Denver teacher coaching study finds no gains in retention or test scores

A large new study of Denver Public Schools’ Teacher Leadership & Collaboration model found no measurable effect on either teacher retention or student test scores, even after the program grew into a districtwide system that reached more than 1,000 teacher-coaches and more than 6,000 mentee teachers. The paper, published in Educational Evaluation and Policy Analysis, examined one of the country’s biggest efforts to let accomplished teachers stay in the classroom part time while also coaching and evaluating peers. (journals.sagepub.com)

For district leaders planning next year’s staffing and professional-development budgets, the practical lesson is not that coaching is useless. It is that leaders should ask tougher design questions before expanding teacher-leader roles: Is coaching protected time or just another duty added onto evaluation, supervision, and school operations? Are coaches able to give frequent, individualized feedback, or is the role spread too thin to change practice? Those questions matter because prior research has found positive average effects for coaching overall, while also warning that results often weaken as programs get larger and more standardized. (journals.sagepub.com)

What the Denver study actually found

The study by Michelle Doughty, Mimi Engel, and Allison Atteberry evaluated Teacher Leadership & Collaboration, or TLC, in Denver Public Schools, a district that an earlier federal partnership description said serves more than 90,000 students with nearly 6,000 teachers across more than 200 schools. The researchers examined a five-cohort rollout of TLC and used a difference-in-differences design tailored to staggered implementation. Their bottom line was straightforward: they found no effects on teacher retention or student test scores. (journals.sagepub.com)

That finding carries weight because TLC was not a boutique pilot. Denver Public Schools still describes Teacher Leadership & Collaboration as a core part of teacher growth and development, saying teachers benefit from instructional leaders who provide feedback, coaching, and support while remaining connected to classroom teaching. Current district recruiting materials continue to advertise roles such as Senior Team Lead, Team Lead, and Team Specialist. In other words, this was a real operating model, not a short-lived experiment on the margins. (careers.dpsk12.org)

Just as important, the program the researchers studied was a bundled intervention. The paper’s abstract says teachers were released from the classroom part time to “coach and evaluate” other teachers. That distinction matters. A model that combines developmental coaching with formal evaluation asks one person to do at least two jobs at once: help colleagues improve instruction and participate in judging performance. (journals.sagepub.com)

Why this is not a verdict on all coaching

School leaders should resist the easy headline that “coaching doesn’t work.” A 2018 meta-analysis of 60 causal studies by Matthew Kraft, David Blazar, and Dylan Hogan found positive average effects of teacher coaching on both instructional practice and student achievement. But that same review also noted that much of the evidence came from literacy coaching for prekindergarten and elementary teachers, and that scaling coaching while preserving quality is a central challenge. (education.umd.edu)

That is why the Denver result is more useful as a design warning than as a blanket rejection of coaching. The prior literature suggests coaching is strongest when it is individualized and built around repeated observation-feedback cycles. Denver’s model, by contrast, asked current teachers to split time between their own classrooms and a broader leadership role that included evaluation responsibilities. It is reasonable to infer that, as systems scale, those added responsibilities can dilute the very intensity and trust that make coaching work in smaller settings. That inference is grounded in the Denver paper’s description of the role and in the broader coaching literature’s warnings about larger programs. (journals.sagepub.com)

The decision guide for districts

So what evidence should districts demand before they expand teacher-leader coaching roles? At minimum, they should want proof that the model preserves the ingredients the coaching literature treats as central: regular observation and feedback, enough protected time to work intensively with teachers, and a manageable enough scope that coaching does not become occasional check-ins. Districts should also ask whether the same person is expected to both coach and formally evaluate peers, because the Denver study suggests that bundled model may not produce the hoped-for gains at scale. That last point is an evidence-based caution, not a settled rule for every setting. (journals.sagepub.com)

Leaders should also insist on local outcome checks before scaling up. Denver’s program is unusually informative because researchers could study a staggered rollout across a large urban district over time. Many districts expand leadership pathways without building in any similarly credible test of whether the model improves retention, instruction, or student learning. If a district cannot answer basic implementation questions such as coach caseload, frequency of coaching cycles, and what outcomes will be tracked, it is probably not ready to call the work “evidence-based.” (ies.ed.gov)

There is one more nuance here. Denver Public Schools continues to present TLC as part of its teacher-career pathway, which suggests districts may value these roles for reasons beyond the outcomes highlighted in this study, such as leadership development or principal capacity. But the new paper sharpens the trade-off: a role can be organizationally appealing and still fail to move the outcomes superintendents most often cite when justifying the investment. For school systems building 2027 budgets, that is the question worth carrying forward: not whether to fund coaching in the abstract, but whether the version being funded is protected instructional support or an overloaded hybrid job whose ambitions have outpaced its evidence. (careers.dpsk12.org)