dissenting:
I respectfully dissent. The panel opinion assumes that the film implied that “Spacecon was involved with the Stranded Workers and responsible for the Glenwood Springs incident,” and that this implied message was false. Op. at 1036 (brackets and internal quotation marks omitted). But it holds that Spacecon produced insufficient evidence that Bensinger acted with malice. I disagree. In my view, the record would support a finding under the clear-and-convincing-evidence standard that Bensinger published the message with reckless disregard for its truth.
There is, of course, much in the panel opinion with which I agree. In particular, I share the panel opinion’s view that the Colorado Supreme Court would require the plaintiff to prove actual malice to sustain a defamation claim based on a statement involving a matter of public concern, even if the plaintiff is a private individual. And I agree that the allegedly false messages published by Bensinger involved matters of public concern. But I believe it worthwhile to discuss two areas of disagreement. I have already mentioned that I think there was sufficient evidence of malice. Before I turn to that matter, however, I will address what I view as a flaw in the panel opinion’s analysis of whether the film was on a matter of public concern.
I. THE PUBLIC-CONCERN ANALYSIS
I cannot join the panel opinion’s analysis of the public-concern issue. The panel *1051opinion correctly states Colorado’s definition of public concern, which apparently is not intended as a departure from the use of the term by federal courts:
Colorado courts have indicated a matter is of public concern whenever it embraces an issue about which information is needed or is appropriate, or when the public may reasonably be expected to have a legitimate interest in what is being published. More specifically, a matter is of public concern when it can be fairly considered as relating to any matter of political, social, or other concern to the community, or when it involves the use of names, likenesses or facts in giving information to the public for purposes of education, amusement, or enlightenment when the public may reasonably be expected to have a legitimate interest in the subject.
Op. at 1035 (citation, brackets, and internal quotation marks omitted).
The opinion then, however, goes on to indicate, without any analysis of the merits of the issue, that a statement that would otherwise be on a matter of public concern may not be so classified if the publisher of the statement knew of its falsity or published it in reckless disregard of whether it was true. Under the heading “Context,” the public-concern discussion states, “Ben-singer would not have been so intimately familiar with the underlying factual circumstances that he would have known or should have known the allegations were baseless.” Id. at 1037. And under the heading “Motivation,” it says, “Any pro-union bias on Bensinger’s part does not alone indicate he was unlikely to be concerned with the truth of what he was publishing.” Id. at 1038.
These observations are irrelevant to whether statements are on matters of public concern. In common parlance the “subject matter” of a statement does not encompass either the truth of the statement or whether the publisher of the statement thought it to be true. The case law analyzes whether a statement is on a matter of public concern by looking to its content, form, and context. See, e.g., Connick v. Myers, 461 U.S. 138, 147-48, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983); McIntyre v. Jones, 194 P.3d 519, 525 (Colo.App.2008). Or, as the Supreme Court recently said in Snyder v. Phelps, — U.S. -, 131 S.Ct. 1207, 1216, 179 L.Ed.2d 172 (2011), “[I]t is necessary to evaluate all the circumstances of the speech, including what was said [the content], where it was said [the context], and how it was said [the form].” These circumstances are what one refers to in deciding what a statement conveys or “says.” To be sure, the context may include “the motivation or ‘point’” of the statements. McIntyre, 194 P.3d at 525. But the motivation or point concerns only the purpose of the statements; for example, did an employee of the district attorney complain about how the office was run because the public’s business was not being performed properly or only because she felt mistreated? See Connick, 461 U.S. at 147-49, 103 S.Ct. 1684. Knowing the motive or point helps one to assess whether the statement related only to a private concern or to one that would engage public interest. In contrast, a statement’s truthfulness, or the culpability of the publisher of a false statement, does not change what it conveys.
Moreover, it makes no sense to say that the culpability of the publisher of a false statement should be a factor in determining what level of culpability must be shown for the publisher to be liable. In this case the only reason for assessing whether a statement was on a matter of public concern is to determine whether it is necessary that the plaintiff prove that the publisher acted with malice. If the published *1052statement was'on a matter of public concern, the plaintiff must show that the publisher knew the statement was false or acted in reckless disregard of whether it was false; if it was not on a matter of public concern, the publisher can be liable if he was merely negligent. See McIntyre, 194 P.3d at 523-24. Declaring that malice is a factor in the public-concern calculus must mean that sometimes the presence of malice makes a difference in that calculus; that is, it must mean that on occasion a statement that would otherwise be on a matter of public concern is not on a matter of public concern because it was published with malice. But why say that if the publisher acted with malice (so the statement was not on a matter of public concern), then the plaintiff does not need to prove malice? Including malice in the public-concern analysis accomplishes nothing.
More troublesome, the panel opinion appears to consider a level of culpability less than malice in its analysis of whether a false statement was on a matter of public concern. It uses the phrases “would have known or should have known” and “was unlikely to be concerned with the truth.” These phrases suggest something short of malice. But free-speech protection could be undermined if courts were to say that a statement is not on a matter of public concern if it was made, say, negligently (or with any culpability less than malice). Instead of having to prove that a statement on a matter of public concern was made with malice, the person claiming to have been defamed could simply prove negligence, because then the statement would not be on a matter of public concern. The panel opinion suggests that other circuits are divided on whether malice is a factor in determining whether a statement is on a matter of public concern. The suggestion, however, is based on a misreading of the cases. The only out-of-circuit case cited is the Sixth Circuit’s opinion in Westmore-land v. Sutherland, 662 F.3d 714 (2011). It involved a suit by a firefighter complaining that he was fired for criticizing the city council. The opinion must be read in context.
In analyzing a claim that the government has improperly retaliated against an employee based on the employee’s speech, the federal courts undertake a five-element analysis derived from a line of Supreme Court opinions originating with Pickering v. Board of Education, 391 U.S. 563, 88 S.Ct. 1731, 20 L.Ed.2d 811 (1968). We ask:
(1) whether the speech was made pursuant to an employee’s official duties; (2) whether the speech was on a matter of public concern; (3) whether the government’s interests, as employer, in promoting the efficiency of the public service are sufficient to outweigh the plaintiffs free speech interests; (4) whether the protected speech was a motivating factor in the adverse employment action; and (5) whether the defendant would have reached the same employment decision in the absence of the protected conduct.
Dixon v. Kirkpatrick, 553 F.3d 1294, 1302 (10th Cir.2009). Ordinarily, the first three elements are matters of law for the court and the last two are for the factfinder. See id. As in defamation actions like this one, public concern is one of the elements. Unlike defamation actions, however, the role of malice has not been specified by the Supreme Court. In Pickering the Court wrote that it had “no occasion to pass upon the ... question whether a statement that was knowingly or recklessly false would, if it were neither shown nor could reasonably be presumed to have had any harmful effects, still be protected by the First Amendment.” 391 U.S. at 574 n. 6, 88 S.Ct. 1731. That occasion has yet to arise in a government-employee retaliation case *1053before the Supreme Court, but cf. United States v. Alvarez, — U.S. -, 132 S.Ct. 2537, 183 L.Ed.2d 574 (2012) (invalidating Stolen Valor Act); so lower courts have had to decide for themselves how to take into account that the employee made a false statement with malice.
In Westmoreland the court addressed the first three elements of the Pickering analysis. It wrote:
To establish that his speech was constitutionally protected, a public employee must show that he was speaking as a private citizen, rather than pursuant to his official duties; that his speech involved a matter of public concern; and, if so, that his interest as a citizen in commenting on the matter outweighed the interest of the State, as an employer, in promoting the efficiency of the public services it performs through its employees.
662 F.3d at 718-19 (internal quotation marks omitted). Only after determining that the firefighter’s speech “undoubtedly involved matters of public concern,” id. at 720, did the court consider whether a public employee has no First Amendment claim if his statements were intentionally or recklessly false, or whether such malicious falsehood is merely a factor to be considered in assessing whether the employee’s interest in speaking outweighs the government employer’s interest. See id. at 720-21. The court read its precedent to say that “Pickering balancing [the third element] is not required if it is determined that the employee made statements with knowledge of, or reckless indifference to, their falsity.” Id. at 721.
The split in the circuits that Westmore-land referred to is not the same as the split between the panel opinion and this dissent. Westmoreland spoke of the circuit differences “over whether intentionally or recklessly false statements fall outside the protection of the First Amendment, or whether such falsity is to be weighed as part of the ... balancing [test].” Id. at 720. In other words, the divide is about whether the presence of malice should be dispositive in public-employment retaliation cases, not whether malice is relevant to the public-concern analysis. The panel opinion accurately quotes Westmoreland, but it misinterprets the statement. West-moreland did not view malice as a factor in the public-concern analysis, nor did it say that there was a circuit split on that issue. As far as I can tell, no other circuit court has held that malice is a proper consideration in assessing whether a statement is on a matter of public concern.
That brings us to precedent in this circuit. It cannot be denied that Quigley v. Rosenthal, 327 F.3d 1044, 1061 (10th Cir.2003), contains language that supports the position of the panel opinion. But it also cannot be denied that Quigley does not analyze the issue or provide any rationale for considering malice in assessing whether a statement involved a matter of public concern. It relies on citations to two cases with a “see generally” signal and one with a “see” signal. The first two cases do not address anything close to malice. As for the third, it concerned a state employee who had filed a grievance claiming race and sex discrimination. She requested a closed hearing but she and her husband wrote her senator complaining of the proceedings and suggesting an investigation of “ ‘possible civil rights violations.’ ” Kemp v. State Bd. of Agric., 803 P.2d 498, 500 (Colo.1990) (en banc). The state agency then informed her that her violation of secrecy “ ‘voided’ ” the grievance proceedings, and she sued, claiming retaliation on the basis of her speech. Id. at 501. The court wrote the following in addressing whether the employee’s speech was on a matter of public concern:
*1054Although the context within which [the employee’s] letter arose was one in which she had a personal stake, it was nevertheless a context in which charges of racial discrimination had been filed. The very purpose of the grievance was to address alleged discrimination, which the grievance manual defines as any act or policy that violates any of a multitude of state and federal anti-discrimination laws, all of which are public matters. Moreover, the form of the speech — a letter to a United States senator, only lends credence to the argument that the speech touched on a matter of public concern.
When an employee alleges a colorable claim that a university is guilty of racial discrimination, it is a matter of public concern, and the fact that she has a personal stake in the outcome does not alchemize that claim into a purely private matter in which the public has no interest.
Id. at 504 (footnote omitted) (emphasis added).
The panel opinion argues that the Colorado court’s use of the term colorable claim indicates that the court thought the veracity of the plaintiff was a factor in its public-concern analysis. But nothing in Kemp suggests that the court was concerned with the employee’s veracity. It was apparently using the term “colorable claim” in the sense in which we use it when assessing whether an allegation suffices to convey jurisdiction — our concern is not the veracity of the pleader but whether the allegations (deemed to be true) state the necessary elements. See, e.g., Iowa Tribe of Kan. & Neb. v. Salazar, 607 F.3d 1225, 1231-32 (10th Cir.2010) (Federal Quiet Title Act prohibits suits challenging United States title in Indian trust land. “[T]he Secretary [of the Interior] need only make a colorable claim that the land is held in trust on behalf of an Indian tribe.... The very purpose of the doctrine ... is to prevent a judicial examination of the merits of the government’s position.” (internal quotation marks omitted)). (We now use the term plausible claim in a similar sense in assessing a motion to dismiss for failure to state a claim. See, e.g., Keith v. Koemer, 707 F.3d 1185, 1190 (10th Cir.2013).) Kemp does not support the proposition that malice plays a role in the public-concern analysis.
As for the panel opinion’s reliance on the ■footnote in Wulf v. City of Wichita, 883 F.2d 842 (10th Cir.1989) — another case dealing with retaliation against a government employee rather than defamation— we are hardly bound by the undeveloped dictum in that opinion. Moreover, our later opinions have not read Wulf as standing for the proposition advanced by the panel opinion. Rather, they view Wulf as treating malice not as a factor in assessing public concern but as an independent ground for denying relief. In Lancaster v. Independent School District No. 5, 149 F.3d 1228, 1234 (10th Cir.1998), we wrote:
Our finding that the speech was not of public concern makes it unnecessary for the court to address whether the interests of the defendant outweighed the plaintiff’s interests, whether plaintiff’s speech was knowingly false, and whether plaintiffs speech was a motivating or a substantial factor for the Board’s employment decision. See Wulf, 883 F.2d at 856-57 (noting these are factors to be considered if speech is found to be of public concern).
(emphasis added). And in Moore v. City of Wynnewood, 57 F.3d 924, 932-33 (10th Cir.1995), we said:
Defendants’ claim that Moore’s accusations were false does not alter our conclusion that his speech was of public concern. We may assume that deliber*1055ately or recklessly false statements by public employees are either unprotected by the First Amendment or, at least, that such intentional falsity would weigh heavily against protection. See Pickering, 391 U.S. at 574, 88 S.Ct. at 1737-38, (suggesting that a public employee might lose the protection of the First Amendment for knowingly or recklessly making false statements); Wulf, 883 F.2d at 858 & n. 24 (suggesting that malicious or reckless false statements would receive less protection). However, here Defendants have failed to show either that Moore’s statements were false or that Moore made them knowing that they were false or with reckless disregard for their veracity.
So what is to be done with Quigley? Although we cannot violate precedent, we have no obligation to cite it. There is ample precedent to support this panel’s (unanimous) view that the statements here were on a matter of public concern without relying on Quigley. I would think it better practice not to cite precedent for a highly questionable proposition unless that proposition appears to require a result that would not otherwise obtain. Then the issue is clearly postured for en banc review, and correction. Or if the highly questionable proposition must be cited, it is helpful to point out concerns about the proposition so that future litigants and courts can anticipate possible overruling of the precedent. Particularly in a diversity case in which we are “predicting” state law, it could be useful to alert the state court that it may not wish to adopt our precedent interpreting state law.
II. EVIDENCE OF MALICE
The basis for my dissent, however, is the panel opinion’s determination that there was insufficient evidence of malice to overcome the motion for summary judgment. To be sure, a defamation plaintiff seeking to resist summary judgment on the issue of malice — whether the publisher of the statement knew it to be false or acted with reckless disregard of its truth — bears a heavy burden. “[T]he evidence in the record [must] support a reasonable jury finding ... that the plaintiff has shown actual malice by clear or convincing evidence.... ” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255-56, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986); see Air Wisconsin Airlines Corp. v. Hoeper, No. 09SC1050, — P.3d -, -, 2012 WL 907764, at *9 (Colo. March 19, 2012) (court must “undertake an independent review of the entire record to ensure that clear and convincing evidence supports a finding of actual malice”). When defamation doctrine protects First Amendment interests, courts must be particularly diligent in reviewing the sufficiency of the evidence of liability (although, as noted by the panel opinion, the United States Supreme Court has held that proof of malice is not necessary in a case like the one before us).
Nevertheless, I think that the panel opinion’s view of Spacecon’s evidence is too jaundiced. We should not forget that “[c]redibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions.... The evidence of the non-movant is to be believed, and all justifiable inferences are to be drawn in his favor.” Anderson, 477 U.S. at 255, 106 S.Ct. 2505 (reviewing summary judgment in defamation case brought by public figures). On several points I question whether the opinion applies that standard. When it says, “Even assuming ... that knowledge of Wilson’s and Stratton’s biases did give Bensinger obvious reasons to doubt their veracity or the accuracy of their reports, that these two sources corroborated each other’s allegations sufficiently establishes Bensinger dispelled those doubts,” Op. at *10561046, the opinion is apparently foreclosing the possibility that a reasonable juror could think that two dubious -sources do not add up to one reliable source. And, contrary to the opinion’s discussion on page 36, I would also think that a reasonable juror could infer that the film was presenting the statements of those complaining of Spacecon as truth, not merely as allegations. The film was a far cry from a newscast that reports allegations and denials without trying to take sides.
In addition, the approach of the opinion strikes me as misconceived. It pursues a strategy of divide and conquer — addressing separately each item of evidence of malice and saying that it is not sufficient in itself to establish malice — a strategy derided by the Supreme Court in another context (but also addressing the reasonableness of a fact-finding) as contrary to the proper approach of looking at the totality of the circumstances. See United States v. Arvizu, 534 U.S. 266, 274, 122 S.Ct. 744, 151 L.Ed.2d 740 (2002) (reversing decision that officers did not have reasonable suspicion for stopping vehicle). The jury is entitled to look at the evidence holistically. Cf. Tellabs, Inc. v. Makor Issues & Rights, Ltd., 551 U.S. 308, 326, 127 S.Ct. 2499, 168 L.Ed.2d 179 (2007) (When assessing sufficiency of pleading, “the court’s job is not to scrutinize each allegation in isolation but to assess all the allegations holistically.”)
Most importantly, it seems to me that the panel opinion misses a critical point regarding the reliability of Bensinger’s sources for the false statement that “Spacecon was involved with the Stranded Workers and responsible for the Glenwood Springs incident.” Op. at 1033-34 (brackets and internal quotation marks omitted). The opinion, in essence, states that Ben-singer did not act with reckless disregard of the truth in relying on two of the workers — and Union representative Wilson, who purported to be merely repeating what workers had told him — who said that they had been told by Leno that they were brought to Colorado to work for Spacecon. But even assuming the veracity of the workers and Wilson, they were providing only hearsay; the reliability of the statement about Spacecon depends on the veracity of Leno. The workers did not say that they had worked for Spacecon. They were relying on what Leno had told them.
It is therefore essential to consider what Bensinger knew about Leno’s veracity. And the record indicates that Bensinger not only should have known, but actually believed, that Leno had lied to the workers about whom they were going to be working for in Colorado. The original complaint filed by the Stranded Workers named Midwest Drywall Company as a defendant, alleging that the labor brokers recruited the workers to work for Midwest and were .acting as Midwest’s agents. Bensinger was familiar with this complaint; indeed, it is shown in the film. The workers’ amended complaint, however, withdrew Midwest as a defendant and made it a plaintiff. It alleged that the brokers, including Leno, “falsely asserted to the Department of Labor (DOL) and to the individual Plaintiffs that Midwest was to employ individual Plaintiffs.” Aplt. App., Vol. IV at 753. A second amended complaint added specifics about these false representations, including allegations that the defendants had submitted forged documents. The evidence would readily permit a jury to find that Bensinger knew about this change in the lawsuit. An article in the Aspen Times reported Midwest’s response to the original claims against it and the statement by 'the Stranded Workers’ attorney that “Midwest appears to be stone-cold innocent, and these 69 men have no beef with an innocent local business.” Id. at 732. Bensinger read the article. *1057He testified in his deposition: “Midwest was named; however, as I say, Midwest was switched from a defendant in the suit because their name was used fraudulently to get those people here. The lawyer in the newspaper says that Midwest was cold-stone innocent in this whole situation.” Id. at 738-39.
Thus, Bensinger accused Spacecon of misconduct in reliance solely on the credibility of a source — Leno—that he knew had falsely attributed to Midwest the same conduct (hiring Leno to bring the Stranded Workers to Colorado) that Leno allegedly attributed to Spacecon. One would think that a jury could be clearly convinced that Bensinger had acted in reckless disregard of the truth when he did not corroborate Leno’s statements about Spacecon despite knowing that Leno had falsely said the same thing about Midwest. Indeed, the Supreme Court has written, “[Rjecklessness may be found where there are obvious reasons to doubt the veracity of the informant or the accuracy of his reports.” Harte-Hanks Commc’ns, Inc. v. Connaughton, 491 U.S. 657, 688, 109 S.Ct. 2678, 105 L.Ed.2d 562 (1989) (internal quotation marks omitted). I read this statement as saying that if a reasonable juror could find (under the clear-and-convincing-evidenee standard) that the publisher had obvious reasons to doubt the veracity or accuracy of its source and nevertheless published as true what the source said, then the juror can properly find malice. That is the situation here.
The panel opinion suggests that Ben-singer could credit the allegation against Spacecon because it was supported by allegations concerning mistreatment of workers at the Little Nell site. That suggestion is misguided. It is important to keep in mind that Leno, as the labor broker, paid the workers (and was responsible for any withholding from the payments); contractors such as Spacecon paid Leno, not the workers themselves. Thus, complaints about pay and withholding concerned Leno, not Spacecon. The specific complaints about Spacecon were limited to racial discrimination in working conditions. However despicable such discrimination is, it is not evidence of Spacecon’s “in-volvefinent] with the Stranded Workers [who did not work at the Little Nell site]” or its “responsibility] for the Glenwood Springs incident.” Op. at 1033-34 (brackets and internal quotation marks omitted). With respect to that involvement and responsibility, the only thing implicating Spacecon was that it employed Leno as a labor broker in the Glenwood Springs area. But so did Midwest, and Bensinger believed that Midwest had no connection with (much less responsibility for) the Stranded Workers. There was no rational basis for Bensinger to distinguish Space-con and Midwest in that respect.
In contrast, there is significant evidence that truth was not what motivated Ben-singer in producing the film. First, there was his motive. Bensinger prepared the film as part of his $7,500 monthly retainer from a union engaged in an organization campaign against Spacecon. The obvious purpose of the union in financing the film was to make Spacecon look bad; and the obvious motive of Bensinger was to please his employer by achieving that goal.
Also supporting a recklessness finding is Bensinger’s failure to check obvious sources, such as the Stranded Workers’ attorney or the reporter who wrote the stories about the workers, that could have undermined his attack on Spacecon. In particular, the absence of a claim against Spacecon in any of the complaints filed by the Stranded Workers would have struck any but the most biased observer as highly suggestive. No one would have had more incentive than counsel for those workers to *1058assert that Spacecon was responsible for their plight if there was sufficient evidence of responsibility to support a good-faith claim.
Bensinger’s failure to check other sources was not for lack of time. He testified that he began work on the film in “early '08,” R.', Vol. 11 at 331-32, and he interviewed and filmed Stranded Workers in late June and early July 2008. Yet he did not screen his film until March 2009. During those months (perhaps a year) he made no effort to contact the reporter or the Stranded Workers’ attorney. See Masson v. New Yorker Magazine, 501 U.S. 496, 521, 111 S.Ct. 2419, 115 L.Ed.2d 447 (1991) (noting that author “was not working under a tight deadline. Unlike a case involving hot news, [the author] cannot complain that she lacked the practical ability to [check the quotations].”)
And Bensinger’s efforts to get information from Spacecon hardly displayed a sincere search for the truth. The film shows that on one occasion Bensinger or an assistant showed up at the Little Nell project with a video camera for an ambush interview of a Spacecon supervisor who responded “No comment,” when the interviewer said, “I have a question about worker abuse allegations.” R., Vol. IV at 637. Later, Bensinger sent a letter from Virginia dated February 23, 2009 (after invitations had been sent for the film’s initial screening 17 days later) to Spacecon vice president John Banks in Colorado and the president of Spacecon’s parent corporation in Pennsylvania, requesting an on-camera interview about allegations “that concern the abuse of immigrant labor by Spacecon.” Id, Vol. VIII at 1837. The letter, stated that Bensinger was producing a documentary film that would be released in the summer of 2009 with “several screenings prior to that.” Id Banks sent a return letter dated March 3, noting that he had learned that the film had already been scheduled for showing on March 12 and expressing doubts about its objectivity, but stating a willingness to provide information in response to questions if given the opportunity to review the film. Bensinger e-mailed questions on March 10. Banks, who was not provided a preview of the film, responded. Among other things, he denied that Spacecon had any involvement with the Stranded Workers. The responses were provided to those who attended the March 12 screening. But Ben-singer did not refer to them in his remarks. (Of course, “publication of a denial by the defamed subject does not absolve a defendant from liability for publishing knowing or reckless falsehoods [or] ... in any way vitiate the value of the evidence that defendant published the charge with reckless disregard for truth or falsity.” Connaughton v. Harte Hanks Commc’ns, Inc., 842 F.2d 825, 837-38 n. 6 (6th Cir.1988) (brackets and internal quotation marks omitted), aff'd 491 U.S. 657, 109 S.Ct. 2678, 105 L.Ed.2d 562 (1989).)
Certainly, the evidence does not compel a finding that Bensinger acted with reckless disregard for the truth. A jury might well vindicate him. It seems to me, however, that a jury could also properly find malice. I respectfully dissent from the ground on which the panel opinion affirms summary judgment for Bensinger.