Selective Prosecution in Colorado: “Similarly Situated” Is Fact-Intensive, and General Statistics Alone Rarely Prove Discriminatory Effect or Purpose

I. Introduction

In Mitchell v. People, 2026 CO 8, the Colorado Supreme Court reviewed a selective prosecution challenge arising from a juvenile felony-murder prosecution in Arapahoe County (the Eighteenth Judicial District). The petitioner, Demarea Deshawn Mitchell (Black), and three juvenile codefendants were originally charged in district court with felony murder, aggravated robbery, and conspiracy to commit aggravated robbery after a planned robbery of vaping products ended in the victim’s death.

Two non-Black codefendants (J.S., a Hispanic female, and D.S.) cooperated with law enforcement; the prosecution refiled their cases in juvenile court (dropping the murder charge), later offering plea bargains that resulted in juvenile dispositions. Mitchell and another Black codefendant remained in district court facing adult consequences, including life with parole eligibility after forty years. Mitchell moved to dismiss for selective prosecution, claiming the charging/plea decisions were race-based.

The key issues were whether Mitchell established (1) discriminatory effect—that similarly situated non-Black individuals were treated more favorably—and (2) discriminatory purpose—that the prosecution acted “because of” race. The Supreme Court affirmed denial of dismissal.

II. Summary of the Opinion

  • The court reaffirmed that a selective prosecution claim requires clear evidence of both discriminatory effect and discriminatory purpose.
  • Mitchell failed to prove discriminatory effect because he was not similarly situated to J.S. and D.S.: unlike them, he confronted the victim, escalated the encounter, and shot and killed the victim.
  • Mitchell’s statistical evidence about racial disparities in arrests/prosecutions generally did not establish selective prosecution in his case because it did not identify similarly situated non-Black comparators who could have been prosecuted for the same offense but were not.
  • Even assuming some showing of discriminatory effect, Mitchell failed to prove discriminatory purpose; the record supported race-neutral explanations (culpability, strength of evidence, cooperation, and enforcement strategy).

III. Analysis

A. Precedents Cited

The court grounded its analysis in established federal and Colorado selective-prosecution doctrine:

1. Charging discretion and its constitutional limits

  • People v. MacFarland, 540 P.2d 1073 (Colo. 1975): Prosecutors have broad charging discretion, but equal protection forbids selectivity based on unjustifiable standards such as race.
  • United States v. Armstrong, 517 U.S. 456 (1996): Sets the modern framework—selective prosecution requires “clear evidence” of both discriminatory effect and discriminatory purpose; to show effect, a defendant generally must identify similarly situated persons of another race who were not prosecuted.

2. What “similarly situated” means in practice

  • Wayte v. United States, 470 U.S. 598 (1985) (quoted in Armstrong): Legitimate prosecutorial factors include case strength, deterrence value, enforcement priorities, and relationship to an overall plan.
  • United States v. Smith, 231 F.3d 800 (11th Cir. 2000): Provides a functional comparator definition—same basic crime in substantially the same manner, with comparable deterrence/enforcement value and equally strong evidence.
  • Keene v. Mitchell, 525 F.3d 461 (6th Cir. 2008): Differences in evidence strength, number of victims, and cooperation can defeat “similarly situated” status; notably, being the triggerman distinguishes culpability.
  • Frederick Douglass Found., Inc. v. District of Columbia, 82 F.4th 1122 (D.C. Cir. 2023): “Similarly situated” is a fact-intensive and case-specific comparative inquiry, not a formal comparison of labels alone.
  • In re United States, 397 F.3d 274 (5th Cir. 2005): Sharing the same charge does not, by itself, make defendants similarly situated.
  • United States v. Wilson, 123 F.4th 1021 (9th Cir. 2024): “Similarly situated” means more than committing the same crime in the same place.

3. Statistical proof and its limits

  • Chavez v. Ill. State Police, 251 F.3d 612 (7th Cir. 2001): Statistics must address the “crucial question” of differential treatment of one class versus an otherwise similarly situated class.
  • United States v. Mesa-Roche, 288 F.Supp.2d 1172 (D. Kan. 2003): Recognizes contexts (like traffic stops) where identifying non-stopped comparators may be uniquely difficult, making statistics more central; the court contrasted that context with prosecutorial decisions here.
  • United States v. Armstrong, 517 U.S. 456 (1996): The court treated Armstrong as the cautionary example—statistics listing only prosecuted Black defendants did not show similarly situated non-Black individuals who were not prosecuted.

4. Discriminatory purpose standard

  • McCleskey v. Kemp, 481 U.S. 279 (1987), quoting Pers. Adm'r v. Feeney, 442 U.S. 256 (1979): Discriminatory purpose requires action taken “because of,” not merely “in spite of,” adverse effects on an identifiable group; awareness of consequences is not enough.
  • United States v. DeBerry, 430 F.3d 1294 (10th Cir. 2005): Discriminatory purpose can be shown by direct or circumstantial evidence—but must be evidence, not speculation.

5. Appellate posture and review

  • Martinez v. People, 2024 CO 6M: The court adopted the standard split—factual findings reviewed for clear error; legal conclusions de novo.
  • The court also referenced the intermediate decision People v. Mitchell, 2024 COA 7M, which had affirmed on the “not similarly situated” ground and did not reach discriminatory purpose.

B. Legal Reasoning

1. Discriminatory effect failed: Mitchell was not “similarly situated”

The court’s reasoning turned on a comparative culpability and evidence-strength assessment consistent with Armstrong, Smith, and Keene v. Mitchell.

  • Distinct role in the killing: The record supported that Mitchell exited the vehicle with a loaded gun, confronted the victim as the victim walked away, struggled with him, and shot him in the chest—making him the principal escalator and the triggerman. J.S. and D.S. did not engage in conduct that escalated to lethal force.
  • Strength of evidence: Two cooperating codefendants implicated Mitchell; Snapchat messages suggested consciousness of wrongdoing (“I did some stuff,” “Shit didn’t go as planned”); and evidence suggested Mitchell disposed of the gun.
  • Legitimate enforcement strategy: Prioritizing prosecution of the most culpable actor (the person directly responsible for the death) fit the legitimate enforcement considerations described in Wayte v. United States and Armstrong.
  • Plea/cooperation differences: The court treated cooperation as a legitimate differentiator: offering deals to less culpable cooperators to secure testimony against the alleged killer was a permissible prosecutorial strategy, not evidence of discrimination.

Critically, the court rejected Mitchell’s argument that initial identical charging (felony murder for all four) established comparability. Citing Frederick Douglass Found., Inc. v. District of Columbia, In re United States, and United States v. Wilson, the court emphasized that “similarly situated” is not a charge-label test; it is an individualized, fact-driven comparison.

2. Statistics did not supply discriminatory effect

The court held Mitchell’s statistics—purporting to show general racial disparities in arrests, adult prosecution, and imprisonment—did not bridge the essential comparator gap. Echoing Armstrong, the court observed that the evidence did not identify non-Black individuals who (a) were similarly situated on facts and culpability and (b) could have been prosecuted for felony murder but were not (or were treated materially better).

3. Discriminatory purpose also failed

Applying McCleskey v. Kemp and Pers. Adm'r v. Feeney, the court required evidence that prosecutors acted at least in part because of race.

  • Statistics were untethered to decisionmakers: General disparity data did not show that prosecutors in this case acted with race-based motive.
  • Plea bargaining disparity was explained by culpability and cooperation: The court treated differential pleas among codefendants as commonplace and, here, justified by legitimate factors.
  • Reverse transfer questioning was statutory and issue-specific: Prosecutorial questioning about prior contacts and weapon use aligned with factors the court must consider under § 19-2.5-801(4)(b), C.R.S. (2025).
  • “Changed emphasis” between hearings was not proof of bias: The court distinguished the reverse transfer inquiry from trial proof; at trial, intent to kill is immaterial to felony murder under § 18-3-102(1)(b), C.R.S. (2019) (and similarly under § 18-3-103(1)(b), C.R.S. (2025) as cited).
  • Speculation about prosecutorial knowledge of jury outcomes was unsupported: The claim that prosecutors expected easier convictions against Black juveniles lacked evidentiary support and thus could not establish purpose.

C. Impact

This decision consolidates and operationalizes the selective prosecution framework in Colorado in three practical ways:

  1. Comparator rigor is central: Defendants must develop comparator evidence showing non-Black individuals who are genuinely comparable on culpability, conduct, evidence strength, and prosecutorial priorities—mere shared charging decisions at case initiation will not suffice.
  2. General disparity statistics are usually insufficient without case-level linkage: Statistics may contextualize claims, but without underlying facts enabling a “similarly situated” comparison, they will rarely establish discriminatory effect or purpose.
  3. Plea/cooperation and “triggerman” status are legitimate differentiators: The opinion signals that prosecutors may lawfully distinguish among codefendants based on escalatory conduct, direct responsibility for a death, evidentiary posture, and cooperation—factors likely to recur in multi-defendant felony-murder prosecutions involving juveniles.

For future litigants, Mitchell increases the premium on early factual development: obtaining discovery, identifying comparator cases with comparable fact patterns, and linking alleged disparities to the specific decisionmakers and decisions at issue.

IV. Complex Concepts Simplified

Selective prosecution
A constitutional claim that the prosecutor chose to charge (or charge more harshly) for an improper reason (like race). It is difficult to prove because prosecutors have broad discretion.
Discriminatory effect
Proof that people of another race who are similarly situated were treated better (e.g., not charged, charged less severely, diverted, or offered materially different outcomes).
Discriminatory purpose
Proof the prosecutor acted at least partly because of race. It is not enough to show a policy had predictable racial impacts; there must be evidence of race-based motivation.
Similarly situated
Not “charged with the same offense,” but comparable in the ways prosecutors legitimately care about: who did what, how severe the conduct was, evidence strength, cooperation, and enforcement priorities.
Felony murder
A murder charge that attaches when a death occurs during certain felonies (like robbery), even if no one intended to kill. That is why the court noted it was “immaterial” whether participants meant for anyone to die.
Reverse transfer hearing
A proceeding where a juvenile whose case is in district court asks to move it to juvenile court. The statute requires consideration of factors like the juvenile’s record and weapon involvement.

V. Conclusion

Mitchell v. People affirms a demanding, evidence-focused approach to selective prosecution claims: defendants must prove both discriminatory effect and purpose with clear evidence, and “similarly situated” requires a granular comparison of culpability, conduct, cooperation, and evidentiary strength—not merely common charges or broad statistical disparities. The opinion’s practical rule is that, absent comparator-specific proof tying differential treatment to race-based intent, courts will treat charging and plea-bargaining differences—especially where one defendant is the triggerman and others cooperate—as legitimate prosecutorial discretion rather than unconstitutional selectivity.