Larry Alan Warch v. Ohio Casualty Insurance Company, and Ohio Casualty Group
435 F.3d 510
4th Cir.
2006
Check Treatment
Docket

*512 OPINION

TRAXLER, Circuit Judge:

Larry Alan Warch brought this action against his former employer, The Ohio Casualty Insurance Company (“OCIC”), alleging that OCIC unlawfully discriminated against him based on his age when it fired him. The district court granted OCIC’s motion for summary judgment. For the reasons discussed below, we affirm.

I.

OCIC fired Warch from his position as fraud investigator on April 26, 2002, when he was age 59. Warch’s termination should not have been a surprise. Since at least 2000, Warch’s supervisors had been giving him negative feedback. On March I, 2000, the director of Warch’s unit, Bill Johansen, sent Warch a letter following up on “several issues pertaining to [Warch’s] performance and efforts.” J.A. 274. The letter explained how one of Warch’s supervisors had discussed company procedures with him “on several occasions” and how there continued to be “shortcomings” that “adversely impact[ed his] personal performance and that of the unit.” J.A. 274. Johansen advised Warch to “adopt immediately all procedures and guidelines.” J.A. 274.

Shortly thereafter, OCIC changed its internal management structure, but Warch’s performance continued to be a source of concern, even under new supervision. The new head of Warch’s unit, Bruce Montgomery, met with Warch and advised him that the files he had turned in “lacked any investigative effort,” that he “had little or no communication with the claims reps assigned to the case,” that he was not leaving his home to conduct field assignments, that he provided “inappropriate” responses to the claims staff that created complaints, and that he needed to “listen more carefully.” J.A. 110. Two months later, Warch received a “Formal Counseling Form” stating that he had failed to conduct thorough, concise, and proper investigations, made recommendations to pay claims without gathering all the facts and using all available tools, and failed to “communicate with claims.” J.A. 279. OCIC informed Warch that he would be placed on probation if he failed to improve.

Improvement did not come and, at the end of 2001, OCIC put Warch on probation for six months due to his unacceptable work product. OCIC required Warch to take corrective action and informed him that he would be terminated if he failed to improve. Although Warch received positive feedback in some areas on his 2001 employee evaluation form, the form also stated that he was “not effective,” had failed to produce “the product ... expected of him,” was having a “difficult time with the changes that have been made,” and needed to “address his work performance and his interpretation of policy provisions.” J.A. 286-87. Warch apparently recognized that his performance was problematic, offering family and personal problems as “an excuse for [his] performance.” J.A. 289.

OCIC extended Warch’s probation. In response to a question from Warch about details of OCIC’s concerns, Montgomery explained to Warch:

We have identified the problems to you many times and each time you state that you understand and then we receive files with the same problems. I feel at this point that we have been more than fair in identifying the problems, explaining what is needed to correct them and the rest will be up to you....
*513 [T]he problems that we see in your work ... are across the board. You must improve these issues during this extended probationary period.

J.A. 291. OCIC terminated Warch on April 26, 2002, while he was still on probation.

Warch filed a complaint with the Equal Employment Opportunity Commission (“EEOC”) in July 2002. Shortly thereafter, the EEOC issued Warch a right to sue letter and Warch commenced this action. Warch sued under the Age Discrimination in Employment Act (“ADEA”), which forbids employers from “discharging] any individual ... because of such individual’s age.” 29 U.S.C.A. § 623(a)(1) (West 1999). Warch pursues his claim under both the “pretext” approach and the “mixed motive” approach. Specifically, he asserts that OCIC’s proffered reason for terminating him was pretext for illegal age discrimination and that, even if OCIC had a nondiscriminatory reason to terminate him, age was a motivating factor in his termination.

The district court granted summary judgment to OCIC. On the pre-text approach, the district court held that Warch failed to create any genuine dispute that he was meeting his employer’s legitimate job expectations or that he was replaced by a substantially younger employee. On the mixed motive approach, the district court ruled that Warch failed to create any genuine dispute, whether based on direct or circumstantial evidence, that age was a motivating factor in his termination.

II.

We review the grant of summary judgment de novo. See Higgins v. E.I. DuPont de Nemours & Co., 863 F.2d 1162, 1167 (4th Cir.1988). Summary judgment is appropriate “if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed.R.Civ.P. 56(c); see Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). We construe the evidence in the light most favorable to Warch and draw all reasonable inferences in his favor. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986).

III.

Warch’s claim under the “pretext” approach derives from McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-05, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), and its progeny. McDonnell Douglas established a burden-shifting framework for a plaintiff to demonstrate that the purported reason for an adverse employment action, such as termination, was actually a pre-text for unlawful discrimination. See Mereish v. Walker, 359 F.3d 330, 334 (4th Cir.2004). Under this framework, a plaintiff must first establish a prima facie case of unlawful discrimination by a preponderance of the evidence. See id.

Generally speaking, to establish a prima facie case of unlawful age discrimination, Warch must show that (1) he is a member of the protected class; (2) he was qualified for the job and met OCIC’s legitimate expectations; (3) he was discharged despite his qualifications and performance; and (4) following his discharge, he was replaced by a substantially younger individual with comparable qualifications. See O’Connor v. Consolidated Coin Caterers Corp., 517 U.S. 308, 312-13, 116 S.Ct. 1307, 134 L.Ed.2d 433 (1996); Causey v. Balog, 162 F.3d 795, 802 & n. 3 (4th Cir.1998).

After the plaintiff establishes a pri-ma facie case, the burden shifts to the employer to produce a legitimate, non-dis *514 criminatory reason for the termination. See Mereish, 359 F.3d at 334. The employer’s burden at this stage “is one of production, not persuasion; it can involve no credibility assessment.” Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 142, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000) (internal quotation marks omitted). If the employer meets this burden, “the presumption of discrimination created by the prima facie case disappears from the case” and the plaintiff must prove that the “proffered justification is pretextual.” Mereish, 359 F.3d at 334; see also Texas Dept, of Cmty. Affairs v. Burdine, 450 U.S. 248, 253, 255, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981).

In this case, the district court granted summary judgment to OCIC on Warch’s claim under the pretext approach by finding that he failed to produce sufficient evidence to prove that he was meeting his employer’s legitimate job expectations or that he was replaced by a substantially younger employee. We address each conclusion in turn.

A. Legitimate Job Expectations

We first address Warch’s claim that the district court erred in considering the question of whether he was meeting his employer’s legitimate job expectations during the prima facie case. According to Warch, he should only be required to prove that he was “qualified” for the job. Inquiring whether the employee met his employer’s legitimate job expectations, Warch argues, improperly collapses the second stage of the McDonnell Douglas framework, where the employer brings forth its legitimate, non-discriminatory reason for the termination, into the prima facie case.

Warch’s argument that he only needs to show that he is “qualified” for the job, rather than meeting his employer’s legitimate job expectations, is foreclosed by circuit precedent. See Hill v. Lockheed Martin Logistics Mgmt., Inc., 354 F.3d 277, 298 (4th Cir.2004) (en banc) (“Hill has failed to establish a prima facie case of ... age discrimination because, by her own admissions ... she has failed to demonstrate that she was performing her job duties at a level that met Lockheed’s legitimate expectations.”); Causey, 162 F.3d at 802 (requiring proof that the employee “was qualified for the job and met the employer’s legitimate expectations”) (emphasis added).

Furthermore, considering an employer’s legitimate expectations comports with the purpose of requiring the establishment of a prima facie case — to screen out those cases whose facts give rise to an inference of nondiscrimination, in other words, to eliminate the most common, nondiscriminatory reasons for the employer’s conduct. See Miles v. Dell, 429 F.3d 480, 488 n. 5 (4th Cir.2005) (citing Burdine, 450 U.S. at 253-54, 101 S.Ct. 1089). Where an employer hires a new employee, the focus typically is on the qualifications of the candidates. Where, however, an employer fires an employee, the employer is more likely focused on other aspects of the employment, such as poor job performance or infractions of company rules.

Once a person is hired, the distinction between qualifications and job expectations tends to blur. An employee may be qualified when hired, but could fail either to maintain his qualifications or, more commonly, to meet his employer’s legitimate expectations for job performance. See Loeb v. Textron, Inc., 600 F.2d 1003, 1013 n. 10 (1st Cir.1979) (noting that the fact that a plaintiff was initially hired generally indicates that he has the “basic qualifications for the job, in terms of degrees, certificates, skills and experience”). In such cases, the prima facie case requires the employee to demonstrate “that *515 he was ‘qualified’in the sense that he was doing his job well enough to rule out the possibility that he was fired for inadequate job performance, absolute or relative.” Id. at 1013; see also La Montague v. American Convenience Prods., Inc., 750 F.2d 1405, 1409 n. 2 (7th Cir.1984) (“When an employee of long standing is discharged, the more appropriate concern would appear to be his job performance, into which the question of his qualifications might be seen to merge.”). Thus, we hold that courts are not required to focus solely on “qualifications” to the exclusion of the employer’s legitimate job expectations when evaluating a plaintiffs prima facie case of unlawful termination.

We also reject Wareh’s contention that consideration of the employer’s legitimate job expectations at the prima facie stage improperly allows consideration of evidence the employer would typically present in the second stage of the McDonnell Douglas framework, that is, where the employer offers the legitimate, non-discriminatory reason for the termination. In support of this argument, W31^*1 relies on the approach announced in Cline v. Catholic Diocese of Toledo, 206 F.3d 651 (6th Cir.2000), which held that, “when assessing whether a plaintiff has met her employer’s legitimate expectations at the prima facie stage of a termination case, a court must examine plaintiffs evidence independent of the nondiscriminatory' reason ‘produced’ by the defense as its reason for terminating plaintiff.” Id. at 660-61. Because OCIC’s proffered reason for terminating him was poor performance, Warch argues the district court erred in considering his job performance at the prima facie stage of the inquiry.

In Cline, a Catholic school refused to renew the contract of one of its second-grade teachers after it discovered that she had become pregnant a few months before getting married, in violation of the Church’s policy against sexual intercourse outside of marriage. The district court ruled' that the teacher had failed to establish a prima facie case of sex and pregnancy discrimination. Specifically, the district court concluded that the teacher could not show she was “qualified” because she had violated her promise to live according to the principles of the Church. The Sixth Circuit reversed, holding that the district court had “conflated the distinct stages of the McDonnell Douglas inquiry by using [the school’sl’nondiscriminatory reason’ as a predicate for finding Cline to have failed to make a prima facie case.” Id. at 660. Warch contends that, as his case involves a termination where the employer asserted unsatisfactory job performance as the basis, for the discharge, the district court made the same error as the district court in Cline by considering OCIC’s evidence of Warch’s poor work performance at the prima facie stage. Warch argues that the district court should have only allowed introduction of this evidence at the second stage of the McDonnell Douglas inquiry.

We disagree. Although the plaintiffs burden is “not onerous,” it nevertheless requires him to “prov[e] by the preponderance of the evidence a prima facie case of discrimination.” Burdine, 450 U.S. at 252-53, 101 S.Ct. 1089; see also St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 506, 113 S.Ct. 2742, 125 L.Ed.2d 407 (1993) (“The plaintiff ... must first establish, by a preponderance of the evidence, a ‘prima facie’ case....”); Mackey v. Shalala, 360 F.3d 463, 468 (4th Cir.2004) (“ ‘[T]he plaintiff-employee must first prove a prima facie case of discrimination by a preponderance of the evidence.’ ”) (quoting Evans v. Technologies Applications & Service Co., 80 F.3d 954, 959 (4th Cir.1996)); Mereish, 359 F.3d at 334 (“[Ajppellants must first establish a prima facie case of discrimination by a preponderance of the evidence.”). And, because a plaintiff must show by a prepon *516 derance of the evidence that he met the employer’s legitimate job expectations to prove his prima facie case, the employer may counter with evidence defining the expectations as well as evidence that the employee was not meeting those expectations. To require otherwise would turn the plaintiffs burden at the prima facie stage into a mere burden of production, making it “difficult to imagine a case where an employee could not satisfy the ‘qualified’ [or legitimate expectation] element as defined in Cline.” Nizami v. Pfizer Inc., 107 F.Supp.2d 791, 801 n. 11 (E.D.Mich.2000). Diluting this element of the prima facie case “defeats the whole purpose of this first stage of the McDonnell Douglas inquiry, which is, after all, to ‘create[ ] a presumption that the employer unlawfully discriminated against the employee’ by ‘eliminat[ing] the most common nondiscriminatory reasons’ for the employer’s action.” Id. (quoting Burdine, 450 U.S. at 254, 101 S.Ct. 1089). In short, we find no impermeable barrier that prevents the employer’s use of such evidence at different stages of the McDonnell Douglas framework.

Despite our disagreement with Cline, however, we are cognizant of the danger that courts might apply the “expectations” or “qualification” element of the prima face too strictly in some cases, resulting in the premature dismissal of potentially meritorious claims of unlawful discrimination. See Cline, 206 F.3d at 660 (“While the discrete stages [of the McDonnell Douglas framework] are meant to facilitate litigants and courts in reaching and resolving that ultimate question of discrimination, when misapplied, they tend to distract courts from the central issue.”). A hypothetical raised in Cline illustrates the point quite well. See id. at 663 n. 7.

In the Cline hypothetical, a truck driver who loses her driver’s license is terminated. A narrow application of the “expectation” or “qualification” element would appear to foreclose the driver from proving her prima facie ease, since with no driver’s license she would not be able to show that she met the job qualifications or legitimate expectations of her employer for a position as truck driver. Yet, even though the driver’s case would never get past the prima facie stage, the employer could have still used the loss of the license as a pretext for illegal discrimination. Evidence tending to show this pretext might be that similarly situated men who lost their licenses were not terminated but, instead, were temporarily suspended until they received new licenses or were transferred to other jobs within the company. Cline attempted to remedy this anomaly by looking “at whether an employee met her employer’s legitimate expectations prior to the event(s) that sparked the termination.” Id. (emphasis added). Following this lead, Warch argues that prior to the onset of events that OCIC cites as its reasons for terminating him, he was meeting OCIC’s performance expectations.

We find this approach to be unworkable, especially where there is no one “event” that “sparked the termination,” but instead a long string of performance problems leading up to firing. The further back in time a court goes to evaluate an employee’s performance, the more removed the evidence is from the time of the termination. As we explained in Lovelace v. Sherwin-Williams Co., 681 F.2d 230 (4th Cir.1982):

The rationality, hence fairness, of this inference [of unlawful discrimination] obviously decreases as the time gap between last proven satisfactory performance and challenged employment action lengthens. Here, the time lag was almost two years. As common experience in such matters teaches, and as the full record reveals the case here to have *517 been, a great deal can happen to alter things in such a time.

Id. at 244. See also O’Connor v. Consolidated Coin Caterers Corp., 56 F.3d 542, 547 (4th Cir.1995) (holding that a review of an employee’s 1989 performance was irrelevant to a determination of whether his performance was satisfactory at the time of his termination in August of 1990), rev’d on other grounds, 517 U.S. 308, 116 S.Ct. 1307, 134 L.Ed.2d 433 (1996).

We think the flexibility of the McDonnell Douglas inquiry, when properly applied, already protects plaintiffs from the feared injustice of the Cline hypothetical. See Brinkley v. Harbour Recreation Club, 180 F.3d 598, 611 (4th Cir.1999) (holding that “the McDonnell Douglas framework should not be applied in a ‘rigid, mechanized, or ritualistic’ manner”) (citations omitted); cf. Moore v. City of Charlotte, 754 F.2d 1100, 1105 (4th Cir.1985) (recognizing that the McDonnell Douglas framework is “less useful” in the context of an alleged discriminatory disciplinary decision than in the context of an alleged discriminatory hiring decision). Although on summary judgment an employer is free to assert that the job expectation prong has not been met, nothing prohibits the employee from countering this assertion with evidence that demonstrates (or at least creates a question of fact) that the proffered “expectation” is not, in fact, legitimate at all. Thus, where application of the qualification or expectation element of the prima facie case seems to preclude an otherwise meritorious claim, the plaintiff is free to demonstrate that the employer’s qualifications or expectations are not, in fact, “legitimate.”

This flexibility, inherent in the McDonnell Douglas framework, is true to the goal of “progressively ... sharpening] the inquiry into the elusive factual question of intentional discrimination,” Hicks, 509 U.S. at 506, 113 S.Ct. 2742 (internal quotation marks omitted) (alteration in original), and the recognition that the shifting burdens of McDonnell Douglas are “meant only to aid courts and litigants in arranging the presentation of evidence,” Watson v. Ft. Worth Bank & Trust, 487 U.S. 977, 986, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988). Cf. Miles, 429 F.3d at 488-89 (recognizing that there may be circumstances where application of the traditional elements of the prima facie case might improperly screen out cases where a presumption of discrimination could still exist). The same considerations apply to the legitimate job expectations and qualifications prongs of the prima fa-cie case.

We now turn to the question of whether Warch has produced sufficient evidence from which a jury could conclude that he met OCIC’s legitimate job expectations. As early as March 2000, before Bruce Montgomery took over as lead of Warch’s unit and over two years before Warch was terminated, Warch had already been reprimanded concerning “several issues pertaining to [his] performance and efforts.” J.A. 274. His supervisors had advised him “on several occasions concerning [company] procedures” and informed him that he continued to suffer from “shortcomings” that “adversely impacted his] personal performance and that of the unit.” J.A. 274. Furthermore, after Montgomery took over as the leader of the unit, Warch acknowledged that his own performance failed to meet OCIC’s expectations. See J.A. 289 (“I feel an excuse for my performance is necessary. My wife has been ill.... ”).

Warch argues that OCIC’s criticisms were too subjective to be considered at the prima facie stage. This complaint will not avail Warch because, if for no other reason, the facts do not support it. OCIC reprimanded Warch based on concrete, specific observations and accompanied its *518 reprimands with explicit instructions on how to improve. For example, although Warch claims the March 2000 letter he received (before Montgomery was his supervisor) dealt only with deficiencies in “file organization” and “office filing procedures, not performance,” Brief of Appellant at 23, the letter explicitly addresses “several issues pertaining to [Warch’s] performance and efforts” and discusses Warch’s “shortcomings.” J.A. 274 (emphasis added).

Warch similarly contends that counseling he received in May 2001 was not specific and used only general and subjective language. We disagree. The Formal Counseling Form states that Warch was, among other things, “making recommendations to pay claims without gathering all the facts and using all tools given to conduct such investigations” and failing to “communicate with claims.” J.A. 279. The form states that Warch’s files “lacked any investigation and there was little or no communication with claims.” J.A. 279. These are observations of objective facts.

Warch refers to opinions of those who thought he was doing a good job and points to savings he brought the company, arguing that this evidence demonstrates he was meeting OCIC’s legitimate performance expectations. As to the opinions, they were offered either by those whose employment with OCIC ended well before Warch’s or by third parties who were never employed by OCIC at all. Thus, they lack probative value as to whether Warch was meeting his employer’s legitimate expectations at the time he was fired. See, e.g., Hawkins v. PepsiCo, Inc., 203 F.3d 274, 280 (4th Cir.2000) (disregarding plaintiffs opinions of her own performance because “[i]t is the perception of the decision maker which is relevant, not the self-assessment of the plaintiff,” and finding the opinions of her co-workers “similarly close to irrelevant”) (citations and internal quotation marks omitted); O’Connor, 56 F.3d at 547 (holding that a review of an employee’s 1989 performance was irrelevant to a determination of whether his performance was satisfactory at the time of his termination in August of 1990); cf. King v. Rumsfeld, 328 F.3d 145, 149-50 (4th Cir.2003) (explaining that co-workers’ opinions might be relevant in certain situations, but not where they fail to establish what expectations the employer had and whether the employee met them).

Warch also claims he was in the upper third of performers in 2001 based on total savings by investigator. This evidence, even if true, is simply not enough to genuinely dispute the considerable evidence of Warch’s repeated failures and negative performance. Faced with such abundant evidence, Warch cannot create a genuine dispute concerning his prima facie case by cherry-picking the record to find one isolated instance where he arguably performed better than the average employee.

Nor has Warch shown that OCIC’s expectations were a “sham designed to hide the employer’s discriminatory purpose,” Brummett v. Lee Enterprises, Inc., 284 F.3d 742, 745 (7th Cir.2002), or were somehow not “legitimate.” Thus, even when viewing the evidence in the light most favorable to Warch, we conclude that no reasonable jury would find that he was meeting OCIC’s legitimate performance expectations.

B. Replacement

Warch’s prima facie case also fails because he did not produce sufficient evidence from which a reasonable jury could conclude that he was replaced by a substantially younger employee. See Causey, 162 F.3d at 802 & n. 3. Warch argues for a “flexible analysis” of this element of the prima facie case, claiming that his “replacement” could have been at any location. He also contends that, even if no one *519 physically replaced him, he can satisfy this element because he was a victim of a discriminatory reduction in force.

Warch initially admitted that he was not replaced by anyone. J.A. 90 (“I was not replaced by anybody. [My position] was eliminated.”); J.A. 262 (statement in post-termination job application). The EEOC similarly explained that Warch “was disciplined prior to his discharge and not replaced.” J.A. 302. 1 Nevertheless, Warch argues that David Hasler, one of Warch’s supervisors, suggested that a “ ‘replacement’ for the SIU unit might be at any location.” Brief of Appellant at 40. Hasler also testified, however, that Warch’s work “was probably spread out among various investigators” after he was let go. J.A. 166. Warch has failed to present any evidence genuinely disputing this fact, and, in light of his own admissions, he cannot prevail on his argument that he was replaced at all, much less by a substantially younger employee.

Warch submits that Montgomery used the sale of OCIC’s New Jersey unit to substantially reduce the number of older employees. Warch points to Causey v. Balog, which stated that a plaintiff terminated as part of a reduction in force “could potentially satisfy the fourth element of a prima facie case by introducing other probative evidence that indicates the employer did not treat age and race neutrally when making its decision.” 162 F.3d at 802.

When OCIC terminated Warch in 2002, it had recently concluded the sale of its line of New Jersey personal automobile insurance policies. OCIC’s David Hasler admitted that over time there would probably be less of a need for investigators in Voorhees, New Jersey. Warch argues that OCIC used its need to reduce investigators as a means to eliminate older workers and points to statistics he contends show a discriminatory pattern. He also submits that “logic and fairness” required the termination of two younger New Jersey employees who were less experienced and needed more guidance.

There are several problems with Warch’s argument. Perhaps most importantly, Warch has presented no evidence that he was actually terminated as a part of a reduction in force. OCIC never told Warch he was “downsized” or “laid off’ because of a loss of business or budgetary needs. Cf. Causey, 162 F.3d at 800 (explaining that Causey’s employer told him that his position was being abolished due to budgetary constraints). In fact, OCIC’s Hasler concluded that, because there were still previously issued policies in place, “there will still be claims on those, and there were open claims on those at the time, so it’s hard to estimate how long it would take for all of that business to roll off.” J.A. 182. Moreover, Warch’s statement that two younger New Jersey investigators should have been terminated instead of him completely ignores his own long string of performance problems.

Warch’s argument is based on nothing more than speculation concerning statistically insignificant information. For example, Warch points to numbers showing the firing, retirement, and termination from his unit of selected investigators over the age of forty. His evidence, however, says nothing about how the turnover rate of over-forty employees differed from that of under-forty employees. A reasonable juror simply could not draw the inferences of age discrimination that Warch claims from much of the statistical evidence he submits. For example, Warch lists the names *520 of six investigators he alleges were over the age of 40 who have either been fired, retired with a company recommendation not to hire, or terminated per mutual agreement. But, one of the listed employees was not 40 when terminated, and two of the remaining investigators left OCIC before control of Warch’s unit was assumed by Montgomery, allegedly the source of the discriminatory animus. Thus, the evidence is simply not probative enough to create a genuine issue of fact. See Fisher v. Asheville-Buncombe Tech. Community College, 857 F.Supp. 465, 470 (WD.N.C.1993) (finding plaintiffs statistical evidence “derived from too small a sample” and “simply too speculative” to be probative of alleged age discrimination).

In light of Warch’s own admissions that he was not replaced, OCIC’s well-documented problems with Warch’s performance, and the absence of any evidence showing that he was the victim of a reduction in force, much less a discriminatory one, no reasonable jury could find that Warch could establish the fourth element of his prima facie case.

IV.

Under a “mixed motive” approach, Warch can establish a claim of age discrimination by demonstrating that his age motivated OCIC’s decision to terminate him, even though it was not the sole motivating factor. See Hill, 354 F.3d at 284. The district court observed that there is some question as to whether a plaintiff claiming age discrimination can use circumstantial evidence when proceeding under a mixed-motive framework. See id. at 285 n. 2; Mereish, 359 F.3d at 339-40 (expressing doubt as to whether the more lenient standard of proof allowed in Title VII claims applied to ADEA claims, but deciding that, even under this more lenient standard, the appellants’ claims failed). Like the district court and the court in Mereish, we conclude that Wareh’s claim fails even when both direct and circumstantial evidence are considered.

Direct evidence must be “evidence of conduct or statements that both reflect directly the alleged discriminatory attitude and that bear directly on the contested employment decision.” Taylor v. Virginia Union Univ., 193 F.3d 219, 232 (4th Cir.1999) (en banc) (citation and internal quotation marks omitted). Even if there is a statement that reflects a discriminatory attitude, it must have a nexus with the adverse employment action. See Brinkley, 180 F.3d at 608 (“To survive summary judgment on the basis of direct and indirect evidence, Brinkley must produce evidence that clearly indicates a discriminatory attitude at the workplace and must illustrate a nexus between that negative attitude and the employment action.”).

The direct evidence presented by Warch on appeal consists of the comments of Montgomery and Robert Burgess, another OCIC employee, that a job candidate who happened to be similar in age and experience to Warch would have a hard time getting a job because “hiring people at that age, they didn’t get the work out of them that they did younger people.” J.A. 211. Warch presents no evidence showing this comment was more than an isolated event or that it had any nexus with the decision to terminate him. 2

*521 As to his circumstantial evidence, Warch claims that Montgomery purposefully rigged a performance audit that Warch originally passed so that it would instead show that he failed. This evidence, however, does not create any inference that age was a motivating factor in Warch’s termination. Furthermore, there is no evidence that the other well-documented performance problems Warch encountered were somehow linked to the failed audit score.

Warch also argues that his statistical evidence revealed a pattern of unlawful age discrimination, and alleged that two older employees were “forced out” around the same time that Montgomery made his age-biased comments. On this argument, the district court properly concluded that Warch had not presented “probative circumstantial evidence to show that OCIC acted with discriminatory animus.” J.A. 59. Warch’s statistics do not show complete information about where the data came from, how many investigators OCIC employed, who supervised them, and, most importantly, as the district court observed, “how many investigators under 40 have left the company during the same time period.” J.A. 59. Any suggestion that the two older employees who left OCIC around the same time as Montgomery’s alleged comments were victims of age discrimination is pure speculation and, without more evidence, simply cannot create an inference of illegal age discrimination.

Accordingly, even when we consider Warch’s circumstantial evidence, he fails to create a genuine dispute on his claim under the mixed motive approach. 3

V.

Under either his pretext approach or his mixed motive approach, Warch failed to create a genuine dispute that he was the victim of illegal age discrimination. As a result, we affirm the decision of the district court.

AFFIRMED

Notes

1

. The district court ruled that this letter was admissible over Warch's hearsay objection. This ruling is not challenged on appeal.

2

. Warch also submitted to the trial court statements allegedly made by Montgomery to Frank Kortyka, an OCIC employee, that he wanted to eliminate older workers and have a younger staff. Warch has not relied on these statements on appeal. As to all of the age-related comments Warch introduced, the district court explained that the exact statements, when they were made, and their context were unclear from the record. The record was clear, however, that none of the *521 statements were directed to Warch or referred directly to him.

3

. Our ruling that Warch failed to submit sufficient evidence on his claim under either the pretext or mixed motive approaches makes it unnecessary for us to address the parties’ arguments concerning who was the “deci-sionmaker.”