CARDAMONE, Circuit Judge:

Christopher Graham, a black employee of the Long Island Rail Road (LIRR), appeals from the grant of summary judgment in the United States District Court for the Eastern District of New York (Trager, J.) entered February 12, 1999 in favor of LIRR. The district court dismissed plaintiff's Title VII claim, which alleged the railroad discriminated on the basis of race when it terminated him from its employ. In a paraphrase of George Orwell's observation that all employees are equal, but some are more equal than others, plaintiff complains that LIRR treated white employees better than it treated him. Even-handed justice in the workplace surely means that an employer should deal with all its employees equally. Because we think there are some material questions of fact as to whether there was equal treatment in this case, we reverse and remand.

BACKGROUND

The LIRR hired Graham in 1975 as a car appearance maintainer and six years later promoted him to the position of car repairman. Graham was not a model employee of the LIRR. He was suspended three times — in 1983 for loafing, failure to follow a direct order, and leaving his job assignment without permission; the following year for disobeying a supervisor's orders and being off assignment without permission; and in 1988 for sleeping while on duty.

The LIRR issued a revised policy in May 1987 entitled "Control of Alcohol and Drug Use," under which a supervisor with reasonable suspicion that an employee was under the influence of drugs or alcohol could order the employee to take a urine test. On August 3, 1988, Chuck Mabie, the General Foreman of the Maintenance of Equipment Car Shop where Graham worked, ordered plaintiff to submit to a urine test. The sample tested positive for cocaine. LIRR held a disciplinary trial at which plaintiff was found guilty of drug use. He was dismissed on September 21, 1988 without being given a "last chance waiver."

The last chance waiver (or last chance agreement) is the means by which LIRR allows a worker who fails a substance abuse test to be reinstated after passing a fitness exam, and on condition that he submit to future random drug or alcohol testing at LIRR's request. Any subsequent positive test or failure to submit to a test is grounds for immediate dismissal. According to Graham, LIRR customarily offered last chance waivers to employees upon their first drug or alcohol offense.

For that reason, he appealed his dismissal to a mediation board consisting of representatives of LIRR and the United Transportation Union. On April 15, 1991 the board restored Graham to work under a last chance waiver. Two months later, on June 17, 1991, General Foreman Mabie again ordered Graham to submit to drug and alcohol testing. That same morning, Graham submitted a urine sample both to the LIRR medical center and to Dr. Abraham, his personal physician. LIRR had its sample tested by the Princeton Diagnostic Laboratories of America Inc. (Princeton Laboratory), which reported the presence of 61 mg/dl of alcohol in Graham's urine. Graham's own physician sent his sample to Roche Biomedical Laboratories (Roche), which reported no alcohol in Graham's urine.

On the basis of the Princeton Laboratory report, LIRR effectively terminated Graham's employment on June 19, 1991 by issuing him an "out of service" notice, and then held a disciplinary trial on August 6, 1991. At this disciplinary trial, Graham offered the results of the sample submitted to Dr. Abraham to show that the result found in the LIRR sample was inaccurate. In support of the Princeton Laboratory report, a doctor and a nurse from the LIRR medical center testified that the center had handled its sample of Graham's urine properly. Nurse Eileen Stocker also stated that Graham told her on the morning the urine sample was taken that he had consumed alcohol as recently as six hours earlier, a statement Graham denies making. The LIRR Maintenance of Equipment Department on October 17, 1991 upheld Graham's dismissal.

Graham unsuccessfully appealed to the mediation board, and after receiving a "right to sue" letter from the Equal Employment Opportunity Commission, filed the instant suit alleging that LIRR violated his rights when it disciplined him. His principal claim is that LIRR intentionally discriminated against him because he is black, in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-5. His complaint also alleged a violation of Title VII based on retaliation and state law causes of action for negligent infliction of emotional distress and fraud.

After discovery that produced the disciplinary records of other LIRR employees who had tested positive for alcohol or drugs and depositions of several LIRR officials, LIRR moved for summary judgment, contending that Graham could not make out a prima facie case of discrimination, and that his other claims lacked merit. Finding plaintiff produced insufficient evidence to establish a genuine issue of material fact that he was discharged under circumstances giving rise to an inference of discrimination, the district court granted LIRR's motion. In particular, the court observed that the disciplinary records revealed that black and non-black employees were treated alike in disciplinary proceedings involving alcohol or drug infractions. Individuals whom Graham alleged had been more favorably treated, it ruled, were not similarly situated. The record before us is in many respects confusing or incomplete, but we think it contains sufficient information set out clearly enough for our purposes, to decide the issues upon which we focus in this appeal.

DISCUSSION Standard of Review

We review a grant of summary judgment de novo applying the same standard as the district court. See Belfi v. Prendergast, 191 F.3d 129, 135 (2d Cir. 1999). Such relief should be granted by the district court only when it determines there is no genuine issue as to any material fact and the moving party is entitled to judgment as a matter of law. See Fed.R.Civ.P. 56(c); see also Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). In deciding the motion, the trial court must first resolve all ambiguities and draw all inferences in favor of the non-moving party, and then determine whether a rational jury could find for that party. See Gallo v. Prudential Residential Servs., Ltd. Partnership, 22 F.3d 1219, 1223-24 (2d Cir. 1994).

At the same time, the non-moving party must offer such proof as would allow a reasonable juror to return a verdict in his favor, see Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 256, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986), and only when that proof is slight is summary judgment appropriate, see Dister v. Continental Group, Inc., 859 F.2d 1108, 1114 (2d Cir. 1988). The trial court's function at this stage is to identify issues to be tried, not decide them. Summary judgment is sparingly used where intent and state of mind are at issue, see Montana v. First Fed. Sav. Loan Ass'n, 869 F.2d 100, 103 (2d Cir. 1989), because, as we have emphasized, careful scrutiny of the factual allegations may reveal circumstantial evidence to support the required inference of discrimination, see Belfi, 191 F.3d at 135; Chertkova v. Connecticut Gen. Life Ins. Co., 92 F.3d 81, 87 (2d Cir. 1996); Chambers v. TRM Copy Ctrs. Corp., 43 F.3d 29, 37 (2d Cir. 1994); Gallo, 22 F.3d at 1224.

Title VII I Prima Facie Case

Graham appeals only the grant of summary judgment dismissing his claim against LIRR for discrimination on the basis of race in violation of Title VII. Under the burden-shifting analysis outlined in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), a plaintiff first must establish a prima facie case of discrimination based on race. See Texas Dep't of Community Affairs v. Burdine, 450 U.S. 248, 252-53, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981); McDonnell Douglas, 411 U.S. at 802, 93 S.Ct. 1817. If he does so, the burden shifts to the employer to articulate a legitimate, non-discriminatory reason for the employee's dismissal. See McDonnell Douglas, 411 U.S. at 802, 93 S.Ct. 1817. If such a reason is proffered, the burden shifts back to the plaintiff to prove that discrimination was the real reason for the employment action. See id. at 804; Fisher v. Vassar College, 114 F.3d 1332, 1336 (2d Cir. 1997) (en banc), cert. denied, 522 U.S. 1075, 118 S.Ct. 851, 139 L.Ed.2d 752 (1998).

To meet the burden of production required for a prima facie case of discrimination, a plaintiff must show that he (1) is a member of a protected class; (2) was performing his duties satisfactorily; (3) was discharged; and that (4) his discharge occurred under circumstances giving rise to an inference of discrimination on the basis of his membership in the protected class. See Chambers, 43 F.3d at 37. The burden a plaintiff, alleging that he was discriminated against by his employer, carries to survive a summary judgment motion at the prima facie stage is a minimal one. See St. Mary's Honor Ctr. v. Hicks, 509 U.S. 502, 506, 113 S.Ct. 2742, 125 L.Ed.2d 407 (1993); Fisher, 114 F.3d at 1340 n. 7.

In this case, the parties agree that Graham can establish the first three elements of his prima facie case, disputing only whether Graham's discharge took place under circumstances giving rise to an inference of discrimination. A plaintiff may raise such an inference by showing that the employer subjected him to disparate treatment, that is, treated him less favorably than a similarly situated employee outside his protected group. See International Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977); Norville v. Staten Island Univ. Hosp., 196 F.3d 89, 95 (2d Cir. 1999) (inference of discrimination arises when individual of one race treated less favorably than those of another race who are similarly situated); Shumway v. United Parcel Serv., Inc., 118 F.3d 60, 63 (2d Cir. 1997) (woman may prove inference of discrimination by showing similarly situated man treated differently); Montana, 869 F.2d at 106 (inference of discrimination arises when female employees are treated less favorably than comparable male employees).

A. Similarly Situated Employees for Title VII Purposes

Whether two employees are similarly situated ordinarily presents a question of fact for the jury. See Taylor v. Brentwood Union Free Sch. Dist., 143 F.3d 679, 684 (2d Cir. 1998) (explaining that jury was asked to decide whether the plaintiff was treated differently from similarly situated white employees), cert. denied, 525 U.S. 1139, 119 S.Ct. 1027, 143 L.Ed.2d 37 (1999); Hargett v. National Westminster Bank, USA, 78 F.3d 836, 839-40 (2d Cir. 1996) (noting that jury was asked to decide "similarly situated" issue); cf. Tomka v. Seiler Corp., 66 F.3d 1295, 1312 n. 11 (2d Cir. 1995) (whether two positions are "substantially equal" for Equal Pay Act claim is a question of fact). The district court's discussion of this subject set forth no legal standard for determining who is a similarly situated employee, and instead confined itself to pointing out factual differences between Graham and the white employees with whom he sought to compare himself. To place the facts of this case in proper context therefore, we must review the standard for deciding whether employees are similarly situated.

When considering whether a plaintiff has raised an inference of discrimination by showing that she was subjected to disparate treatment, we have said that the plaintiff must show she was "similarly situated in all material respects" to the individuals with whom she seeks to compare herself. Shumway, 118 F.3d at 64. In Shumway, we adopted without discussion the Sixth Circuit's "all material respects" standard. See id. (citing Mitchell v. Toledo Hosp., 964 F.2d 577, 583 (6th Cir. 1992)). We then went on to find that the female plaintiff in Shumway, who claimed she was forced to resign for violating an anti-fraternization policy, was not similarly situated to the male co-workers she claimed violated the same policy but who were not disciplined. These co-workers and plaintiff had different supervisors, and at least one male manager who reported to the same supervisor as plaintiff was interviewed and ultimately resigned for violation of the policy. Moreover, plaintiff presented no evidence that the alleged conduct of the comparators was similar to her conduct, which involved a long-term relationship with an hourly employee, harassing behavior and lying. For these reasons, with little elaboration of what it meant to be similarly situated "in all material respects" and with no additional citation to Mitchell, we determined that plaintiff and the comparators were not similarly situated. See Shumway, 118 F.3d at 64.

We have said since that to satisfy Shumway's "all material respects" standard for being similarly situated, a plaintiff must show that her co-employees were subject to the same performance evaluation and discipline standards. See Norville, 196 F.3d at 96. In addition, the standard we used in Shumway requires plaintiff to show that similarly situated employees who went undisciplined engaged in comparable conduct. See 118 F.3d at 64. That an employee's conduct need not be identical to that of another for the two to be similarly situated is also reflected in the language of McDonnell Douglas, where the Supreme Court used the phrase "comparable seriousness" to identify conduct that might help to support an inference of discrimination. 411 U.S. at 804, 93 S.Ct. 1817; see also McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273, 283 n. 11, 96 S.Ct. 2574, 49 L.Ed.2d 493 (1976) ("precise equivalence in culpability between employees is not the ultimate question"); Ricks v. Riverwood Int'l Corp., 38 F.3d 1016, 1019 (8th Cir. 1994) (employing comparable seriousness standard); McAlester v. United Air Lines, Inc., 851 F.2d 1249, 1261 (10th Cir. 1988) (stating that in a disparate treatment case the fact that other employees did not commit the exact same offense as the plaintiff does not prohibit consideration of their testimony as long as their acts were of comparable seriousness).

The Sixth Circuit itself has backed away from a rigid interpretation of Mitchell, cautioning that such precedent should not "invite a comparison between the employment status of the plaintiff and other employees in every single aspect of their employment. . . . Courts should not assume . . . that the specific factors discussed in Mitchell are relevant factors in cases arising under different circumstances." Ercegovich v. Goodyear Tire Rubber Co., 154 F.3d 344, 352 (6th Cir. 1998). Instead, courts should make an "independent determination" of the factors relevant to each case. Id.

What constitutes "all material respects" therefore varies somewhat from case to case and, as we recognized in Norville, must be judged based on (1) whether the plaintiff and those he maintains were similarly situated were subject to the same workplace standards and (2) whether the conduct for which the employer imposed discipline was of comparable seriousness. See 196 F.3d at 96. In other words, there should be an "objectively identifiable basis for comparability." Cherry v. American Tel. Tel. Co., 47 F.3d 225, 229 (7th Cir. 1995). Hence, the standard for comparing conduct requires a reasonably close resemblance of the facts and circumstances of plaintiff's and comparator's cases, rather than a showing that both cases are identical. See Conward v. Cambridge Sch. Comm., 171 F.3d 12, 20 (1st Cir. 1999) (explaining that "[r]easonableness is the touchstone" and recognizing that "the plaintiff's case and the comparison cases . . . need not be perfect replicas").

In Lynn v. Deaconess Med. Ctr.-West Campus, 160 F.3d 484, 488 (8th Cir. 1998), the Eighth Circuit explained that a requirement that employees "engage in the exact same offense" in order to find them similarly situated is unwise because it would produce "a scenario where evidence of favorable treatment of an employee who has committed a different but more serious, perhaps even criminal offense, could never be relevant to prove discrimination." The determination that two acts are of comparable seriousness requires — in addition to an examination of the acts — an examination of the context and surrounding circumstances in which those acts are evaluated. See Hargett, 78 F.3d at 839 (noting that "determination of whether acts are of `comparable seriousness'" does not require a "myopic examination of the acts alone").

We turn then to whether Graham and his fellow employees who received last chance waivers were similarly situated and whether plaintiff therefore satisfies the fourth element of his prima facie case.

B. Last Chance Waivers

Graham raised three arguments in the district court to support his position that LIRR gave preferential treatment to non-black employees in administering its drug and alcohol disciplinary policies. First, he contends that at the time of his first dismissal LIRR maintained a policy of automatically giving last chance waivers for first time drug or alcohol offenders, and that while he did not receive one, a white employee did. Second, he asserts that he was subjected to urine testing only six weeks after his reinstatement, while white employees reinstated on last chance waivers were not tested so frequently. Third, he avers that similarly situated non-black employees received multiple last chance waivers before dismissal, while he was fired for allegedly violating the only waiver he ever got. The district court rejected all three of these contentions. On appeal plaintiff does not renew his second argument, which is accordingly deemed waived. See Norton v. Sam's Club, 145 F.3d 114, 117 (2d Cir. 1998) (issues not argued in briefs are waived and will not be addressed on appeal). We therefore analyze his first and third challenges.

1. Does Failing to Give Graham a Last Chance Waiver On First Offense Show Discrimination?

Beyond his bare assertion, Graham offers no proof that LIRR gave last chance waivers automatically for first time drug and alcohol offenders as a matter of policy or custom. Instead, the only evidence he offers is that one white employee, Carl DiPersia, tested positive for cocaine use in 1987 and received a last chance waiver after his first violation without having to file a grievance. Such isolated evidence is insufficient to make a showing that LIRR maintained a policy or custom.

In addition, the chief trial and investigative officer in Graham's department, Ralph J. Domenici, denied that such a policy was in effect during his tenure in the department between 1985 and 1991. Domenici presented evidence that three white car repairmen were fired without a last chance waiver after they failed their first urine test, while one black employee who failed his first test was given a last chance waiver. Consequently, even if plaintiff's claim were allowed to proceed past the prima facie stage, this unrebutted proof from defendant would support a finding on summary judgment that no categorical policy of giving last chance waivers for first offenses was in place when Graham was first dismissed.

Accordingly, Graham's first challenge that the denial of a last chance waiver for his initial offense in 1988 constitutes circumstantial evidence of impermissible discrimination in 1991 cannot succeed.

2. Was Dismissal After Only One Last Chance Waiver Disparate Treatment?

We turn to plaintiff's third argument. Graham offered evidence that at least two white employees who also tested positive for drugs or alcohol received multiple last chance waivers while he was fired after being given only one. He declares that A.J. Elmendorf, who is white, received a last chance waiver in 1992 for testing positive for cocaine, and then, after testing positive again in March 1993, was given a second last chance. He also contends that DiPersia was given three last chance waivers, in 1987, 1988, and 1992.

Before focusing on this evidence however, the district court first concluded in light of Domenici's unrefuted evidence that plaintiff failed to make a prima facie showing. Yet since the burden at this stage of the McDonnell Douglas analysis rests solely on Graham, it was premature to consider LIRR's evidence. The burden to produce evidence does not shift to a defendant in a discrimination claim until a prima facie case has been established. See Hicks, 509 U.S. at 506-07, 113 S.Ct. 2742 (explaining that under McDonnell Douglas, the production by the plaintiff of a prima facie case then shifts the burden to the defendant to "produc[e] evidence" that the adverse employment action was taken for a "legitimate, non-discriminatory reason"). Accordingly, we think only Graham's evidence should be considered when deciding whether plaintiff has met his initial burden.

In this regard, the district court concluded that Elmendorf, DiPersia and plaintiff were not similarly situated. It reasoned that Elmendorf's second last chance waiver for a drug violation occurred after he was demoted from the position held when he received his first last chance waiver, such that his second "last chance" was only his first "last chance" while employed as a car repairman. DiPersia, on the other hand, received an initial last chance waiver after a positive test for drugs and alcohol, but subsequently received two more last chance waivers for excessive absenteeism and was dismissed after violating his third last chance waiver. Thus, LIRR urges that these other last chance waivers were based on different conduct.

As previously stated, to be similarly situated, Graham, Elmendorf, and DiPersia must have been subject to the same disciplinary standards and have engaged in conduct of comparable seriousness. See, e.g., Norville, 196 F.3d at 96. According to the disciplinary letters in the record, the LIRR last chance waivers warned that

any use or possession of controlled substances is absolutely prohibited. . . . I understand that any future violation of the Policy on drugs and alcohol, or of the terms of this agreement, will result in my dismissal.

Furthermore, the following conditions will control:

1. I will maintain a satisfactory attendance record.

2. I will be evaluated by the Employee Assistance Program. . . . If I am found to be in non-compliance with the requirements of EAP I will be dismissed from service. . . .

(emphasis in original document). Nowhere does the policy refer to the position held by the employee. Rather, the language stipulating dismissal for a violation is categorical.

Consequently, the trial court's distinction that Elmendorf's second last chance came after he was demoted is immaterial when viewed against the mandatory dismissal language of the LIRR policy. See Cherry, 47 F.3d at 229 (postulating that company-wide policy uniformly applicable to all members of a group may be basis for comparing employees, regardless of particular job being performed). Hence, it would be reasonable to find that Graham and Elmendorf were similarly situated after receiving a last chance waiver stipulating that "any future violation of the Policy on drugs and alcohol . . . will result in my dismissal," despite the fact that Elmendorf also received a demotion along with his first last chance waiver. That Elmendorf, a white employee, then received a second last chance waiver upon his second violation of the Policy, while Graham, a black employee, did not, raises an inference that Graham's dismissal for the second violation was a result of race discrimination.

Further, the district court's conclusion that DiPersia's second and third last chances were not for drug or alcohol violations and therefore that he and Graham were not similarly situated improperly resolved a fact question. While the conduct charged against plaintiff and DiPersia clearly was not identical, both employees nevertheless violated the express terms of the waiver they each signed. DiPersia had agreed "that any future violation . . . of the terms of this agreement, will result in my dismissal." One such term was the maintenance of a satisfactory attendance record. On the face of the waivers, dismissal follows automatically from a violation of the last chance agreement.

A rational jury could find that everyone who has been given last chance waivers listing the same conditions mandating dismissal is subject to the same standards, i.e., that any violation of the last chance agreement would result in dismissal. A jury could further rationally find that a violation for excessive absenteeism and one for alcohol use were of comparable seriousness in light of the categorical standard contained in the last chance agreements. See Lynn, 160 F.3d at 488 (finding that employee dismissed for sleeping on the job and employee dismissed for failing to assist a patient in attaching a piece of therapeutic equipment and failing properly to fill out forms engaged in conduct of comparable seriousness).

In deciding whether a prima facie case was established, the district court applied an unduly inflexible standard in dismissing Graham's third claim. By determining that Graham, Elmendorf, and DiPersia were not similarly situated, the trial court erred. Whether these employees were similarly situated presented a material question of fact. Were a jury to find that Graham was similarly situated to DiPersia and Elmendorf, it could conclude that the disparate treatment he received completes a prima facie showing of discrimination under Title VII. Thus, the dismissal of plaintiff's complaint for failure to make out a prima facie case on the last chance waiver issue must be reversed.

II Pretext

Assuming arguendo that LIRR could meet its burden to put forth a legitimate, non-discriminatory reason for terminating plaintiff's employment, we examine whether Graham has proffered sufficient evidence to survive summary judgment on the issue of pretext.

A. Last Chance Waiver

A showing that similarly situated employees belonging to a different racial group received more favorable treatment can also serve as evidence that the employer's proffered legitimate, non-discriminatory reason for the adverse job action was a pretext for racial discrimination. See Hargett, 78 F.3d at 839. Were a jury to find on these facts that there was disparate treatment in LIRR's disciplining of plaintiff when compared to similarly situated employees, it could also find that LIRR's proffer of a non-discriminatory reason for Graham's dismissal was pretextual. See Greenway v. Buffalo Hilton Hotel, 143 F.3d 47, 51 (2d Cir. 1998) (inconsistent application of company disciplinary policy was sufficient for jury to find that employer's defense was pretext for discrimination). Such would preclude the grant of summary judgment to the employer under the third prong of McDonnell Douglas.

Nor is Graham's cause of action undermined by LIRR's evidence. In our earlier discussion of plaintiff's prima facie case on this claim, we noted that the district court credited evidence from LIRR that similarly situated non-black employees were treated in the same manner as plaintiff and that one other black employee was treated more favorably. Although the record does not clearly substantiate such a finding, even if it were true, it would be insufficient to sustain summary judgment for LIRR. Since Title VII's principal focus is on protecting individuals, rather than a protected class as a whole, an employer may not escape liability for discriminating against a given employee on the basis of race simply because it can prove it treated other members of the employee's group favorably. See Connecticut v. Teal, 457 U.S. 440, 453-55, 102 S.Ct. 2525, 73 L.Ed.2d 130 (1982); EEOC v. Staten Island Sav. Bank, 207 F.3d 144, 151 n. 5 (2d Cir. 2000). Only where overwhelming evidence exists that an employer acted for reasons other than discrimination will the proof submitted by the plaintiff as to similarly situated employees be negated. Cf. Montana, 869 F.2d at 106-07 (finding that any inference of discrimination plaintiff claimed flowed from the retention of the six men while she was dismissed was nullified by evidence that a total of 45 managers, both men and women, were discharged pursuant to a reduction-in-force, while four men and four women were retained). LIRR's proof therefore serves only to create a question of fact.

B. Results of Drug Test

Graham additionally insists that even if his dismissal were based on the facially non-discriminatory finding that he had alcohol in his system on the morning of June 17, 1991, the factual basis for this finding was incorrect. He maintains that LIRR's allegation that he failed the test was merely pretext for a discriminatory discharge. The district court did not credit plaintiff's assertion of pretext because it concluded that Graham's evidence with respect to the drug test result could not support a reasonable jury's verdict in his favor.

We agree with the district court's conclusion on this issue. To sustain its employment decision, the LIRR need not show that the result of the drug test was actually correct, but only that it reasonably relied on the Princeton Laboratory's test result. Even assuming a jury were later to find that the laboratory's drug test was either improperly administered or inaccurate does not change the correctness of the district court's conclusion. The key question is whether it was reasonable for the employer to rely on the test result in making its employment decision. Considering that the record contains no evidence to suggest LIRR purposefully interfered with the testing process in order to discriminate against plaintiff, a reasonable jury could conclude only that it was reasonable for the LIRR to rely on the test that showed plaintiff positive for alcohol in his blood. Reasonable reliance by an employer on a laboratory test, even where misplaced, does not provide any basis for a jury to find pretext. See Fisher, 114 F.3d at 1338. As a consequence, plaintiff's claim that the Princeton Laboratory's test result was a pretext for a discriminatory discharge does not raise a triable issue of fact, and that claim was properly dismissed.

CONCLUSION

The district court's conclusions regarding the similarity of Graham, Elmendorf and DiPersia improperly resolved factual questions. Since we also find questions of fact with respect to plaintiff's ultimate burden on the issue of pretext, the grant of summary judgment is reversed, and the case remanded to the district court for further proceedings limited solely to the merits of plaintiff's claim pertaining to multiple last chance waivers.