COLJI-Crim. E:03 (2022) “Firmly Convinced / Real Possibility” Reasonable-Doubt Instruction Is Constitutional and Does Not Shift or Lower the Burden of Proof

1. Introduction

This Colorado Supreme Court decision addresses a recurring and high-stakes criminal-procedure question: how courts may explain “proof beyond a reasonable doubt” to jurors without diluting the prosecution’s constitutional burden or implying the defense must prove innocence. The petitioner, Alexis Teran Sanchez, challenged his alcohol-impaired-driving-related convictions on the ground that the county court used COLJI-Crim. E:03 (2022)—the then-newly revised model instruction (the “2023 Instruction”)—which defines reasonable doubt using the phrases “firmly convinced” and “real possibility.”

The dispute centered on whether those phrases (1) lower the burden of proof, (2) shift the burden toward the defendant, and (3) when combined with certain voir dire and closing-argument remarks, created a constitutionally defective reasonable-doubt framework.

2. Summary of the Opinion

The court emphasized (a) broad federal and state judicial acceptance of the “firmly convinced / real possibility” formulation, (b) ordinary-language meanings consistent with that acceptance, and (c) record context showing repeated, explicit instructions that the burden “entirely” rested with the prosecution and never shifted to the defense.

3. Analysis

3.1 Precedents Cited

A. Colorado burden-of-proof and instruction review framework

• Garcia v. People: Cited for de novo review of jury-instruction constitutionality and whether the jury was properly instructed. The court uses Garcia to frame the review posture: this is a pure legal question, not a discretionary evidentiary one.
• Pettigrew v. People: Reinforces that due process requires proof beyond a reasonable doubt of each element, while also recognizing trial courts retain “reasonable discretion” in how they define reasonable doubt, so long as the constitutional demand is met.
• People v. Flockhart and People v. Hoskin: Establish the status of pattern instructions as “nonbinding” but “time-trusted guideposts,” while warning that a pattern instruction that misstates or lowers the standard is unconstitutional. These cases support the court’s method: treat COLJI as persuasive but test it against due process.
• Tibbels v. People: Supplies two critical components. First, it reaffirms that an instruction lowering the burden is unconstitutional structural error requiring automatic reversal. Second—and most importantly here—it articulates the operative inquiry for contextual remarks: whether there is a “reasonable likelihood” the jury understood the court’s statements, in the context of instructions as a whole and the trial record, to allow conviction on a standard lower than beyond a reasonable doubt. The court applies this test to voir dire remarks.
• Johnson v. People: Provides a contrast case where a confusing, stray judicial remark did not warrant reversal because correct instructions were repeatedly given. Johnson is used to classify the “best guess” exchange as non-dispositive when surrounded by correct burden-of-proof reiterations.
• Domingo-Gomez v. People: Governs review of the prosecutor’s unobjected-to remarks under plain error—whether the alleged error undermined fundamental fairness and cast serious doubt on the verdict’s reliability. The court relies on Domingo-Gomez to conclude the prosecutor’s “need[ed] to find him guilty” comment, brief and followed by correct burden language, was not plain error.

B. Federal constitutional permissibility of wording choices

• Victor v. Nebraska: Supplies the baseline: no particular “form of words” is constitutionally required if the jury is instructed on the necessity of proof beyond a reasonable doubt. The opinion also leverages Justice Ginsburg’s endorsement (concurrence) of the Federal Judicial Center formulation that uses “firmly convinced” and “real possibility,” treating it as a strong signal of constitutional acceptability.
• Boyde v. California: Used to justify considering jurors’ commonsense understanding rather than lawyerly parsing of semantic nuances. This supports the court’s rejection of the petitioner’s argument that everyday uses of “firmly convinced” necessarily imply a lower threshold.
• Cage v. Louisiana (and note to Estelle v. McGuire): Cited as an example of wording that did lower the burden (“grave uncertainty” and “moral certainty”), distinguishing the 2023 Instruction from historically problematic formulations.

C. Federal appellate reception of “firmly convinced” and “real possibility”

• United States v. Hunt and United States v. Alonzo: Support the proposition that “firmly convinced” is materially consistent with constitutionally adequate expressions of reasonable doubt, and that “abstract semantics” challenges should fail when the instruction accurately conveys the burden.
• United States v. Velasquez and United States v. Barrera-Gonzales: Reinforce that “firmly convinced” conveys at least as much certainty as other accepted formulations; Velasquez equates it with “hesitate to act,” while the Colorado opinion notes that “hesitate to act” has fallen out of favor for comprehension and conceptual reasons.
• United States v. Porter and United States v. Taylor: Address scrutiny of “real possibility.” Porter flags confusion when a court fails to clarify the difference between “possibility” and “real possibility” and does not explicitly tie the concept back to the prosecution’s burden; yet even there, the burden was not deemed shifted, and other instructions “compensated.” Taylor supports the idea that correct explanations of presumption of innocence and the prosecution’s burden dispel any implication the defendant must prove a “real possibility” of innocence.
• United States v. Conway, United States v. Brand, United States v. Petty, United States v. Watkins (and Harris v. Bowersox): Cited to demonstrate that multiple circuits accept the “firmly convinced” formulation and, in combination with “real possibility,” regard it as comprehensible and correct.
• United States v. Williams: Used for the proposition that “real” serves to prevent acquittals based on any conceivable or fanciful possibility of innocence, consistent with reading the instruction as a whole.

D. Colorado Court of Appeals decisions approving the instruction

• People v. Schlehuber and People v. Melara: The Supreme Court aligns with both divisions that upheld the twin phrases as an accurate statement of the reasonable-doubt standard. Schlehuber is particularly central: “firmly convinced” conveys “near certitude” (more than “highly probable,” less than absolute certainty), and “real possibility” prevents acquittal based on “fanciful” doubt.

E. Contrary authority and how the court treated it

• State v. Perez: The petitioner’s principal state-court counterexample, suggesting “real possibility” invites abandonment of the presumption of innocence and that one might be “firmly convinced” yet still have reasonable doubt. The Colorado Supreme Court acknowledged the criticism but treated it as outweighed by broader authority and, critically, by reading the instruction in full—especially its explicit burden allocation to the prosecution.
• State v. Jackson and State v. Putz: Cited as examples of state-court acceptance of the formulation, with Putz emphasizing that language must not be judged “in artificial isolation,” echoing Colorado’s context-driven approach.
• United States v. Artero: Adds additional support for permissibility, again relying on Justice Ginsburg’s approval in Victor.

F. Handling of social science “jury studies” introduced on appeal

• Stoltie v. California and Stoltie v. Tilton: Invoked by the petitioner to argue that empirical work casts doubt on the FJC wording. The court, however, emphasized limitations in such studies and did not treat them as conclusive evidence that the instruction is unconstitutional in actual courtroom conditions.
• Mun. Subdistrict, N. Colo. Water Conservancy Dist. v. OXY USA, Inc. and CRE 201(b): Used to reject the implicit request for judicial notice of disputed, methodology-dependent empirical findings.
• Daubert v. Merrell Dow Pharms., Inc., Hines v. Oliver, and Davis v. Pursel: Support the court’s point that trial courts—via gatekeeping and factfinding—are the proper forum for weighing expert methodologies and statistical studies, potentially through hearings and cross-examination.
• Petition of Edilson: Provides the decisive appellate constraint—evidence not presented to the trial court will not be considered on review. This becomes an important procedural teaching of the case: litigants should not expect to build a new empirical record for the first time in the Supreme Court.

3.2 Legal Reasoning

The court’s reasoning proceeds in three moves.

Move 1: The instruction’s text allocates the burden correctly and expresses the required level of certainty

The court begins with the instruction’s structure. It does not merely define reasonable doubt in isolation; it opens by unambiguously placing the “burden of proof” on the prosecution to prove “each and every element” beyond a reasonable doubt, and clarifies that the standard is more than “highly probable” but less than “absolute certainty.” That framing, in the court’s view, forecloses the central burden-shifting concern: jurors are not asked to decide whether the defense has proved a “real possibility” of innocence; they are told what it means for the prosecution to have failed to meet its burden.

Move 2: “Firmly convinced / real possibility” is an endorsed, complementary pairing—not a dilution

Relying on the Federal Judicial Center lineage and extensive federal circuit approval—bolstered by Justice Ginsburg’s praise in Victor v. Nebraska—the court characterizes “firmly convinced” (guilt) and “real possibility” (not guilty) as complementary instructions that work “in tandem.” “Firmly convinced” communicates a high level of confidence (“near certitude”) while “real possibility” ensures acquittal is not premised on merely “fanciful” or “imaginary” doubt. The court expressly rejects the petitioner’s “tautology” argument, viewing the phrases as performing different clarifying functions.

Move 3: Contextual remarks did not create a “reasonable likelihood” of a lower standard

Applying Tibbels v. People, the court holds the “best guess” voir dire exchange did not lower the burden because it was immediately followed by repeated, emphatic statements that the burden was “entirely” on the prosecution and “never” shifted—indeed, the court even underscored that the defense could do nothing and the burden would remain with the People. The Supreme Court treats this as a Johnson v. People-type scenario: an isolated remark outweighed by correct, repeated statements and the final written charge (the 2023 Instruction) given verbatim.

As to the prosecutor’s closing argument, the court applies Domingo-Gomez v. People plain-error review because no contemporaneous objection was made. The court treats the “need[ed] to find him guilty” line as brief and contextualized by reminders that jurors must be “firmly convinced,” concluding it did not undermine trial fairness or cast serious doubt on the verdict’s reliability.

3.3 Impact

• Stabilization of Colorado’s modern reasonable-doubt instruction: The decision supplies statewide, binding confirmation that COLJI-Crim. E:03 (2022)’s “firmly convinced / real possibility” language is constitutional. Post-Tibbels, this matters because trial judges and litigants were on heightened alert that reasonable-doubt explanations could trigger structural error.
• Context-first approach to alleged burden dilution: The opinion reinforces that courts must evaluate contested remarks in the full trial context using the “reasonable likelihood” test from Tibbels v. People—distinguishing between a judge’s undermining, alternative definition (as in Tibbels) and a stray or ambiguous comment surrounded by correct instructions (as here and in Johnson).
• Procedural discipline for “jury study” challenges: By invoking CRE 201(b), Daubert, and Petition of Edilson, the court signals that empirical attacks on jury instructions must be developed in the trial court, with adversarial testing. Appellate courts will be reluctant to treat such studies as indisputable “facts,” especially when offered for the first time on review.
• Guidance for prosecutors and trial judges: Although the prosecution prevailed, the decision implicitly cautions against language that sounds like a directive to convict. The analysis shows such remarks are survivable under plain error when fleeting and counterbalanced by correct burden explanations—but that is not an endorsement of the phrasing as best practice.

4. Complex Concepts Simplified

• “Beyond a reasonable doubt”: The highest proof standard in American law. It does not require absolute certainty, but it does require jurors to reach a very high level of confidence before convicting.
• “Structural error”: A fundamental trial defect (such as lowering the prosecution’s burden of proof) that requires automatic reversal without asking whether the error likely affected the outcome.
• “Reasonable likelihood” test (from Tibbels v. People): The question is not whether a phrase could be misunderstood in the abstract, but whether it is reasonably likely the jury actually understood the instructions and statements—taken as a whole—to allow conviction on too low a standard.
• “Plain error” (from Domingo-Gomez v. People): When no timely objection is made, reversal is warranted only for especially serious errors that undermine the trial’s fundamental fairness and create serious doubt about the verdict’s reliability.
• Judicial notice (CRE 201(b)): A court may accept some facts without proof only when they are not reasonably disputable. The court held contested empirical “jury studies” do not qualify.
• Daubert gatekeeping: Trial judges screen expert evidence for reliability and relevance. The court underscored that appellate courts are not the forum to “try out” new expert studies for the first time.

5. Conclusion

Terran Sanchez v. People cements a key post-2023 rule in Colorado criminal practice: using COLJI-Crim. E:03 (2022)’s “firmly convinced / real possibility” definition of reasonable doubt is constitutionally permissible and does not, by itself, lower or shift the prosecution’s burden. The court also clarifies how to evaluate ancillary trial remarks—through Tibbels context-based “reasonable likelihood” review for judicial statements and Domingo-Gomez plain-error review for unobjected-to prosecutorial comments. Finally, the decision warns that empirical challenges to jury-instruction wording must be built in the trial court, not debuted on appeal.