2026 Ballot - Judiciary

Somewhere on your ballot, apart from the races you've been hearing about, there's a long list of names with a yes and a no next to each one. Those are judges, and thirteen of them are on Boulder ballots this year. Most people either vote yes down the line or skip the section entirely, and I don't blame them. The official evaluations give all thirteen the same grade, "Meets Performance Standards," and every one of those votes was unanimous. That tells you they show up and do the job competently. It doesn't tell you much about how they rule.

So I went looking for the part the blue book leaves out. (Twenty years of audit and security work will do that to a person.) The question was simple: in a criminal case, when it's close, does this judge tend to side with the prosecution or with the defendant? I asked the same thing about civil cases, but as you'll see, that turned out to say a lot less.

How retention works

Colorado judges are appointed, not elected. Once they've served a first term, and at the end of every term after that, they go on the ballot and you decide whether they keep the job. A judge needs a majority of yes votes to stay. If a judge loses, the governor appoints a replacement from a short list chosen by a nominating commission, so a no vote removes a judge but doesn't pick the next one. You can vote on some judges and leave others blank.

The short version

A few definitions first. "Harsh" here means siding with the prosecution: upholding convictions and sentences, or setting tougher terms. "Lenient" means siding with defendants: reversals, suppressed evidence, or lighter outcomes. Neither one is a compliment or an insult. Which one you want more of is up to you.

Judge by judge

Sorted from most lenient to most harsh. Each name links to the judge's official 2026 evaluation. Confidence reflects how much case evidence exists, not how sure the label sounds.

The appeals judges: the solid part

Seven of the thirteen sit on the Court of Appeals or the Supreme Court, and they're the easiest to read because they write opinions. The ones that say the most are split decisions and separate opinions, since a unanimous ruling reflects the whole panel rather than one judge. For context, the Court of Appeals fully affirms roughly 70–82% of criminal appeals, depending on the sample, so a judge who reverses a lot more or a lot less than that stands out. The Court of Appeals counts below come from the court's weekly opinion announcements from January 2024 through October 2026, about 14–21 criminal opinions per judge.

Justice William W. Hood III (Supreme Court, appointed 2014)

• Wrote the 4-3 majority in People v. McKnight (2019), holding that a sniff by a marijuana-trained dog is a search under the state constitution.
• Was the lone dissent in People v. Nelson (2015), arguing exonerated defendants should get their fees back. The U.S. Supreme Court later agreed.
• Dissented in Clark v. People (2024), seeking a new trial after a racially biased juror stayed on the panel, and wrote 2026 majorities ordering a new murder trial and barring a retrial on double jeopardy grounds.
• On the other side of the ledger: wrote the opinion upholding a Google reverse-keyword warrant in People v. Seymour (2023), and opinions upholding murder convictions as harmless error in Ray and Townsell.

Judge Katharine E. Lum (Court of Appeals, 2022)

• The defendant won in full or in part in 6 of the 14 criminal opinions she wrote in the sample, about 43%.
• Wrote the reversal in People v. McRae (2025), tossing drug convictions and a 25-year sentence because police had no reasonable suspicion to extend a traffic stop, and reversed a felony murder conviction for the second time in 2026.
• Counterweight: in People v. Eddington (2026) she held that GoFundMe donations to a victim don't reduce what the defendant owes in restitution.

Judge Karl L. Schock (Court of Appeals, 2022)

• Affirmed 14 of 21 authored criminal opinions (67%), near or below the court's rate.
• Dissented in People v. Al-Jihad (2025) from a ruling for the Jefferson County DA on a self-defense instruction, and wrote a concurrence in Rodriguez-Ortiz calling a cell-phone warrant unconstitutionally broad.
• Wrote new-trial orders in a child abuse case with a 32-year sentence and a second-degree murder case.

Judge Elizabeth L. Harris (Court of Appeals, 2015)

• The defendant won in 4 of 18 criminal opinions she wrote in the sample, about 22%, which is near average.
• Wrote 2-1 rulings for defendants over dissents in Ez-Zahir and Xiong, and a concurrence arguing a three-strikes defendant deserved resentencing.
• Counterweight: in People v. Jacobs (2026) she reinstated dismissed murder charges.

Judge Rebecca R. Freyre (Court of Appeals, 2015)

• Affirmed 18 of 19 criminal opinions she wrote in the sample, the most pro-prosecution routine record of the six appeals judges.
• Her notable reversals protect structural rights: closed courtrooms and a denied suppression hearing.
• In People v. Shockey (2024) she wrote a 2-1 ruling barring a murder retrial; the Supreme Court reversed her, and she then affirmed the conviction.

Judge Pax L. Moultrie (Court of Appeals, 2024)

• Affirmed about 15 of 20 authored criminal opinions (75–83%).
• In People v. Simms (2026): reversed a felony menacing conviction (a fist isn't a "bludgeon") but upheld the reasonable-doubt instruction over a dissent. Also ordered a new trial on drug convictions carrying a 36-year sentence.

Judge Grant T. Sullivan (Court of Appeals, 2024)

• Affirmed 15 of 18–19 authored criminal opinions (79–83%).
• Dissented in People v. Abdul-Rahman (2024), saying courts had no jurisdiction over a parolee who skipped the internal appeal; the Supreme Court later agreed. Dissented again in Valles-Dominguez (2025) to keep a $349,999 restitution order the majority struck as late.
• Counterweight: reversed robbery convictions over a pro-police juror and a sentence on right-to-counsel grounds.

The trial judges: mostly informed guesses

The six local trial judges are much harder to score, and I want to be upfront about that. Sentencing and bond data by judge aren't published, and the evaluation surveys don't separate what prosecutors think from what defense attorneys think. Thus a lot of what follows is a handful of reported cases plus survey numbers.

Nancy W. Salomone (20th District, appointed 2017, Chief Judge since July 2025) is a former public defender who had a criminal docket through September 2024 and now runs the treatment court, probation violations, and the competency docket. In May 2026 she sentenced the Pearl Street firebomber to life without parole plus hundreds of years; the life term was mandatory, and how much discretion she had on the rest is unclear. In 2025 she vacated Michael Clark's 2012 murder conviction after a DNA retest, with prosecutors no longer objecting. Attorneys rated her above the district average for fairness to both sides and for consistent sentencing.

Michael T. Kotlarczyk (20th District, appointed 2024) came from the civil side of the Attorney General's office and has only heard civil cases. He dismissed the TABOR challenge to the CU South flood project but refused the city's bid to make residents pay about $46,000 of its legal fees. Attorneys rated him 3.8 out of 4 for fairness.

J. Chris Larson (20th District, appointed 2024) is a former Deputy U.S. Attorney who heard civil cases until this rotation and now has a criminal docket. The best-known ruling blocked the city's four-hour-notice closure of Ash House on University Hill, calling it "well-meaning but legally deficient."

Zachary I. Malkinson (Boulder County Court since 2019) was a public defender in Alameda County and then a Boulder defense lawyer. In a 2024 DUI case he excluded a deputy's unrecorded observations under the 2020 police-accountability law; the Boulder DA went to the Supreme Court, which dismissed the appeal on procedure. His attorney score for being fair to both sides went from 3.5 in 2022 to 3.2 in 2025 to 2.4 this year.

Elizabeth Brodsky (Boulder County Court since 2015) ran Boulder County Legal Services before becoming a magistrate. On the 2020 eviction docket she was reported saying she is "reluctant to default" tenants. Attorneys rate her 3.6 for fairness to both sides.

Monica Haenselman (Boulder County Court since 2023, Presiding Judge this year) was a magistrate on criminal, family, and juvenile cases. I found no reported rulings. Her 2026 survey is the weakest on the ballot: 2.5 out of 4 overall, with only 33% of non-attorneys saying she meets standards and 47% saying she doesn't. Attorneys are kinder; 82% say she meets standards. Her "leans harsh" label is an inference from those litigant ratings, not from anything she's ruled.

The civil side says less

I also scored non-criminal rulings, from liberal (siding with civil-rights claimants, workers, tenants, consumers, injured plaintiffs, regulators, and disclosure) to conservative (siding with businesses, insurers, landlords, property rights, TABOR limits, and narrow liability). Nobody lands far from the center. Nearly all Court of Appeals civil rulings are unanimous, and many are routine appeals the court almost always affirms, so the differences between judges are small.

• Most liberal: Elizabeth Harris. Every contested civil opinion I found went that way, including a longer window to sue insurers in Weatherill v. State Farm.
• Leaning liberal: William Hood (wrote the 5-2 transgender care ruling against Children's Hospital, but leans conservative on insurance contracts) and Rebecca Freyre.
• Middle: Katharine Lum, Pax Moultrie, and Grant Sullivan, whose civil record reads as pro-government more than ideological (including joining the ruling upholding Boulder's camping ban).
• Leaning conservative: Karl Schock, whose opinions read duty and access rights narrowly, for example finding no duty for Lime to a cyclist hit by a scooter rider.
• Too little to say: all six trial judges.

Two judges split across the axes, which I found interesting: Lum and Schock are among the most defendant-friendly judges in criminal appeals, but moderate to slightly conservative on civil cases.

Other things you'll hear

• Hood and Sullivan. In an October 9 opinion column in Colorado Politics, Dennis Maes urges a no vote on both, arguing that they knew about the Judicial Department's pay-for-silence scandal before it was reported and didn't report it, and that the official evaluations left this out. That's an advocate's argument, not an official finding, and it has nothing to do with harsh or lenient.
• Haenselman's survey numbers. The lowest on the ballot, especially from non-attorneys. Boulder Reporting Lab's voter guide flags it too.
• Malkinson's fairness drop. Given his defense background and the DUI ruling, the critical attorneys are plausibly prosecutors, but the survey doesn't say.
• Harris's timeliness. A median of 145 days from argument to opinion, with eight cases still undecided, per Colorado Politics.

What this can and can't tell you

The appeals scores are the solid ones, because they rest on opinions the judges actually wrote. Even so, the samples are 14–21 opinions per judge and mostly unpublished, and Hood's count leans on splits that made the news. The trial-court scores are mostly informed guesses. Kotlarczyk and Larson have no criminal record to score, and the labels for Haenselman and Brodsky come from surveys and eviction-court remarks rather than sentencing.

One more thing I didn't expect: prior jobs predicted very little. Former prosecutors Hood and Schock both lean lenient, and Freyre, who spent 25 years as a public defender, has the highest affirmance rate of the bunch. So I'd skip the shortcut of voting on résumés.

What I'm asking

This isn't an endorsement list, and I'm not going to tell you how to vote on any of these thirteen. What you want from a judge is your call. What I am asking is that you don't skip that part of the ballot. If something above catches your eye, click through to the official evaluations and read the cases for yourself. Then get your ballot back: drop boxes are open around the clock, and ballots have to be received by 7 p.m. on November 3.

And while you're filling it out, the council race is on there too. I've written about how I'd vote that part, too.