss stage, it may do so if the defense is apparent on the face of the complaint.” (citation omitted)).

17 and Simon, 14 but I part company with their conclusion that the pleadings are sufficient to establish AHA, Saba, and Berlacher’s allegedly defamatory statements were made with actual malice.

As a limited purpose public figure, Wang had to plead that the allegedly defamatory statements were made with “actual malice.” Marcone, 754 F.2d at 1081-82. “‘Actual malice’ is a term of art that does not connote ill will or improper motivation.” McCafferty, 955 F.3d at 359; Masson v. New Yorker Mag., Inc., 501 U.S. 496, 510 (1991) (“Actual malice under the New York Times standard should not be confused with the concept of malice as an evil intent or a motive arising from spite or ill will.”). Instead, “[t]he actual malice standard requires that the defendant make the defamatory remarks with ‘knowledge that [the statement] was false or with reckless disregard of whether it was false or not.’” McDowell v. Paiewonsky, 769 F.2d 942, 951 (3d Cir. 1985) (alteration in original) (quoting N.Y. Times Co. v. Sullivan,

14 A limited public figure is typically one who “actively participates in the public issue in a manner intended to obtain attention.” Marcone v. Penthouse Int’l Mag. for Men, 754 F.2d 1072, 1083 (3d Cir. 1985); Schiavone Constr. Co. v. Time, Inc., 847 F.2d 1069, 1077 (3d Cir. 1988) (“[A]lthough they are not well known throughout the country or on every issue, [they] are nonetheless sufficiently involved in one particular arena to qualify as public figures for that purpose.”). “For example, a meat producer that aggressively advertised its product in the media becomes a limited purpose public figure for comment on the quality of its product.” Marcone, 754 F.2d at 1084 (citing Steaks Unlimited, Inc. v. Deaner, 623 F.2d 264, 273-74 (3d Cir. 1980)).

18 376 U.S. 254, 280 (1964)). “[R]eckless disregard for the truth means that the defendant ‘in fact entertained serious doubts as to the truth’ of the statement,” id. (quoting St. Amant v. Thompson, 390 U.S. 727, 731 (1968)), “or that the defendant had a ‘subjective awareness of probable falsity,’” id. (quoting Gertz v. Robert Welch, Inc., 418 U.S. 323, 335 n.6 (1974)). This “standard is a subjective one, based on the defendant’s actual state of mind,” though a court may rely on “objective circumstantial evidence” to determine actual malice. Schiavone Constr. Co. v. Time, Inc., 847 F.2d 1069, 1089-90 (3d Cir. 1988); see also Kendall v. Daily News Publ’g Co., 716 F.3d 82, 93 (3d Cir. 2013) (“The subjective nature of this inquiry requires that there be some evidence showing, directly or circumstantially, that the defendants themselves understood the potential defamatory meaning of their statement.”).

Actual malice exists where a “publisher’s allegations are so inherently improbable that only a reckless man would have put them in circulation,” or “where there are obvious reasons to doubt the veracity of the informant or the accuracy of his reports.” St. Amant, 390 U.S. at 732; see also Schiavone, 847 F.2d at 1090 (“Where the defendant finds internal inconsistencies or apparently reliable information that contradicts its libelous assertions, but nevertheless publishes those statements anyway, the New York Times actual malice test can be met.”). For example, in McCafferty we reasoned that actual malice could be established where a publisher “made up [a commentator’s] statements, used them to accuse him of breaking the law when it knew he was innocent, or published facts contrary to information it otherwise knew.” 955 F.3d at 360. In Tucker v. Fischbein, we recognized actual malice where an attorney publicly mischaracterized a filed

19 complaint and only later attempted to explain his false statements by claiming that he had failed to read the complaint before speaking on it. 237 F.3d 275, 285 (3d Cir. 2001). In Schiavone, we concluded the defendant acted with actual malice where reliable, independent sources contradicted the veracity of a defamatory claim, and the publisher deliberately omitted “exculpatory” information and “add[ed] editorial comment to draw attention” from readers. 847 F.2d at 1091- 92.

Absent circumstances that indicate a publisher was or should have been on notice of the falsity of their statements, we have generally declined to conclude that defamatory statements were made with actual malice. See, e.g., Marcone, 754 F.2d at 1091 (concluding no actual malice was shown where false statements were not based on facts so “inherently improbable that [publisher] should have been put on notice of the probability of falseness”).

Applying this framework to Wang’s factual allegations, and assuming that Defendants’ statements are false and otherwise defamatory, Wang’s allegations do not show that any of the statements were made with knowledge or reckless disregard of their falsity. See Marcone, 754 F.2d at 1091 (“Although the lack of truthfulness of a statement is a prerequisite to liability, mere falsity, without more, is generally not sufficient to establish actual malice.”). Because I agree

20 with the Majority that Wang failed to establish actual malice as to Wiley and Simon, I will not discuss those claims. 15

As to AHA, the Majority holds that Wang alleged actual malice by relying on: (1) assertions in the AHA Statement, and (2) AHA’s failure to give Wang an opportunity to respond. Majority at 20-21. Neither establishes actual malice.

First, the AHA Statement—“denounc[ing]” Wang’s views and claiming he engaged in “deliberate misinformation or misrepresentation,” App. 166-67—was strongly worded, but not sufficient to show actual malice. Passionate opposition to a speaker’s views does not constitute actual malice, even if motivated by “spite or ill will.” Masson, 501 U.S. at 510. Moreover, AHA’s claim that Wang relied on “deliberate” falsehoods does not show that it acted with knowledge of falsity or reckless disregard for truth. Rather, it is reasonable for AHA to have surmised that if Wang misrepresented sources

15 I would not grant Wang leave to amend his defamation claim against AHA because (1) he already had an opportunity to amend the complaint as to this allegation, and (2) AHA was passing along the hospital’s views, thus undermining an actual malice determination, see, e.g., Hatfill v. N.Y. Times Co., 532 F.3d 312, 325 (4th Cir. 2008) (no actual malice where defendant relied on source “considered an expert in [her] field”), and hence amendment would be futile.

21 in making his argument, he did so “deliberate[ly]” to bolster his conclusions. 16

Second, that AHA did not to provide Wang with notice of the retraction or an opportunity to rebut the claims against him does not constitute actual malice. Indeed, we have previously determined no actual malice existed where a publisher failed to offer the subject of a news report the opportunity to respond. McCafferty, 955 F.3d at 359. Even if Wang had alleged that giving an author notice and an opportunity to respond was standard practice in the industry (which he did not), “an extreme departure from professional

16 Wang characterized the AHA Statement as arguing the Article “was factually inaccurate, affirming it ‘take[s] [its] responsibility to ensure factual accuracy seriously [in the publication of the JAHA]. AHA assured the public and the medical community that ‘[t]he Journal can and will do better[,’] in the future, to prevent the publication of other articles like Plaintiff’s article, which ‘contain[ed] deliberate misinformation’ and ‘misrepresentation,’” App. 144 (FAC ¶ 39) (quoting App. 166-67 (alterations of quoted material in original)), but this distorts the AHA Statement. For example, the AHA Statement provides that it was “reviewing the journal’s peer-review and publication processes to ensure future submissions containing deliberate misinformation or misrepresentation are never published.” App. 167. It does not necessarily characterize the Article as an example of “deliberate misinformation and misrepresentation,” as Wang pleaded, App. 144 (FAC ¶ 39). Even with Wang’s framing of this statement, however, his claim still fails because AHA’s statement does not show that it acted with knowledge of falsity or reckless disregard for truth.

22 standards, without more, will not support a finding of actual malice.” Tucker, 237 F.3d at 286 (citing Harte-Hanks Commc’ns, Inc. v. Connaughton, 491 U.S. 657, 665 (1989)). 17

Wang also failed to plead actual malice as to Saba and Berlacher. Wang contends that actual malice was shown by: (1) the Article’s easily verifiable assertions, (2) Saba and Berlacher’s strong opposition to Wang’s position on race- based admissions, and (3) their orchestrated “attack campaign” against Wang, App. 143 (FAC ¶ 38). Even if true, none constitute actual malice.

First, Saba and Berlacher’s alleged failure to review all of the sources in Wang’s article does not demonstrate actual malice. See Marcone, 754 F.2d at 1089 (“Failure to investigate, without more, does not demonstrate actual malice.”); see also Harte-Hanks, 491 U.S. at 688 (“[F]ailure to investigate before publishing, even when a reasonably prudent person would have done so, is not sufficient to establish reckless disregard.”). “[T]he purposeful avoidance of the truth is” distinguishable from a mere failure to investigate, Harte- Hanks, 491 U.S. at 692, but this case is not one in which the publisher of the statement failed to interview a key witness or necessarily did not consult other objective evidence. Id. at 692- 93. Furthermore, the complaint does not plead facts to suggest

17 By the Majority’s logic, Wiley—as a fellow publisher of JAHA, which distributed the Retraction Statement and A Path Forward—would also have failed to provide Wang with notice and an opportunity to respond. The Majority properly recognized the claim against Wiley should fail, and that the claim against AHA should fail for the same reasons.

23 Saba and Berlacher purposefully avoided learning whether their claims were false.

The Majority insists—without citation—that Berlacher and Saba must have checked all of Wang’s sources and determined they were “substantially accurate,” but still decided to “trash[]” the Article. Majority at 18. This inference is not supported by the complaint. There is no allegation that Saba and Berlacher checked each source. Rather, the allegations show that Saba and Berlacher claimed some sources Wang quoted or relied on were not accurately conveyed and their mere failure to fully investigate does not constitute actual malice. Marcone, 754 F.2d at 1089.

Second, Wang argues that he has pleaded actual malice by pointing at Saba and Berlacher’s opposition to his views and their characterization of the Article as “scientifically invalid and racist.” App. 145 (FAC ¶ 44). That does not show they made their statements with knowledge or reckless disregard of their falsity. See Masson, 501 U.S. at 510 (“Actual malice under the New York Times standard should not be confused with the concept of malice as an evil intent or a motive arising from spite or ill will.”). Although it “cannot be said that evidence concerning motive or care never bears any relation to the actual malice inquiry,” Harte-Hanks, 491 U.S. at 668, the vitriol of a challenged statement alone does not show that it was made with actual malice for defamation purposes. In a similar vein, the Majority adds its view that Saba and Berlacher’s criticism was “a stark departure from normal academic practice.” Majority at 19. Even if this were true, as explained above, “an extreme departure from professional standards, without more, will not support a finding of actual

24 malice.” Tucker, 237 F.3d at 286 (citing Harte-Hanks, 491 U.S. at 665).

Third, even if Saba and Berlacher attempted to induce others to criticize Wang’s work, that conduct would not establish actual malice because it does not suggest they knew or reasonably should have known their claims were false. See McCafferty, 955 F.3d at 359 (“‘Actual malice’ is a term of art that does not connote ill will or improper motivation.”).

Because Wang’s complaint does not establish that Saba and Berlacher acted with actual malice, the public figure doctrine precludes recovery even if Wang’s defamation claims against them were otherwise actionable. 18

III

I agree with the Majority that our First Amendment standard “is supposed to be a shield for free expression, not a sword to silence disfavored speech,” Majority at 23, and that is the very reason why I would dismiss these defamation claims. Thus, I respectfully dissent from Part II of the Majority’s opinion.

18 Because Wang’s defamation claims against Saba, Berlacher, and Simon fail, so too do his defamation claims against Pitt and UPMC predicated on vicarious liability. See Mamalis v. Atlas Van Lines, Inc., 560 A.2d 1380, 1383 (Pa. 1989) (“A claim of vicarious liability is inseparable from the claim against the agent since any cause of action is based on the acts of [an agent].”).