Carlo Croce v. David Sanders

NOT RECOMMENDED FOR PUBLICATION File Name: 21a0067n.06

Case No. 20-3577

UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT FILED Feb 03, 2021 CARLO M. CROCE, ) DEBORAH S. HUNT, Clerk ) Plaintiff-Appellant, ) ) ON APPEAL FROM THE UNITED v. ) STATES DISTRICT COURT FOR ) THE SOUTHERN DISTRICT OF DAVID SANDERS, ) OHIO Defendant-Appellee. )

BEFORE: KETHLEDGE, THAPAR, and READLER, Circuit Judges.

THAPAR, Circuit Judge. Carlo Croce’s name appears on over 1,000 scientific research articles. Sometimes all he contributed to the article was an idea, while another scientist conducted the research and wrote up the results. A different scientist, David Sanders, discovered that some of these papers contained manipulated data and plagiarized text. When Sanders went to the press with his discovery, Croce sued him for defamation. The district court granted Sanders’s motion for summary judgment. We affirm.

I.

Dr. David Sanders is a biological-sciences professor at Purdue University. He makes a practice of discovering and reporting instances of data falsification and fabrication in scientific papers. So when he received a tip about manipulated images in a scientific article about lung cancer, he took a look. One of the images depicting a protein analysis appeared to have been manipulated. Case No. 20-3577, Croce v. Sanders

Among the paper’s authors was Dr. Carlo Croce, a celebrated cancer researcher and professor at the Ohio State University. Croce’s name appeared last—suggesting that the paper came from researchers at Croce’s lab but that Croce did not himself conduct the experiment. Sanders, concerned about what appeared to be intentional manipulation of data, kept digging. He ultimately discovered problems in about thirty articles that listed Croce as a co-author.

Sanders reported his concerns to the respective journals. But he found their responses unsatisfactory, so he contacted a reporter from the New York Times, James Glanz. He told Glanz about the problems he’d discovered in the articles, and Glanz investigated. As part of his investigation, Glanz sent a letter to Ohio State and Croce, asking for comments. The letter described the alleged problems in “Croce’s papers”—papers that Croce had co-authored. In the letter, Glanz named Sanders as the source of the allegations. Glanz’s investigation led to a New York Times article about Croce: Years of Ethics Charges, but Star Cancer Researcher Gets a Pass.

The New York Times article prompted a follow-up report by Meghan Holden of the Lafayette Journal & Courier, a paper local to Sanders’s university. The article, Purdue Biologist Calls Out Cases of Scientific Misconduct, described the thankless and risky work of identifying research misconduct in the scientific field. The piece referenced the New York Times article and said that the costs of whistleblowing “didn’t stop Sanders from alleging that [Croce] falsified data or plagiarized text in more than two dozen articles Croce has authored.”

Croce sued Sanders for defamation and intentional infliction of emotional distress based on statements in the two newspaper articles and the Glanz letter. He also sued the New York Times in a separate action. Croce v. New York Times Co., 930 F.3d 787 (6th Cir. 2019). But his claims against the paper failed because the article, as a whole, was not defamatory. It was instead a

-2-Case No. 20-3577, Croce v. Sanders “standard piece of investigative journalism . . . present[ing] newsworthy allegations made by others, with appropriate qualifying language.” Id. at 790.

II.

The parties agree that Ohio law governs Croce’s defamation claims. “To establish defamation, the plaintiff must show: (1) that a false statement of fact was made, (2) that the statement was defamatory, (3) that the statement was published, (4) that the plaintiff suffered injury as a proximate result of the publication, and (5) that the defendant acted with the requisite degree of fault in publishing the statement.” Am. Chem. Soc. v. Leadscope, Inc., 978 N.E.2d 832, 852 (Ohio 2012) (citation omitted). The parties contest only the first, second, and fifth elements. Because Croce cannot satisfy the first element—that Sanders made a false statement of fact—we do not consider the others.

A statement cannot support a defamation claim if it is an expression of opinion, or if it is “substantially true.” Susan B. Anthony List v. Driehaus, 779 F.3d 628, 633 (6th Cir. 2015). Croce identifies six allegedly defamatory statements—two from each document.1 Of the six, five are either statements of opinion or substantially true. And Croce has offered no admissible evidence in support of the sixth statement, only hearsay. Thus, the district court correctly granted summary judgment to Sanders on each of his claims.

A.

An opinion cannot give rise to a defamation claim under Ohio law. See Wampler v. Higgins, 752 N.E.2d 962, 971 (Ohio 2001). Whether a statement is fact or opinion is a question of law for the court to decide. Id. at 976–78.

1 Croce broke one of the passages up into three separate defamatory statements. But both parties analyze the passage as a whole, so we will count it as one statement and evaluate all of its components.

-3-Case No. 20-3577, Croce v. Sanders

Four considerations guide our analysis: “the specific language used, whether the statement is verifiable, the general context of the statement, and finally, the broader context in which the statement appeared.” Vail v. The Plain Dealer Publ’g Co., 649 N.E.2d 182, 185 (Ohio 1995). These considerations turn on “the reasonable reader’s perception of the statement—not on the perception of the publisher.” McKimm v. Ohio Elections Comm’n, 729 N.E.2d 364, 371 (Ohio 2000).

At least three of the six statements that Croce identified are expressions of opinion—the first statement of the New York Times article, and both statements in the Glanz letter.

1.

Of the two statements Croce identifies in the New York Times article, the first is an expression of opinion. The statement is a direct quote from Sanders. It appears in a paragraph of its own, so we include the prior paragraph for context:

In 2013, an anonymous critic contacted Ohio State and the federal authorities with allegations of falsified data in more than 30 of Dr. Croce’s papers. Since 2014, another critic, David A. Sanders, a virologist who teaches at Purdue University, has made claims of falsified data and plagiarism directly to scientific journals where more than 20 of Dr. Croce’s papers have been published.

“It’s a reckless disregard for the truth,” Dr. Sanders said in an interview. R. 11-1, Pg. ID 69–70 (emphasis added). Although the article does not explicitly identify what the “it” in Sanders’s quote means, a reasonable reader would understand his statement to refer to the incidence of falsified data and plagiarism in Croce’s papers.2

The statement expresses Sanders’s opinion on the matter—that he considered the problems he observed in Croce’s papers to reflect a reckless disregard for the truth. Croce says “reckless disregard” has a “precise meaning,” but he never identifies what that meaning is. That is

2 Sanders said in a deposition that he was referring to a specific instance of image manipulation, but we must construe the meaning of the statement as a reasonable reader would. Vail, 649 N.E.2d at 185.

-4-Case No. 20-3577, Croce v. Sanders understandable. The term is imprecise, in law and otherwise. See St. Amant v. Thompson, 390 U.S. 727, 730 (1968) (explaining that the legal “reckless disregard” standard “cannot be fully encompassed in one infallible definition”). Outside of the legal context, imprecise adjectives signal to the listener that the speaker is expressing a subjective point of view. There is no clear point at which careless conduct becomes reckless, and the reasonable reader understands that. In everyday speech, the word “reckless” is “value-laden and represents a point of view that is obviously subjective,” and thus does not support a cause of action. Vail, 649 N.E.2d at 186.

2.

The two statements Croce identified from the letter that Glanz wrote to Ohio State and Croce are also expressions of opinion. Both statements appear in a single sentence:

Dr. Sanders argues—because in his observation [1] the image fabrication, duplication and mishandling, and plagiarism in Dr. Croce’s papers is routine, and because the authors routinely dispute those allegations when confronted with them through the journals—that [2] Dr. Croce is knowingly engaging in scientific misconduct and fraud.

R. 59-11, Pg. ID 474 (emphasis added). Sanders does not disagree that this passage reflects what he said to Glanz.

The first statement, as identified by Croce, is that “the image fabrication, duplication and mishandling, and plagiarism in Dr. Croce’s papers is routine.” Id. Croce does not dispute that there have been some instances of data manipulation and plagiarism in papers bearing his name— Sanders identified around thirty. He objects instead to the characterization of these problems as “routine.” And understandably so: The number of problematic articles is small compared to the 1,265 he has co-authored. But no matter how much Croce may disagree with Sanders’s assessment, what counts as routine is a matter of opinion—and thus protected under Ohio law.

To say something is routine is to make an imprecise characterization that “lacks a plausible

method of verification.” Vail, 649 N.E.2d at 186 (citation omitted); see also Wampler, 752 N.E.2d -5-

Case No. 20-3577, Croce v. Sanders at 979 (“We seek to determine whether the . . . statements are objectively capable of proof or disproof, for a reader cannot rationally view an unverifiable statement as conveying actual facts.” (cleaned up)). How many problems make something routine? We can’t say, because there is no objective line.

Instead, the line varies from speaker to speaker and from context to context. For example, a shopkeeper might say that shoplifting is a routine problem among teenage customers (even if most teenagers who come to the store don’t steal). Or a person might say that her spouse routinely forgets his keys in the front door (even if that happens only rarely). Both speakers know the relative numbers are small, but that’s not the point. Neither is using “routine” to say that the rate of problems exceeds any particular threshold. Nor are they suggesting that thievery is a routine part of growing up, or that leaving keys in the door is a routine part of coming home. They are saying that the rate of problems—whatever it is and whatever the cause—is too high.

“In [Sanders’s] observation,” the rate of image manipulation and plagiarism is high enough to be called routine. R. 59-11, Pg. ID 474. That is an expression of opinion based on Sanders’s own subjective assessment of the problems in Croce’s papers. And as an opinion, it cannot support a defamation suit.

The second statement Croce identifies in the letter is, “Dr. Croce is knowingly engaging in scientific misconduct and fraud.” Id. Although it might look like a statement of fact standing alone, the full sentence makes clear that this statement is an expression of Sanders’s opinion.

To start, the sentence opens with, “Dr. Sanders argues . . . that,” thus identifying the statement as a conclusion of Sanders’s own argument. Id.; see also Wampler, 752 N.E.2d at 980 (“[T]he language surrounding the averred defamatory remarks may place the reasonable reader on notice that what is being read is the opinion of the writer.”); Scott v. News-Herald, 496 N.E.2d

-6-Case No. 20-3577, Croce v. Sanders 699, 707 (Ohio 1986) (describing the phrase “[Author] Says” in the headline of an article as an “[o]bjective cautionary term[]” that “would indicate to even the most gullible reader that the article was, in fact, opinion”). To argue something is to make a case for it (ordinarily, one that can be disputed). And to announce that a statement follows from an argument is to acknowledge that the statement is not a straightforward declaration of fact.

Of course, the label is not decisive—the “potential for abuse” would be too high. Scott, 496 N.E.2d at 707. A speaker who “impl[ies] that he has first-hand knowledge that substantiates the opinions he asserts” can be on the hook for defamation even if he cloaks his assertion in the language of argument or opinion. Vail, 649 N.E.2d at 186; Mehta v. Ohio Univ., 958 N.E.2d 598, 609 (Ohio Ct. App. 2011) (same); see also Scott, 496 N.E.2d at 707 (“[O]ne should not escape liability for accusations of crime simply by using, explicitly or implicitly, the words ‘I think.’” (cleaned up)). But here, there is no reason to suspect abuse. “Sanders argues” is not just a label.

How do we know? First, the full sentence identifies the premises of the argument, alerting the reader to the limitations of Sanders’s knowledge. See Restatement (Second) of Torts § 566 (Am. L. Inst. 1977) (“A simple expression of opinion based on disclosed or nondefamatory facts is not itself sufficient for an action of defamation . . . .”). Sanders based his argument on two observations: (1) the presence and frequency of problems in Croce’s papers, and (2) the authors’ defensive responses to notifications about the problems. A reasonable reader would assume that these two observations were Sanders’s best pieces of evidence, and that Sanders had nothing more damning connecting Croce to the misconduct—certainly not a first-hand account.

Second, reasonable readers would see that there is ample room for a different interpretation of the evidence Sanders presented. As a general matter, whether a set of facts amounts to misconduct and fraud is likely to be a subject of genuine disagreement; we would expect people

-7-Case No. 20-3577, Croce v. Sanders to have different opinions on the question. See id. (“The simple expression of opinion . . . occurs when the maker of the comment states the facts on which he bases his opinion of the plaintiff and then expresses a comment as to the plaintiff’s conduct, qualifications or character.”). That’s especially true here, where Sanders’s evidence falls short of establishing that Croce was knowingly engaged in scientific misconduct and fraud. Indeed, neither observation identifies Croce’s role in any sort of misconduct. Taken together, they could lend themselves to any number of interpretations; Sanders made an argument in favor of one. His allegedly defamatory statement is neither an assertion of fact nor a conclusion that follows incontrovertibly from asserted facts as a matter of logic. It is instead a subjective take that is up for debate.

Finally, the broader context reinforces our conclusion that a reasonable reader would understand the statement to be an expression of Sanders’s own opinion. The statement appeared in a letter that a journalist sent to Ohio State and Croce as part of his investigation. The journalist described the statement as an “allegation,” and he asked if Croce “disagree[d]” with it. R. 59-11, Pg. ID 474. The letter’s preliminary and investigatory nature puts the reader on notice that the statements did not result from a “thorough investigation[] which yielded impartial results.” Mehta, 958 N.E.2d at 611–12 (explaining that readers would understand allegations of plagiarism that appeared in a university’s official press release to be true). A reasonable reader—an Ohio State professor or administrator—would understand the letter to be an inquiry about the truth of the allegations, and not a conclusive affirmation of their truth.

Thus, “while [Sanders’s] mind [was] certainly made up, the average reader viewing the words in their internal context would be hard pressed to accept [this statement] as impartial reporting.” Wampler, 752 N.E.2d at 980 (cleaned up). Not only does the letter qualify the

-8-Case No. 20-3577, Croce v. Sanders statement as a conclusion of a genuine argument, but the broader context supports that view. As a properly qualified expression of opinion, the statement cannot support an action for defamation.

B.

Having determined that three of the six statements are expressions of opinion, we turn to two others: one in the New York Times article, and one in the Lafayette Journal & Courier article. These statements also cannot support a defamation claim because they are substantially true.

A true statement cannot be defamatory. Ed Schory & Sons, Inc. v. Soc’y Nat’l Bank, 662 N.E.2d 1074, 1083 (Ohio 1996). Under Ohio defamation law, a statement can be misleading and fail to disclose relevant facts but still be “true” as long as it has “some truth in it.” Susan B. Anthony List, 779 F.3d at 633. If the “‘gist’ or ‘sting’ of the statement is substantially true,” that is sufficient. Id.

New York Times. Croce identified a second allegedly defamatory Sanders quote in the New York Times article: “A lab that is engaging in violating scientific norms is being rewarded for that very effort.” R. 11-1, Pg. ID 81. Here, Croce argues that the specific language in the subject of the sentence could amount to a factual assertion that people in Croce’s lab have violated scientific norms. But even so, the statement does not support a defamation claim because, under Ohio law, it qualifies as substantially true.

Journals have found research problems and plagiarism in articles coming from Croce’s lab. Sometimes, the problems were severe enough for the journals to publish corrections or expressions of concern (and sometimes to withdraw the paper). However you define “scientific norms,” we know that academic journals felt some responsibility to alert the scientific community about problems in some of Croce’s papers. That suggests the papers contained problems outside the

-9-Case No. 20-3577, Croce v. Sanders range of acceptable research and publishing practices. Thus, the statement that people in Croce’s lab have violated scientific norms is substantially true.

Lafayette Journal & Courier. One of the allegedly defamatory quotes Croce identifies in the local paper is also substantially true. Sanders said: “If you wanted to just make up data you could do it in a way that’s much more difficult to detect, but they didn’t because they were able to get away with this relatively simple manipulation . . . . They continued to do it over and over again.” R. 59-16, Pg. ID 485. After considering each component of the quote, we find that it contains no false statements of fact.

The passage contains two factual assertions: “they were able to get away with this relatively simple manipulation,” and “they continued to do it over and over again.” There is no explicit antecedent for “they,” and this passage stands alone as its own paragraph. But the preceding paragraph discusses the “corrections, retractions, or editors’ notices” that journals have published on Croce’s articles. Id. A reasonable reader who then reads the Sanders quote would think “they” refers either to authors of these papers with manipulated data, or to the scientists who manipulated the data themselves.

Thus, one interpretation of this passage is that (1) some scientists manipulated data and published articles based on that data; (2) Croce appeared as a co-author on these articles; and (3) this happened multiple times. These assertions are substantially true: Journals have found and corrected instances of data manipulation in several papers with Croce’s name on them. And the scientists who succeeded in getting the papers published got away with it (for a time). Finally, the journals have found multiple instances of data manipulation, also supporting the statement that “they continued to do it over and over again.” That interpretation is substantially true. Because a

- 10 -Case No. 20-3577, Croce v. Sanders statement susceptible of a true interpretation is not false under Ohio defamation law, this passage is not actionable. Susan B. Anthony List, 779 F.3d at 633; McKimm, 729 N.E.2d at 371.

Croce disagrees. He says that the “they” refers to him specifically, and so the quotation can be read only as an accusation that Croce personally and repeatedly manipulated data. Croce recognizes that the plural pronoun “they” conflicts with this interpretation, but he maintains that his understanding is the only plausible one because he is the only scientist besides Sanders named in the article. But Croce’s interpretation—syntax problem and all—is not a plausible reading of the passage. And even if it were, we could not adopt his interpretation: Under Ohio law, “if allegedly defamatory words are susceptible of two meanings, one defamatory and one innocent, the defamatory meaning should be rejected, and the innocent meaning adopted.” McKimm, 729 N.E.2d at 372 (cleaned up).

C.

Of the six defamatory statements that Croce identified, only one remains. The Lafayette Journal & Courier article says that the risks of whistleblowing “didn’t stop Sanders from alleging that Dr. Carlo Croce . . . falsified data or plagiarized text in more than two dozen articles Croce has authored.” R. 59-16, Pg. ID 484 (emphasis added).

But Sanders denies ever having made that accusation. He says he has focused his criticisms on the problems with the articles themselves, and he has never accused Croce himself of plagiarizing or manipulating data.

Croce has not supplied admissible evidence to rebut this denial. His only evidence that Sanders made those allegations is the news article, and the district court properly concluded that Croce cannot introduce the article for that purpose. It is hearsay—that is, an out-of-court statement offered to prove the truth of what is asserted in the statement. Fed. R. Evid. 801(c). Croce offers

- 11 -Case No. 20-3577, Croce v. Sanders the news article (an out-of-court statement) to prove that Sanders made the allegations. So the article is inadmissible, and Croce submitted no other materials to show that the evidence could be presented in an admissible form at trial. See Fed. R. Civ. P. 56(c)(2). Thus, the district court did not err in refusing to consider the article at summary judgment.

In short, because none of the six statements Croce identified can support a defamation action, the district court correctly granted summary judgment to Sanders.

III.

Along with the claims for defamation, Croce brought a claim for intentional infliction of emotional distress. This claim is derivative of his defamation claims. Because Croce loses on the defamation claims, he cannot prevail on this claim either. Vail, 649 N.E.2d at 186.

We affirm.

- 12 -