Five cohorts. Five years. Roughly 150 schools. More than 6,000 people coached. And when the researchers finally ran the numbers, the program hadn't moved anything they could measure.

The largest evaluation of an instructional coaching program to date, published this month in the journal Educational Evaluation and Policy Analysis, found no benefits to either teacher retention or student test scores from an intensive, expensive teacher-coaching model used in hundreds of schools. Researchers led by Michelle Doughty, a postdoctoral fellow at the Tennessee Education Research Alliance, tracked the rollout of the Instructor Coaching Program in a large urban district serving a majority of students of color.

The title of the paper is blunt: "Null Effects of an Intensive Teacher Leadership and Coaching Program at Scale."

This is a schools story. It is also the single most uncomfortable piece of evidence currently sitting in front of anyone turning a one-to-one coaching practice into a group program, a certification, or a bench of associate coaches.

The model was the good one

This wasn't cheap PD delivered by a vendor on a Zoom link. The Instructor Coaching Program aimed to move effective teachers into part-time instructional coaching positions, so they could support other teachers while remaining in the classroom. Coaches were released from teaching part time to coach assigned mentees, lead professional learning communities, and contribute to teacher evaluations.

Practitioners coaching practitioners. Peer credibility. Embedded, ongoing, not a workshop. If you have ever pitched a group program, you have pitched this design.

It was also expensive. Doughty and her colleagues found ICP cost about $270 per student in a participating school, with roughly $52,500 per coach.

The program was rolled out in five cohorts of schools and eventually reached over 6,000 mentee teachers. Cohort 1 started in 2014; Cohort 5 adopted in 2018. By the end of the rollout, all of the approximately 150 schools in the district were running it.

Then the outcomes. The program did not increase teacher retention at the school or district level, and had no measurable effect on student achievement in English/Language Arts or math — estimated effects on math leaned negative, though not significantly.

The talent math nobody puts in the business plan

Here is the sentence every coach scaling a team should read twice.

"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.

Swap the nouns. It's very hard to find people who are both excellent at the thing and excellent at coaching other people to do the thing. Those are two different skills, and the second one is rarer.

The paper's own reading of the data points there. Because more effective teachers tend to become higher-rated coaches, the authors write, this may indicate the program struggled to recruit sufficient numbers of effective coaches, leaving the program with null effects overall — an explanation with supportive evidence in the heterogeneity analysis but not definitive causal proof.

The first ten coaches you hire are your best people. The next fifty are the market.

The dosage quietly drains out

For schools that joined in early years, the mentoring was intense, Doughty said. It did not stay that way.

One published breakdown of the study's data notes that early cohorts saw coaches conducting about five observations per mentee annually, dropping to two or three per year after 2018, and that the program initially recruited the district's most highly effective teachers, with the talent pool thinning as it expanded universally.

Nobody voted to cut the program in half. It just got stretched. Same name, same brochure, same price, less of it per person.

That is exactly how a $20,000 one-to-one engagement becomes a $4,000 group program with a monthly "hot seat" and a Slack channel nobody reads.

A mediocre coach is not a neutral coach

The most useful finding for the coaching business is buried in the supplemental analysis. Using coach effectiveness measures built from teacher surveys, the researchers found suggestive evidence that the program may increase teacher retention in schools where teachers were more positive about their coaches, while it seems to decrease retention in schools where teachers were more negative about their coaches.

A weak coach didn't produce zero. It pointed the wrong way. Average out enough of both and you get a null result — which is precisely what a scaled program looks like from the outside: fine, nothing wrong, nobody complaining, nothing happening.

The role conflict most certifications ignore

There's a structural trap here too. The same analysis of the study notes that the program formalized a dual role rather than separating coaching from formal evaluation, and by 2022 roughly 80% of mentee teachers received their summative professionalism ratings directly from their coaches. The literature cited in the study warns that combining high-stakes evaluation with peer coaching often generates defensiveness, limiting willingness to reflect openly and act on feedback.

If your certification program's mentor is also the person who signs off on whether the client passes, you have built the same machine. People do not get vulnerable in front of the person holding the grade.

The questions to ask before you add seats

Doughty said leaders who have or are planning instructional coaching should consider a set of questions, starting with how coaches are chosen and trained, and how coaching fits into overall staffing and career options.

This is not an isolated result either. Though the study looks at one program, ICP's model mirrors that used in many districts nationwide, said David Blazar, a professor of urban education policy at the University of Maryland. And the warning has been on the table since 2018, when a meta-analysis of 60 randomized trials found that the effectiveness of coaching programs declines as the number of teachers involved increases. "It's very hard to cut corners financially and maintain quality in teacher professional-development programs," Brown University's Matthew Kraft said at the time.

The kicker is financial. The researchers point out that putting that budget toward a general 10% salary increase for all teachers would likely have reduced turnover far more effectively than the coaching program did.

The boring, unscalable, unsexy intervention beat the designed one.