join, Concurring in Part and Dissenting in Part:
With all due respect to our colleagues, one of Judge Barksdale’s opening statements puts this case in perspective:
[I]t is simply nothing short of absurd to hold that the police chiefs and .sheriffs are not vested with discretion in choosing which teachers to use (and pay) for training the police chiefs and sheriffs’ own student-officers — the very persons the police chiefs and sheriffs are responsible for training.- This cannot be the law.
In holding otherwise, as he says, the majority “has turned its back on the fundamental, compelling reasons for qualified immunity; it ignores the discretionary element that lies at the heart of that doctrine.” The majority has rendered a very un-balaneed analysis of the balancing tests required in this case. This portion of our dissent will address the qualified immunity claim of the police officials as it relates to the teachers’ claims for. First Amendment retaliation. Judge Barksdale’s portion of the dissent discusses the police officials’ potential liability for violating 42 U.S.C. § 1985 and their corresponding immunity claim.
*382 I. Background
To set the stage for the police officials’ actions against Kinney and Hall, it is useful to recount undisputed facts concerning their expert testimony and the officials’ concerns. They agreed, without following ETPA instructions to obtain prior approval, to become paid experts in 1997 on behalf of the plaintiffs in Gonzales v. City of Kerrville. A year later, at trial, the Kerrville newspaper reported that eyewitnesses testified the suspect had fired in excess of forty shots while standing on the Guadalupe River Dam hitting objects including an apartment window, a garbage can and a patrol car window. The defendant police sniper testified that he first told the suspect to drop his rifle, and when the suspect lifted the rifle and pointed it at him, the officer killed the suspect in self-defense. Kinney’s and Hall’s expert conclusions were that the sniper’s failure to apply his training and defendant, City of Kerrville’s lack of a proper policy were proximate causes of the tragic shooting and that the sniper’s use of deadly force amounted to excessive force.
Rejecting these expert opinions, the jury found in favor of the Kerrville police officer, and the federal district judge overturned the award against the city. After Kinney’s and Hall’s opinion was rejected, the take-nothing judgment was affirmed by this court on appeal. See Gonzales v. City of Kerrville, 205 F.3d 1337 (5th Cir.1999).
The police officials have deposed or attested, inter alia, that appellees’ expert testimony hurt the close working relationship required between academy instructors and representatives of the cities and counties; damaged teamwork required among those involved in training officers; threatened the confidentiality of information city and county officers share with Kinney and Hall about their procedures and practices; undermined feelings of loyalty and confidence; and represented an improper use of the instructors’ affiliation with ETPA.
II. Standard of Review
While the majority correctly cites the general standards of review for summary judgment and qualified immunity appeals, they repeatedly mischaracterize the court’s function in free speech cases and thus would send to the jury issues that it is our obligation to decide. This case is, we are agreed, governed by the balancing test framed by the Supreme Court in Pickering v. Bd. of Educ., 391 U.S. 563, 88 S.Ct. 1731, 1734-35, 20 L.Ed.2d 811 (1968), and refined and extended by Connick v. Myers, 461 U.S. 138, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983), and Board of County Commissioners v. Umbehr, 518 U.S. 668, 116 S.Ct. 2342, 135 L.Ed.2d 843 (1996). The relevant issues are: (1) whether an employee’s or contractor’s speech constituted a matter of public concern; (2) whether the public employer’s legitimate countervailing government interests outweigh the value of the protected speech; (3) whether the protected speech was a substantial or motivating factor in the discipline or termination; and (4) whether the employer would have acted against the employee for some other conduct regardless of the speech. See Umbehr, 518 U.S. at 675, 116 S.Ct. at 2347. The first two issues are matters for the court to decide de novo while the last two may comprise jury issues. See Williams v. Seniff, 342 F.3d 774, 782 (7th Cir.2003); Melton v. City of Oklahoma City, 879 F.2d 706, 713 (10th Cir.1989). Courts, not juries, determine the extent of protection accorded to First Amendment conduct as matters of policy and uniformity. Melton, 879 F.2d at 713 (concluding that “the trial court improperly submitted to the jury the ques-
*383tion of whether [the plaintiffs] speech was constitutionally protected”).
The majority, unfortunately, appears to have confused the second and third issues and thus would leave to the jury the fundamental question of First Amendment protection that is ours to decide. The majority holds that because a fact dispute exists as to whether Kinney and Hall were “blackballed” or “boycotted” to enforce a “code of silence” (the third Pickering issue), this court may not take into account the police officials’ proffered institutional reasons for disenrolling their officers from appellees’ classes (the second issue). The majority reasons because it must give Kinney and Hall the benefit of drawing all inferences in their favor on summary judgment review, a trial is required to determine the legitimacy of the governmental interests. The weight those interests receive in the Pickering balance is, however, for this court to decide. Accordingly, sending to the jury issues crucial to the Pickering balance would be improper.
That this court alone decides the Pickering balance is reinforced by several facts. First, all discovery is complete, and there is no real dispute about the operative facts. Second, whether one characterizes the police officials’ actions as merely “disenroll-ing” students from appellees’ classes or as “blackballing” or “boycotting” the instructors is a matter of semantics, not motive. Third, there is no evidence that the officials themselves used the term “blackball” or “boycott” to describe their actions; those pejoratives were used by ETPA President Holda and pervade the appel-lees’ complaint and the district court opinion. Finally, the majority opinion itself concludes that the police officials advanced no legitimate interests to place in the Pickering balance, and it freely evaluates the disputed evidence. See, e.g., Kinney v. Weaver, 301 F.3d 253, 258 n.3, 259 n.4, 282 n.25 (5th Cir.2002) (en banc). In other words, while purporting to rest on the existence of disputed fact issues, the majority has rendered its conclusion on the first and second Pickering issues listed above. The majority’s de facto balancing is additionally undermined, not only by its failure to take the entire record into account, but by its erroneous requirement that the police officials prove actual disruption, to the exclusion of potential disruption, caused in their departments by the protected speech. The Supreme Court has held, to the contrary, that an employer’s legitimate concern about potential disruption arising from protected speech is entitled to deference. Umbehr, 518 U.S. at 676, 116 S.Ct. at 2348 (recognizing that the Court has “consistently given greater deference to government predictions of harm used to justify restriction of employee speech”) (citations and quotations omitted).
The majority’s miscalculation of Pickering balancing necessarily affects its conclusion on qualified immunity, as the majority reiterates that there are no legitimate governmental interests on the police officials’ side of the balance.
Unlike the majority, we neither wash our hands of the crucial responsibility to determine the extent of protection owed to Kinney’s and Hall’s voluntary expert testimony, nor obscure the Pickering determination with erroneous or unsupported fact issues. Thus, while deferring balancing at this point, we must acknowledge the existence of legitimate governmental interests on the police officials’ side.
III. Qualified Immunity
The doctrine that confers qualified immunity from suit on public officials performing discretionary functions is not an “insignificant aberration.” See Pierce v. Smith, 117 F.3d 866, 882 (5th Cir.1997). *384For over twenty years, the Supreme Court has explained that qualified immunity strikes a balance between providing redress to individuals for abuses of public office and protecting society against claims that “frequently run against the innocent as well as the guilty[.]” Harlow v. Fitzgerald, 457 U.S. 800, 814, 102 S.Ct. 2727, 2736, 73 L.Ed.2d 396 (1982). Society bears the cost of unfounded lawsuits in “the expenses of litigation, the diversion of official energy from pressing public issues, and the deterrence of able citizens from acceptance of public office.” Id. There is also the “danger” that “fear of being sued will ‘dampen the ardor of all but the most resolute, or the most irresponsible [public officials], in the unflinching discharge of their duties.’ ” Id. (citation and quotation omitted).
For these reasons, qualified immunity shields discretionary official conduct to prevent lawsuits that do not allege violations of clearly established constitutional law of which a reasonable person would have known. Harlow, 457 U.S. at 819, 102 S.Ct. at 2739. The standard of conduct embodies objective legal reasonableness. So measured, qualified immunity affords “ample protection to all but the plainly incompetent or those who knowingly violate the law.” Malley v. Briggs, 475 U.S. 335, 341, 106 S.Ct. 1092, 1096, 89 L.Ed.2d 271 (1985). To disentitle public officials to qualified immunity, the unlawfulness of their conduct “must be apparent,” Anderson v. Creighton, 483 U.S. 635, 640, 107 S.Ct. 3034, 3039, 97 L.Ed.2d 523 (1987), and “all reasonable officials would have realized the particular challenged conduct violated the constitutional provision sued on[.]” Pierce, 117 F.3d at 871 (citations omitted). Indeed, if “officers of reasonable competence could disagree on th[e] issue, immunity “ should be recognized.” Malley, 475 U.S. at 341, 106 S.Ct. at 1096. The law is clearly established only where “it would be clear to a reasonable officer that his conduct was unlawful in the situation he confronted.” Saucier v. Katz, 533 U.S. 194, 202, 121 S.Ct. 2151, 2156, 150 L.Ed.2d 272 (2001) (emphasis added).
No doubt, the test of objective legal reasonableness does not always require immunity in the absence of an identical or even “materially similar” case to guide official conduct. See Hope v. Pelzer, 536 U.S. 730, 741, 122 S.Ct. 2508, 2516, 153 L.Ed.2d 666 (2002). In Hope, the Court held that Alabama prison officials could have readily inferred from pre-existing authority that it was unconstitutional to chain recalcitrant prisoners painfully and long to a “hitching post.” Id. As a context-specific denial of qualified immunity, Hope does not spring eternal for Kinney and Hall. The contrasts between the two eases are plain.
First, the Eighth Amendment proscribes “unnecessary and wanton infliction of pain” on prisoners, Whitley v. Albers, 475 U.S. 312, 319, 106 S.Ct. 1078, 1084, 89 L.Ed.2d 251 (1986). With only two paragraphs of discussion, the Court in Hope found in the prisoner’s allegations an “obvious” Eighth Amendment violation. Hope, 536 U.S. at 741, 122 S.Ct. at 2516. In this case, however, rather than dealing with an “obviously cruel” practice (compare Hope, 536 U.S. at 745, 122 S.Ct. at 2518), the court confronts a First Amendment protection of free speech that is not unequivocal; courts must accommodate the public interest in effective provision of government services when the speaker works for or on behalf of the government. No rigid rule of liability exists. See Pickering, 391 U.S. at 568, 88 S.Ct. at 1734-35. Thus, the majority requires well over 20 pages of legal reasoning to explain why the police officials could not constitutionally *385disenroll their students from Kinney’s and Hall’s classes.
Just as the governing constitutional standard in Hope was simpler, so was the determination that the law was clearly established. An earlier circuit court case had specifically held unconstitutional, inter alia, the practice of handcuffing prisoners to “the fence and to cells for long periods of time.... ” Gates v. Collier, 501 F.2d 1291, 1306 (5th Cir.1974).1 Another case had held it unconstitutional to deny water to a prisoner as punishment for his refusal to work, explaining that conduct which jeopardizes the prisoner’s health or inflicts physical abuse after he stops resisting authority is actionable. Ort v. White, 813 F.2d 318, 325 (11th Cir.1987). Finally, a Department of Justice report to Alabama authorities condemned exactly the corporal punishment at issue in Hope.
Despite the majority’s creative review of Fifth Circuit government employee free speech precedents, none of our cases had remotely conducted the free speech balancing inherent in the relation between law enforcement departments and police academy instructors. As will be seen, the only related authorities were decided outside this Circuit and uniformly denied liability or granted immunity.
Thus, that “fair warning” could be given to the prison officials in Hope does not modify the general test for qualified immunity applicable in this case. As the Court acknowledged, “in some circumstances, as when an earlier case expressly leaves open whether a general rule applies to the particular type of conduct at issue, a very high degree of prior factual particularity may be necessary.” Hope, 536 U.S. at 741-42, 122 S.Ct. at 2516 (quoting United States v. Lanier, 520 U.S. 259, 269, 117 S.Ct. 1219, 137 L.Ed.2d 432 (1997))(internal citations and quotations omitted). Nor did Hope cast doubt on the Court’s decision that to determine whether the law is clearly established, public officials should consider controlling cases in their own jurisdiction or, alternatively, refer to a consensus of persuasive authority outside it. Wilson v. Layne, 526 U.S. 603, 617, 119 S.Ct. 1692, 1700, 143 L.Ed.2d 818 (1999).
Only recently, this Court expressed en banc our caution toward denying qualified immunity in novel factual cases. In McClendon v. City of Columbia, 305 F.3d 314, 332 (5th Cir.2002)(en banc), the author of today’s majority opinion found it compelling that no court in 1993 had applied the state-created danger theory of § 1983 liability to a similar factual context. This court held that “qualified immunity should be granted ‘if a reasonable official would be left uncertain of the law’s application to the facts confronting him.’ ” Id. at 332 (quoting Salas v. Carpenter, 980 F.2d 299, 311 (5th Cir.1992))(other citation omitted).2 Further, despite the adoption of the state-created danger theory of liability by nearly all other circuit courts at the time of the conduct in question, this court denied that they comprised a consensus of eases of persuasive authority sufficient to provide “fair warning,” because the constitutional right was not defined with “sufficient clarity to enable a reasonable official to assess the lawfulness of his conduct.” Id. at 332-33. This court concluded that:
*386The fact that the state-created danger theory was recognized at a general level in [other courts’] precedents did not necessarily provide Officer Carney with notice that his specific actions created such a danger.... [T]his is not a situation where “a general constitutional rule already identified in the decisional law ... applfied] with obvious clarity to the specific conduct in question.”
Id. (internal citation and quotation omitted). McClendon then states: “Indeed, general principles of the law are less likely to provide fair warning where, as here, applicability of the doctrine is highly context-sensitive.” Id. at 332 n. 13 (citation omitted).
As the foregoing authorities suggest, for immunity purposes, the question “is not whether other reasonable or more reasonable courses of action were available” to public officials. See Pierce, 117 F.3d at 883. Immunity shields officials so long as their conduct is reasonable, even though wrong in hindsight. Saucier, 533 U.S. at 205, 121 S.Ct. at 2158. The question here is whether, among police chiefs and sheriffs similarly situated to the appellants, “all but the plainly incompetent” would have realized at the time that what they did violated Kinney’s and Hall’s First Amendment rights to testify voluntarily as expert witnesses. Pierce, 117 F.3d at 883 (citing Hunter v. Bryant, 502 U.S. 224, 228, 112 S.Ct. 534, 537, 116 L.Ed.2d 589 (1991)).
To apply these principles of qualified immunity, the Supreme Court’s two-step test normally begins by considering whether, on the facts alleged by the plaintiffs, any constitutional violation occurred; if a violation could be made out, “the next, sequential step is to ask whether the right was clearly established,” i.e., whether it would be clear to a reasonable officer that his conduct was unlawful in the situation he confronted. Saucier, 533 U.S. at 201-02, 121 S.Ct. at 2156.
The majority’s errors become evident by inverting the process here. Consequently, let us assume arguendo that the police chiefs and sheriffs violated the First Amendment by disenrolling their students from Kinney’s and Hall’s classes. Assume, that is, that the appellees engaged in some level of protected speech, and Pickering/Connick balancing applies. The qualified immunity question is as framed by the majority:
We must ask whether it was clearly established at the time of the Police Officials’ conduct that the First Amendment forbade them from retaliating against Kinney and Hall, the employees of their contractor, on account of the instructors’ Kerrville testimony.
Maj. Opn. at 369. Our answer is resoundingly that the law was not clearly established.
The law was not clearly established for three reasons. First, this court and seven other circuits have recognized that public officials are more likely entitled to qualified immunity when the underlying constitutional law depends on balancing tests enforced by the judiciary, and no factually similar case exists. Second, the Fifth Circuit cases relied on by the majority are critically different from this case, while other circuits’ more relevant precedents either found no liability or qualified immunity for law enforcement officials. Third, in an unprecedented approach to Pickering/Connick balancing, the majority inflates the value of the appellees’ “speech,” while discounting from the balancing test the police officials’ legitimate interests; no “clearly established law” supports the majority’s approach.
*387 A. Qualified, Immunity and Constitutional Balancing.
At the heart of Kinney’s and Hall’s First Amendment claim is the case- and context-specific Pickering!Connick balancing test. In Pickering and its progeny, the Supreme Court has balanced the interest of each plaintiff as a citizen in commenting on matters of public concern against the interests of the state, as an employer or contractor, in promoting the efficiency of the public services it performs. Id. at 568, 88 S.Ct. at 1735; see also Connick, 461 U.S. at 140, 103 S.Ct. at 1686. Pickering emphasized, however, that in view of the “enormous variety of fact situations” in which critical statements by public employees may be thought by their superiors to furnish grounds for dismissal, it was not “appropriate or feasible to attempt to lay down a general standard” for resolving free-speeeh claims of public employees and that it could only “indicate some of the general lines along which an analysis of the controlling interests should run.” Pickering, 391 U.S. at 569, 88 S.Ct. at 1735. Subsequently, the Court acknowledged that the particularized balancing required by Pickering is difficult even for judges to accomplish. See Connick, 461 U.S. at 150, 103 S.Ct. at 1692. In short, “while it may have been clear since 1968 that a citizen does not forfeit First Amendment rights entirely when he becomes a public employee [or contractor], the scope of those rights in any given factual situation has not been well defined.” Benson v. Allphin, 786 F.2d 268, 276 (7th Cir.1986).
For immunity determinations, the implications of this rule-avoiding constitutional standard seem obvious. The Supreme Court has alluded to the enhanced likelihood of granting qualified immunity in First Amendment cases:
Even when the general rule has long been clearly established (for instance, the First Amendment bars retaliation for protected speech), the substantive legal doctrine on which the plaintiff relies may facilitate summary judgment. ... [TJhere may be doubt as to the illegality of the defendant’s particular conduct (for instance, whether a plaintiffs speech was on a matter of public concern).
Crawford-El v. Britton, 523 U.S. 574, 592-93, 118 S.Ct. 1584, 1594, 140 L.Ed.2d 759 (1998). Fifteen years ago, this court explained that:
One consequence of case-by-case balancing is its implication for the qualified immunity of public officials whose actions are alleged to have violated an employee’s First Amendment rights. There will rarely be a basis for a priori judgment that the termination or discipline of a public employee violated “clearly established” constitutional rights.
Noyola v. Texas Dep’t of Human Res., 846 F.2d 1021, 1025 (5th Cir.1988). Noyola’s “self-evident tenet of qualified immunity jurisprudence,” see Moran v. Washington, 147 F.3d 839, 846 (9th Cir.1998), has been embraced by at least seven other circuits.3 Even before Noyola, the Sev*388enth Circuit held that when a constitutional rule involves the balancing of competing interests, the standard may be clearly established, but its application is so fact dependent that the “law” can rarely be considered “clearly established.” Benson, 786 F.2d at 276. In such cases, “the facts of the existing case law must closely correspond to the contested action before the defendant official is subject to liability under Harlow [v. Fitzgerald]” Id. Noyola, Moran and Benson express the consensus view among circuit courts.
While the majority relegates Noyola to a footnote, that case remains the law of this Circuit. Judge Higginbotham, for instance, cited Noyola when observing that, “the fact-specific balancing test of Pickering complicates the question of whether an act violated clear law. This is because the question is not only the clarity of the standard but its clarity in application.” Boddie v. City of Columbus, 989 F.2d 745, 750 (5th Cir.1993). Judge Garwood added that qualified immunity
principles have particular force where, as here, resolution of whether the defendant’s conduct violated the constitutional provision sued on is heavily dependent on a balancing or weighing against each other of different factors according to
the degree they are present in the matrix of facts constituting the particular context in which the asserted violation occurred.
Pierce, at 882. Noyola has been frequently cited in our court.4
Noyola counsels judicial reticence toward abrogating qualified immunity in government employee First Amendment cases, but it does not act as a dispensation of the duty to examine each case carefully. In several cases, after citing Noyola, this court has denied the defense.5 When, as here, the First Amendment case law is dissimilar from the precedents, we must echo the caution expressed in Noyola, Boddie, Pierce, and among the circuits that fact-sensitive balancing of “the matrix of factors constituting the particular context in which the asserted violation occurred” gives particular force to an immunity defense. In such cases, “a very high degree of prior factual particularity may be necessary.” Hope, 536 U.S. at 741, 122 S.Ct. at 2516; see also McClendon, 305 F.3d at 332 n. 13.
B. Finding Similar Cases for Immunity Comparison.
Against the backdrop of Noyola and the First Amendment balancing standards, we *389conclude that no reasonable police chiefs and sheriffs could have clearly understood in October, 1998 that they were violating the First Amendment by refusing to enroll their recruits in Kinney’s and Hall’s classes. No compelling or compellingly analogous Fifth Circuit case law gave the officials “fair warning” in this context-sensitive balancing area of constitutional law. Cases from other circuits uniformly granted extra deference to law enforcement officials’ decisions.
The majority apparently overlooks the requirement that there be a higher degree of similarity between cases to satisfy the clearly established law prong of qualified immunity. The majority concedes that “our past cases do not include one that has specifically addressed retaliation against instructors at a police academy.” Instead, the majority relies exclusively on ordinary whistleblower cases6 and one case brought by a college professor who testified as an expert witness.7 Such cases entail, however, a significantly different mix of interests for balancing purposes than the one before us.
Consider first the “ordinary” whistle-blower cases. This court has consistently held that a public employee is “speaking out on a matter of public concern” when he becomes a “whistleblower” and thus complains of, or testifies against, fellow employees’ misconduct or against his employer’s practices. This court has protected a wide variety of whistleblower conduct, some of it emanating from within law enforcement agencies. All of these cases, however, concerned fact witnesses, employees who had personal knowledge of misconduct within their own governmental units.
In holding that ordinary whistleblower cases afford “clearly established law” for this case, the majority elides several critical distinctions. Foremost, Kinney’s and Hall’s testimony did not equate with whis-tleblower conduct. Their opinions were valuable only insofar as they correspond with someone else’s account of the underlying facts. But it is the eyewitness who “blows the whistle,” not the expert who simply synthesizes and interprets the factual testimony. Qualified expert testimony is fungible, not irreplaceable. The majority implies, nevertheless, that without Kinney’s and Hall’s expert testimony, the plaintiff in the Kerrville case would have been unable to pursue his claim. Thus the public has a special interest in receiving expert opinions. This suggestion blinks reality. Our litigious culture affords well-qualified experts in every conceivable specialty, including law enforcement practices and training. If the majority intends, not so subtly, to hint that these experts had unique credibility because of their affiliation with ETPA, their implication proves the police officials’ contention: Kinney and Hall created a conflict of interest by taking advantage of their job titles in the courtroom.
Not only is the speech in whistleblower cases generically different from appellees’ expert testimony, but the corresponding interests of public employers are different. This court has been unsympathetic to employer retaliation against government whistleblowers, since their unorthodox conduct may furnish the public’s only protection against internal misconduct. A *390public employer has little, if any, legitimate interest in hiding dirty linen from the taxpaying public. The case before us is not, however, so easily pigeonholed. The police officials are not concealing misdeeds within their departments. Indeed, since the Kerrville plaintiff on whose behalf Kinney and Hall testified left court empty-handed, the police officials’ “retaliation” did not ultimately stifle the exposure of wrongdoing. The majority’s facile analogy with ordinary whistleblower cases is simply wrong. We have here assumed that the police officials’ actions would not satisfy the Pickering/Connick balancing test after careful analysis, but such a legal conclusion does not so ineluctably follow from a few citations to whistleblower cases as to “clearly establish” the guiding law.
The majority’s analogy to cases involving educators is also weak. In Rainey, this court concluded that a college teacher’s contracts were unconstitutionally breached because of his testimony as a defense expert witness in a pornography case. Rainey v. Jackson State Coll., 481 F.2d 347, 349 (5th Cir.1973). Holding that the breach violated Rainey’s First Amendment rights, this court did not engage in Pickering balancing. Id. at 349-50. By its nature, Rainey’s testimony could not have conflicted with the interests of his employer. No countervailing employer interests were advanced by the college against Rainey’s right to testify. Legally and factually, Rainey is a poor fit with this case.
Closer factually to the instant case is the policy of Texas A&M University (and a state legislative appropriation provision), implemented before the police officials took action directed at Kinney and Hall, that broadly forbade university employees from testifying as expert witnesses for parties adverse to the state’s interests. See Hoover v. Morales, 164 F.3d 221, 223-24 (5th Cir.1998). The police officials cite the policy as reflecting clearly established law in the Fifth Circuit. The majority discounts appellants’ reliance, because the policies were under federal court challenge, and ultimately did not survive. We agree that Hoover’s context is sufficiently different as not to furnish controlling authority in support of the police officials.
By the same token, however, the majority ought to concede that Hoover reinforces the principle that in this context-sensitive balancing area of constitutional law, what is clearly established must be closely related factually and legally to a case at hand. Significantly, this Court in Hoover “assumed that there will be occasions when the state’s interests in efficient delivery of public services will be hindered by a state employee acting as an expert witness or consultant-” 164 F.3d at 226 (emphasis added). Hoover concludes by stating:
But our task in this case requires us to apply a Pickering case-by-case analysis, and in doing so we conclude that the expert witness rider and TAMUS policy No. 3105 are impermissibly overbroad. Our opinion does not foreclose consideration of rules and regulations aimed at limiting expert testimony of faculty members or other state employees which adhere to our First Amendment jurisprudence.
164 F.3d at 227. Unlike the majority opinion, Hoover does not oversimplify Pickering balancing and in its way lends powerful support to the officials’ plea that no clearly established Fifth Circuit law condemned their actions regarding Kinney and Hall.
While the majority has strained to find that clearly established Fifth Circuit law was contrary to the police officials’ conduct, they ignore or minimize, in the immunity discussion, three circuit court cases involving alleged retaliation by law en*391forcement agencies for non-whistleblower testimony.
The case most closely on point is Tedder v. Norman, 167 F.3d 1213 (8th Cir.1999), decided only a few months after the events at issue here. Tedder was the Deputy Director of the Arkansas Law Enforcement Training Academy. After voluntarily testifying as an expert for the plaintiff in an excessive use of force case, Tedder was demoted. The Eight Circuit affirmed a summary judgment in Tedder’s First Amendment lawsuit against his supervisor. The court held that:
Testimony concerning possible misconduct of public officials is speech on a matter of public concern that warrants constitutional protection, ... but, as the district court stated, “it is not the place for an employee of ALETA, let alone its Deputy Director to volunteer to give such testimony without a subpoena.”
167 F.3d at 1215. Further, the court found a “significant threat of disruption to the relationships between the [academy] and the law enforcement agencies that it trains.” Id. at 1215. On balancing the relevant interests, the court ruled for the defendant against Tedder’s claim of unconstitutional retaliation. Id.
The majority would distinguish Tedder because the defendant there testified against an officer employed by a law enforcement agency actually trained by the Arkansas academy. The Tedder court never specifically emphasizes this fact, however, and it found that the testimony caused “actual disruption and potential further disruption” to the academy. Id. Tedder not only undercuts the majority’s First Amendment analysis, but clearly supports a finding of qualified immunity. It would be a strange constitutional rule indeed that protects a public employer’s adverse action against an employee for expert testimony, but punishes the non-employer for concerns over the very same activity. Even stranger would be the denial of qualified immunity to the non-employer whose internal relations are most affected by the expert testimony, while Tedder’s supervisor was granted qualified immunity. Id.
The majority also ignores a Third Circuit case, decided well before the events here, which exonerated a law enforcement agency that demoted one of its officers for voluntarily appearing as a character witness (for a friend’s son) at a bail bond hearing. Green v. Philadelphia Hous. Auth., 105 F.3d 882 (3rd Cir.1997). In Green, the officer left the hearing, declining to take the stand, when he learned that the son was charged with involvement in a drug ring. The court found that the officer’s decision to testify constituted First Amendment protected activity, but also that the public’s interest in his voluntary court appearance is “somewhat more limited than it would be if his appearance were subpoenaed.” 105 F.3d at 888 (citing cases).
Ultimately, the Pickering/Connick balancing test weighed in the department’s favor, as an employer, because of its significant interests in protecting the department’s reputation and in successfully fighting drugs and crime. The court held that “any risk of departmental injury or disruption weighs heavily under the Pickering balancing test.” Id. Green thus found for the police department even though Green’s supervisor had previously approved his court appearance.
For immunity purposes, Green is closely related contextually to the present case. Green attributed significant weight to the police department’s justification for its disciplinary action, and it carefully explains why not all court testimony is equivalent for First Amendment purposes. In these ways, Green furnishes a backdrop for the *392police officials’ conduct just the opposite of the synthetic “clearly established law” concocted by the majority.
The third case relevant for immunity purposes was brought against an Oklahoma district attorney and agents for the Oklahoma Bureau of Narcotics and Dangerous Drugs, alleging that the rescission of an offer of employment to coordinate the DA’s drug task force was based on the plaintiffs previous expert witness testimony for a murder defendant. Worrell v. Henry, 219 F.3d 1197 (10th Cir.2000). The murder trial in which the would-be employee testified involved the killing of one of the narcotics bureau’s agents. The court granted summary judgment for the district attorney, who was the prospective employer, but it denied summary judgment and qualified immunity to the chief narcotics bureau agent. The Tenth Circuit held that Pickering/Connick balancing was appropriate to evaluate the First Amendment consequences of the district attorney’s refusal to hire Worrell, but it did not find Pickering appropriate to analyze the alleged retaliation by the non-employer, non-contractor agents of the narcotics bureau. In reaching the latter conclusion, the court acknowledged “that there may be instances in which the operations of a third party agency are so intertwined with the operations of the employing agency that the Pickering balancing should be applied.” Worrell, 219 F.3d at 1212, n. 3.
Worrell demonstrates that if the police officials’ role is viewed through the Pickering/Connick lens, their claim to immunity should be ironclad. Even if their position more closely resembles that of the narcotics bureau agents, however, they could argue that they fall under Worrell’s caveat because their operations are closely intertwined with ETPA.
We may end this section where we began. Because this case involves constitutional balancing, the “clearly established law” must have existed at a higher level of specificity than might be required in other types of immunity eases. This is hardly an extraordinary conclusion. It follows as a negative implication from this court’s en banc holding in McClendon that, absent controlling circuit authority, “a ‘consensus of cases of persuasive authority1 might, under some circumstances, be sufficient to compel the conclusion that no reasonable officer could have believed that his or her actions were lawful.” McClendon, 305 F.3d at 329. In this case, the Fifth Circuit precedents cited by the majority involve fundamental distinctions in the nature of the speech as well as the public employer’s interests. While useful, such cases hardly compel the conclusion that the police officials could not properly disenroll their officers from Kinney’s and Hall’s classes. The majority overlooked other circuits’ cases that discussed Pickering balancing in the specific context of law enforcement agencies and various types of testimony. Whether or not the majority would agree with the outcome of those cases, two of them predate the police officials’ conduct here, and they should be regarded as constituting a consensus of persuasive authority arrayed against the majority’s conclusion. At best, one must conclude that there was no “clearly established law” that gave the police officials “fair warning” of the unconstitutionality of their conduct. See McClendon, 305 F.3d at 332-33 (no consensus of cases from other jurisdictions where those cases applied the constitutional rule differently and facts were insufficiently similar; qualified immunity granted to police officer).
C. Novelty in the Majority’s Balancing Exercise.
The third proof of error in the majority’s qualified immunity analysis arises from the *393way it strikes the Pickering/Connick balance. In October, 1998, no court had held that a law enforcement employee’s right to testify voluntarily as an expert witness outweighed the interests of the agency. See Worrell, 219 F.3d at 1206-07 (discussing prior circuit court cases and noting that even where Pickering balancing favored the employee, a different result might be reached where an agency could show a disruption in its operations). And to this day, no cases have, in the law enforcement context, elevated non-whistle-blower testimony so high, or rated the department’s interests so low, as the majority does here. This is not to say (at this point) that the majority is incorrect, but the novelty of this balance cuts against any conclusion that “clearly established law” proscribed the police officials’ conduct. The appellants’ position thus resembles that of the county supervisors in Umbehr, whose qualified immunity was upheld on appeal while the Supreme Court approved the application of Pickering/Connick balancing to the relations between independent contractors and government entities. See generally, Umbehr, 518 U.S. 668, 116 S.Ct. 2342.
Because the majority has exaggerated the analogy between voluntary expert testimony and whistleblower testimony, it elevated Kinney’s and Hall’s interests in testifying as voluntary expert witnesses to almost absolutely protected status. The majority has thus extended or partially overruled Hoover v. Morales, which rejected such an absolutist approach. See Hoover, 164 F.3d at 227. From the perspective of other circuits, too, the majority’s conclusion is unprecedented. The Tenth Circuit has specifically held otherwise: “First Amendment protection of public employees’ testimony is not absolute. There are instances in which government entities’ interests as employers outweigh employees’ interests in free expression and the policy of encouraging truthful and uninhibited testimony.” Worrell, 219 F.3d at 1205. Worrell then described with approval the way in which the courts in Green and Tedder evaluated the clash between law enforcement officers’ rights to testify and their agency’s significant interests. Worrell, 219 F.3d at 1206-07. In Green, as noted above, the Third Circuit held that the officer’s voluntary appearance at a bail hearing, although constitutionally significant, was entitled to less weight. Green, 105 F.3d at 888-89. The Eighth Circuit in Tedder also decided that the voluntariness of the deputy director’s expert testimony in a police brutality case lessened its First Amendment protection. Tedder, 167 F.3d at 1215. No other case has ascribed to voluntary expert witness testimony like that of Kinney and Hall such elevated First Amendment status.
Likewise, in unprecedented fashion, the majority holds for naught the police officials’ description of their institutional interests in controlling the education of department officers.8 The majority’s hostility toward the police officials’ position is contrary to Waters v. Churchill, which described the government’s “significant” interest as an employer as follows:
When someone who is paid a salary so that she will contribute to an agency’s effective operation begins to do or say things that detract from the agency’s effective operation, the government employer must have some power to restrain her.
*394511 U.S. 661, 675, 114 S.Ct. 1878, 1887-88, 128 L.Ed.2d 686 (1994). Waters further noted:
[W]e have consistently given greater deference to government predictions of harm used to justify restrictions of employee speech than to predictions of harm used to justify restrictions on the speech of the public at large. Few of the examples we have discussed involve tangible, present interference with the agency’s operation. The danger in them is mostly speculative.
511 U.S. at 673, 114 S.Ct. at 1887. In Umbehr, too, the Court reminded that, “Pickering requires a fact-sensitive and deferential weighing of the government’s legitimate interests.” Umbehr, 518 U.S. at 677, 116 S.Ct. at 2348.
As was previously explained, the expert testimony caused an uproar because police officials and student-officers feared Kinney and Hall might use information gleaned in their classes to testify against the sponsoring agencies; that their testimony interfered with unfettered classroom discussions; and that they misused their affiliation with ETPA to enhance their testimony. One can certainly understand the sensitivity of the police officials about Kinney’s and Hall’s testimony. Far from exhibiting wanton police brutality, the Gonzales case portrayed an officer’s life endangered by a deranged shooter. Such situations are the stuff of law officers’ nightmares and domestic tragedies. There is no evidence that Kinney and Hall were lawyers, and they are entitled to their professional opinions as law enforcement instructors. Nevertheless, feelings of loyalty, confidence and teamwork between the agencies and the instructors were understandably strained by this testimony. Further, it is evident that the instructors enhanced their credibility because of their association with ETPA, and that the police officials might legitimately question whether the instructors’ impartiality was undermined because they initially agreed to be paid experts.
The majority might refuse to defer and throw all these institutional concerns to the winds in its First Amendment analysis. But in doing so, not only does it abuse the general cautions expressed by the Supreme Court, but it contradicts authorities from several circuits. In Green, the Third Circuit described as “very significant” the interests of the housing authority police department as an employer where the officer’s voluntary testimony created a “risk of departmental injury based on the ‘potential disrúptiveness of the speech.’ ” (Green, 105 F.3d at 888, quoting Waters, 511 U.S. at 680, 114 S.Ct. at 1890.)
In Tedder, the court concluded that Ted-der’s testimony caused actual disruption and potential further disruption between ALETA and the law enforcement agencies that it was charged with training. Tedder, 167 F.3d at 1215. The court was concerned that students’ loss of faith in the ALETA’s Deputy Director, who had the authority to approve or veto lesson plans, could spread to every class taught there. See id. The Eighth Circuit has elsewhere recognized in emphatic terms the heightened interests of law enforcement agencies. See, e.g., Shands v. City of Kennett, 993 F.2d 1337, 1344-45 (8th Cir.1993); Tindle v. Caudell, 56 F.3d 966, 971-73 (8th Cir.1995).
In Worrell, the Tenth Circuit reiterated that “personal loyalty and confidence among employees are especially important in law enforcement” and noted that “[tjhese concerns are heightened in smaller offices and departments, where relatively minor disturbances in morale may create significant problems.” Worrell, 219 F.3d at 1208. The court adds that the district attorney was not obliged to *395wait for an actual breakdown in the functioning of his taskforce before taking action. Id. at 1208-09. He was entitled to rely on reasonable predictions of workplace disruption. Id.; see also Waters, 511 U.S. at 673, 114 S.Ct. at 1887 (noting the “substantial weight [afforded] to government employers’ reasonable predictions of disruptions”).
The Seventh Circuit has repeatedly held that “ ‘[djeference to the employer’s judgment regarding the disruptive nature of an employee’s speech is especially important in the context of law enforcement.’ ” Williams v. Seniff, 342 F.3d 774, 783 (7th Cir.2003) (quoting Kokkinis v. Ivkovich, 185 F.3d 840, 845 (7th Cir.1999)).
Finally, the Sixth Circuit has held that police officials are entitled to qualified immunity for taking reasonable administrative action to preclude one of their officers from exploiting his uniform and his position in the police department to advocate on behalf of the National Rifle Association. See generally, Thomas v. Whalen, 51 F.3d 1285 (6th Cir.1995). While acknowledging the protected status of the officer’s political speech, the court pointed out that “no court has recognized a right to exploit one’s rank in public employment solely for the purpose of enhancing credibility for personal or political gain.” Whalen, 51 F.3d at 1291.
In addition, the majority wholly overlooks that public employers are entitled to deference in dealing with employees whose trust and loyalty are essential to the functioning of a public office. See, e.g., Connick, supra, 461 U.S. at 151-52, 103 S.Ct. at 1692 (“When close working relationships are essential to fulfilling public responsibilities, a wide degree of deference to the employer’s judgment is appropriate.”); Kinsey v. Salado Indep. Sch. Dist., 950 F.2d 988, 994 (5th Cir.1992) (en banc). There can hardly be any dispute that law enforcement instructors occupy a sensitive and extremely important position with respect to the agencies’ mission. If training fails because the trainees have lost confidence in their instructors and are unwilling to discuss issues frankly with them, the consequences can be disastrous. The above-cited cases specifically refer to institutional loyalty as a quality especially required in law enforcement agencies, yet the majority ignores it.
At this point, we have assumed the correctness of the majority’s final conclusion that a First Amendment violation occurred. But in its novel approach to balancing the instructors’ interests against those of the law enforcement agencies, the majority is making new law, not simply expounding “clearly established law.” Kinney and Hall may choose to pursue their suits against the municipal entities, but the individual defendants are entitled to qualified immunity.
TV. Was There a Violation of the First Amendment ?
Although we have demonstrated that the police officials were entitled to qualified immunity regardless of whether their conduct violated the First Amendment, we would also hold that their actions were not, under the Pickering/Connick balancing test, unconstitutional.
Pickering “requires a fact-sensitive and deferential weighing of the government’s legitimate interests.” Umbehr, 518 U.S. at 677, 116 S.Ct. at 2348. In holding that Pickering balancing applies to cases in which a governmental entity appears to condition the provision of contracts on a third party’s constitutionally protected expression, the Supreme Court observed that the “nuanced” Pickering approach, “which recognizes that a variety of interests may arise in independent contractor cases, is superior to a bright-line rule distinguish*396ing independent contractors from employees.” Id. at 678, 116 S.Ct. 2342. The Court also found it “far from clear, as a general matter, whether the balance of interests at stake is more favorable to the government in independent contractor cases than in employee eases.” Id. at 680, 116 S.Ct. at 2350.
Whether Kinney and Hall are classified as third-party independent contractors or as employees is not as significant as how their overall function, and their voluntary expert testimony, affected the law enforcement agencies’ performance of a public mission. Compare Umbehr, id. at 679, 116 S.Ct. at 2349 (noting that a bright-line rule that “would leave First Amendment rights unduly dependent on whether a state law labels a government service provider’s contract” as one of employment or a contract for services is “a very poor proxy for the interests at stake”). To the extent the majority opinion depends on labeling Kinney and Hall as independent contractors rather than employees, its analysis is oversimplified and inconsistent with Umbehr,9 Further, the majority’s reliance on this court’s cases involving government contractors is hollow, since neither the speech at issue in those eases nor the governmental interests at stake is comparable to the present case.10
Balancing the interests in this case on a clean slate, the appellees’ testimony constituted speech on a matter of public concern and was entitled to some level of constitutional protection. For reasons previously discussed, we, unlike the majority, do not characterize the protection as “extremely strong.” Other circuits’ opinions have properly distinguished voluntary testimony from testimony under subpoena. See Green, 105 F.3d at 888; Tedder, 167 F.3d at 1215. Further, voluntary expert witness testimony is distinct from standard whistleblower conduct, and on the facts of this case, Kinney’s and Hall’s expert opinions were not essential to exposing wrongdoing by a policeman or a police department.
The interests of the law enforcement agencies have been well-documented in other cases, where it has been held that there is a “special need for deference to the employment decisions of those responsible for insuring public safety.” Kokkinis, 185 F.3d at 845. Law enforcement agencies have “a more significant interest than the typical government employer in regulating the speech activities of [their] employees in order to promote efficiency, foster loyalty and obedience to superior officers, maintain morale, and instill public confidence.” Tyler v. City of Mountain Home, 72 F.3d 568, 570 (8th Cir.1995) (internal quotation marks and citations omitted). The police officials here deposed or attested that appellees’ expert testimony hurt the close working relationship required between academy instructors and representatives of the cities and counties; damaged teamwork required among those involved in training officers; threatened the confidentiality of information officers share with their instructors about procedures and practices; undermined feelings of loyalty and confidence; and represented an improper use of the *397instructors’ affiliation with ETPA. The police officials offered evidence of actual and potential disruption to their training programs and departments.
Two factors detract, in the majority’s view, from the strength of these articulated interests and the deference they are due. First, it is contended that because appellees testified “400 miles away” from ETPA and in a case not involving a trainee or department sponsor of the ETPA program, the police officials’ concerns are misplaced. The distance between Kerrville and Kilgore, Texas, is a red herring. Both cities are in Texas, both are governed by the same regime of state and federal law, and there is no showing that police officers trained in east Texas rarely or never migrate to other areas of the state, or that ETPA’s influence spreads no further than the boundaries of its sponsoring department. The “400 mile argument” is disingenuous.11
That the expert testimony posed a conflict of interest with Kinney’s and Hall’s status as instructors, despite its taking place outside ETPA’s formal jurisdiction, is a conclusion entitled to deference. As the majority notes, the police officials did not, in events leading up to this case, explain how they used the term conflict of interest, but the officials’ ineloquence does not mean their judgment is entitled to no weight. Moreover, it is evident that in testifying before a Kerrville jury, the ap-pellees’ status as instructors at a Texas police academy would enhance their credibility and lend the prestige of ETPA to their words. By their status, the instructors necessarily implicated the sponsoring departments (despite any professional disclaimers) in the legitimacy of their expert opinions. Even if other possible standards for conflicts of interest, e.g., detracting from their time available to prepare for and conduct ETPA classes or “double-dipping” on salary and witness fees, are excluded, a conflict existed in this sense. See, e.g., Thomas v. Whalen, 51 F.3d at 1292 (discussing a government entity’s interest in preserving the appearance of impartiality).
More consequential is the majority’s criticism that the police officials could not legitimately discipline Kinney and Hall to enforce an “unwritten code” of silence, whereby police officers do not testify against each other. We agree that enforcing the “code of silence” to stifle speech concerning police misconduct is not a legitimate governmental interest. The evidence shows, however, that this was not a contemporaneous justification formally offered for the police officials’ conduct, and, in fact, reference to an “unwritten code” was made by only one of the appellants, during a television interview. As can be seen from the wealth of detail in the majority opinion, this case contained an abundance of contemporary oral and documentary evidence as well as post-litigation depositions that explored the police officials’ reasons for their action. That only one reference appears throughout the record to an “unwritten code” is significant. This stray remark should not be blown out of proportion.
The dominant theme in cases that have considered the Pickering balance in the context of law enforcement is the need for a high degree of personal loyalty and confidence, esprit de corps, harmony and good morale within departments and between instructors and trainees. Tedder, for instance, relied on the interests of the actual employer, the police training academy, which has only an indirect stake in the *398results of training. Tedder, 167 F.3d at 1215. Tedder reflects the even stronger interests of the officials in this case, whose departments and officers rely on the academy’s training for the sake of the public. In Worrell, the Tenth Circuit held that a district attorney’s interest in preventing disruption with other law-enforcement agencies outweighed the applicant’s interest in avoiding retaliation for his testimony as an expert witness against a police officer. Worrell, 219 F.3d at 1208-09. In Green, the reputation and law enforcement capacity of the housing authority police were held to prevail over an officer’s decision to testify voluntarily. Green, 105 F.3d at 888-89. The Seventh Circuit has on several occasions held that protected speech was subordinate to the institutional interests of law-enforcement agencies. See Seniff, 342 F.3d at 783-85; Kokkinis, 185 F.3d at 844-45.
In stark contrast to those cases, the majority here ignores the paramilitary interests of the law-enforcement agencies and reduces their “legitimate” concerns to the instructors’ competence and teaching ability. As a matter of law, and based on this record, those interests are too narrowly defined.
We conclude that this is a closer case under Pickering balancing than others in the law enforcement area. While Kinney and Hall engaged in protected conduct, their voluntary expert testimony did not carry such a high degree of public importance, in general or in the facts of this case, as ordinary whistleblower testimony. Further, we owe special deference to the law enforcement agencies’ legitimate interests in maintaining discipline, harmony, confidentiality and morale in their departments and training programs. We also note that while the police officials opined that Kinney and Hall should be fired, they succeeded only in disenrolling their students from the instructors’ classes. They did not and could not directly terminate appellees’ employment.
On balance, we conclude that the police officials did not violate the First Amendment by disenrolling their students from appellees’ classes. The officials have the discretion to decide, consistent with the First Amendment, by whom their officers will be taught.
V Conclusion
For the foregoing reasons, I respectfully dissent from the denial of qualified immunity on the appellees’ First Amendment claims, and I join Judge Barksdale’s dissent.
. As the Supreme Court noted, cases decided by the Fifth Circuit before the split that created the Eleventh Circuit remain binding in the Eleventh Circuit.
. Likewise, there is no case in this circuit that denied qualified immunity under the circumstances presented here. The majority's failure to point to such a case is persuasive evidence that these officers were “uncertain of the law’s application to the facts confronting [them].” Id.
. Moran v. Washington, 147 F.3d 839, 847 (9th Cir.1998) (stating that in Pickering balancing cases, "the law regarding such claims will rarely, if ever, be sufficiently 'clearly established’ to preclude qualified immunity"); Kincade v. City of Blue Springs, Mo., 64 F.3d 389, 398 (8th Cir.1995) ("[W]hen PickeringH . is at issue, the asserted First Amendment right can rarely be considered 'clearly established’.... ”); DiMeglio v. Haines, 45 F.3d 790, 806 (4th Cir.1995) (noting that only "infrequently" will the law be clearly established when a balancing of interests is involved); O’Connor v. Steeves, 994 F.2d 905, 917 n. 11 (1st Cir.1993) (same); McDaniel v. Woodard, 886 F.2d 311, 314 (11th Cir.1989) (quoting Noyola and finding qualified immunity applicable because the constitutional right was un*388clear); Melton v. City of Oklahoma City, 879 F.2d 706, 729 (10th Cir.1989) ("In some circumstances, the fact-specific nature of the Pickering balancing may preclude a determination of 'clearly established law' ...."), vacated on other grounds, 928 F.2d 920 (10th Cir.1991) (en banc); Benson v. Allphin, 786 F.2d 268, 276 (7th Cir.1986) (stating that the application of fact-dependent law "can rarely be considered ‘clearly established’ ").
. See Keenan v. Tejeda, 290 F.3d 252 (5th Cir.2002); Harris v. Victoria Indep. Sch. Dist., 336 F.3d 343, 345 (5th Cir.1999) (on petition for reh'g) (per curiam) (specifically stating that Noyola "reflects the law of this circuit”); Pierce v. Smith, 117 F.3d 866 (5th Cir.1997); Wallace v. Texas Tech Univ., 80 F.3d 1042 (5th Cir.1996); Vander Zee v. Reno, 73 F.3d 1365 (5th Cir.1996); Gunaca v. Texas, 65 F.3d 467 (5th Cir.1995); Brady v. Fort Bend County, 58 F.3d 173 (5th Cir.1995); Boddie v. Columbus, 989 F.2d 745 (5th Cir.1993); Caine v. Hardy, 943 F.2d 1406 (5th Cir.1991) (en banc); Kinsey v. Salado Indep. Sch. Dist., 916 F.2d 273 (5th Cir.1990), vacated by, 950 F.2d 988 (5th Cir.1992) (en banc); Connelly v. Comptroller of Currency, 876 F.2d 1209 (5th Cir.1989); Price v. Brittain, 874 F.2d 252 (5th Cir.1989); Evans v. Dallas, 861 F.2d 846 (5th Cir.1988) (per curiam); Brawner v. Richardson, 855 F.2d 187 (5th Cir.1988).
. See Harris, 168 F.3d at 225; Boddie, 989 F.2d at 750; Brawner, 855 F.2d at 193. These cases have little bearing on the application of Piclcering balancing here.
. Brawner, 855 F.2d at 191-92; Matherne v. Wilson, 851 F.2d 752, 761 (5th Cir.1988).
. Rainey v. Jackson State Coll., 481 F.2d 347 (5th Cir.1973). Reeves v. Claiborne County Bd. of Educ., 828 F.2d 1096 (5th Cir.1987), is also cited by the majority as a case concerning an educator, but it does not involve expert testimony and protects truthful fact testimony against one's employer's interest.
. The majority relies on two cases in concluding the appellants have no legitimate interests in the Pickering balancing. Both are inappo-site. In one of these, the policeman was disciplined for associating with a union. See Boddie, 989 F.2d at 747. The other involved -a prison nurse's whistleblower activity. Frazier v. King, 873 F.2d 820, 826 (5th Cir.1989).
. In the foregoing qualified immunity discussion, this opinion, like the majority's, necessarily focuses on government employee cases, since those are the most common. Proper Pickering analysis, however, balances the relevant interests without regard to labels.
. See N. Miss. Communications v. Jones, 792 F.2d, 1330 (5th Cir.1986) (county board retaliated against newspaper for critical editorials and stories) and Blackburn v. City of Marshall, 42 F.3d 925 (5th Cir.1995) (wrecker service denied permission to use police radio frequency after complaint to police chief; no Pickering analysis at all).
. We need not and do not speculate whether expert testimony given outside the State of Texas would have a different impact on the evaluation of the agencies' interests.