ELCRA Discrimination: “Qualified” Must Be Assessed Independently of the Employer’s Stated Reasons; Workforce-Matched Termination Statistics Do Not Support an Inference of Bias

1. Introduction

In Oneka Spicer v. Harvard Maintenance, Incorporated (6th Cir. May 1, 2026) (unpublished), the Sixth Circuit affirmed summary judgment for an employer on Michigan Elliott-Larsen Civil Rights Act (“ELCRA”) claims alleging race discrimination, sex discrimination, and hostile work environment.

Plaintiff-Appellant Oneka Spicer, an African American woman, worked as a cleaner for Harvard Maintenance at a multi-building site. During her employment, tenants complained about her cleaning, and Harvard issued discipline culminating in an incident involving alleged unauthorized use of client property (sitting in a client chair to charge a phone) and alleged insubordination. Spicer’s employment was governed by a Collective Bargaining Agreement (“CBA”) with progressive discipline and a grievance process; “Last Chance Agreement” (“LCA”) practice was also discussed. After a grievance meeting, Harvard terminated Spicer, and the union did not pursue a CBA challenge.

The appeal raised two central questions: (1) whether Spicer produced sufficient circumstantial evidence to satisfy the McDonnell Douglas framework under ELCRA (including comparators and statistics), and (2) whether the alleged comments and conduct were “severe or pervasive” enough to constitute a hostile work environment.

2. Summary of the Opinion

  • Disposition: Affirmed summary judgment for Harvard on race discrimination, sex discrimination, and hostile work environment; affirmed denial of reconsideration.
  • Key holdings:
    • The district court improperly evaluated “qualification” by relying on Harvard’s asserted performance/discipline reasons, contrary to Sixth Circuit guidance; however, the error was harmless because Spicer failed the “inference of discrimination” element.
    • Comparator evidence failed because the proposed comparator’s misconduct (repeated minor performance issues) was not of “comparable seriousness” to Spicer’s admitted “major” rule violations.
    • Statistical evidence failed because termination rates closely tracked workforce composition; therefore, the statistics did not show a “significant disparity” or eliminate nondiscriminatory explanations.
    • Hostile work environment failed because the cited evidence was either unknown to Spicer during employment, speculative, or too isolated, and it did not alter the conditions of employment.
    • Reconsideration was properly denied under Rule 59(e) principles; the local-rule “palpable defect” framing did not change the outcome.

3. Analysis

3.1 Precedents Cited

A. Summary judgment and evidentiary sufficiency

  • Smith v. P.A.M. Transp., Inc., 154 F.4th 375 (6th Cir. 2025) and Med. Mut. of Ohio v. k. Amalia Enters. Inc., 548 F.3d 383 (6th Cir. 2008): The panel applied de novo review to both summary judgment and the denial of a Rule 59(e) motion.
  • Anderson v. Liberty Lobby, Inc., 477 U.S. 242 (1986) and Fisher v. Nissan N. Am., Inc., 951 F.3d 409 (6th Cir. 2020): The court reaffirmed the “genuine dispute of material fact” standard and the obligation to view evidence favorably to the nonmovant.
  • Viet v. Le, 951 F.3d 818 (6th Cir. 2020): Conclusory assertions, without supporting facts, cannot defeat summary judgment—an important theme in rejecting Spicer’s broad claims of “systemic” discrimination.

B. ELCRA’s alignment with Title VII methodology

  • Wasek v. Arrow Energy Servs., Inc., 682 F.3d 463 (6th Cir. 2012) and Rouch World, LLC v. Dep't of C.R., 987 N.W.2d 501 (Mich. 2022): The Sixth Circuit reiterated that federal Title VII precedent informs ELCRA analysis.

C. The circumstantial-evidence framework and prima facie elements

  • McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973): The court applied the familiar three-step burden-shifting model because Spicer relied on circumstantial evidence.
  • Lowe v. Walbro, LLC, 147 F.4th 601 (6th Cir. 2025), Hazle v. Ford Motor Co., 628 N.W.2d 515 (Mich. 2001), and DeBrow v. Century 21 Great Lakes, Inc., 620 N.W.2d 836 (Mich. 2001): These authorities supplied the articulation of ELCRA’s prima facie test and the inference-of-discrimination requirement.

D. “Qualified” cannot be collapsed into the employer’s asserted reason for discharge

  • Wexler v. White's Fine Furniture, Inc., 317 F.3d 564 (6th Cir. 2003) (en banc): “Qualified” is assessed by “minimum objective criteria,” not by the employer’s asserted reasons for termination.
  • Cicero v. Borg-Warner Auto., Inc., 280 F.3d 579 (6th Cir. 2002) and Loyd v. Saint Joseph Mercy Oakland, 766 F.3d 580 (6th Cir. 2014): The panel emphasized repeated Sixth Circuit warnings against conflating the prima facie “qualification” prong with later-stage employer justifications.
  • Williams v. Dep't of Health & Hum. Servs., No. 355203, 2021 WL 5027962 (Mich. Ct. App. Oct. 28, 2021) and Town v. Mich. Bell Tel. Co., 568 N.W.2d 64 (Mich. 1997): Michigan authority was used in parallel to reinforce the staged nature of the analysis and the “legitimate expectations” framing.

The court’s noteworthy move was to identify the district court’s methodological mistake—using Harvard’s performance and rule-violation evidence to decide “qualification”—yet still affirm because Spicer failed the separate requirement that the termination occurred under circumstances giving rise to an inference of discrimination.

E. Forfeiture at the summary judgment stage

  • Grand v. City of Univ. Heights, 159 F.4th 507 (6th Cir. 2025): Arguments not adequately developed in response to summary judgment may be forfeited on appeal. The panel flagged that Spicer’s “qualification” evidence was largely not presented below.

F. Comparator and statistical proof

  • Tennial v. United Parcel Serv., Inc., 840 F.3d 292 (6th Cir. 2016), Wright v. Murray Guard, Inc., 455 F.3d 702 (6th Cir. 2006), and Ercegovich v. Goodyear Tire & Rubber Co., 154 F.3d 344 (6th Cir. 1998): These cases drove the “similarly situated in all relevant respects” and “comparable seriousness” analysis used to reject Spicer’s proposed comparator.
  • Peeples v. City of Detroit, 891 F.3d 622 (6th Cir. 2018), Barnes v. GenCorp Inc., 896 F.2d 1457 (6th Cir. 1990), Amini v. Oberlin Coll., 440 F.3d 350 (6th Cir. 2006), and Thompson v. Fresh Prods., LLC, 985 F.3d 509 (6th Cir. 2021): These authorities informed the court’s insistence that statistics must show a meaningful disparity and must be methodologically/explanatorily strong enough to support an inference of discrimination.

G. Pretext stage not reached

  • Martinez v. Cracker Barrel Old Country Store, Inc., 703 F.3d 911 (6th Cir. 2013): Because Spicer failed the prima facie case, the panel did not proceed to pretext—reinforcing that each McDonnell Douglas step matters.

H. Hostile work environment

  • Nathan v. Great Lakes Water Auth., 992 F.3d 557 (6th Cir. 2021): Supplied the five elements for a hostile-work-environment claim.
  • Bruce v. Adams & Reese, LLP, 168 F.4th 367 (6th Cir. 2026) and Harris v. Forklift Sys., Inc., 510 U.S. 17 (1993): Anchored the “severe or pervasive” standard.
  • Hawkins v. Anheuser-Busch, Inc., 517 F.3d 321 (6th Cir. 2008): The panel used this to exclude hostile-environment evidence that Spicer was unaware of during employment (e.g., after-the-fact descriptions).
  • Kellar v. Yunion, Inc., 157 F.4th 855 (6th Cir. 2025) and McNeal v. City of Blue Ash, 117 F.4th 887 (6th Cir. 2024): Cited for the principle that isolated comments typically do not establish an environment “permeated” with discriminatory intimidation, ridicule, or insult.

I. Reconsideration practice and local-rule interaction

  • Miller v. William Beaumont Hosp., 121 F.4th 556 (6th Cir. 2024) (order): Explained that, after the local-rule amendment, reconsideration of final orders in the Eastern District of Michigan proceeds under Rule 59(e) or Rule 60(b), not merely local-rule standards.

3.2 Legal Reasoning

A. The “qualification” prong—error identified, but not outcome-determinative

The panel agreed with Spicer on a doctrinal point: a court may not use the employer’s asserted nondiscriminatory reasons (here, alleged poor performance and major rule violations) to negate the plaintiff’s prima facie “qualification.” Relying on Wexler v. White's Fine Furniture, Inc., Cicero v. Borg-Warner Auto., Inc., and Loyd v. Saint Joseph Mercy Oakland, the court framed “qualification” as an objective inquiry into minimum criteria, distinct from the later inquiry into whether the employer’s stated reason is legitimate or pretextual.

But the court did not remand because it found an independent, dispositive deficiency: Spicer failed to produce evidence that her termination occurred under circumstances giving rise to an inference of discrimination. In other words, even if she could clear the “qualification” prong, she could not clear the “inference” prong.

B. The inference-of-discrimination prong—deposition concessions and thin circumstantial proof

The panel emphasized that Spicer’s own deposition testimony substantially undercut her claims: she could not identify non-Black employees treated better than Black employees, could not provide specific incidents showing sex-based termination, and largely relied on subjective impressions (e.g., “tone”) rather than concrete discriminatory content.

C. Comparator evidence—“comparable seriousness” is a gatekeeper

Spicer’s comparator was a male coworker (Robert Hood) who received an LCA before termination for repeated poor performance. The court held this was not comparable to Spicer’s situation because her admitted “major” violations (unauthorized use of customer property and refusal to heed supervisory instruction) were categorically more serious than repeated “minor” performance issues. Applying Tennial v. United Parcel Serv., Inc. and Wright v. Murray Guard, Inc., the panel treated the different severity level as a “differentiating or mitigating circumstance” that explained the different outcome and defeated comparability.

D. Statistical evidence—workforce composition can defeat inference

Spicer’s statistical proof showed approximately 94% of terminated employees were Black, while the workforce was approximately 93% Black (with the court noting a similar 1% gap under its own calculations). The panel held that such near-parity fails the Peeples v. City of Detroit requirement of a “significant disparity” and does not “eliminate the most common nondiscriminatory explanations”—namely, that the workforce is overwhelmingly Black, so terminations will be as well.

This portion of the opinion effectively operationalizes a practical rule for litigants: statistics that merely mirror the underlying workforce composition—without a meaningful disparity, controls, or explanatory power—will not carry the inference burden under ELCRA/Title VII logic. The court’s reliance on Amini v. Oberlin Coll. and Thompson v. Fresh Prods., LLC underscores that courts will look beyond the headline number to methodology and what the statistic actually explains.

E. Pretext—skipped (and, alternatively, unsupported)

Citing Martinez v. Cracker Barrel Old Country Store, Inc., the panel declined to reach pretext because the prima facie case failed. It added that the same comparator/statistical evidence would not have established pretext in any event.

F. Hostile work environment—awareness, speculation, and isolation problems

The court affirmed summary judgment because Spicer could not show “severe or pervasive” harassment. It excluded evidence she learned only after the fact (e.g., post-meeting descriptions) under Hawkins v. Anheuser-Busch, Inc.. It also deemed the “chair” remark insufficient: Spicer’s belief that it was racially motivated was speculative, and a one-time comment generally does not create a work environment “permeated” with discriminatory abuse under Kellar v. Yunion, Inc. and McNeal v. City of Blue Ash. The panel further noted a practical causation point: the alleged remark could not have altered the conditions of employment where Spicer did not return to the workplace environment after it.

G. Reconsideration—Rule 59(e) framing, same result

While the district court referenced the “palpable defect” standard from E.D. Mich. L.R. 7.1(h), the panel—consistent with Miller v. William Beaumont Hosp.—treated the motion as one under Rule 59(e) and affirmed because the motion simply reargued issues and did not cure the evidentiary deficiencies in the prima facie case.

3.3 Impact

  • Discipline-based terminations under ELCRA: The case highlights that even where a district court misapplies the “qualification” prong, an ELCRA plaintiff still must produce concrete, non-speculative evidence supporting an inference of discriminatory motive to avoid summary judgment.
  • Comparator rigor: The “comparable seriousness” requirement can be dispositive where workplace rules categorize conduct as “major” versus “minor.” Plaintiffs should expect courts to treat those categories—and the underlying misconduct— as key “relevant respects” for similarity.
  • Statistics must outperform baseline demographics: Where termination percentages approximate workforce percentages, the statistics may affirmatively support a nondiscriminatory explanation rather than an inference of bias unless bolstered by additional analysis (controls, disparity measures, decision-maker slicing, job-category stratification, etc.).
  • Hostile environment proof limits: Post hoc descriptions unknown to the plaintiff during employment are typically unusable to prove the environment’s hostility; isolated comments without clear protected-class linkage remain unlikely to satisfy “severe or pervasive.”
  • Practice pointer on reconsideration: In E.D. Mich., parties should frame reconsideration as Rule 59(e) or Rule 60(b) and present new law, new facts, or clear error—reargument will not suffice.

4. Complex Concepts Simplified

ELCRA
Michigan’s primary employment discrimination statute, analogous in many applications to federal Title VII.
Summary judgment
A pretrial ruling where the court decides there is no genuine dispute of material fact and the moving party wins as a matter of law.
McDonnell Douglas framework
A three-step method for circumstantial discrimination cases: (1) plaintiff establishes a prima facie case; (2) employer gives a legitimate, nondiscriminatory reason; (3) plaintiff shows that reason is pretext.
“Qualified” (prima facie)
An objective, minimum-criteria inquiry (skills/experience/basic ability), not a backdoor way to accept the employer’s asserted reason for firing as true.
Comparator evidence
Evidence that similarly situated employees outside the protected class were treated better. Similarity typically requires the same supervisor/standards and misconduct of comparable seriousness.
Statistical evidence
Numbers can suggest discrimination only if they show meaningful disparity and rule out obvious nondiscriminatory explanations (like workforce composition).
Hostile work environment (“severe or pervasive”)
Harassment must be so extreme or frequent that it changes the conditions of employment and creates an abusive environment; isolated or speculative incidents usually do not qualify.
CBA / Grievance / LCA
A CBA is a union contract setting discipline and dispute procedures. A grievance is the contract’s challenge process. An LCA is a last-opportunity deal allowing continued employment under strict conditions.

5. Conclusion

Oneka Spicer v. Harvard Maintenance, Incorporated underscores two practical evidentiary lessons in ELCRA litigation: (1) courts must keep the prima facie “qualification” inquiry analytically separate from the employer’s asserted nondiscriminatory reasons, but (2) plaintiffs still must produce concrete comparator, statistical, or other circumstantial proof that supports an inference of discrimination. On hostile work environment, the opinion reinforces that actionable hostility must be known to the plaintiff during employment and must be severe or pervasive—isolated, speculative, or post hoc material will not do.