Opinion by Judge FISHER; Concurrence by Judge BERZON.

This appeal involves an alleged retaliatory discharge of an employee after she complained about co-workers and one of her supervisors and presents a question that this circuit has not yet answered: Can a final decision maker’s wholly independent, legitimate decision to terminate an employee insulate from liability a lower-level supervisor involved in the process who had a retaliatory motive to have the employee fired? We conclude that, on the record in this case, the answer must be yes, because the termination decision was not shown to be influenced by the subordinate’s retaliatory motives.

The plaintiff-appellee, Lea Lakeside-Scott (“Scott”), was fired from her position as an information systems specialist at Multnomah County’s Department of Community Justice (“DCJ”), ostensibly for her improper use of DCJ’s computers and email system. Scott then brought this lawsuit alleging that her termination was actually in retaliation for her engaging in speech protected under the First Amendment and by Oregon’s whistleblower protection statute. While she was employed at DCJ, Scott had complained about co*800workers’ violations of County policies, including by one of her supervisors — Jann Brown — whom she also accused of favoring gay and lesbian employees in hiring and promotion decisions. Brown played a role in the process that led to Scott’s termination, although the ultimate decision was made independently by Joanne Fuller, director of DCJ’s information systems department. Scott contends that Brown wanted to retaliate against Scott for her accusations against Brown, and thus unlawfully influenced Fuller’s decision to fire Scott.

Scott filed her retaliatory discharge claim against the County and Brown in federal district court. After a trial, a jury found in Scott’s favor, awarding her $650,000 in compensatory and punitive damages against Brown.1 The district court denied Brown’s motion for judgment as a matter of law (“JMOL”), and this appeal followed. We conclude there was insufficient evidence to support the verdict against Brown, given the evidence that it was Fuller’s independent decision to terminate Scott. We therefore reverse the district court’s denial of Brown’s JMOL and remand for entry of judgment in her favor.

BACKGROUND2

Scott began her employment in DCJ’s information services unit in August 1997. During the relevant time period, her direct supervisor was Monna Hogue. Hogue reported to Dan Gorton, who reported to Brown, who, in turn, reported to the department’s director, Ms. Fuller.

Scott frequently complained to Gorton and Hogue about her perceived problems in the office. Her grievances included personality conflicts with other DCJ employees, promotions she did not receive and alleged misuse of the County computer system by co-workers and managers. In October 2001, Scott filed a formal complaint with the Oregon Bureau of Labor and Industries (“BOLI complaint”) alleging, among other things, that Brown gave preferential treatment to gays and lesbians in hiring and promotions. Brown learned about the BOLI complaint shortly thereafter; she was shocked by its allegations of favoritism, which she took personally.

In November 2001, Fuller ordered Brown to search the email of an employee, David Landis, as part of an investigation of another DCJ employee who had allegedly sent racially discriminatory emails at work. Lacking the technical ability to do the search herself, Brown directed Tami Williams to do it. Williams sent the emails and attachments she recovered during her search to the human resources department. Attached to one of these emails was a journal, written by Scott and sent by her to Landis, that contained discriminatory comments and excerpts of other employees’ work documents. It is unclear whether Williams knew about the journal when she sent the emails to human resources. After human resources personnel discovered the journal, either they or Fuller instructed Brown to look for additional material from Scott.3

*801In the meantime, either someone in human resources or perhaps Williams informed Brown about Scott’s journal having been found among Landis’ email documents. When Brown read the journal, which included excerpts of personal emails and documents from co-workers and supervisors as well as several apparently derogatory remarks about homosexuals, she immediately showed it to Fuller. At an ensuing meeting attended by Fuller, Brown and a County counsel, Fuller decided to place Scott on administrative leave (standard practice during an employee investigation) and directed Brown to write a letter to Scott informing her of this decision. With the assistance of the human resources department, Brown prepared and signed a standardized letter advising Scott she was being placed on administrative leave. After consulting with the human resources department and Fuller about how to present the letter to Scott, Brown had Scott come to her office the next morning along with two other managers, Gorton and Rich Scott, in case there “was any trouble.” After the meeting, Brown instructed the two managers to unplug Scott’s computer, which they placed in Brown’s office, where it was “made operational.”

Once Scott was placed on administrative leave, Fuller directed John Turner, an investigator on her staff, to conduct an internal inquiry into Scott’s possible violations of County work rules or policies. Brown was not involved in framing the charges to be investigated, outlining the direction of the investigation or providing a list of witnesses. Rather, Turner met with Collette Umbras, the human relations department manager, to outline which official work rules were implicated by Scott’s supposed misconduct. The charges ultimately included misusing County property, conducting personal business on County time, inappropriately accessing the emails and documents of other employees and engaging in prohibited workplace harassment and prejudicial acts. Fuller sent Scott a letter to notify her of these charges as Turner began his investigation.

Over the course of his investigation, Turner interviewed 22 witnesses, including Brown and Scott. He also reviewed Scott’s journal and several of her emails. Brown’s role in the investigation was limited to answering Turner’s questions; she did not provide him with any written materials. For her own part, Scott admitted to Turner that she had engaged in the conduct that had led to the charges against her, and she conceded that she was fully aware of the policies, procedures and rules governing the use of County property— particularly those prohibiting harassment and discrimination and accessing databases for personal or non-business related reasons. She said, however, that she did not know her behavior violated any of these rules and claimed that Hogue knew about and had authorized many of her actions.

At the conclusion of his investigation, Turner produced a report to Fuller detailing his findings and recommending that all of the charges against Scott be sustained. Fuller sent a letter to Scott describing the report’s findings and, after meeting with Scott to provide her with an opportunity to explain her actions, decided to terminate her employment. Fuller testified that although Scott’s journal was the reason she decided to initiate the investigation, she based her decision to fire Scott on all of the evidence that Turner procured during *802his investigation. As detailed in her termination letter to Scott, Fuller “removed” certain charges but nevertheless concluded that the remainder of the sustained charges-misusing County property, conducting personal business on County time, accessing other employees’ emails and documents and engaging in workplace harassment and prejudicial acts-warranted Scott’s termination. The magnitude of Scott’s misconduct was on a scale that was completely different from what Fuller had seen in other employees and destroyed her ability to trust Scott to uphold DCJ’s policies in the future. Throughout her trial testimony, Fuller reiterated that Brown played no role in her decision to fire Scott. Scott did not produce any evidence to the contrary.

Scott brought a retaliatory discharge lawsuit in federal district court against the County and Brown, alleging she was wrongfully terminated because she had filed the BOLI complaint and openly criticized both DCJ and Brown, and claiming these were protected activities under the First Amendment and 42 U.S.C. § 1983 as well as under Oregon’s Whistle-blower Act. After Brown moved unsuccessfully for summary judgment, a jury found in Scott’s favor and awarded her economic damages of $140,000, noneconomic damages of $10,000 and punitive damages of $500,000. The district court denied Brown’s motion for JMOL and this timely appeal followed.4

STANDARD OF REVIEW

‘We review the denial of a motion for a judgment as a matter of law de novo.” Ostad v. Oregon Health Sci. Univ., 327 F.3d 876, 881 (9th Cir. 2003). We view the evidence in the light most favorable to the party in whose favor the jury returned a verdict and draw all reasonable inferences in her favor. See id; Gilbrook v. City of Westminster, 177 F.3d 839, 847-48 (9th Cir. 1999). “Judgment as a matter of law is proper when the evidence permits only one reasonable conclusion and the conclusion is contrary to that reached by the jury.” Ostad, 327 F.3d at 881. Nevertheless, a reasonable inference “cannot be supported by only threadbare conclusory statements instead of significant probative evidence.” Barnes v. Arden Mayfair, Inc., 759 F.2d 676, 680-81 (9th Cir. 1985) (internal quotation marks omitted); see also Genthe v. Lincoln, 383 F.3d 713, 716 (8th Cir. 2004) (noting within the context of a motion for JMOL that an inference is reasonable “when it may be drawn from the evidence without resort to speculation” (internal quotation marks omitted)); Willis v. Marion County Auditor’s Office, 118 F.3d 542, 545 (7th Cir. 1997) (noting within the context of a motion for JMOL that a “mere scintilla is not enough” to sustain a verdict for the prevailing party).5 Conse*803quently, JMOL is appropriate when the jury could have relied only on speculation to reach its verdict.

I.

“To state a First Amendment claim against a public employer, an employee must show: (1) the employee engaged in constitutionally protected speech; (2) the employer took ‘adverse employment action’ against the employee; and (3) the employee’s speech was a ‘substantial or motivating factor for the adverse action.’ ” Marable v. Nitchman, 511 F.3d 924, 929 (9th Cir. 2007) (quoting Coszalter v. City of Salem, 320 F.3d 968, 973 (9th Cir. 2003), citing Bd. of County Comm’rs, Wabaunsee County, Kan. v. Umbehr, 518 U.S. 668, 675, 116 S.Ct. 2342, 135 L.Ed.2d 843 (1996)); see also Garcetti v. Ceballos, 547 U.S. 410, 126 S.Ct. 1951, 164 L.Ed.2d 689 (2006); Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 287, 97 5.Ct. 568, 50 L.Ed.2d 471 (1977); Settlegoode v. Portland Pub. Sch., 371 F.3d 503, 510 (9th Cir. 2004). “If the plaintiff makes those showings, then the burden shifts to the defendant to show ‘by a preponderance of the evidence that it would have reached the same decision ... even in the absence of the [plaintiffs] protected conduct.’ ” Gilbrook, 177 F.3d at 854 (quoting Mt. Healthy, 429 U.S. at 287, 97 S.Ct. 568).

Brown expressly assumed in her appellate briefs that Scott’s BOLI complaint was protected speech.6 Scott’s termination plainly qualifies as an adverse employment action. See Umbehr, 518 U.S. at 675, 116 S.Ct. 2342; Ray v. Henderson, 217 F.3d 1234, 1237 (9th Cir. 2000).7 Thus the first two prongs of Mt. Healthy’s pri-ma facie test are not at issue. Instead, Brown argues that Scott’s protected conduct was not a “substantial or motivating factor” in her termination, which would have occurred even in the absence of this conduct, because even assuming that the jury could have reasonably found that Brown harbored animosity toward Scott, Fuller made the final and independent decision to terminate Scott.

We have assessed the liability of a subordinate supervisor who was not the final decision maker under Mt. Healthy’s “substantial or motivating” standard as well as its mixed motive approach. Compare Ostad, 327 F.3d at 882-83 (analyzing facts *804under “substantial or motivating” standard), with Gilbrook, 177 F.3d at 853-56 (analyzing facts under mixed motive standard). Under either rubric, however, the focus is on the question of causation. See Ostad, 327 F.3d at 882-83 (relying on Gil-brook in its analysis); Gilbrook, 177 F.3d at 855 (citing opinions addressing prima facie cases of discrimination in its analysis). Causation is the dispositive issue here as well. Consequently, the critical questions are: (1) whether a final decision maker’s independent investigation and termination decision, responding to a biased subordinate’s initial report of misconduct, can negate any causal link between the subordinate’s retaliatory motive and an employee’s termination; and, if so, (2) whether the record here compels the conclusion that Fuller conducted an independent investigation and made a wholly independent decision to terminate Scott such that Brown cannot be held liable for causing her to be fired.

In Gilbrook, we established that a “subordinate cannot use the non-retaliatory motive of a superior as a shield against liability if that superior never would have considered a dismissal but for the subordinate’s retaliatory conduct.” 177 F.3d at 855. We expressly declined, however, to decide “what the result should be, as a matter of law, if the facts showed that the final decision-maker made a wholly independent, legitimate decision to discharge the plaintiff, uninfluenced by the retaliatory motives of a subordinate.” Id.; see also Ostad, 327 F.3d at 883. The record before us requires us to answer this heretofore open question.

II.

Most of this circuit’s retaliatory motivation jurisprudence has arisen from cases in which the issue has been whether the final decision maker was liable because she was retaliating against the employee or her decision was tainted by the retaliatory motivation of a subordinate. See, e.g., Poland v. Chertoff, 494 F.3d 1174 (9th Cir. 2007). In the latter situation, the subordinate’s unlawful motivation has been imputed upstream to the final decision maker. In the present case it is the supposedly biased subordinate, Brown, who is herself being charged with liability. Nonetheless, these “upstream imputation” cases are relevant to our analysis. See Gilbrook, 177 F.3d at 855 (citing imputation cases in its discussion of subordinate liability).8 If a final decision maker is not liable when her decision to terminate a plaintiff is sufficiently independent from a subordinate’s unlawful motive, the circumstance proving that independence may also show that the subordinate did not cause the plaintiffs termination.

Gilbrook, in specifically addressing a subordinate supervisor’s liability, did observe that as a general matter the nature of § 1983 liability is such that “the ‘requisite causal connection can be established not only by some kind of direct personal participation in the [termination], but also by setting in motion a series of acts by others which the actor knows or reasonably should know would cause others to inflict the constitutional injury.’ ” 177 F.3d at 854 (citing and quoting John*805son v. Duffy, 588 F.2d 740, 743-44 (9th Cir. 1978) (holding sheriff could be liable for his omission to perform duties imposed by state law that deprived the plaintiff of his property without due process)). This general principle, as applied within the context of the employment setting, cannot be taken so literally as to convert into a constitutional tort a subordinate supervisor’s mere participation in, while performing her normal supervisory responsibilities, the initiation of a disciplinary process that results in an employee’s otherwise appropriate and lawful termination. As we recognized in Gilbrook itself, a superi- or’s nonretaliatory employment decision “does not automatically immunize a subordinate against liability for her retaliatory acts”; subordinate liability instead turns on the “intensely factual” determination of whether the superior never would have made this decision “but for the subordinate’s retaliatory conduct.” 177 F.3d at 854-55 (emphasis added).

Here, it is not clear that Scott relies on Brown’s role in bringing the journal to Fuller’s attention or her participation in the administrative leave decision as alone sufficient to support the jury’s verdict finding Brown liable for Scott’s termination.9 Even if we consider those events, however, the evidence negates any inference that Fuller would not have taken any action against Scott but for Brown’s retaliatory motivations. The journal surfaced in a workplace investigation of possible employee misconduct, and several DCJ employees became aware of its existence contemporaneously with Brown. Brown was engaged in activities typical and appropriate for her position when she became aware of the journal. The journal itself was accidentally discovered during an investigation of another employee’s conduct, and Scott did not allege that Brown targeted her for investigation or selectively reported her misconduct.10 Brown was at most a part of a process that included several other employees who were focused on disciplining violations of workplace rules and policies. Given the numerous potential rules violations revealed in the journal and the actions taken by the human resources department, it is unreasonable to conclude that Fuller — who had already initiated the inquiry into another employee’s misuse of emails-would not have been informed of or reacted to the journal but for Brown’s animus against Scott.

Given the evidence that Fuller made an independent, principled decision for her own reasons to investigate and eventually terminate Scott, there was by definition no “constitutional injury.” Id. at 854. Scott was terminated for violating County and DCJ rules through actions she admitted she had committed. As we emphasized in Gilbrook, “[w]e do not hold that a final decision-maker who lacks any improper motive never can absolve a subordinate of liability for his or her retaliatory acts, any more than we hold that such a decision-*806maker always can absolve the subordinate.” 177 F.3d at 855 (emphasis in original). On the facts of this case, where the evidence shows “that the final decision-maker made a wholly independent, legitimate decision to discharge the plaintiff, uninfluenced by the retaliatory motives of’ Brown, we hold that the neutrality of the decisionmaking process eliminated any “causal” link to Brown’s bias. Id.

A recent “upstream” case in which a biased subordinate’s involvement in an adverse employment action rose to the level of tainting the final decision to terminate is illustrative. See Poland, 494 F.3d 1174. Poland held that the Custom Service’s administrative investigation of the supervisor’s retaliatory charges did not “shield[ ]” the Customs Service from imputed liability because “the allegedly independent adverse employment decision was not actually independent” due to the subordinate supervisor’s significant involvement in the decisionmaking process. Id. at 1182. We held that while an initiation of an investigation would not “on its own” be enough to impute a subordinate’s animus, the biased subordinate “had a pervasive influence on the administrative inquiry that led to the adverse employment action.” Id. at 1183. The supervisor specifically requested the investigation, sent a lengthy memo and supporting documentation outlining numerous incidents of the plaintiffs alleged malfeasance and provided the list of 21 witnesses who were contacted during the investigation. See id.; see also Dominguez-Curry v. Nevada Transp. Dep’t, 424 F.3d 1027, 1032-33, 1039-40 (9th Cir. 2005) (imputing animus in a Title VII failure-to-promote claim when the subordinate was one of two employees who interviewed and ranked the petitioner); Perez v. Curcio, 841 F.2d 255, 258 (9th Cir. 1988) (imputing animus in an age discrimination action when the final decision maker relied on reports written by the subordinate both in initiating the investigation and ultimately in demoting the plaintiff).11 We have likewise affirmed the liability of subordinates when they wielded a similarly significant degree of influence over the final decision maker’s adverse employment decision. For example, in Gilbroolc, we held that a final, unbiased decision maker did not eliminate the liability of two subordinates where the subordinates had “participated directly” in a multitiered termination process. Gilbrook, 177 F.3d at 853. There, one subordinate was the “driving force” behind the investigation of the terminated *807employees and made the initial decision to terminate them. The second subordinate then conducted hearings on the employees’ appeals and affirmed the terminations. We held that a third and final decision maker with the power to ratify, reject or modify the termination decisions did not cut off the liability of the biased subordinates because there was evidence that the employees would not have been disciplined but for the subordinates’ actions in bringing the charges in the first place. See id. at 850-51, 853; see also Ostad, 327 F.3d at 883 (holding subordinate liable when he provided the primary source of information on which the final decision maker “substantially relied ... to reach the decision to terminate”).

The facts before us here show a workplace in which the initial report of possible employee misconduct came from a presumably biased supervisor, but whose subsequent involvement in the disciplinary process was so minimal as to negate any inference that the investigation and final termination decision were made other than independently and without bias. We must not “place an employee in a worse ‘position as a result of the exercise of constitutionally protected conduct than he would have occupied had he done nothing.’ ” Gilbrook, 177 F.3d at 855 (citing Mt. Healthy, 429 U.S. at 285, 97 S.Ct. 568). But concomitantly the Supreme Court has admonished that we must not “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. ... [T]hat [employee] ought not to be able, by engaging in such conduct, to prevent his employer from assessing his performance record.... ” Mt. Healthy, 429 U.S. at 285-86, 97 S.Ct. 568.

III.

With the foregoing principles and spectrum of cases as our guide, and recognizing “that the ultimate question of the subordinate’s liability ‘is an intensely factual one, the results of which will vary depending on the circumstances,”’ Ostad, 327 F.3d at 883 (quoting Gilbrook, 177 F.3d at 855), we hold that Brown cannot be found liable based on her limited involvement in the chain of events that ultimately led to Scott’s termination. The jury could have reasonably found that Brown was involved in initiating the investigation of Scott after the discovery of her journal and in the decision to place her on administrative leave. But neither Brown’s role in the events leading up to the investigation nor the evidence of her participation in that inquiry rises to the level of involvement in or influence on Fuller’s termination decision that would allow the jury reasonably to find that Brown’s animus was a “ ‘substantial’ or ‘motivating’ factor” in Fuller’s decision to fire Scott. Gilbrook, 177 F.3d at 853-54 (quoting Mt. Healthy, 429 U.S. at 287, 97 S.Ct. 568). Instead, the record of Fuller’s independent actions and judgments compels the conclusion that she was not influenced by any retaliatory motive on Brown’s part.

As to the discovery of Scott’s apparently improper journal that triggered the inquiry into her conduct, it is relevant that Brown became aware of the document only incidentally through an investigation of another employee initiated by Fuller and the human resources department. Moreover, although Brown was present at the meeting where, according to standard practice, it was decided that Scott would be placed on administrative leave pending the investigation, that decision did not make Scott’s termination a foregone conclusion. Instead, senior management, aware of Scott’s apparent misconduct, pursued an inquiry into the alleged violations. There is no evidence that but for Brown, Fuller would *808have ignored the journal and let the matter drop. That other DCJ employees, including the human resources manager, became aware of and reacted to the journal suggests the very opposite. Accordingly, the relevant question remains whether Brown improperly influenced the subsequent investigation or the decision to terminate Scott itself such that she can be held liable for Fuller’s decision. On this record, the only reasonable finding is that she did not.12

There is no evidence that Brown outlined possible reasons for Scott’s discharge or recommended her termination, either at the initial meeting or thereafter. Cf. Gilbrook, 177 F.3d at 850, 853-55; Ostad, 327 F.3d at 880, 883. To the contrary, Umbras outlined potential violations Turner was to investigate, and it was Turner who recommended at the conclusion of his investigation which charges should be sustained. Scott asks us to hold that a reasonable juror could have inferred that Brown “was the investigation” because (1) Scott’s computer was placed in Brown’s office after Scott was placed on administrative leave and (2) Williams completed the follow up research ordered by the human resources department “to see if ... there were any other e-mails from [Scott] or jobs or anything else that pertained to her” while Turner was still conducting his investigation. Scott appears to argue that these circumstances alone permitted the jury to find that Brown was an influential player in Turner’s investigation who searched for and provided documentary evidence that actually led to Scott’s termination. But such an inference would be pure speculation, as Scott presented no evidence that Brown discovered any information from Scott’s computer or from Williams, much less that she provided any such information to Turner or Fuller. Significantly, Scott deposed Williams but neither introduced her deposition testimony nor called her as a witness to tell the jury what, if anything, she had turned over to Brown. Scott’s suppositions are “only threadbare conclusory statements” that cannot support a reasonable inference that Brown influenced the decision to terminate Scott. Cf. Barnes, 759 F.2d at 680-81 (internal quotation marks omitted).

Instead, the record shows that Brown played a very limited role in the investigation. She did not provide a witness list or any documentary evidence during Turner’s investigation, and there is no evidence that Brown told Turner anything inappropriate during their interview. Cf. Poland, 494 F.3d at 1183. Nor did the investigation “substantially rel[y]” or “depend[] heavily” on Brown, who was one of 22 witnesses Turner interviewed. Cf. Ostad, 327 F.3d at 880, 883 (holding subordinate liable when the decision maker “depended heavily” on his testimony, which “took up 217 pages of the hearing’s 403-page transcript”); Stacks v. Sw. Bell Yellow Pages, Inc., 27 F.3d 1316, 1323 (8th Cir. 1994) (noting a final decision maker cannot escape liability when the facts on which he “reifies] have been filtered” by a subordinate with illegitimate motives). Brown’s minimal participation was too limited and constrained to have tainted the investigative process, particularly when it was part *809of Brown’s job to cooperate with Turner’s inquiry.

Fuller’s substantial role in the process that resulted in Scott’s termination is as important as Brown’s minimal participation. Fuller authorized the thorough investigation of the charges against Scott. Once she received Turner’s report, she critically examined its contents, meeting with Scott to allow her to present any mitigating evidence and ultimately rejecting two of Turner’s recommendations. Scott herself admitted, “I actually thought that she was listening, and she said she would consider everything.” As outlined in the termination letter composed and signed by Fuller, in the final analysis she fired Scott for violations of formal work rules, executive orders and contractual agreements. Scott admitted she was fully aware of these policies and rules, and both Fuller and Umbras emphasized that they had never before seen violations of this magnitude. Under these circumstances, no reasonable juror could have concluded that the investigative process was a “sham or conduit” for Brown’s animosity or that Fuller was “duped” into terminating Scott. Willis, 118 F.3d at 547-48. To the contrary, the record shows that Fuller’s decision was based on her own analysis that “was not jaded by anyone else’s subjective and possibly [illegitimate] evaluation.” Id. at 547. Accordingly, we hold that, as a matter of law, Fuller’s “wholly independent, legitimate decision to discharge [Scott], uninfluenced by the retaliatory motives of a subordinate” prohibited the jury from finding Brown liable for Scott’s termination. Gilbrook, 177 F.3d at 855.

Finally, practical considerations also animate our conclusion. As discussed above, all aspects of Brown’s limited involvement in this case were at the express direction or under the authority of her superiors in the management structure. To say that the County’s investigative process was fatally and irrevocably tainted by Brown’s overall limited involvement would stymie legitimate corporate management and discipline, which must necessarily involve and rely upon supervisory staff. Cf. City of San Diego v. Roe, 543 U.S. 77, 125 S.Ct. 521, 160 L.Ed.2d 410 (2004) (noting the “common-sense realization that government offices could not function if every employment decision became a constitutional matter”); Brooks v. City of San Mateo, 229 F.3d 917, 928 (9th Cir. 2000) (limiting the type of behavior that constitutes an adverse employment action based on the concern “that employers will be paralyzed into inaction once an employee has [engaged in protected conduct], making such [conduct] tantamount to a ‘get out of jail free’ card for employees engaged in job misconduct”). Absent evidence that a supervisor abuses her participation to advance a retaliatory agenda, we decline to endorse such a result in this case.

CONCLUSION

On the evidence presented to the jury, we hold as a matter of law that Fuller’s wholly independent decisionmaking negated any causal connection between Brown’s retaliatory bias and Scott’s termination. Brown therefore is not liable for any damages Scott suffered as a result of her discharge, and the district court incorrectly denied Brown’s motion for a JMOL. We reverse the court’s judgment and remand for entry of judgment for Brown.

REVERSED and REMANDED.

concurring in the judgment:

Although I concur in the result of this case, I do not agree with the majority’s conclusion that there was insufficient evidence from which a reasonable jury could have found that Brown’s retaliatory ani*810mus was a “substantial or motivating factor,” see Gilbrook v. City of Westminster, 177 F.3d 839, 853 (9th Cir. 1999), in Fuller’s decision to terminate Scott. As the majority intimates, a supervisor who, with retaliatory intent, selectively reports an employee to a superior is liable for the employee’s termination even if the supervisor does not influence the resulting investigation or termination decision. See Maj. Op. at 804-05. In this case, the majority concludes that Brown discovered Scott’s journal and reported it to Fuller while “engaged in activities typical and appropriate for her position,” and then states that Scott does not allege that Brown “targeted her for investigation or selectively reported her misconduct.” See id. But Scott obviously does so allege, she has maintained from the outset — and the jury must have found — that Brown reported her to Fuller for a retaliatory reason. The difference between targeting someone for investigation and retaliating against them by reporting them, knowing they would thereupon be investigated, entirely escapes me. So, the majority can only be holding that any retaliatory animus that Brown harbored against Scott did not motivate her decision to report Scott to Fuller, and that a reasonable jury could not have concluded otherwise.

I disagree. I would hold that Scott met her burden of establishing that Brown’s retaliatory animus was a “substantial or motivating factor” in Fuller’s decision to terminate Scott. See Gilbrook, 177 F.3d at 853. The majority’s implicit conclusion to the contrary — that Brown was not motivated by retaliatory reasons but was just doing her job — disregards our mandate to view “the evidence in the light most favorable to the party in whose favor the jury returned a verdict and draw all reasonable inferences in her favor,” and fails to accord the required deference to a reasonable jury decision. Maj. Op. at 802 (citing Os-tad v. Oregon Health Sci. Univ., 327 F.3d 876, 881 (9th Cir. 2003)); see also id. (“Judgment as a matter of law is proper when the evidence permits only one reasonable conclusion and the conclusion is contrary to that reached by the jury.”).

Alternatively, the majority can possibly be read as holding that a supervisor, even if motivated by retaliatory animus, may not be held liable for instigating an investigation likely to result in an adverse action, as long as the superior who makes the final decision does so independently. If so, the holding squarely conflicts with Gil-brook, which held to the contrary.

At the same time, the facts of this case indicate that Fuller’s investigation would have occurred even without Brown’s instigation. I would therefore hold that Brown also met her burden, under part two of the two-part Gilbrook/Mt. Healthy burden-shifting framework, of showing that Scott’s discharge would have occurred “even in the absence of the protected conduct.” See Gilbrook, 177 F.3d at 854; Maj. Op. at 803, 807-08. My decision to reverse the judgment, in other words, rests on Brown’s successful affirmative defense, not on Scott’s failure to establish her initial case.

A. The “Substantial or Motivating Factor” Requirement

Under the two-part burden-shifting framework outlined by the Supreme Court in Mt. Healthy City School District Board of Education v. Doyle, 429 U.S. 274, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977), and applied by this Court in Gilbrook, the plaintiff bears the initial burden of demonstrating that her protected conduct was a “ ‘substantial’ or ‘motivating’ factor in the defendant’s employment decision.” Gilbrook, 177 F.3d at 853-54 (quoting Mt. Healthy, 429 U.S. at 287, 97 S.Ct. 568). *811Once the plaintiff makes this showing, the burden shifts to the defendant to show “ ‘by a preponderance of the evidence that it would have reached the same decision ... even in the absence of the plaintiffs protected conduct.’ ” Id. (quoting Mt. Healthy, 429 U.S. at 287, 97 S.Ct. 568).

In this case, as in Gilbrook, the defendant is the supervisor who initially reported the plaintiffs misconduct, not the employer or the individual who made the ultimate decision to terminate. See Gilbrook, 177 F.3d at 853-54. For this reason, vicarious liability eases such as Poland v. Chertoff, 494 F.3d 1174 (9th Cir. 2007), are, contrary to the majority’s reliance on them, not analogous. Those cases are concerned with identifying situations in which it would be equitable to hold an otherwise faultless superior liable for the acts of an admittedly culpable subordinate. See SpeiseR et. al, The AMERICAN Law of Torts § 4:1 (“[Vicarious liability] is an interesting form of social policy adjustment under which, even though a person ... is himself personally without fault, legal public policy renders him nevertheless liable.”). As a result, the focus of vicarious liability cases is properly on the actions of the superior, and therefore on whether those actions were tainted by the subordinate’s animus.

In addition, vicarious liability cases implicate very different policy considerations than subordinate liability cases, particularly with respect to the protection of an employee’s First Amendment rights. In the vicarious liability context, the concern is that, in the absence of a rule imposing vicarious liability, an employer could evade liability by isolating the final deeision-maker and, in essence, willfully ignoring the bias of a subordinate. Thus, the inquiry in a vicarious liability case is into whether the superior’s investigation and termination decision were truly independent.

By contrast, the concern in a subordinate liability case is that too lenient a standard will permit supervisors deliberately to target individuals who have engaged in protected activity, knowing that their actions will lead superiors lacking retaliatory animus to harm the individual. The issue is not, in other words, whether the employer is acting with willful ignorance (which an “independent investigation” requirement is good at measuring), but rather whether the subordinate supervisor is deliberately placing the subordinate in the way of harm (which the “independent investigation” requirement only partly measures). Accordingly, as will appear, the inquiry in a subordinate liability case is broader, requiring not only a determination as to whether the investigation and termination decision were influenced by the biased subordinate, but also whether the subordinate initiated those proceedings in retaliation for the employee’s protected activity.

The focus in a case such as this one, then, must be on the supervisor’s conduct — specifically on whether that conduct (a) was a “substantial or motivating factor” in the employee’s termination, and (b) was motivated by retaliatory animus.

I agree with the majority that, on this record, Brown did not in any substantial way influence Fuller’s investigation once it began, or Fuller’s ultimate decision to terminate Scott.1 See Maj. Op. 807-08. *812Brown’s conduct after she reported the journal to Fuller was therefore not a cause of Scott’s termination, regardless of its motivation.

But Brown’s lack of participation in Fuller’s investigation and decision to terminate does not end the causation inquiry.2 Rather, we must also consider whether Brown, while acting with retaliatory animus, “set[] in motion a series of acts by others which [she knew] or reasonably shouldfhave known] would cause others to inflict the constitutional injury.” See Gilbrook, 177 F.3d at 854 (noting that the “requisite causal connection can be established not only by some kind of direct personal participation in the deprivation, but also by setting in motion a series of acts by others which the actor knows or reasonably should know would cause others to inflict the constitutional injury”) (quoting Johnson v. Duffy, 588 F.2d 740, 743-44 (9th Cir. 1978)); see also Restatement (seoond) of ToRts § 433 (1965) (“The following considerations are ... important in determining whether the actor’s conduct is a substantial factor in bringing about harm to another: ... (b) whether the actor’s conduct has created a force or series of forces which are in continuous and active operation up to the time of the harm.”).3

It is undisputed that Fuller launched the investigation that led to Scott’s termination immediately after Brown presented her with Scott’s journal. Because Scott’s journal contained significant evidence of misconduct, Brown either knew or should have known that presenting the journal to Fuller would “set[] in motion a series of acts” by Fuller that “would cause [Fuller] to inflict the constitutional injury,” i.e., to terminate Scott. See Gilbrook, 177 F.3d at 854. As a result, Brown’s act of reporting Scott to Fuller was a “cause” of Scott’s termination, even though she did not “directly] ... participate]” in the ultimate decision to terminate. See Gilbrook, 177 F.3d at 854; see also Restatement (Second) of Torts §§ 433, 441, 442A.

*813The majority appears to suggest, however, that Brown’s decision to report the journal to Fuller cannot support Scott’s claim, because Brown discovered the journal and brought it to Fuller’s attention while she was “engaged in activities typical and appropriate for her position.” Maj. Op at 805. As a result, the majority suggests, any retaliatory animus Brown harbored against Scott could not be, as a matter of law, a “substantial or motivating factor” in Scott’s discharge. Maj. Op. at 807-08.

Although the evidence could support the majority’s interpretation, it does not foreclose the opposite conclusion. See Ostad, 327 F.3d at 881 (“Judgment as a matter of law is proper when the evidence permits only one reasonable conclusion and the conclusion is contrary to that reached by the jury.”). “The causal link between a protected activity and the alleged retaliatory action ‘can be inferred from timing alone’ when there is a close proximity between the two.” See Thomas v. City of Beaverton, 379 F.3d 802, 812 (9th Cir. 2004) (quoting Villiarimo v. Aloha Island Air, Inc., 281 F.3d 1054, 1065 (9th Cir. 2002)). In this case, Brown reported Scott’s journal to Fuller approximately one month after Scott filed her BOLI complaint. This close temporal proximity was probably sufficient evidence on its own to support the jury’s conclusion that Brown was motivated by retaliatory animus when she reported Scott. See Thomas, 379 F.3d at 812 (holding that sufficient evidence of causation existed where adverse employment action occurred seven weeks after protected activity); see also Yartzoff v. Thomas, 809 F.2d 1371, 1376 (9th Cir. 1987) (three month time lapse between protected conduct and alleged retaliatory act provided sufficient evidence of causal connection). In addition, Brown admitted at trial that Scott’s BOLI complaint surprised and upset her, and that she took Scott’s allegations personally. Also, Brown was asked to look for other material from Scott, but, before she finished doing so, she reported Scott to Fuller based on the journal alone. These combined facts provided sufficient evidence — albeit, I concede, not with much room to spare — from which a reasonable jury could have concluded that Brown’s decision to report Scott and thereby set in motion the chain of events that led to Scott’s termination was motivated by retaliatory animus, not just routine practice.

Moreover, the majority is wrong if it is suggesting that the fact that Brown’s action in bringing the journal to Fuller’s attention occurred while “Brown was engaged in activities typical and appropriate for her position,” Maj. Op. at 805, precludes the jury’s verdict that she nonetheless acted with a retaliatory motive. Supervisors have the power to retaliate precisely because their responsibilities include the authority to take steps that will foreseeably harm their subordinates’ employment circumstances. If we absolve supervisors from liability for First Amendment violations whenever they take actions while engaged into their normal responsibilities, then they will almost never be liable, even if the motive for an action challenged was retaliatory.

Here, the jury necessarily found that at least one substantial reason Brown brought the journal to Fuller’s attention was that she was angry about the BOLI charge, and, as I have shown, that factual conclusion, although quite debatable, is supported by adequate evidence. That bringing work infractions to Fuller’s attention was part of Brown’s job does not negate the finding as a matter of law, any more than a supervisor with the responsibility to hire and fire for rule infractions would be absolved of liability as a matter of law if he fires a subordinate who en*814gages in speech in part because of that speech.

Because a reasonable jury could have found that Brown acted with a retaliatory motive when she brought Scott’s journal to Fuller’s attention, and because that act was a substantial and foreseeable link in the causal chain that led to Scott’s termination, I cannot agree with the majority that Scott failed, as a matter of law, to show that her protected activity was “a substantial or motivating factor” in her termination.

It is possible that rather than concluding that Brown did not act with a retaliatory intent when she reported Scott to Fuller, the majority is instead holding that, even if Brown acted with a retaliatory motive, Fuller’s independent investigation and termination decision absolved Brown of any liability as a matter of law. If so, the majority’s holding is in tension with Gil-brook.

In Gilbrook, the plaintiffs-employees sued both their subordinate supervisors and the superior who made the ultimate decision to terminate. Gilbrook, 177 F.3d at 852-53. The jury concluded that “the plaintiffs’ protected conduct had played a ‘substantial or motivating’ role in [the subordinate supervisor’s] actions against plaintiff, but had not played such a role in the actions of [the superior who made the final decision to terminate].” Id. at 853. Despite the fact that the jury absolved the final decision-maker of liability — presumably because he made an independent decision to terminate — the court upheld the verdict against the subordinate supervisors. Id. at 855-56. In other words, the fact that the final decision maker acted independently and without a retaliatory motive did not automatically negate the subordinate supervisors’ liability for the employees’ termination. In fact, the court expressly rejected the argument, made by the subordinate supervisors, that the superior’s “legitimate, nonretaliatory motive ‘cuts off the liability of his subordinates” as a matter of law. Gilbrook, 177 F.3d at 853.

B. Brown’s Affirmative Defense

Because I would hold that Scott presented sufficient evidence from which a jury could have concluded that her protected conduct was a “substantial or motivating factor” in her termination, I would reach the question whether Brown established the affirmative defense outlined in Mt. Healthy. That defense required Brown to demonstrate, by a preponderance of the evidence, that “the disciplinary action would have been taken against [Scott] even in the absence of the protected conduct.” Gilbrook, 177 F.3d at 855 (citing Mt. Healthy, 429 U.S. at 287, 97 S.Ct. 568); see also Ostad, 327 F.3d at 883 (holding a subordinate supervisor liable where the supervisor instigated an investigation, participated actively in that investigation, and failed to prove that the employee “would have been terminated ... even in the absence of his protected speech”).4

Brown met that burden here. She introduced evidence that the journal initially was uncovered through an unrelated investigation, instigated by Fuller, of another employee; that Brown herself did not discover the journal; that employees in the human resources department were independently aware of the journal; and that the journal, by virtue of its length and inflammatory content, was likely to attract attention from HR employees and to war*815rant a report to senior management. Given that evidence, I would hold that a reasonable jury could only have concluded that Fuller would have learned of the journal, instigated the investigation of Scott and reached the same conclusion even without prompting from Brown, and therefore Scott’s termination would have occurred “even in the absence of the protected conduct.” See Gilbrook, 177 F.3d at 855.

In sum, I would reverse the jury’s verdict under part two, not part one, of the Mt. Healthy burden-shifting framework on the grounds that, based on the facts presented, Fuller’s investigation — and hence Scott’s termination — was inevitable. Such a holding acknowledges the deference we must afford to a reasonable decision by a jury, and avoids placing too great an evi-dentiary burden on a plaintiff seeking to vindicate her First Amendment rights.

For all these reasons, I concur in the judgment, but not in the majority opinion.