Anthony Martin v. Susan Duffy

PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT No. 18-7122 ANTHONY FRED MARTIN, Plaintiff – Appellant, v. SUSAN DUFFY, Defendant – Appellee. Appeal from the United States District Court for the District of South Carolina, at Florence. David C. Norton, District Judge. (4:15-cv-04947-DCN) Submitted: September 11, 2020 Decided: October 13, 2020 Before GREGORY, Chief Judge, WYNN, and HARRIS, Circuit Judges. Reversed and remanded by published opinion. Chief Judge Gregory wrote the opinion, in which Judge Wynn and Judge Harris joined. Matthew A. Fitzgerald, Caroline S. Burton, MCGUIREWOODS LLP, Richmond, Virginia; Nicholas Klaiber, Indira T. Edwards, Aaron Kleinmann, CAPITAL ONE, N.A., Richmond, Virginia, for Appellant. Andrew F. Lindemann, LINDEMANN, DAVIS & HUGHES, P.A., Columbia, South Carolina, for Appellee. GREGORY, Chief Judge: Captain Susan Duffy placed Anthony Fred Martin in administrative segregation the day after he accused a corrections officer of sexual assault. Duffy said it wasn’t her decision. But 110 days later, she authorized Martin’s release. To date, the parties have spent nearly five years litigating whether Duffy segregated Martin because he accused her colleague of misconduct or because safety and investigatory concerns required it. Applying Mt. Healthy City School District Board of Education v. Doyle, 429 U.S. 274 (1977), the magistrate judge presiding over Martin’s case found legitimate penological interests, not retaliatory animus, led to Martin’s segregation. The district court adopted the magistrate judge’s recommendation, denying Martin’s motion for summary judgment and granting Duffy’s cross-motion. We agree that Mt. Healthy provides the appropriate framework for reviewing inmates’ First Amendment retaliation claims. But because the court below improperly resolved genuine disputes of material fact in Duffy’s favor, we reverse the district court’s grant of summary judgment and remand the case for further proceedings consistent with this opinion. I. Martin is an inmate at Perry Correctional Institution (“Perry CI”). In September 2014, Martin filed a grievance against Sergeant B. Rogers, an officer at the facility. Martin said Rogers pulled him aside as he was leaving the cafeteria earlier that day. Under the guise of conducting a “proper shakedown procedure,” Rogers assaulted Martin, “rubbing and touching [Martin’s] anus and penus [sic] area in a lingering and excessive manner.” 2 The next day, Duffy “questioned [Martin] relentlessly” and placed him in administrative segregation. Duffy gave Martin a “Notice of Placement in PHD (Pre- hearing Detention)” form. Beforehand, Duffy had checked a box on the form, indicating that Martin’s segregation was designed to “maintain the integrity of an investigation.” Two months went by without any updates on the status of Martin’s investigation. Still segregated, Martin completed a “Request to Staff Member” form on November 18, 2014. He accused Duffy of placing him in “lock up” as a “reprisal” for his grievance against Rogers and argued that no one since Duffy had asked him about Rogers’ misconduct. Duffy responded a month later: “You were placed under investigation by the Division of Investigations. You are no longer under investigation and are currently on the yard list.” On December 31, 2014, a prison official told Martin he could rejoin the general prison population. Fearing “further and greater acts of retaliation,” Martin refused to return to the yard and requested a transfer. The prison charged him with an offense for failing to obey orders. It found Martin guilty of the charge and imposed sanctions. Proceeding pro se, Martin filed a complaint against Duffy in the District of South Carolina. A magistrate judge screened the complaint under 28 U.S.C. § 1915(e)(2)(B) and concluded that Martin failed to state a cognizable § 1983 claim. See R. & R. 3–5, Martin v. Duffy, No. 4:15-cv-4947-DCN-TER (D.S.C. Dec. 30, 2015). Over Martin’s objections, the district court adopted the magistrate judge’s recommendation and dismissed the suit. Order Affirm. R. & R. 1–2, No. 4:15-cv-4947-DCN-TER (D.S.C. Jan. 20, 2016). Martin appealed. This Court affirmed in part and reversed in part the district court’s judgment. Martin v. Duffy, 858 F.3d 239, 243 (4th Cir. 2017), cert. denied, 138 S. Ct. 738 3 (2018). We agreed that Martin failed to state claims for a violation of his equal protection or due process rights, but held that Martin stated a cognizable claim for retaliation under the First Amendment. We also held that Duffy was not entitled to qualified immunity at this stage because Martin sufficiently alleged that Duffy had violated a clearly established right. Id. at 251 (citing Booker v. S.C. Dep’t of Corr., 855 F.3d 533, 546 (4th Cir. 2017)). Martin’s First Amendment claim was remanded to the district court for consideration on the merits. Id. at 249–54. On remand, Martin filed an amended complaint. First Am. Compl., Martin v. Duffy, No. 4:15-cv-4947-DCN-TER (Oct. 2, 2017). Martin and Duffy both moved for summary judgment. 1 In reviewing the parties’ motions, the magistrate judge concluded that (1) Martin engaged in protected First Amendment activity by filing a grievance, (2) Duffy thereafter placed Martin in segregation—an action that adversely affected Martin’s First Amendment rights, and (3) there was a causal connection between Martin’s grievance and his placement in segregation. Martin v. Duffy, No. 4:15-cv-4947-DCN-TER, 2018 WL 9850161, at *3 (D.S.C. July 25, 2018). Martin’s prima facie case notwithstanding, the magistrate judge determined that “prison officials ‘may still prevail by proving that they would have made the same decision absent the protected conduct for reasons reasonably related to a legitimate penological interest.’” Id. (quoting Rauser v. Horn, 241 F.3d 330, 334 (3d Cir. 2001)). Applying Rauser’s (i.e., Mt. Healthy’s) same-decision test, the 1 Martin filed a second amended complaint during the summary-judgment briefings. Sec. Am. Compl., Martin v. Duffy, No. 4:15-cv-4947-DCN-TER (Feb. 14, 2018). Neither party supplemented their summary-judgment briefings in response to this amendment. But on appeal, both parties treat the second amended complaint as the operative complaint. 4 magistrate judge determined that Martin’s claim failed because the administrative segregation was rationally related to legitimate penological interests—namely, Duffy’s concern for Martin’s safety and the integrity of a pending investigation. Id. Over Martin’s objections, the district court adopted the magistrate judge’s recommendation, granted Duffy summary judgment, and dismissed Martin’s complaint. Martin v. Duffy, No. 4:15-cv-4947-DCN, 2018 WL 9850164, at *1 (D.S.C. Aug. 6, 2018). Martin filed a Fed. R. Civ. P. 59(e) motion, which the district court denied. Martin then timely filed a notice of appeal. J.A. 284–85. This Court appointed Martin appellate counsel. II. We review a district court’s grant of summary judgment de novo. Carter v. Fleming, 879 F.3d 132, 139 (4th Cir. 2018). Summary judgment is only appropriate when, viewing the facts in the light most favorable to the nonmoving party, “there is no genuine dispute as to any material facts and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. Proc. 56(a). At the summary-judgment stage, we draw “all justifiable inferences . . . in [that party's] favor.” Anderson v. Liberty Lobby, 477 U.S. 242, 255 (1986). We read the pleadings of a pro se plaintiff liberally and interpret them “to raise the strongest arguments that they suggest.” Burgos v. Hopkins, 14 F.3d 787, 790 (2d Cir. 1994). 5 III. To recover damages under 42 U.S.C. § 1983, a plaintiff must show (1) “the conduct complained of was committed by a person acting under color of state law,” and (2) this conduct deprived a person of rights, privileges or immunities secured by the Constitution or laws of the United States. Adickes v. S.H. Kress & Co., 398 U.S. 144, 150 (1970). Here, Martin argues that Duffy violated his First Amendment rights by “repris[ing], harass[ing,] and retaliat[ing] [against him] simply because [he] had attempted to informally resolve a grievance.” To state a colorable First Amendment retaliation claim, a plaintiff “must allege that (1) []he engaged in protected First Amendment activity, (2) the defendant[] took some action that adversely affected [his] First Amendment rights, and (3) there was a causal relationship between [his] protected activity and the defendant[’s] conduct.” Martin, 858 F.3d at 249. A. This case presents us with a matter of first impression for this circuit: whether Mt. Healthy applies to prisoners’ retaliation claims. We find that it does. In the employment context, courts use a burden-shifting framework to evaluate causation in First Amendment retaliation claims. Mt. Healthy, 429 U.S. at 287. Mt. Healthy involved a city school board’s recommendation to not rehire one of the district’s untenured teachers. Id. at 282. Leading up to the teacher’s release, he was involved in several altercations—with fellow teachers; with cafeteria workers; with students. Id. He also publicly criticized his school’s administration on a local radio station. Id. at 282. The district accepted the school board’s recommendation and notified the teacher of its 6 decision. Id. The notice referenced the radio station incident, a crude interaction the teacher had with two students, and “a notable lack of tact in handling professional matters.” Id. at 282–83. The teacher sued, alleging the school board retaliated against him for levying public criticism against the school. Id. at 283. The district court agreed, finding the undisputed facts showed that the teacher’s protected conduct was a substantial factor in the board’s release decision. Id. The Court of Appeals upheld the district court’s grant of summary judgment. The Supreme Court, however, changed course. Specifically, the Supreme Court disavowed “[a] rule of causation which focuse[d] solely on whether protected conduct played a part, ‘substantial’ or otherwise, in a decision not to rehire.” Id. Such a rule, the Court feared, “could place an employee in a better position as a result of the exercise of constitutionally protected conduct than he would have occupied had he done nothing.” Id. The Court found that, standing alone, the “substantial factor” test failed to properly distinguish “between a result caused by a constitutional violation and one not so caused.” Id. The solution? The same-decision test. After an employee establishes a prima facie case of retaliation, the same-decision test allows an employer to defeat the claim by proving “it would have reached the same decision . . . in the absence of the protected conduct.” Id. An employer must make this showing by a preponderance of the evidence. Id. “If the defendant fails to carry that burden, the inference is that ‘but for’ causation . . . has been shown: the plaintiff would not have been harmed had his rights not been violated by the defendant.” Greene v. Doruff, 660 F.3d 975, 979 (7th Cir. 2011) (Posner, J.). 7 B. We first acknowledge the circuit split this question has created. Although several of our sister circuits apply Mt. Healthy to prisoners’ retaliation claims, others have rejected this burden-shifting approach. Compare Smith v. Mosley, 532 F.3d 1270, 1278 (11th Cir. 2008) (applying same-decision test to prisoner’s First Amendment retaliation claim); Rauser, 241 F.3d at 333 (same); Thaddeus-X v. Blatter, 175 F.3d 378, 399 (6th Cir. 1999) (same); Graham v. Henderson, 89 F.3d 75, 80 (2d Cir. 1996) (same); Babcock v. White, 102 F.3d 267, 275 (7th Cir. 1996) (same), with DeMarco v. Davis, 914 F.3d 383, 388 (5th Cir. 2019) (rejecting Mt. Healthy’s burden-shifting framework for prisoner’s First Amendment retaliation claim); Spencer v. Jackson County Mo., 738 F.3d 907, 911–12 (8th Cir. 2013) (same); Peterson v. Shanks, 149 F.3d 1140, 1144 (10th Cir. 1998) (same). The causation element in retaliation claims asks whether the considerations which animated the defendant’s conduct were permissible or impermissible. Smith v. Mosley, 532 F.3d 1270, 1278 (11th Cir. 2008). Courts adopting the same-decision test have determined that the burden of answering that question is best shared. We agree. The alternative approach—requiring plaintiffs to bear the entire burden of establishing “but for” causation in the first instance—places too weighty a burden on the individual arguing he was punished for exercising his constitutional rights. See Woods v. Smith, 60 F.3d 1161, 1166 (5th Cir. 1995) (recognizing that “but for” causation “places a significant burden on the inmate”); Goff v. Burton, 7 F.3d 734, 738 (8th Cir. 1993) (“The district court appeared to recognize . . . that the “but for” standard “would be extraordinarily difficult for a prisoner to show”). This rings particularly true in the prison context, where inmates, often 8 proceeding without the assistance of counsel, are poorly positioned to collect and present evidence of a prison official’s subjective intent. Martin, as a prime example, was placed in administrative segregation for nearly four months after he engaged in protected conduct. J.A. 11. The transfer placed Martin in a separate, solitary-confinement like unit, apart from the general prison population. J.A. 75. This type of confinement undoubtedly curbed Martin’s ability to collect direct or circumstantial evidence of retaliatory animus in the days and weeks immediately following his grievance. But even when inmates are part of the general population, prison policies strictly circumscribe their schedule, placement, and ability to interact with those around him. See McKune v. Lile, 536 U.S. 24, 39 (2002) (“It is well settled that the decision where to house inmates is at the core of prison administrators’ expertise.”); Turner v. Safley, 482 U.S. 78, 85–87 (1987) (collecting cases) (warning that undue scrutiny of prison officials’ “day-to- day judgments” would “seriously hamper their ability to anticipate security problems and to adopt innovative solutions to the intractable problems of prison administration”); Freeman v. Tex. Dep’t of Crim. Just., 369 F.3d 854, 864 (5th Cir. 2004) (“Prison officials may legitimately punish inmates who verbally confront institutional authority without running afoul of the First Amendment.”). These restrictions often further legitimate penological interests, but they also create a stark asymmetry in inmates’ and prison officials’ access to information about the prison’s decision-making process. If, notwithstanding these hurdles, an inmate shows that protected conduct was a substantial or motivating factor in a prison guard’s decision to take adverse action, it is appropriate that 9 the burden of proving a permissible basis for taking that action then shifts to the person who took it. The circuits that reject this approach insist forsaking Mt. Healthy is necessary to both temper the flood of prison litigation and preserve the “wide latitude” afforded prison officials “in [controlling] and disciplining . . . inmates.” See, e.g., Woods, 70 F.3d at 1166; Goff, 7 F.3d at 737–38. We do not find either justification persuasive. First, it is not the role of the judiciary to erect substantive barriers to prisoners’ civil rights actions. See Crawford-El v. Britton, 523 U.S. 574, 596 (1998). That’s Congress’s job, and it has already acted. The Prison Litigation Reform Act requires inmates to pay filing fees; denies in forma pauperis status to most prisoners with three or more prior “strikes”; bars suits for mental or emotional injury unless there is a prior showing of physical injury; limits attorney’s fees; directs district courts to screen prisoners’ complaints before docketing; authorizes courts to dismiss “frivolous,” “malicious,” or meritless actions sua sponte; permits the revocation of good time credits for federal prisoners who file malicious or false claims; and encourages hearings by telecommunication or in prison facilities to make it unnecessary for inmate plaintiffs to leave prison for pretrial proceedings. 42 U.S.C. § 1997e. Prisoner suits that surmount these obstacles must still survive the pretrial hurdles all federal civil rights litigants face. See Kisela v. Hughes, 138 S. Ct. 1148, 1152 (2018) (“Qualified immunity attaches when an official’s conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known.”) (internal quotations omitted); Ashcroft v. Iqbal, 556 U.S. 662, 678–79 (2009) (“Rule 8 . . . does not unlock the doors of discovery for a plaintiff armed 10 with nothing more than conclusions.”); Liberty Lobby, Inc., 477 U.S. 242, 252 (1986) (“The mere existence of a scintilla of evidence in support of the plaintiff’s position will be insufficient [to survive a motion for summary judgment].”). We find neither reason nor authority to further winnow these claims. Second, Mt. Healthy does not encroach upon the deference granted to prison officials. See Rauser, 241 F.3d at 334 (explaining that the Mt. Healthy framework “do[es] not disregard the deferential standard articulated in Turner”). To be sure, courts must take care to “afford appropriate deference and flexibility to prison officials in the evaluation of proffered legitimate penological reasons for conduct alleged to be retaliatory.” Pratt v. Rowland, 65 F.3d 802, 807 (9th Cir. 1995) (citing Sandin v. Conner, 515 U.S. 472, 482 (1995)). But the obligation to view prison officials’ explanations charitably does not free officials from providing any explanation at all. Once a plaintiff establishes his protected conduct was a substantial or motivating factor in the defendant’s decision to take adverse action, the defendant is appropriately tasked with explaining why her decision was not animated by retaliatory motives. C. Martin, without proposing an alternative, argues that the Court should reject the same-decision test in cases where a prisoner is not charged with misconduct. Appellant Br. at 14. Martin correctly recognizes that prisoners’ retaliation claims often arise after an inmate has engaged in both protected conduct and misconduct. See, e.g., O’Bryant v. Finch, 637 F.3d 1207, 1219–20 (11th Cir. 2011) (inmate sanctioned after filing grievances and violating prison rules); Smith v. Mosley, 532 F.3d 1270, 1272–75 (11th Cir. 2008) 11 (same); Greene, 660 F.3d at 976–77 (inmate fired from prison library job after threatening to file a grievance against librarian and being accused of violating library rules); Mitchell v. Horn, 318 F.3d 523, 527–28 (3d Cir. 2003) (inmate sent to restrictive housing after filing grievances and being accused of a drug infraction); Graham, 89 F.3d at 77–78 (inmate sent to restrictive housing after filing grievances and being accused of organizing a “work slowdown”). In these cases, Mt. Healthy applies with intuitive ease. If a prison official shows she would have taken the same actions if the inmate engaged only in misconduct, courts logically infer legitimate reasons caused the adverse action, not retaliatory ones. See Greene, 660 F.3d 975, 979–80. The same-decision test is a clumsier endeavor in cases like this one, where the permissible and impermissible reasons for taking an adverse action are inseparably intertwined. Martin contends that Duffy placed him in segregation as punishment for his grievance. Duffy, on the other hand, argues Martin’s grievance simply alerted prison officials of the need to protect both Martin and the integrity of an impending investigation. Either way, it’s undisputed that Martin’s grievance gave rise to his segregation. But unlike situations involving misconduct, Duffy cannot show that she would have taken the same actions absent Martin’s grievance—not because safety and investigatory concerns cannot justify the use of administrative segregation, but because her concerns, even if legitimate, were a byproduct of Martin’s complaint. The same-decision test, as originally phrased, fails to conceive of a scenario where a defendant’s reasons for taking adverse action are both legitimate and wholly bound up in a plaintiff’s protected conduct. 12 Nevertheless, it is more prudent to adjust Mt. Healthy than to abandon it. We find persuasive the Second Circuit’s approach to this dilemma in an analogous context. See Greenwich Citizens Comm., Inc. v. Cntys. of Warren & Washington Indus. Dev. Co., 77 F.3d 26, 33 (2d Cir. 1996). There, local residents and community groups filed a lawsuit against a local government agency opposing a proposed construction project. Id. at 28. The agency filed several counterclaims in response. Id. at 29. The plaintiffs then filed a § 1983 suit in federal court, arguing that the agency filed their counterclaims in retaliation for the plaintiffs’ original suit. Id. At summary judgment, the district court found that (1) the plaintiffs engaged in constitutionally-protected conduct by filing their original lawsuit, (2) the plaintiffs established the agency’s counterclaims were motivated by or substantially caused by the plaintiffs’ decision to file a lawsuit, and (3) the agency failed to prove it would have filed its counterclaims absent plaintiffs’ original suit. Id. The question of whether the plaintiffs suffered a chilling effect such that the agency’s counterclaims amounted to an “adverse action” was left for the jury. Id. The district court rejected the agency’s argument that plaintiffs had to prove retaliatory intent to prevail. Id. On appeal, the Second Circuit held that this was in error. It reasoned that, in some contexts, rigidly applying Mt. Healthy improperly creates a “standard of strict liability based purely on cause and effect”—a standard the Supreme Court has rejected for First Amendment retaliation claims. 2 Id. at 30 (citing Younger v. Harris, 401 U.S. 37, 51 (1971) 2 We, like the court in Greenwich, emphasize this decision does not address “whether the state of mind of a governmental defendant is relevant to all First Amendment claims.” 77 F.3d at 31. 13 (“[T]he existence of a ‘chilling effect,’ even in the area of First Amendment rights, has never been considered a sufficient basis, in and of itself, for prohibiting state action.”)); see also Heffernan v. City of Paterson, 136 S. Ct. 1412, 1419 (2016) (“To win, the employee must prove an improper [] motive.”); Bell v. Sch. Bd. of City of Norfolk, 734 F.2d 155, 157 (4th Cir. 1984) (upholding district court’s Rule 12(b)(6) dismissal of retaliatory- counterclaim claim because defendant lacked ‘evil motive’ when filing its suit). After explaining why Mt. Healthy did not intend to “dispens[e] with the traditional state-of-mind requirement for retaliatory First Amendment claims,” Greenwich concluded: “[A]lthough the language in Mt. Healthy refers to the plaintiff’s conduct, the Court’s analysis, properly understood, attempts to weigh the impact of the defendant’s impermissible reason on the defendant’s decision to act.” 77 F.3d at 32 (emphasis in original). The causation inquiry does not solely consider “the causative effect of the plaintiff’s protected conduct”; it considers “the combined causative effect of the plaintiff’s conduct and the defendant’s impermissible reason.” Id. Greenwich then identified two types of First Amendment retaliation claims: (1) claims where the principal dispute is whether protected or unprotected conduct caused a defendant to take adverse action, and (2) claims where the principal dispute is whether a defendant took adverse action for retaliatory purposes. Id. at 33. Mt. Healthy exemplifies the first category of cases: dual-motive cases. There, the school board fired an employee after he engaged in both protected and unprotected conduct. The board did not dispute that its discharge decision was designed to punish. Rather, it argued that the employee’s protected conduct, albeit a factor in the decision, was not the “but for” cause of the decision. 14 Mt. Healthy, 429 U.S. at 285–87. In dual-motive cases like Mt. Healthy, “the defendant can properly be said to be acting on the basis of either ‘protected conduct’ or ‘impermissible reason’” because “a finding that the defendant acted because of the protected conduct is tantamount to a finding that the defendant acted with retaliatory intent.” Id. at 33 (emphasis in original). But there are other cases, like Greenwich and the one here, “where the distinction between ‘protected conduct’ and ‘impermissible reason’ becomes important.” Id. In this second category of cases—“unitary event” cases—the plaintiff’s protected conduct is a single event “that could prompt either a permissible or an impermissible reason on the part of the defendant to act.” Id. (citing Bell, 734 F.2d at 155). In Greenwich, the defendant’s adverse action—filing counterclaims—undisputedly flowed from the plaintiff’s protected conduct—filing a lawsuit. Id. at 32. So too here, Martin’s placement in segregation undisputedly flowed from his grievance. J.A. 170–71, 187–88. But as Greenwich explained: In these situations, “protected conduct” does not necessarily equal “impermissible reason,” and if the test of “but for” causation is phrased in terms of the impact of the “protected conduct,” then this phrase becomes an inadequate proxy for the proper inquiry into whether the defendant acted with retaliatory intent. In cases involving unitary events, claims of alleged retaliation for the exercise of a constitutional or statutory right require focusing precisely on whether the defendant acted for an impermissible reason, and not merely in response to the plaintiff’s conduct. Id. at 33. To wit, in “unitary event” retaliation claims, the same-decision test asks not whether the defendant would have reached the same decision absent the plaintiff’s protected 15 conduct, but whether the defendant would have reached the same decision absent a retaliatory motive. Id. Other courts have sub silencio adopted this reformulation when the facts so require. See, e.g., Texas v. Lesage, 528 U.S. 18, 21 (1999) (“The government can avoid liability by proving it would have made the same decision without the impermissible motive.”); Paterek v. Vill. of Armada, Mich., 801 F.3d 630, 646 (6th Cir. 2015) (“[T]he burden shifts to the defendant to put forth evidence showing it would have taken the adverse action absent any retaliatory motive.”); Babcock, 102 F.3d at 275 (“[T]he ultimate question is whether events would have transpired differently absent the retaliatory motive.”) (citing Mt. Healthy, 429 U.S. at 287). We pause to recognize a third category of First Amendment retaliation claims: claims where the parties equally dispute what conduct caused defendant’s adverse action and whether the defendant bore retaliatory animus. Take, for example, a case where a prison decides to transfer one of its inmates. The day before the inmate learns of his transfer, he files a lawsuit against the prison for an unrelated incident. When the inmate learns of his transfer, he files another lawsuit, alleging the prison is transferring him in retaliation for his original suit. Consider, too, the example of an inmate who files a grievance against a prison guard. The next day, the guard places the inmate in disciplinary segregation for an infraction the inmate did not commit. The inmate files a suit against the guard, alleging she manufactured disciplinary charges against him in retaliation for his original suit. In these examples, both parties argue that a single event led to the adverse action but disagree about what event prompted the action. As Spiegla v. Hull, 371 F.3d 928, 933–34, 943 (7th Cir. 2004) and Graham, 89 F.3d at 80–81 illustrate, the dual-motive 16 framework will often provide an effective way to ferret out causation and retaliatory animus in these types of cases. Ultimately, we find Mt. Healthy strikes the right balance in allocating the parties’ burdens and allowing courts flexibility in evaluating causation. The district court correctly concluded that Mt. Healthy’s burden-shifting framework governed Martin’s retaliation claim. IV. We nonetheless reverse the district court’s grant of summary judgment. In reviewing both parties’ motions, the magistrate judge determined that Martin established a prima facie case as a matter of law. Martin, 2018 WL 985016, at *3 (citing Martin, 858 F.3d at 249). Neither Martin nor Duffy challenge this finding on appeal. The magistrate judge then proceeded to the same-decision test, explaining that “even if a prisoner asserting a First Amendment retaliation claim establishes the three elements of such a claim, prison officials may still prevail by proving that they would have made the same decision absent the protected conduct . . . .” Id. (quoting Rauser, 241 F.3d at 334 (internal quotations omitted)). 3 To meet this burden, Duffy filed an affidavit, stating that she placed Martin in a holding cell because an investigator told her to. J.A. 170. She explained that Martin’s segregation was designed to “protect the integrity of the[ir] investigation” and ensure 3 As discussed above, the proper inquiry in “unitary event” cases like this is whether the defendant would have reached the same decision absent her retaliatory motives. That said, we do not reverse the decision below simply because the district court used the wrong words. Rather, the existence of material factual disputes made summary judgment improper. 17 Martin’s “safety and protection.” Id. at 171. The court determined that these reasons were both true and rationally related to legitimate penological interests. Martin, 2018 WL 985016, at *3 (citing Hewitt v. Helms, 459 U.S. 460, 473 (1983)). A. The district court erred in crediting Duffy’s reasons for segregating Martin. At the summary-judgment stage, a court must view all facts in the light most favorable to the non- moving party. Fed. R. Civ. P. 56(a). “Credibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge.” Liberty Lobby, 477 U.S. at 255. Whether Duffy would have placed Martin in segregation absent a retaliatory motive is a question of material fact. To be sure, Duffy filed an affidavit stating that a prison investigator instructed her to segregate Martin pending an investigation. J.A. 171. Duffy insisted Martin “was not placed in an administrative holding cell as punishment or retaliation for making [] sexual misconduct allegations against Sgt. Rogers” and that she informed Martin that the goal of his placement was “to protect the integrity of the investigation” and ensure “his [] safety and protection.” Id. But Duffy failed to identify any specific threats to either Martin’s safety or the integrity of an investigation. Moreover, Duffy’s affidavit is, at times, inconsistent. Most notably, she insists Martin’s segregation “was not her decision” but also states that she was the one to end it. See J.A. 171–72. This inconsistency further casts doubt on Duffy’s version of events. Martin also filed an affidavit. He explained that Duffy “questioned [him] relentlessly” before moving him to administrative segregation. J.A. 82. Duffy kept him 18 segregated for 110 days, preventing him from filing a “step 1 10-5 grievance” against Rogers. J.A. 84. Finally, Martin presented circumstantial evidence that this segregation was arbitrary, and that the impetus for his eventual release was not completion of an investigation but a “Request to Staff Member” form he submitted after two months of unexplained confinement. J.A. 83, 119. Indeed, it is unclear whether any investigation ever occurred. See J.A. 119. Viewing these conflicting accounts in the light most favorable to Martin, there is a genuine dispute of material fact regarding whether Duffy would have placed Martin in segregation for 110 days absent a retaliatory motive. Because a reasonable juror could find that Duffy placed Martin in segregation and kept him segregated for impermissible reasons, summary judgment was improper. B. Martin also argues that the district court erred in allowing Duffy to prevail under the same-decision test by pointing to a policy of placing all inmates who file grievances in administrative segregation. A prison policy that—either in text or in practice—segregates all complaining inmates, he contends, is unconstitutional. We have doubts about whether the type of blanket policy Martin describes could survive Turner, 482 U.S. at 89. See Shepard v. Quillen, 840 F.3d 686, 693–94 (9th Cir. 2016). But we reserve them because this question is not directly before the Court. Duffy did not argue in her cross-motion for summary judgment that she placed Martin in administrative segregation pursuant to a blanket policy of segregating complaining inmates. The district court did not rest its decision on a finding that Duffy was acting pursuant to prison policy. And there is no 19 evidence in the record that Perry CI retains this type of policy. To the contrary, in Martin’s original complaint, he alleged, “Other inmates in the general population statewide have attempted to informally resolve grievances of inappropriate an[d] unwanted touching ‘battery’ and were not placed on segregation as the plaintiff was.” We nonetheless take this opportunity to reemphasize the well-settled principle that “[a]n action motivated by retaliation for the exercise of a constitutionally protected right is actionable, even if the act, when taken for a different reason, might have been legitimate.” Woods, 60 F.3d at 1165; see also Maben v. Thelen, 887 F.3d 252, 262–63 (6th Cir. 2018); Allah v. Seiverling, 229 F.3d 220, 224–25 (3d Cir. 2000); Williams v. Meese, 926 F.2d 994, 998 (10th Cir. 1991). Courts view retaliation claims in the prison context with an eye toward avoiding “excessive judicial involvement in prison administration.” Pratt, 65 F.3d at 807 (citing Sandin, 515 U.S. at 482). But the broad deference afforded to corrections officers is not without limits. We cabin that deference when failing to do so would “unfairly tempt corrections officers to enrobe themselves and their colleagues in what would be an absolute shield against retaliation claims.” Maben, 887 F.3d at 263 (quoting Woods, 60 F.3d at 1165). C. In sum, the district court correctly invoked Mt. Healthy but erred in making credibility determinations at the summary-judgment stage. Because genuine disputes of material fact preclude summary judgment, the Court reverses and remands for further proceedings consistent with this opinion. REVERSED AND REMANDED 20