OPINION

CLAY, Circuit Judge.

Petitioner Karon Jackson ("Jackson") appeals from the district court's order granting Defendant Detroit Receiving Hospital's ("DRH") motion for summary judgment and dismissing Jackson's claim that DRH terminated her employment because of her sex in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e. For the reasons set forth below, we REVERSE the district court's judgment and REMAND for further proceedings.

BACKGROUND

In 1998, Jackson began working at DRH's Mental Health Crisis Center ("Crisis Center") as a mental health technician ("MHT"). As an MHT, Jackson's duties generally involved assisting registered nurses ("RN") with the treatment and processing of psychiatric patients. Particularly relevant here were the MHTs' duties with regard to patient discharge. Although RNs were required to authorize the discharge of a patient, MHTs assisted in the discharge process by, inter alia, gathering paperwork and retrieving the patient's personal property. Once the discharge was fully processed, MHTs were responsible for physically escorting the patient out of the Crisis Center. At some point prior to the patient being escorted out of the Center, both the RN and the MHT were required to check the patient's identification wristband ("ID band") and bloodwork to verify that the correct patient was being discharged.

On September 6, 2013, Jackson was assigned to work the morning shift as an MHT in the Crisis Center. At the beginning of the shift, Jackson's manager, Leorea Heard, held a meeting with Crisis Center staff. The purpose of the meeting, as indicated on a sign-in sheet circulated at the beginning of the meeting, was to emphasize the importance of checking patients' ID bands prior to discharge. Although it is undisputed that Jackson signed the sign-in sheet for the meeting, the parties disagree on the extent to which Jackson actually participated in that meeting. At her deposition, Jackson testified that she did not read the sign-in sheet, and that she missed the meeting entirely because a disruptive patient required assistance. Heard, on the other hand, recalled that Jackson attended the entire meeting.

In any case, it is undisputed that later that same morning, Jackson assisted an RN named Christine Moore with the discharge of an incorrect patient; both Moore and Jackson failed to check the patient's ID band prior to his discharge. Fortunately, the wrongly discharged patient, who had checked himself into the Crisis Center the previous night complaining of depression and suicidal thoughts, returned to the Center on his own roughly seven hours after being mistakenly discharged. At an emergency meeting with DRH administrators held that afternoon, Jackson admitted that she had made a mistake. She explained, however, that she made the mistake while acting on the instructions of RN Moore, and that "[i]t was a fast, busy morning." (R. 14-11, PageID 161.) At her deposition, Jackson clarified that she had just finished assisting a different agitated patient when RN Moore instructed her to escort the wrong patient out of the Crisis Center.

Three days after the discharge incident, DRH issued a memorandum terminating Jackson's employment.1 The memorandum stated that Jackson had violated "major infractions" "k" and "q" of DRH's disciplinary policy, which prohibit:

(R. 14-11, PageID 160.) DRH's disciplinary policy defines major infractions as "those violations of policy or misconduct which are generally considered more serious and could therefore result in immediate dismissal or disciplinary suspension without having a prior disciplinary record." (R. 15-5, PageID 230.) The termination memorandum was signed by four persons in Jackson's supervisory chain of command, all of whom are female.

During her employment as an MHT at DRH, Jackson consistently received high ratings on her performance evaluations. In her final evaluation before her termination, Jackson received an overall rating of 3.63 on a four-point scale. On this scale, a score of three corresponded to "very good/above expectations;" a score of four corresponded to "exceptional/exceeds expectations." (R. 14-1, PageID 101.) The record contains no evidence of Jackson having been disciplined for any infraction prior to the incident leading to her termination. At her deposition, Heard testified that she could not recall Jackson ever receiving a "final warning" or being placed on a "last chance agreement." (R. 14-2, PageID 122.)

At the time of Jackson's termination, fourteen of the Crisis Center's eighteen nurses were female; nine of its twelve social workers were female. However, Jackson was the only female out of fourteen MHTs. As explanation for this disparity, Jackson alleges that the Crisis Center's staff preferred male MHTs for their ability to physically handle unruly patients.

On April 23, 2014, Jackson filed suit against DRH alleging a single claim of discrimination on the basis of sex in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e. In the litigation that followed, Jackson argued that although her failure to check the wrongly discharged patient's ID band constituted a terminable infraction, other circumstances surrounding her termination indicated that she was fired because she is a woman. Specifically, Jackson argued that DRH did not terminate two male MHTs who made mistakes of comparable seriousness. These male MHTs (collectively "comparators") were Ronald Duncan and Lester Little.

Ronald Duncan was employed by DRH as an MHT beginning in August 2008. In July 2011, Duncan was placed on "final warning" status for being absent from his work area (a minor infraction) and for improper personal use of DRH computers (a major infraction). In February 2012, Duncan was again placed on "final warning" status, this time for testing positive for drug use (a major infraction). As a result of that drug infraction, Duncan entered into a year-long "last chance agreement," under which he could be terminated for violating any of DRH's "standards of conduct and work performance, including any minor or major infraction. . . ." (R. 15-8, PageID 246.) At her deposition, Leorea Heard testified that under this agreement, any infractions would result in Duncan's termination.

In April 2012, while still subject to the last chance agreement, Duncan walked the wrong patient out of the Crisis Center because he failed to check the patient's ID band. The patient that Duncan walked out was also disabled and required crutches, but Duncan walked him out without his crutches. Leorea Heard cited Duncan's conduct as constituting two major infractions, including the health and safety violation for which Jackson was terminated. That same day, Duncan walked off the job without punching out, which was cited as a minor infraction. As a result of these infractions, Duncan was suspended pending an investigation. During the investigation, Leorea Heard watched a video of the discharge incident that showed that Duncan

(R. 14-2, PageID 128.) Based on this video, Heard and other DRH administrators determined that discharge of the wrong patient was not Duncan's fault. Duncan received no discipline for his actions that day apart from the initial five-day suspension.

Lester Little was employed by DRH as an MHT beginning in January 2011. On October 22, 2012, Little was issued a written warning for violating DRH's attendance policy. On April 16, 2013, Little was given "verbal counseling" for walking a patient out of the Center on the instructions of a social worker. Little had been informed the previous week that he was not to discharge any patient without instructions from the attending nurse. On July 24, 2013, Little failed to perform a complete search of a newly admitted patient. It was later discovered that the patient was wearing a waist pouch containing three knives. Little's disciplinary write-up for the incident determined that his actions violated the same two major infractions for which Jackson was terminated—infractions k and q.2 Little's supervisor, Leorea Heard, was on vacation at the time of this incident, but later signed off on the disciplinary form that suspended Little for five days and placed him on "final warning" status; Little was not terminated.

On February 26, 2015, DRH filed a motion for summary judgment in Jackson's civil suit. After determining that oral argument was unnecessary, the district court entered an opinion and order granting DRH's motion for summary judgment and dismissing Jackson's case with prejudice. The court reasoned that, although Jackson had established a prima facie case of discrimination, Jackson had not proffered evidence sufficient to allow a reasonable jury to infer that DRH's purported reason for terminating her was pretextual. Specifically, the court held that the comparators' mistakes were not similar enough to Jackson's to raise an inference that her differential treatment was due to discrimination. Jackson timely appealed.

DISCUSSION

I. Standard of Review

We review a district court's decision to grant summary judgment de novo. Little v. BP Expl. & Oil Co., 265 F.3d 357, 361 (6th Cir.2001). A grant of summary judgment will be upheld only where no genuine dispute of material fact exists and the moving party is entitled to judgment as a matter of law. See Fed.R.Civ.P. 56(a). A dispute of a material fact is genuine so long as "the evidence is such that a reasonable jury could return a verdict for the non-moving party." Ford v. Gen. Motors Corp., 305 F.3d 545, 551 (6th Cir.2002) (quoting Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986)).

"The district court, and this Court in its review of the district court, must view the facts and any inferences reasonably drawn from them in the light most favorable to the party against whom judgment was entered." Kalamazoo Acquisitions, L.L.C. v. Westfield Ins. Co., 395 F.3d 338, 342 (6th Cir.2005) (citing Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986)). In other words, "at the summary judgment stage[,] the judge's function is not himself to weigh the evidence and determine the truth of the matter but to determine whether there is a genuine issue for trial." Anderson, 477 U.S. at 249, 106 S.Ct. 2505; see also Arban v. W. Publ'g Corp., 345 F.3d 390, 400 (6th Cir. 2003) ("This court does not weigh the evidence, evaluate the credibility of witnesses, or substitute its judgment for that of the jury.").

II. Analysis

Under 42 U.S.C. § 2000e-2(a)(1), an employer may not "discharge any individual. . . because of such individual's race, color, religion, sex, or national origin.. . ." A plaintiff alleging that her employer violated this prohibition may prove her claim using either direct or circumstantial evidence. White v. Baxter Healthcare Corp., 533 F.3d 381, 391 n.5 (6th Cir.2008). Where a plaintiff relies on circumstantial evidence, courts typically apply the three-part burden-shifting framework developed by the Supreme Court in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), and Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981), to determine whether the plaintiff has proffered sufficient evidence to survive summary judgment. White v. Columbus Metro. Hous. Auth., 429 F.3d 232, 240 (6th Cir. 2005).

Under the McDonnell Douglas/Burdine framework, the plaintiff bears the initial burden of establishing a prima facie case of discrimination. White, 533 F.3d at 391. "Once the plaintiff establishes this prima facie case, the burden shifts to the defendant to offer evidence of a legitimate, non-discriminatory reason for the adverse employment action." Id. If the defendant meets this burden, "the burden shifts back to the plaintiff to show that the defendant's proffered reason was not its true reason, but merely a pretext for discrimination." Id. at 391-92. Importantly, at the summary judgment stage of litigation, courts should not allow "th[is] burden-shifting analysis [to] obfuscate the appropriate question—whether there exists a genuine issue of material fact." Provenzano v. LCI Holdings, Inc., 663 F.3d 806, 813 (6th Cir.2011); Cline v. Catholic Diocese of Toledo, 206 F.3d 651, 661 (6th Cir.2000) ("On a motion for summary judgment, a district court considers whether there is sufficient evidence to create a genuine dispute at each stage of the McDonnell Douglas inquiry.").

Below, the district court held that Jackson was sufficiently similarly situated to MHT Duncan—but not MHT Little—to establish a prima facie case. Even so, the district court granted DRH's motion for summary judgment on the grounds that Jackson had not proffered sufficient evidence to create a genuine dispute as to whether DRH's justification for firing her was pretextual. Before us, therefore, are two issues: (1) whether Jackson's proffered evidence was sufficient to establish a prima facie case as to MHTs Duncan and Little; and (2) whether Jackson's evidence was sufficient to create a triable dispute of fact as to whether DRH's justification for Jackson's termination was pretextual. We address these issues below.

A. Prima Facie Case

In Burdine, the Supreme Court stated that the purpose of the prima facie stage of the burden-shifting framework is to

Burdine, 450 U.S. at 254, 101 S.Ct. 1089 (internal citations and quotation marks omitted). Because "[t]he prima facie case is meant simply to force a defendant to proceed with its case," Provenzano, 663 F.3d at 813, we have admonished that the plaintiff's burden "at the prima facie stage is `not onerous' and `poses a burden easily met.'" Id. (quoting Cline, 206 F.3d at 660).

To establish a prima facie case of sex discrimination, a plaintiff must demonstrate that: "(1) she is a member of a protected group; (2) she was subjected to an adverse employment decision; (3) she was qualified for the position; and (4) . . . similarly situated non-protected employees were treated more favorably." Peltier v. United States, 388 F.3d 984, 987 (6th Cir. 2004). In this case, DRH does not dispute that Jackson meets the first three elements of the prima facie case. We therefore focus on the final element: whether Jackson has established that similarly situated, non-protected employees were treated more favorably.

In Mitchell v. Toledo Hosp., 964 F.2d 577 (6th Cir.1992), we noted three factors relevant to determining whether employees are "similarly situated" in the context of cases alleging differential disciplinary action:

Id. at 583; see also Johnson v. Kroger Co., 319 F.3d 858, 867 (6th Cir.2003) ("the weight to be given to each [Mitchell] factor can vary depending upon the particular case"); Ercegovich v. Goodyear Tire & Rubber Co., 154 F.3d 344, 352 (6th Cir. 1998) (observing that the Mitchell factors "generally are all relevant considerations in cases alleging differential disciplinary action"). As we later explained in Ercegovich, the Mitchell factors do not require a plaintiff to "demonstrate an exact correlation with the employee receiving more favorable treatment in order for the two to be considered `similarly-situated;' rather,. . . the plaintiff and the employee with whom the plaintiff seeks to compare himself or herself must be similar in `all of the relevant aspects.'" 154 F.3d at 352 (quoting Pierce v. Commonwealth Life Ins. Co., 40 F.3d 796, 802 (6th Cir.1994)).

Jackson and DRH largely focus on Mitchell's third factor: whether Jackson's conduct was sufficiently similar to her comparators' such that DRH's differential discipline creates an inference of discriminatory motive.3 In evaluating this factor, we look to whether the comparators' actions "were of `comparable seriousness' to the conduct for which Plaintiff was discharged." Mitchell, 964 F.2d at 583 (quoting Lanear v. Safeway Grocery, 843 F.2d 298, 301 (8th Cir.1988)). A plaintiff "is not required to show that his proposed comparator's actions were identical to his own." Colvin v. Veterans Admin. Med. Ctr., 390 Fed.Appx. 454, 459 (6th Cir.2010); see also Bobo v. United Parcel Serv., Inc., 665 F.3d 741, 751-52 (6th Cir.2012). Even so, a plaintiff cannot establish a reasonable inference of discriminatory motive based on her "employer's more severe treatment of more egregious circumstances." Clayton v. Meijer, Inc., 281 F.3d 605, 612 (6th Cir.2002); see also Colvin, 390 Fed.Appx. at 458-59 (holding plaintiff and coworkers' conduct not sufficiently similar to create an inference of discrimination based on differential discipline); Simpson v. Vanderbilt Univ., 359 Fed.Appx. 562, 569 (6th Cir.2009) (holding same); Ruth v. Children's Med. Ctr., 940 F.2d 662, at *7 (6th Cir.1991) (table) (holding same).

We conclude that Jackson sufficiently demonstrated that Duncan's and Little's actions were of "comparable seriousness" to the conduct for which Jackson was discharged to establish a prima facie case. Mitchell, 964 F.2d at 583. Duncan's actions were nearly identical to Jackson's: he escorted the incorrect patient out of the Crisis Center because he failed to check the patient's ID band. It is undisputed that Duncan was subject to the same requirement that he check patients' wristbands before discharge, and that his failure to do so constituted a major infraction. Similarly, as a result of his improper search, MHT Little was cited for violation of the same two major infractions for which Jackson was terminated.4 Leorea Heard herself described "[t]he patient getting an improper search" as a "major incident" that would necessitate a staff-wide training session. (R. 14-2, PageID 119.) And importantly, neither Duncan nor Little was terminated for his mistake.

On appeal, DRH argues that Jackson's mistake was so much more egregious than those of her comparators as to prevent her from being able to establish a prima facie case. See Clayton, 281 F.3d at 612. We disagree. Certainly, differences exist between Jackson, Duncan, and Little. But thorough explication of those differences is unnecessary for "eliminat[ing] the most common nondiscriminatory reasons for the [employer's termination of the plaintiff]." Burdine, 450 U.S. at 253-54, 101 S.Ct. 1089; see also Provenzano, 663 F.3d at 813 ("One common misapplication [of the McDonnell Douglas/Burdine framework] is the tendency to push all of the evidence into the prima facie stage and ignore the purpose for and application of the three stages."). Stated differently, those differences do not render Jackson and her comparators so facially distinguishable as to obviate the need for DRH to provide any explanation for its differential treatment. See Provenzano, 663 F.3d at 813 ("The prima facie case is meant simply to force a defendant to proceed with its case."). Thus, we reserve discussion of those differences for the later stages of the McDonnell Douglas/Burdine framework, in which "the factual inquiry proceeds to a new level of specificity." Burdine, 450 U.S. at 255, 101 S.Ct. 1089.

Based on the obvious similarities between Jackson, Duncan, and Little—and the fact that neither Duncan nor Little was terminated—we conclude that Jackson has sufficiently demonstrated "circumstances giving rise to an inference of unlawful discrimination." Blair, 505 F.3d at 529. Thus, Jackson has met her "not onerous" and "easily met" preliminary burden of establishing a prima facie case. Provenzano, 663 F.3d at 813.

B. Legitimate Reasons and Pretext

"Once the plaintiff establishes [her] prima facie case, the burden shifts to the defendant to offer evidence of a legitimate, non-discriminatory reason for the adverse employment action." White, 533 F.3d at 391. "This is a burden of production; although `[t]he defendant need not persuade the court that it was actually motivated by the proffered reasons,' it must raise `a genuine issue of fact as to whether it discriminated against the plaintiff.'" Cline, 206 F.3d at 666 (quoting Burdine, 450 U.S. at 254, 101 S.Ct. 1089); see also Burdine, 450 U.S. at 256, 101 S.Ct. 1089 ("the employer's burden is satisfied if he simply `explains what he has done' or `produc[es] evidence of legitimate nondiscriminatory reasons.'").

Once the employer has come forward with a nondiscriminatory reason for firing the plaintiff, "the plaintiff must identify evidence from which a reasonable jury could conclude that the proffered reason is actually a pretext for unlawful discrimination." Provenzano, 663 F.3d at 812 (quoting Macy v. Hopkins Cnty. Sch. Bd. Of Educ., 484 F.3d 357, 364 (6th Cir.2007)). Notably, this burden on the plaintiff "merges with the ultimate burden of persuading the court that she has been the victim of intentional discrimination." Burdine, 450 U.S. at 256, 101 S.Ct. 1089. Thus, on summary judgment, "[i]n evaluating pretext and the plaintiff's ultimate burden, the court should consider all [probative] evidence in the light most favorable to the plaintiff, including the evidence presented in the prima facie stage." Provenzano, 663 F.3d at 812; see also Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 147, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000) ("Proof that the defendant's explanation is unworthy of credence is simply one form of circumstantial evidence that is probative of intentional discrimination").

DRH's justification for its allegedly differential treatment of Jackson is that the mistake for which she was fired was more egregious than either Duncan's or Little's mistakes. The parties also offer additional circumstantial evidence that they argue tends to support or weaken any inference of a discriminatory motive for Jackson's termination. We conclude that when all relevant evidence is viewed in the light most favorable to Jackson, and all reasonable inferences are drawn in her favor, the circumstances surrounding her termination would permit a reasonable jury to infer that DRH's justifications for her termination were pretextual, and that she was instead terminated because of her sex.

1. Differences Between Jackson and Her Coworkers

Where an employer argues that the plaintiff's differential discipline was justified by material differences in context, we evaluate whether that justification is pretextual by looking to the same or similar factors as when evaluating the "similarly situated" element of the prima facie case. See, e.g., Johnson, 319 F.3d at 867 (using the Mitchell/Ercegovich "similarly situated" analysis to evaluate pretext); Smith v. Leggett Wire Co., 220 F.3d 752, 762 (6th Cir.2000) (same); see also McCarthy v. Ameritech Pub., Inc., 763 F.3d 469, 482 (6th Cir.2014) ("The district court held that McCarthy failed to meet each of her burdens: She did not adduce evidence of the fourth element of her prima facie case, nor did she adduce evidence of pretext. The same evidence is germane to each inquiry.").

Cases from this Circuit also provide guidance specific to the pretext inquiry. In Dews v. A.B. Dick Co., 231 F.3d 1016, 1021 (6th Cir.2000), we held that "[a] plaintiff can demonstrate pretext by showing that the [employer's] proffered reason [for the adverse action] (1) has no basis in fact, (2) did not actually motivate the defendant's challenged conduct, or (3) was insufficient to warrant the challenged conduct." On appeal, Jackson cites this three-pronged approach and states that she "relies . . . only on the third method." (Pl.'s Br. at 18.) Demonstrating pretext using this method "ordinarily . . . consists of evidence that other employees, particularly employees not in the protected class, were not fired even though they engaged in substantially identical conduct to that which the employer contends motivated its discharge of the plaintiff." Manzer v. Diamond Shamrock Chems. Co., 29 F.3d 1078, 1084 (6th Cir.1994), overruled on other grounds by Geiger v. Tower Auto., 579 F.3d 614 (6th Cir.2009).

We pause to note that Manzer's "substantially identical" language should not be read to increase the weight of Jackson's burden of proof at the pretext stage. To do so would obfuscate the ultimate question: whether Jackson has adduced enough evidence to allow a reasonable jury to draw the inference that she was terminated because of her sex. See Burdine, 450 U.S. at 256, 101 S.Ct. 1089. In other words, requiring Jackson bear a greater burden of proof at the pretext stage "would mistakenly `apply[] legal rules which were devised to govern "the basic allocation of burdens and order of presentation of proof" in deciding this ultimate question.'" Cline, 206 F.3d at 661 (quoting U.S. Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711, 716, 103 S.Ct. 1478, 75 L.Ed.2d 403 (1983)).

Rather, we read Manzer as merely restating Burdine's principle that at the pretext stage, "the factual inquiry proceeds to a new level of specificity." 450 U.S. at 255, 101 S.Ct. 1089. This is a change in the rigor with which we evaluate Jackson's similarity to her comparators; it is not an increase in the weight of her evidentiary burden. See Provenzano, 663 F.3d at 813-14 ("a general weighing of the qualifications of [plaintiff] and [her coworker] is necessary at the prima facie stage; however, this light review must be distinguished from the more rigorous comparison conducted at the later stages of the McDonnell Douglas analysis").

When conducting this more rigorous comparison, we again focus on the severity of the differently treated employees' actions. See Chattman v. Toho Tenax Am., Inc., 686 F.3d 339, 350 (6th Cir.2012) (comparing the "severity" and "commonality" of the employees' conduct). The relative severity of two actions is not determined solely by whether those actions violated the same company rule or policy. See, e.g., Perez v. Illinois, 488 F.3d 773, 777 (7th Cir.2007) ("although in this case the same rules were implicated, the offenses. . . are of significantly different severity such that they are not comparable."). Instead, employers—and therefore courts—are free to consider both the actual and potential consequences of the employee's actions. See, e.g., Clayton, 281 F.3d at 612 (holding differential discipline could not create inference of discrimination where, although several employees violated the same rule, only plaintiff's violation led to the injury of his coworker); Ruth, 940 F.2d 662, 1991 WL 151158, at *7 (holding same where "[plaintiff's] error differed in both type and severity of potential consequences").

DRH argues that substantive differences in the circumstances of Jackson's mistake—including the potential consequences of that mistake—render her mistake more egregious than Duncan's and Little's. Thus, it is DRH's contention that Jackson's harsher punishment could not lead a reasonable jury to infer that she was treated differently on the basis of sex. We address these arguments as they apply to each of Jackson's comparators.

i. MHT Duncan

As discussed above, Jackson argues that male MHT Ronald Duncan was treated more favorably despite committing the same mistake for which she was fired—namely, escorting the wrong patient out of the Crisis Center as a result of failing to check that patient's ID band. DRH provides two justifications for its more severe punishment of Jackson's mistake: first, Jackson's failure to check her patient's ID band was a mistake with more severe potential consequences because her patient was suicidal; second, Jackson had that morning participated in a staff meeting, the sole purpose of which was to reinforce the importance of checking patients' ID bands prior to discharge.

In support of its first argument, DRH cites Ruth, 940 F.2d 662, 1991 WL 151158, at *7, and Colvin, 390 Fed.Appx. at 459-60. Yet, both of those cases involved facially obvious differences in the egregiousness of the employees' mistakes. In Ruth, the plaintiff pharmacist dispensed a drug at ten times the prescribed dosage; the comparator dispensed the proper dose of an incorrect but effectively similar antibiotic. 940 F.2d 662, 1991 WL 151158, at *3. And in Colvin, the plaintiff pharmacist "erroneously dispensed ten times the ordered amount of a controlled substance, failed to include relevant dosage information on a prescription, and neglected to catch a physician's error regarding how a drug should be administered;" the comparator had problems processing orders on time. See 390 Fed.Appx. at 459. Here, unlike Ruth and Colvin, the employees violated the same rule, the results of that violation were the same, and the harm that ultimately befell the patients because of that violation was the same. Any perceived differences in egregiousness, therefore, are based on DRH's speculation over what might have come to pass.

DRH's reliance on such speculation is problematic for two reasons. First, given the record currently before us, we find that speculating on the likelihood and relative severity of injuries that might have befallen Jackson's and Duncan's patients—and whether differences in those hypothetical injuries justified differential discipline—is a task better suited to a jury. See Anderson, 477 U.S. at 255, 106 S.Ct. 2505 ("the drawing of legitimate inferences from the facts are jury functions, not those of a judge"). Second, were we inclined to engage in such speculation, the record would not compel us to come down on DRH's side: Duncan's mistakenly-discharged patient was disabled and required crutches, but Duncan walked him out without his crutches. Viewing these facts in the light most favorable to Jackson, we could reasonably conclude that Duncan's patient might have suffered a serious injury—or even died—had he fallen as a result of not having his crutches. Indeed, the likelihood of this potential harm to Duncan's patient was obvious, given that the patient was wearing a surgical shoe at the time Duncan discharged him without his crutches. Presumably, this is the same potential harm that DRH believes makes Jackson's mistake more egregious.

DRH's second attempt to distinguish Duncan—by noting that Jackson's mistake took place on the same day as a meeting emphasizing the importance of checking ID bands—is similarly unpersuasive. Although the extent of Jackson's participation in the meeting is disputed, it is undisputed that MHT Duncan was subject to the same requirement that he check patients' ID bands before escorting them out of the Crisis Center. Duncan was suspended pending investigation after failing to follow that rule, indicating that even before the September 6, 2013 meeting, DRH administrators considered failure to check ID bands to be a serious offense. On review of the record, it appears to us that the September 6, 2013 morning meeting did not announce any substantive changes in DRH's policy. Rather, drawing all reasonable inferences in Jackson's favor, the meeting appears to have been intended merely to remind employees about the ID band rule and to emphasize its importance.

Certainly, this conclusion still permits the inference that DRH disciplined Jackson more harshly not because of her sex, but to make an example of her. But this is only one of several possible inferences: a jury could also reasonably infer that DRH should have been more lenient towards Jackson, given that DRH had shown considerable leniency when dealing with Duncan's violation of the same rule, despite the numerous aggravating circumstances in his case. At the time of his violation of the ID band rule, Duncan had been working at DRH for only four years; in that time, he had twice been placed on "final warning" status for committing various major and minor infractions. Indeed, at the time of his violation, Duncan was under a "last chance agreement" for drug use, under which any infractions would result in his termination. And on the same day that he violated the ID band rule, Duncan walked off the job without clocking out—itself a minor infraction. Yet, Duncan was not terminated. Jackson, a well-reviewed veteran employee with no disciplinary record, was terminated for making the same mistake as Duncan, but without any of the above aggravating circumstances.

DRH responds that Duncan was shown leniency for his violation because, after an investigation, Manager Heard concluded that Duncan was not at fault. Heard based her conclusion on a video recording of Duncan's incident showing that he discharged his patient "on the word of the nurse" while he was already engaged in escorting a different group of patients out of the Crisis Center. (R. 14-2, PageID 128.) But Jackson relayed a similar story: she was ordered to escort her patient out after she had just finished assisting a different agitated patient; Jackson took it on faith that the RN had checked the patient's ID band. From these facts, a jury could reasonably infer that DRH was pre-disposed to accepting Duncan's excuse and/or rejecting Jackson's.

In sum, we conclude that Duncan was treated differently under circumstances that could allow a reasonable jury to reject DRH's explanation for Jackson's differential treatment.

ii. MHT Little

Jackson argues that male MHT Lester Little was treated more favorably despite committing an infraction of comparable severity: Little failed to perform a complete search of a newly admitted patient, and it was later discovered that the patient was wearing a waist pouch containing three knives. As justification for its more severe punishment of Jackson's mistake, DRH argues that (1) Little's mistake was not comparable because he did not violate the Patient Identification Policy for which Jackson was terminated, and (2) that "there is no record evidence that [Little's] patient presented an imminent risk of harm to himself . . . or to others at the Crisis Center." (Def.'s Br. at 26.)

To begin, DRH cites no authority for the proposition that a plaintiff and her comparator must commit exactly the same mistake in order to permit a reasonable inference of intentional discrimination from their differential discipline. As we noted in the context of our discussion on Jackson's prima facie case, supra, our cases make clear that the relevant inquiry is whether the comparator's conduct was substantially identical "in all of the relevant aspects." Ercegovich, 154 F.3d at 352 (internal quotation marks omitted). Ercegovich's relevancy requirement is no less applicable in the pretext stage of our analysis than it was in the prima facie case stage.5 See Johnson, 319 F.3d at 867 (applying Ercegovich's relevancy requirement in the context of pretext analysis); Weigel v. Baptist Hosp. of E. Tennessee, 302 F.3d 367, 378-79 (6th Cir.2002) (same); Smith, 220 F.3d at 762 (6th Cir.2000) (same). Viewed in the light most favorable to Jackson, the facts indicate that Little's mistake was substantially identical to Jackson's in the relevant aspects.

Several facts in the record can be used to draw the inference that DRH considered Jackson's and Little's infractions substantially identical in terms of severity of potential consequences. That their mistakes were cited as violating the same two major infractions indicates that those mistakes were substantially identical in terms of potential disciplinary consequences: DRH's discipline policy defines major infractions by the potential for immediate termination. Similarly, that Jackson's and Little's mistakes were cited as violations of major infraction "k," which is defined as "[a]ny action or conduct that endangers or may be detrimental to the well being of a patient, co-worker, physician, contractor or visitor" (R. 14-11, PageID 160), suggests that DRH considered the potential harm resulting from those mistakes to be similar. Indeed, Leorea Heard provided "[t]he patient getting an improper search" as an example of an "incident that might have put the unit in danger." (R. 14-2, PageID 119.)

The facts also indicate that the circumstances of the mistakes themselves were substantially identical. Both employees violated well-established rules intended to prevent injury to patients or staff. Both employees' violations resulted in the circumstances that the rule was presumably intended to prevent: in Little's case, a patient was able to possess weapons within the Crisis Center; in Jackson's case, an incorrect patient was escorted out of the Center. Moreover, Jackson's and Little's actions were identical in the severity of their actual consequences: neither employees' violation resulted in harm to patients or staff.

In response, DRH again argues that its decision to terminate Jackson was based on its own speculation regarding what might have come to pass. Specifically, it notes that, whereas Jackson's "patient presented an imminent risk of harm to himself," there was no such evidence regarding Little's patient (see Def.'s Br. at 26); rather, any speculation regarding what might have come to pass as a result of Little's patient sneaking knives into a mental ward would necessarily be "based on conjecture and stereotyping." (Id.) We disagree. A jury could reasonably infer that allowing a mental health patient to carry three knives into a mental health crisis center had self-evidently severe potential consequences. Indeed, substantial evidence in the record, including DRH's own disciplinary policy and Heard's testimony, suggests that DRH itself considers the dangers posed by an improper search to be more than mere "conjecture and stereotyping."

For these reasons, we conclude that Little was treated differently under circumstances that could allow a reasonable jury to reject DRH's explanation for Jackson's differential treatment.

2. Other Circumstances Bearing on the Reasonableness of an Inference of Discrimination

Jackson and DRH argue that other facts in the record bear on the reasonableness of an inference of discrimination. See Reeves, 530 U.S. at 147, 120 S.Ct. 2097. Most notably, both parties cite the ratio of females working in various positions within the Crisis Center in support of their positions. Jackson argues that an inference of discriminatory motive can be drawn from the fact that at the time of her termination, she was the only female out of fourteen MHTs working in the Crisis Center. DRH responds that the opposite inference is more properly drawn from its employment statistics: 77% of the Crisis Center's eighteen nurses and 75% of its twelve social workers were female. In total, DRH argues, the Crisis Center was a majority female workplace, thus reducing the reasonableness of an inference that DRH terminated Jackson because she is female.

We reiterate that on review of a motion for summary judgment, we are required to "view the facts and any inferences reasonably drawn from them in the light most favorable to the party against whom judgment was entered." Kalamazoo Acquisitions, L.L.C., 395 F.3d at 342. So viewed, the composition of the Crisis Center's staff weighs in Jackson's favor. In a majority female workplace, the fact that Jackson was the only female out of fourteen MHTs supports her contention that DRH preferred males for that particular position.

DRH also argues that "[a]ll of the decision-makers involved in Plaintiff's termination were female, like Plaintiff, and it is unreasonable to conclude that gender motivated their decision-making." (Def.'s Br. at 28-29.) But the Supreme Court has explicitly rejected this argument:

Oncale v. Sundowner Offshore Servs., Inc., 523 U.S. 75, 78-79, 118 S.Ct. 998, 140 L.Ed.2d 201 (1998).

Certainly, Oncale does not prohibit any consideration of the fact that the decision-maker was the same sex as the plaintiff. See id. But that case makes clear that a mere conclusory assertion to that effect does not, on its own, render unreasonable an inference of discriminatory intent. We find this maxim especially true here, where the primary inference Jackson attempts to draw from the record is that Crisis Center management preferred males for the position of MHT, perhaps "because they thought females could not physically handle unruly patients." (Pl.'s Br. at 1.)6 Jackson's case is thus largely unaffected by the fact that her managers were female.

Finally, DRH points to Jackson's consistently positive performance reviews as evidence that her managers harbored no ill will against her on account of her being female. As we discussed above, however, a jury could draw the opposite inference from this fact: if Jackson was an exemplary employee, DRH should have given Jackson the same second chance that it provided to her less-than-exemplary male coworkers.

CONCLUSION

In sum, when "the facts and any inferences reasonably drawn from them [are viewed] in the light most favorable to" Jackson, Kalamazoo Acquisitions, L.L.C., 395 F.3d at 342, a reasonable jury could conclude that DRH terminated her because of her sex. See Ford, 305 F.3d at 551. We therefore REVERSE the district court's judgment and REMAND this case for further proceedings.