BOGGS, Circuit Judge.
Will Singleton (“Singleton”), who was employed as a Registered Nurse by the appellees (collectively “Select”), appeals from a decision of the district court granting Select’s motion for summary judgment on Singleton’s claim of unlawful retaliation under Title VII of the Civil Rights Act and its Kentucky state-law equivalent.1 Select alleges that it terminated Singleton for numerous and ongoing errors pertaining to Singleton’s documentation of narcotics administration and patient pain levels, which it discovered during an investigation it initiated following the discovery of a discrepancy in the count of a narcotic Singleton was responsible for administering. Because Singleton cannot establish that this rationale was a pretext for unlawful retaliation, we affirm the judgment of the district court.
I
Singleton, who is white, is a Registered Nurse who began employment with Select on August 30, 2004. Select operates a long-term acute care facility on one floor of Good Samaritan Hospital, in Lexington, Kentucky; among Select’s 42 beds are a four-bed intensive-care unit. During the approximately 19 months he worked for Select, Singleton was generally assigned to the ICU beds.
Singleton’s performance during much of his time at Select was considered at least adequate. Toward the end of his tenure at Select, however, Director of Clinical Services Kim McGowen began noting problems with Singleton’s documentation. The record indicates that she had a “counseling session” with Singleton on December 16, 2005, and recorded the following summary:
It was brought to my attention that Will’s charting needed improvement by Supervisor. Talked with Will on this night about improving his charting. Areas needed to improve Legibility and pain scale correlation. Will stated he had Dyslexia and at times he would get a few of his letters backwards. Told Will he needed to take his time so anyone that reviewed charts could read them. That I could hardly read his writing some nights and his pain documentation was terrible. He was told he needed to correlate the pain meds he was giving with the pain scale and Nurses[’] notes.
A follow-up note, dated February 1, 2006, indicated that McGowen “[tjalked to Will Singleton on the phone Feb 1st about pain documentation. Stated he understood the issue at hand and would improve. Chart[397]*397ing has not improved....” At her deposition, McGowen testified that “I always felt like Will was a good nurse. I didn’t have any problems with Will until probably sometime in November [2005].” In November 2005, McGowen said, Singleton “just sort of changed”; his writing became more illegible, and his attitude worsened to the point that he began having conflicts with other RNs working in his unit.
In the meantime, Singleton had begun making formal complaints. On October 5, 2005, he wrote to McGowen to make various allegations about Pam Clouse, a “Charge Nurse” who supervised shifts. One of those complaints alleged that Clouse had told him two racist jokes. In response, McGowen conducted an investigation that eventually determined that, though no one else claimed to have heard Clouse tell racially offensive jokes, one employee had heard her use the term “wetback” on one occasion. McGowen informed Clouse that the term was offensive, unacceptable, and against Select policy, and could be grounds for termination.
Singleton again complained to McGowen about Clouse on November 5, 2005. On that date, he wrote a letter “to inform you of the racial jokes that Pam Clouse is telling to people at work. Tonight she told 2 jokes about blacks and Mexicans. She is continually talking about blacks.” McGowen investigated, but despite speaking with (she believes) every night-shift employee she again could find no one who had heard Clouse tell these jokes. Nevertheless, she told all employees that racial comments would not be tolerated and that any such comments should be brought to her attention; she also spoke to Clouse and told her that “if there were any further reports that she had engaged in additional conduct of the type Will reported, she would be fired.” No further accusations of racist language or jokes were brought against Clouse by anyone.
Singleton’s next complaint occurred on February 27, 2006, when he faxed a letter to Select’s Regional Human Resources Director, Josceylon Buchs. After making several unrelated complaints, Singleton included a section under the heading “It pays to be Married to Rosa” in which he complained that “Bill is a prn aid here at select. Yet if is Bill-vs-black-tech-vs-woman. Bill works and ‘they' stay home.” He then included a list of dates and names, purporting to demonstrate that “Bill” had worked on instances when the hospital had called off others who should have received priority. The letter itself did not identify which of the listed employees were black, white, male, or female; however, according to Singleton’s brief on appeal, the list showed that four black employees had been called off a total of twelve times, whereas three white employees had been called off only three times.
In response to Singleton’s February 27 letter, McGowen conducted an investigation and wrote a two-and-a-half page email to Select Specialty Hospital Lexington, Inc.’s CEO Rick Daugherty, in which she appears to have reconstructed the scheduling decisions on the dates identified by Singleton. In so doing, her email appears to establish that every such decision either conformed with Select’s policies with respect to the order in which employees were to be called off or, if they deviated from that order, they did so for a legitimate reason (typically because the employees not utilized had requested that they be called off earlier than they would otherwise be). No further action was taken to investigate.
The next significant occurrence happened on March 4-5, 2006, while Singleton was working a 7:00 p.m.-to-7:00 a.m. shift. On that shift, as on his others, Singleton [398]*398had access to a “Med Dispense” machine, which is an electronically locked cabinet that stores and dispenses pharmaceuticals. Nurses administering drugs contained in the Med Dispense machine are required to enter an employee I.D. and password, perform a count of the drug present in the relevant drawer and match that number to the count kept by the machine, and then adjust the machine’s count to account for whatever number of pills/capsules they are removing. Among the pharmaceuticals stored in the Med Dispense machine was Dilaudid, a Schedule II narcotic.
The “end of shift” count of Dilaudid conducted at 7:14 p.m. on March 4, 2006— fourteen minutes after Singleton’s shift began — indicated that there were 36 Dilaudid tabs in the Med Dispense machine. The dispense report generated by the machine indicated that, following the end-of-shift count, Singleton was the next person to access Dilaudid. At 8:42, he was shown on the report as removing three tabs. He then accessed Dilaudid again at 1:15 a.m. to remove another three pills, with no one else having done so in the interim. At that time, Singleton notified Clouse, who was the supervisor on duty, that the physical count prior to his removing medication was 30, whereas the machine’s count was 33. A subsequent “full system count” revealed that a different medication, Versed, was three pills over its machine count. Clouse filled out a Confidential Incident Report in which she reported the discrepancy and which used the word “resolved” in the account; the incident was then investigated by the hospital’s Director of Pharmacy, Keith Thomas, who came to the conclusion that the missing three pills were not accounted for.2
Under a Select policy titled “Guidelines for Suspicion of Controlled Substance Diversion,” any time an employee is suspected of diverting controlled substances intended for patient use, an investigation is required. Such an investigation was initiated in Singleton’s case, which included an examination of his documentation and narcotics control practices over the most recent six weeks. While the investigation was ongoing, Singleton was suspended with pay beginning March 13, 2006. The investigation was conducted by McGowen and her immediate supervisor, Molly Pers-by, the Division Director of Clinical and Quality Services. The results were circulated to Daugherty, Buchs, and Senior Vice President of Clinical and Quality Services Mary Burkett. On March 27, 2006, Daugherty, Persby, Buchs, McGowen, Burkett, and Human Resources Coordinator Janet Whitlock discussed the results of the investigation via conference call. On that occasion, Singleton’s documentation practices were described as “horrific,” in terms of his failure to properly control narcotics, his lack of patient pain assessment, and his poor medical documentation. Specifically, in addition to having nearly-illegible handwriting and failing to properly record medication on patient charts, Singleton was also criticized for giving narcotics (or at least documenting that he gave narcotics) to patients who had not been in pain.
[399]*399Daugherty fired Singleton on March 28, 2006, with Whitlock present. On that date, Daugherty presented Singleton with a dismissal form indicating that ‘Will Singleton has consistently failed to meet nursing practice standards related, but not limited to, control of narcotic medications and proper documentation of pain assessment and narcotic administration. Due to these circumstances, we are terminating his employment with Select Specialty Hospital.” Singleton signed the form, acknowledging that the had reviewed it and had an opportunity to comment on it.
Singleton filed his complaint in the district court on July 18, 2007. The district court granted summary judgment to Select on Singleton’s federal claims on January 27, 2009, at which time the court also declined to retain jurisdiction over his state claims. This timely appeal followed.
II
We review a district court’s grant of summary judgment de novo. Int'l Union v. Cummins, Inc., 434 F.3d 478, 483 (6th Cir.2006). We will uphold such a grant “if the pleadings, depositions, answers to interrogatories, and admissions on file, together with affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to judgment as a matter of law.” Fed.R.Civ.P. 56(c). In considering a motion for summary judgment, we must view the inferences to be drawn from the underlying facts in favor of the non-moving party, in this case the plaintiff. See Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986). The central issue is “whether the evidence presents a sufficient disagreement to require submission to a jury or whether it is so one-sided that one party must prevail as a matter of law.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 251-52, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986).
III
Title VII forbids an employer from “dis-criminat[ing] against any of his employees ... because [the employee] has opposed any practice made an unlawful employment practice by [Title VII], or because [the employee] has made a charge, testified, assisted, or participated in any manner in an investigation, proceeding, or hearing under [Title VII].” 42 U.S.C. § 2000e-3(a).
Where, as here, a plaintiff presents only circumstantial evidence to support a Title VII retaliation claim, we utilize the same McDonnell Douglas/Burdine evidentiary framework that is used to assess claims of discrimination. See Wrenn v. Gould, 808 F.2d 493, 500 (6th Cir.1987). The plaintiff has the initial burden under this framework to establish a prima facie case of retaliation by showing that (1) he engaged in activity protected by Title VII; (2) this exercise of protected rights was known to the defendant; (3) the defendant thereafter took adverse employment action against the plaintiff, or the plaintiff was subjected to severe or pervasive retaliatory harassment by a supervisor; and (4) there was a causal connection between the protected activity and the adverse employment action or harassment. Morris v. Oldham County Fiscal Court, 201 F.3d 784, 792 (6th Cir.2000).
Assuming that the plaintiff establishes a prima facie case, “the burden of production of evidence shifts to the employer to ‘articulate some legitimate, nondiscriminatory reason’ for its actions.” Canitia v. Yellow Freight Sys., 903 F.2d 1064, 1066 (6th Cir.1990) (quoting McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973)). In this case, importantly, the “legitimate, nondiscrimi[400]*400natory reason” for Select’s action was not the Dilaudid count discrepancy of March 5, 2006, but rather the problems with narcotics control and documentation expressed by the disciplinary action form signed by Singleton on March 28, 2006.
Once the defendant in a Title VII retaliation action satisfies the burden to put forth a legitimate nondiscriminatory reason for its conduct, the plaintiff must then demonstrate by a preponderance of the evidence that the legitimate reason offered by the defendant was not its true reason, but instead was a pretext designed to mask retaliation. McDonnell Douglas, 411 U.S. at 802, 93 S.Ct. 1817. It is at this stage of the McDonnell Douglas/Burdine analysis that the district court found Singleton’s claim to be lacking, and so it is with us.3
In Manzer v. Diamond Shamrock Chemicals Co., 29 F.3d 1078, 1084 (6th Cir.1994), we identified three ways in which a plaintiff may rebut a defendant’s legitimate, nondiscriminatory reason and demonstrate pretext. The plaintiff may show that (1) the employer’s stated reason for terminating the employee has no basis in fact, (2) the reason offered for terminating the employee was not the actual reason for the termination, or (3) the reason offered was insufficient to explain the employer’s action. Ibid. Regardless of which rebuttal method is employed, the plaintiff retains the ultimate burden of producing “sufficient evidence from which the jury could reasonably reject [the defendants’] explanation and infer that the defendants intentionally discriminated against him.” Johnson v. Kroger Co., 319 F.3d 858, 866 (6th Cir.2003) (alteration in original).
Singleton attempts to demonstrate pretext under all three Manzer methods. He first argues that the legitimate non-discriminatory reason relied upon by Select was “factually false” because the Dilaudid count discrepancy was “immediately resolved and attributed to a counting error, and was only later seized upon by Select as purported grounds to intensify its scrutiny of Singleton and eventually terminate his employment.” However, this analysis does not address the first Manzer method properly. An attempt to show pretext by demonstrating that the proffered reasons have no basis in fact “is easily recognizable and consists of evidence that the proffered bases for the plaintiffs discharge never happened, i.e., that they are ‘factually false.’ ” Manzer, 29 F.3d at 1084. Here, Select’s proffered nondiscriminatory reason was not that Dilaudid was missing on one particular occasion, but rather that Singleton’s documentation and narcotics control were deficient over a period of several weeks. Hence, a demonstration of pretext under the first method would focus on establishing a question of fact that there were no such deficiencies. Instead, Singleton argues that Select ought not to have been looking at his documentation practices at all, and that it examined them only as a cover for its retaliatory action. This is more properly conceived as going to the second Manzer method.
In fact, Singleton’s deposition contains multiple instances in which he was shown charts prepared by him that contained inconsistencies or errors that he could not account for. With respect to more than one patient to whom Singleton had given morphine, for example, Singleton admitted that there was no place on their charts where he had documented how many milligrams of the drug he had administered. [401]*401Singleton was also unable to explain why, on one chart, he had administered morphine to a patient that had been resting comfortably for several hours, and whom Singleton had rated a “zero” on a pain scale of zero to ten. On another occasion, Singleton charted that he had administered Ativan, an addictive sedative and anti-anxiety drug, an hour earlier than the medication sign-out sheet showed him accessing it.
Singleton does, in a different part of his brief, dispute the factual basis for Select’s conclusion that his documentation was deficient. However, his responses are not sufficient to address the wealth of documentary evidence supplied by Select. In reply to the allegation that his handwriting was often illegible, Singleton points to one instance in which Persby said “from what I can read from Will’s note at eight o’clock, assessment at charted, no signs and symptoms of pain” (emphasis added) — hardly an admission that his handwriting was consistently or even frequently legible. In response to allegations that he failed to document patient pain scales, Singleton argues that his assessment of “zero” constitutes documentation, but does not address Select’s contention that, if such was the case, he was clearly giving narcotics to patients that ought not to have been receiving them. Further, in answer to Select’s evidence demonstrating that he failed to document dosages properly on patients’ charts, Singleton asserts that he was recording dosages on the medication sign-out sheets, which, even if true, is irrelevant to the question of whether he was recording those dosages in other places required by hospital policy. Hence Singleton’s explanations fail to raise a genuine issue of material fact as to the factual falsity of Select’s proffered explanation, and do not show pretext under the first Manzer method.
A plaintiff using the second Manzer method, by contrast, must show that the proffered reason(s) “did not actually motivate the defendant’s challenged conduct.” Johnson, 819 F.3d at 866 (internal quotation marks omitted). In this type of showing,
the plaintiff attempts to indict the credibility of his employer’s explanation by showing circumstances which tend to prove that an illegal motivation was more likely than that offered by the defendant. In other words, the plaintiff argues that the sheer weight of the circumstantial evidence of [retaliation] makes it “more likely than not” that the employer’s explanation is a pretext, or coverup.
Manzer, 29 F.3d at 1084 (emphasis in original). As indicated above, Singleton argues that he has put forth evidence that his documentation practices did not actually motivate Select to fire him, in the form of evidence that Select’s reason for looking at his documentation at all was pretextual. However, Singleton’s evidence does not meet this burden. He claims that the Dilaudid discrepancy was immediately resolved at the time it was discovered, but such a claim ignores the affidavit of the hospital’s Director of Pharmacy, Keith Thomas, who indicated that he personally investigated the incident and that the discrepancy was not due to employee miscount, erroneous inputting of numbers, “or other such human error mistakes,” and was never resolved. Moreover, the affidavit points out that the use of the term “resolved” on the Confidential Incident Report filed by supervisor Clouse on the night in question was mandated by the machine’s software, and referred only to the fact that the machine’s running count had been adjusted to match the number of tabs observed to be physically present in the drawer at the time the discrepancy was reported. Hence there is no evidence [402]*402that this discrepancy was the kind of easily-understood error of the kind that might not have called for a full investigation under Select’s policy.4
Singleton compares his case to Imwalle v. Reliance Medical Products, Inc., 515 F.3d 531 (6th Cir.2008), in which we upheld a jury verdict finding that a defendant had engaged in unlawful discrimination under the Age Discrimination in Employment Act. In Imwalle, we held that the plaintiff had introduced sufficient evidence to permit a jury to disbelieve the employer’s proffered performance-based explanation by demonstrating that (1) the company had been successful for 14 years under Imwalle’s management, and.(2) he had not been responsible for the industry-wide decline that had more recently affected its sales figures. Id. at 547-48. Further, we held that, though it was a “close question,” a rational jury could conclude that retaliation was the true motive for Imwalle’s dismissal because (1) even prior to his termination, other retaliatory actions (such as cutting Imwalle out of important meetings) had occurred, and (2) Imwalle’s discrimination claim was discussed in a prepared statement read at his termination meeting. Id. at 549-51. Singleton argues that his situation is analogous to Imwalle’s, because, he alleges, Select’s CEO falsely told him that he had failed a drug test (which Singleton was required to take as part of the investigation) and refused to give him copies of the investigation reports. Singleton also points to an email written on February 28, 2006 by Daugherty in which Daugherty mentioned that Singleton had “attempted to sue his former employer for issues involving scheduling.”
These comparisons are inapposite. Unlike in Imwalle, the “retaliatory atmosphere” Singleton claims to have existed consisted only of the investigation into his documentation procedures, and hence was part and parcel of the adverse action of which he ultimately complains; to permit him to argue that the investigation makes it more likely that Select’s true motivation was retaliatory would in effect permit him to argue that the adverse action itself is evidence of motive. Singleton presents no evidence that the process by which this investigation went forward was different from investigations into similar incidents, or that the process was inconsistent with hospital policy. Moreover, the email by Daugherty that Singleton references as evidence that Singleton’s complaint was at the forefront of Daugherty’s mind was dated February 28, 2006, weeks before Singleton’s termination; in contrast with Im-walle, where reference was made to the plaintiff’s discrimination complaint at the termination meeting itself, here the reference merely demonstrates the unremarkable fact that, one day after Singleton sent a letter alleging that Select was engaging in unfair scheduling practices, Select’s CEO was thinking about the fact that Singleton had attempted to sue a prior employer on similar grounds.
Singleton additionally relies on an entirely speculative argument that the evidence of his poor documentation procedures was a post-hoc fabrication by McGowen. Singleton himself denies that any of the conversations she alleges that [403]*403she had with him occurred, and argues that McGowen manufactured the handwritten notes of those conversations once her own judgment began to come under scrutiny. As we are required to do on summary judgment, we resolve this question of fact in favor of Singleton, and presume that these conversations never took place. However, McGowen’s handwritten notes merely purported to document occasional attempts by her to correct a problem; the actual evidence relied upon by the management team that decided to terminate Singleton included charts, drug logs, and other documents that were indisputably filled out by him. Hence even if McGowen’s notes were entirely fabricated, they were not essential to the decision-making process. Even if they were, it is also true that “[w]hen an employer reasonably and honestly relies on particularized facts in making an employment decision, it is entitled to summary judgment on pretext even if its conclusion is later shown to be ‘mistaken, foolish, trivial, or baseless.’” Chen v. Dow Chem. Co., 580 F.3d 394, 401 (6th Cir.2009).
Singleton also argues that a “Confidential Incident Report” filled out by nurse supervisor Chris LeMaster on March 16, 2006, is evidence that the investigation into his documentation practices was a sham because it refers to him as already having been terminated. There is no evidence that LeMaster, however, was privy to the decision-making process by which Singleton was terminated, or that he would know if there had been some kind of consensus reached by management on that issue pri- or to the date Select says that decision was reached, March 27; hence, his belief that Singleton had been terminated is not probative with respect to the question of pretext. Instead, it appears highly likely that the reference to Singleton being “terminated” is instead a reference to his having been suspended.
Similarly unavailing are Singleton’s attempts to demonstrate pretext under the third Manzer method, a type of showing that typically consists of evidence that other employees were not terminated even though they engaged in substantially identical conduct. Manzer, 29 F.3d at 1084. Singleton first claims that he was treated differently in that others were not investigated following discrepancies with their Med Dispense machine counts. This misses the point of the test, which is to compare an employer’s treatment of other employees who had engaged in the conduct that is claimed to be the legitimate nondiscriminatory reason for termination, not conduct that allegedly gave rise to the discovery of that reason. Even so, Singleton does not point to any comparable employee that was treated differently. He alleges that, during his tenure at Select, approximately 343 discrepancies were reported with the machine, many of which involved narcotics. Yet he does not indicate how many of these were, as was true in his case, found to be unexplained even after an investigation by the Director of Pharmacy, and indeed the Director of Pharmacy’s affidavit clarifies that Singleton’s discrepancy was not one of the “many” that could be chalked up to a human error mistake. Singleton does point to one incident in which a single Dilaudid pill was unaccounted for on March 19, 2006, which “apparently was not investigated by Select nor any disciplinary action taken against the nurse in question.” However, Singleton does not appear to have obtained any evidence as to the factual circumstances of that case; there is no testimony from the nurse in question or the nurse supervisor on duty that would clarify whether the discrepancy was actually one that ought to have triggered an investigation like the one undertaken in Singleton’s case. “The jury may [404]*404not reject an employer’s explanation ... unless there is a sufficient basis in the evidence for doing so.” Manzer, 29 F.3d at 1083 (emphasis in original).
Singleton also argues that he was treated differently from other employees with respect to his documentation errors, a theory that would properly fall within the third Manzer method of demonstrating pretext. His evidence of that proposition, however, is simply insufficient. He first argues that Select treated him differently because it did not follow its “progressive discipline policy,” which calls for employees to first receive a documented verbal warning, followed by a written warning and then termination of employment. Select points to the deposition of Persby, who testified that the use of the progressive policy was not mandatory, but Singleton responds that in an email concerning the plaintiffs in the case that was originally consolidated with his, Persby wrote that “[w]e do insist that everyone go through the process (verbal, written, then termination).” The employees referenced in that email, however, were not being disciplined for the same kinds of infractions as Singleton; rather, they were “caus[ing] a great deal of disturbance ... creating] a hostile work environment, try[ing] to intimidate the other staff, talking] about staff when they are present.... ”
As Persby testified, the foundational concern with Singleton’s behavior was that it could affect patient safety. While we have held that “[e]vidence that [a] progressive-discipline policy asserted as a rationale for an employee’s termination was not uniformly applied is evidence of pretext,” Lamer v. Metaldyne Co. LLC, 240 Fed.Appx. 22, 33 (6th Cir.2007), in this case it does not appear that the progressive-discipline policy was even “asserted as a rationale” for Select’s action. Singleton appears to read Lamer as holding that, when an employer utilizes a progressive-discipline policy in any circumstances, it must utilize it in all circumstances, which is plainly not the case; rather, Lamer holds that when an employee is discharged pursuant to such a policy, it must have been applied consistently with other such actions. In his deposition testimony, CEO Daugherty indicated that the progressive disciplinary policy at Select was “for issues that are not totally egregious.... For behaviors that are just totally out of line, and we list those in the handbook, you can have immediate termination without going through the process.” Though the record does not appear to contain a list of offenses that would not require progressive discipline before termination, the fact that Singleton’s offenses could have endangered patients, whereas his proposed comparators’ offenses merely disrupted the working environment, renders the comparison Singleton seeks to make inapposite. Indeed, the same Select policy that required an investigation into any suspected diversion of controlled substances further mandates that “[t]he employee will be terminated, regardless of truthful admission if there are clearly documented practice issues.” (emphasis in original).
While Singleton does argue that others who committed documentation errors were given verbal or written warnings, moreover, he fails to show that these other employees had a record of documentation problems similar to his. In reviewing the disciplinary action forms he submits pertaining to these other employees, there is no evidence that any of them reflect ongoing problems as opposed to occasional lapses. Indeed, as discussed above, the decision-makers in this case had ample evidence before them that Singleton had received multiple warnings about his documentation errors.
[405]*405Because Singleton is unable to demonstrate that Select’s proffered reason is a pretext for illegal retaliation, his Title VII claim fads. The judgment of the district court is AFFIRMED.