Norman Wang v. University of Pittsburgh

UNITED S TATES C OURT OF APPEALS FOR THE THIRD CIRCUIT
No. 25-1816
N ORMAN WANG ,
Appellant
v.
U NIVERSITY 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: S HWARTZ , BIBAS , and PHIPPS , Circuit Judges
Argued: Mar. 2, 2026; Filed: July 7, 2026
_____________________________
O PINION 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.

-- 1 of 62 --

2
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 of 62 --

3
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 of 62 --

4
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 of 62 --

5
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 of 62 --

6
“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 of 62 --

7
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 of 62 --

8
“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 of 62 --

9
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 of 62 --

10
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 of 62 --

11
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 of 62 --

12
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 of 62 --

13
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 of 62 --

14
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 of 62 --

15
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 of 62 --

16
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 of 62 --

17
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 of 62 --

18
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 of 62 --

19
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 of 62 --

20
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 of 62 --

21
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 of 62 --

22
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 of 62 --

23
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 of 62 --

24
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 of 62 --

25
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 of 62 --

26
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 of 62 --

27
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 of 62 --

28
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 of 62 --

29
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 of 62 --

30
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 of 62 --

31
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 of 62 --

32
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 of 62 --

33
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 of 62 --

34
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 of 62 --

35
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 of 62 --

Counsel for Appellant
Shawn M. Rodgers
G OLDSTEIN LAW PARTNERS
John R. Renner [Argued]
Michael E. Rosman
Michelle A. Scott
C ENTER FOR INDIVIDUAL RIGHTS
Counsel for Appellee Wiley Periodicals Inc.
Elizabeth Seidlin-Bernstein
Michael L. Berry [Argued]
B ALLARD SPAHR
Counsel for Appellee American Heart Ass’n, Inc.
Howard A. Chajson
Steven L. Ettinger [Argued]
John C. Conti
Jeffrey J. Wetzel
D ICKIE MC CAMEY & CHILCOTE
Counsel for Appellee Marc Simon
Bryon M. Chowka
Robert J. Ridge [Argued]
CLARK H ILL
Counsel for Appellee University of Pittsburgh
Jeremy D. Engle
M ARCUS & SHAPIRA
Jeremy W. Brinster
Brian Boynton [Argued]
Bruce M. Berman
Molly M. Jennings
W ILMER CUTLER PICKERING H ALE & D ORR
Counsel for Appellees University of Pittsburgh Medical Center, University of Pittsburgh
Physicians, Samir Saba, Mark Gladwin, and Kathryn Berlacher
Emilie R. Hammerstein
Taylor Brailey

-- 36 of 62 --

37
Terrence H. Murphy
Theodore A. Schroeder [Argued]
L ITTLER MENDELSON

-- 37 of 62 --

1
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).

-- 38 of 62 --

2
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.

-- 39 of 62 --

3
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.

-- 40 of 62 --

4
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

-- 41 of 62 --

5
Statement”)5 with virtually the same paragraph6 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

-- 42 of 62 --

6
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.

-- 43 of 62 --

7
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

-- 44 of 62 --

8
defamed him in the Retraction Statement and A Path Forward.
II8
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.”).

-- 45 of 62 --

9
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).

-- 46 of 62 --

10
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.

-- 47 of 62 --

11
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

-- 48 of 62 --

12
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).

-- 49 of 62 --

13
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”

-- 50 of 62 --

14
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

-- 51 of 62 --

15
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

-- 52 of 62 --

16
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.

-- 53 of 62 --

17
B
1
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)).

-- 54 of 62 --

18
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)).

-- 55 of 62 --

19
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

-- 56 of 62 --

20
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”).
2
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

-- 57 of 62 --

21
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.

-- 58 of 62 --

22
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.

-- 59 of 62 --

23
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.

-- 60 of 62 --

24
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

-- 61 of 62 --

25
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].”).

-- 62 of 62 --