The bulletin board for America's public schools. Parents, teachers, students, and staff. One community per school.
teacher coachingprofessional developmentteacher retentioneducation research

A $20 million-a-year teacher coaching program did not raise test scores or improve retention

Kate Carter
Former Educator · Nov 14, 2026 · 12:06 PM ET
A $20 million-a-year teacher coaching program did not raise test scores or improve retention
See more of our coverage in your search results.Add allk12 on Google

A district program that moved about 1,000 classroom teachers into part-time coaching jobs, serving more than 6,000 of their colleagues, did not raise student test scores or keep more teachers in their schools, according to a new study in Educational Evaluation and Policy Analysis, a peer-reviewed journal of the American Educational Research Association.

The paper, published online October 13, 2026, is by Michelle Doughty and Mimi Engel of the University of Colorado Boulder and Allison Atteberry of the University of Virginia. The Institute of Education Sciences, the U.S. Department of Education's research arm, and the William T. Grant Foundation paid for it.

The authors call the program the Instructor Coaching Program, or ICP. It ran in a large urban district where about half of students are Hispanic, and the paper does not name the district. Between 2014 and 2022, ICP "hired almost 1,000 coaches to serve over 6,000 teachers in approximately 130 schools," the authors wrote, more than almost any coaching program studied before with cause-and-effect methods.

How the program worked and what it cost

Coaches were teachers with 2 years of effective ratings. They got several thousand dollars in bonus pay and usually spent half their time out of the classroom, coaching assigned teachers, called mentees, leading teacher teams, and evaluating mentees, including the formal yearly rating that goes on a teacher's record.

Schools joined in five waves from 2014 to 2018. Covering coaches' class time plus their stipends alone topped $20 million a year starting in 2020, in a district with a yearly budget of around $1 billion. That is about 2% of the budget. The authors' rough estimates put ICP at about $13,000 per mentee teacher, $52,500 per coach, $5,800 per teacher in a participating school, and $270 per student a year.

The coaching was intense at first. From 2015 to 2018, coaches typically observed each mentee about five times a year. After 2018, it fell to two or three. Even in 2022, the authors wrote, that looked more like routine observations, about one per semester, than the roughly seven a year typical of coaching cycles. By 2022, about 80% of mentees also got their yearly professionalism rating from their coach instead of a principal or assistant principal.

What the study found

Because schools joined at different times, the researchers compared how retention and test scores changed in schools that had just started ICP with schools that had not started yet. This is not a randomized trial, and the authors wrote that the study's "causal warrant" is weaker than for programs that began with random assignment.

Every main estimate landed close to zero. The table shows the overall effect from the paper's two models, with the standard error, a measure of uncertainty, in parentheses. An estimate smaller than about twice its standard error is usually read as no detectable effect, and none of these reach that bar.

OutcomeStandard modelModel built for staggered start dates
Teachers staying in their school (percentage points)-0.77 (1.59)-1.65 (2.28)
Teachers staying in the district (percentage points)-0.39 (1.14)-2.03 (1.47)
Reading/English scores, grades 4 to 8 (SD)0.007 (0.015)-0.010 (0.025)
Math scores, grades 4 to 8 (SD)-0.008 (0.017)-0.030 (0.025)

Source: Doughty, Engel, and Atteberry, Educational Evaluation and Policy Analysis, 2026, Tables 3 and 4. SD means standard deviations of student test scores. Test score results end in 2019 because of COVID testing disruptions.

The authors said the analysis could catch a test score effect as small as 0.05 standard deviations, a unit researchers use to compare results across tests, and a change in school-level retention of about 4.5 percentage points. For math, any effect too small to detect "is more likely to be negative than positive," they wrote.

The study did not measure whether coaching changed how teachers taught, because the researchers lacked enough data on teaching quality from before ICP began. Hints that ICP helped special education teacher retention and reading scores for students of new teachers were, in the authors' words, "preliminary and imprecise."

The benchmark is a 2018 review in Review of Educational Research by Matthew Kraft, David Blazar, and Dylan Hogan. Combining 60 studies, they found coaching improved teaching by 0.49 standard deviations and achievement by 0.18. Programs with fewer than 100 teachers raised achievement by 0.28, while programs with more than 100 teachers raised it by 0.10, about a third as much. Doughty's team said its estimates rule out effects about a quarter the size of the average coaching effect (0.05 against 0.18) and half the average for larger programs (0.05 against 0.10).

The authors said their "strongest (but not conclusive) evidence" points to a shortage of effective coaches and the expectation that coaches do formal evaluations. As ICP grew, fewer coaches came from the top-rated teachers. Coaches who taught reading or math in 2014 had an average value-added score, a measure of a teacher's contribution to test score growth, of 0.06. That fell to 0.04 in 2015 and has hovered around 0.02 since 2016.

"As a program gets bigger, you require more and more high-quality coaches, and it's very hard to get people who are both excellent at teaching children in the classroom and excellent at coaching other adults," Doughty said.

Blazar, a University of Maryland professor, said ICP's model mirrors one used in many districts. "Just because we know that coaching works or tutoring works doesn't mean that it's a good idea to roll it out districtwide," Blazar said. Jessica Garner of the International Society for Transforming Education said the findings speak more to the difficulty of scaling and sustaining new programs than to coaching itself.

The authors also put the cost in pay terms. The roughly $5,800 per teacher equals about a 10% raise for every district teacher, and they concluded ICP did less for retention than they would likely expect from spending that money on teacher pay or bonuses. Using peers as formal raters also feeds a long debate over how teachers should be evaluated.

Coaching is common. In December 2023, 59% of public schools reported having at least one instructional coach, according to the National Center for Education Statistics, and 42% reported adding coach positions since the 2019-20 school year.

Sources
Educational Evaluation and Policy Analysis: Null Effects of an Intensive Teacher Leadership and Coaching Program at Scale (Doughty, Engel, and Atteberry, 2026)
Review of Educational Research: The Effect of Teacher Coaching on Instruction and Achievement (Kraft, Blazar, and Hogan, 2018)
National Center for Education Statistics: School Pulse Panel release, February 15, 2024
Education Week: A Major Teacher-Coaching Program Flopped. What Went Wrong?

Frequently asked questions

Does teacher coaching improve student test scores?
On average, yes. A 2018 review of 60 studies found coaching raised achievement by 0.18 standard deviations. But a 2026 study of one large district program serving more than 6,000 teachers found no detectable effect on reading or math scores.
Why did the large coaching program fail to raise scores?
The authors could not say for sure. Their strongest, though not conclusive, evidence points to a shortage of highly effective coaches as the program grew and to coaches doing formal evaluations of the colleagues they were supposed to be coaching.
How common are instructional coaches in public schools?
In December 2023, 59% of public schools reported having one or more instructional coaches, according to the National Center for Education Statistics. And 42% said they had added coach positions since the 2019-20 school year.
LOADING COMMENTS…
WRITTEN BY
Kate Carter
Kate Carter
Former Educator

Kate Carter spent nearly 20 years in public school classrooms before transitioning to education writing and curriculum consulting. She taught middle and high school English and social studies across two states.

EXPERTISE
K-12 curriculum and instructionEducation Policy