Pretext at Summary Judgment Requires Affirmative Evidence—Credibility Attacks, Speculative Comparator Proof, and “Stellar Performance” Are Insufficient Under the NMHRA

Case: Sousa v. Chipotle Services, LLC, No. 25-2008 (10th Cir. Mar. 2, 2026) (published)
Court: United States Court of Appeals for the Tenth Circuit
Claim: Age discrimination under the New Mexico Human Rights Act, N.M. Stat. Ann. § 28-1-7(A)
Disposition: Summary judgment for employer affirmed (insufficient evidence of pretext)

1. Introduction

Donald Sousa, a 54-year-old field leader for Chipotle in New Mexico, alleged Chipotle terminated him because of age in violation of the New Mexico Human Rights Act. Chipotle asserted it terminated him for food-safety and cleanliness failures—most notably a serious cockroach infestation and “extremely dirty” conditions at Store 2952, plus multiple internal site audits showing other restaurants in his patch failed cleanliness standards.

The central appellate issue was narrow: at the third step of the McDonnell Douglas framework, did Sousa produce enough evidence of pretext to allow a reasonable jury to infer age discrimination? He relied primarily on (i) allegedly “fabricated” final-warning documents delivered at termination, (ii) comparator evidence involving younger field leaders, (iii) the termination of another older field leader, and (iv) his prior strong performance.

2. Summary of the Opinion

The Tenth Circuit affirmed summary judgment for Chipotle. Applying federal summary-judgment procedure in a diversity case, the court held Sousa failed to produce sufficient evidence that Chipotle’s stated reasons were “unworthy of credence.” Key holdings:

  • Final warnings: Absent affirmative evidence contradicting the decisionmaker’s explanation (HR-directed documentation), a plaintiff cannot avoid summary judgment by arguing a jury might disbelieve the employer’s witness.
  • Comparators: Courts may decide at summary judgment whether comparator evidence is sufficiently similar to support an inference of discrimination; the similarity analysis is not automatically reserved for the jury.
  • Decisionmaker knowledge matters: Comparator incidents not known to the decisionmaker do not support pretext.
  • Pattern evidence: A two-termination “pattern” involving older employees did not show pretext where nondiscriminatory explanations remained and comparability was not established.
  • Past performance: Strong historical performance does not establish pretext where the employer honestly believed current food-safety failures warranted termination and no contrary policy/practice evidence was shown.

3. Analysis

3.1 Precedents Cited (and How They Drove the Outcome)

Governing summary-judgment method in diversity cases

  • Padhiar v. State Farm Mut. Auto. Ins. Co., 479 F.3d 727 (10th Cir. 2007): The court anchored the procedural/substantive split—state law governs NMHRA substance, but federal law governs the Rule 56 summary-judgment standard. This foreclosed Sousa’s attempt to rely on a more plaintiff-friendly reading of New Mexico summary-judgment practice.
  • Walkingstick Dixon v. Oklahoma ex rel. Reg'l Univ. Sys. of Okla. Bd. of Regents, 125 F.4th 1321 (10th Cir. 2025), and Anderson v. Liberty Lobby, Inc., 477 U.S. 242 (1986): Provided the core Rule 56 framework (genuine dispute of material fact; reasonable jury standard) and the articulation of pretext as whether the rationale is “unworthy of credence.”
  • Litzsinger v. Adams Cty. Coroner's Office, 25 F.4th 1280 (10th Cir. 2022): Cited for the principle that facts are viewed in the light most favorable to the nonmovant unless contradicted by the record.

NMHRA discrimination framework and relationship to federal law

  • Cates v. Regents of N.M. Inst. of Mining & Tech., 954 P.2d 65 (N.M. 1998), and McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973): Established that New Mexico applies the federal burden-shifting model to NMHRA claims lacking direct evidence.
  • Gonzalez v. N.M. Dep't of Health, 11 P.3d 550 (N.M. 2000), quoting Smith v. FDC Corp., 787 P.2d 433 (N.M. 1990): The panel recognized New Mexico’s caution that federal interpretations do not bind New Mexico law, but still used federal authority for guidance—particularly on the pretext inquiry at summary judgment.
  • Kendrick v. Penske Transp. Servs., Inc., 220 F.3d 1220 (10th Cir. 2000): Supplied the canonical three routes to show pretext (falsity; deviation from written policy; deviation from unwritten policy/practice, often via comparator evidence).

“Honest belief” and decisionmaker-perspective limitations

  • EEOC v. C.R. England, Inc., 644 F.3d 1028 (10th Cir. 2011): The court emphasized that pretext is judged based on the facts as they appeared to the decisionmaker—not the plaintiff’s subjective view.
  • Young v. Dillon Cos., 468 F.3d 1243 (10th Cir. 2006): Reinforced that the key question is not whether the decision was wise or correct but whether the employer honestly believed its reasons and acted in good faith.

Credibility attacks require counter-evidence

  • Helget v. City of Hays, 844 F.3d 1216 (10th Cir. 2017): A centerpiece of the opinion. The panel used Helget to hold Sousa could not survive summary judgment by arguing the jury might disbelieve Hannan’s explanation for the “Final Warning” paperwork; Sousa needed affirmative admissible evidence contradicting it.
  • Nat'l Am. Ins. v. Am. Re-Ins., 358 F.3d 736 (10th Cir. 2004): Confirmed that mere attacks on the credibility of the movant’s evidence do not create a triable issue absent contrary evidence.
  • Simmons v. Sykes Enters., Inc., 647 F.3d 943 (10th Cir. 2011): Supported the court’s approach to conduct de novo review rather than separately addressing allegations that the district court weighed evidence.

Comparator “similarly situated” evidence as a summary-judgment gatekeeping function

  • Riggs v. AirTran Airways, Inc., 497 F.3d 1108 (10th Cir. 2007): The opinion relied heavily on Riggs to reject the argument that comparator similarity is always for the jury. Courts must assess whether the plaintiff has produced enough evidence of similarity to permit an inference of discrimination; otherwise, summary judgment is proper.
  • George v. Leavitt, 407 F.3d 405 (D.C. Cir. 2005), and Mandell v. Cty. of Suffolk, 316 F.3d 368 (2d Cir. 2003): Quoted within the Riggs discussion to frame the “ordinarily a jury question” idea while underscoring the court’s threshold role at summary judgment.
  • Juneau v. Intel Corp., 127 P.3d 548 (N.M. 2005): Raised by Sousa to argue New Mexico would send similarity questions to the jury. The court rejected his reliance: even if state practice differed, Padhiar makes clear federal summary-judgment procedure controls in diversity.

Progressive discipline and policy deviation arguments

  • Lobato v. N.M. Env't Dep't, 733 F.3d 1283 (10th Cir. 2013): Used to reinforce that where written policies do not require progressive discipline, termination without written warnings does not itself imply pretext.

Speculation about trial proof and pattern evidence standards

  • Johnson v. Weld Cty., 594 F.3d 1202 (10th Cir. 2010): The court rejected Sousa’s assertion that he “anticipates” presenting evidence at trial about the severity of EcoSure failures; speculation cannot substitute for admissible summary-judgment evidence.
  • Pippin v. Burlington Res. Oil & Gas Co., 440 F.3d 1186 (10th Cir. 2006): The panel invoked Pippin to dispose of the “pattern” argument: alleged patterns must eliminate nondiscriminatory explanations by showing disparate treatment among comparable individuals.

3.2 Legal Reasoning (How the Court Got to “No Pretext”)

  1. Framework accepted; only pretext contested. The parties did not dispute prima facie case or Chipotle’s legitimate, nondiscriminatory reason. The appellate analysis therefore lived entirely at McDonnell Douglas step three.
  2. “Final Warning” paperwork did not create a triable credibility dispute. Sousa’s theory was that Hannan “cooked up” final warnings to fabricate progressive discipline. The court found the documents did not falsely suggest prior written warnings (the termination notice cited verbal warnings and made no claim of earlier written warnings), and—critically—Sousa produced no admissible evidence contradicting Hannan’s HR-guidance explanation. Under Helget v. City of Hays and Nat'l Am. Ins. v. Am. Re-Ins., that ends the issue at summary judgment.
  3. Comparator analysis failed on “comparable seriousness” and “decisionmaker knowledge.”
    • Chaparro: Evidence was limited to Store 2952 failing at least one EcoSure audit after Sousa’s termination (“Possibly” more). Sousa provided no admissible evidence that one failed EcoSure audit is comparable to (a) a persistent infestation observed by Hannan and (b) multiple separate restaurants failing internal cleanliness standards in the same week, supported by audit photos.
    • Sanchez: Although the record contained troubling details about pests/cleanliness in several stores, the court treated as dispositive that the pretext inquiry is tethered to what Hannan knew at the time (EEOC v. C.R. England, Inc.). The record did not show Hannan’s awareness of most cited incidents; the evidence supported at most limited knowledge (e.g., floor-cleaning issues at Store 898; certain other isolated issues). That was insufficient to show disparate treatment of similarly situated employees known to have committed comparably serious violations.
    The court also rejected Sousa’s legal premise that similarity must always be left to the jury, reaffirming Riggs v. AirTran Airways, Inc. as a summary-judgment “gatekeeping” rule.
  4. Pattern evidence failed because comparability failed. Sousa invoked the termination of another older field leader, Tiffany Rodriguez, as evidence of age-based patterning. The court recognized nondiscriminatory grounds for Rodriguez’s termination: an “extraordinary” infestation involving thousands of cockroaches. Under Pippin v. Burlington Res. Oil & Gas Co., pattern evidence must eliminate nondiscriminatory explanations by showing disparate treatment among comparables; Sousa’s comparator proof (Chaparro/Sanchez) could not make Rodriguez’s situation comparable.
  5. Strong past performance did not rebut honest belief. The court treated Sousa’s “top-performing” recognition and high EcoSure scores as insufficient to show pretext. The question is not fairness or proportionality; it is whether Chipotle honestly believed current food-safety issues justified termination (Young v. Dillon Cos.) and whether the court should refrain from second-guessing discipline severity (Kendrick v. Penske Transp. Servs., Inc.).
  6. Totality still did not permit a reasonable inference of discrimination. Even aggregating all categories, the court concluded no rational factfinder could deem Chipotle’s rationale “unworthy of credence” under Walkingstick Dixon v. Oklahoma ex rel. Reg'l Univ. Sys. of Okla. Bd. of Regents.

3.3 Impact (What This Opinion Changes or Reinforces)

  • Reinforces an evidence-production discipline at pretext stage: Plaintiffs must counter key employer testimony with affirmative admissible evidence; mere “a jury could disbelieve them” arguments will not create a triable issue (explicitly applying Helget v. City of Hays in an employment pretext context).
  • Strengthens “decisionmaker knowledge” as a comparator filter: Evidence of other stores’ problems is not enough; plaintiffs must link those problems to what the decisionmaker knew when acting (EEOC v. C.R. England, Inc.), limiting “gotcha” comparator narratives built on information not shown to be in the decisionmaker’s universe.
  • Comparator disputes are triable only if similarity evidence clears a threshold: The court underscores Riggs v. AirTran Airways, Inc. as a practical gatekeeping rule: courts may (and must) decide whether comparator proof is sufficient to support a discrimination inference, rather than reflexively sending similarity disputes to a jury.
  • Diversity litigants cannot leverage state summary-judgment practice: By invoking Padhiar v. State Farm Mut. Auto. Ins. Co., the opinion signals that NMHRA plaintiffs in federal court must meet the federal Rule 56 evidentiary demands even if they believe state practice is more forgiving (the court rebuffed reliance on Juneau v. Intel Corp. for procedural purposes).
  • Documentation decisions are not inherently suspicious: Delivering “Final Warning” forms contemporaneously with termination—at least on this record and absent counter-proof—was treated as HR documentation rather than evidence of fabrication, potentially insulating certain HR “papering” practices from being characterized as pretext without stronger proof.

4. Complex Concepts Simplified

  • “McDonnell Douglas” framework: A three-step method for cases without direct discriminatory statements. (1) Plaintiff shows basic inference of discrimination; (2) employer gives a legitimate reason; (3) plaintiff must show that reason is a cover (pretext).
  • “Pretext”: Not merely that the employer was wrong or harsh, but that the stated reason is so weak/inconsistent that a jury could infer the real reason was unlawful discrimination.
  • “Honest belief” rule: If the employer honestly believed the stated reason and acted in good faith, courts generally will not treat the decision as discriminatory even if the employer’s conclusion was arguably mistaken.
  • “Similarly situated” comparator: Another employee who is alike in relevant ways (role, supervisor, standards) and engaged in misconduct of comparable seriousness, yet received better treatment—used to infer discrimination.
  • “Decisionmaker perspective” limitation: Facts that the decisionmaker did not know at the time typically cannot show that the stated reason was a lie; pretext is assessed based on what the decisionmaker understood when deciding.
  • Diversity jurisdiction and procedure: Even when state substantive law governs the claim (NMHRA), federal procedural law (Rule 56) governs how evidence must be presented at summary judgment in federal court.

5. Conclusion

Sousa v. Chipotle Services, LLC is a stringent evidentiary decision about the pretext stage of employment discrimination litigation under the NMHRA when the case is in federal court. The opinion’s practical teaching is that plaintiffs must build pretext with admissible proof that (i) contradicts key employer explanations, (ii) ties comparator misconduct to the decisionmaker’s knowledge, and (iii) demonstrates comparable seriousness sufficient to support an inference of discrimination. Without that, even evidence of strong past performance, unusual HR paperwork timing, or generalized “pattern” allegations will not prevent summary judgment.