STEM Coaching vs. Traditional PD: Why Embedded Coaching Outperforms Workshop Models

Most districts spend their professional development budget on workshops: full-day sessions, outside presenters, binders full of strategies, and a shared lunch. Teachers attend, take notes, return to their classrooms — and by the following Monday, the vast majority of what they heard has no effect on what they do.

This isn't cynicism. It's the documented implementation rate for workshop-based PD. The research on what actually changes teacher practice has been consistent for two decades: one-shot training events don't work at scale. Coaching does. The problem is that most districts still default to workshops because they're easier to plan, cheaper to run in the short term, and more familiar. Coaching feels harder to start. But once you understand the mechanism behind the difference — and see the side-by-side numbers — the case becomes hard to argue against.

The Workshop Problem

In 2015, TNTP (formerly The New Teacher Project) published The Mirage, a study of professional development spending across three large US school districts. The findings were stark: districts spent an average of $18,000 per teacher per year on PD, with no measurable effect on teaching effectiveness. Teachers reported feeling improved after PD. Their actual practice didn't change.

The mechanism behind this is well-understood. Joyce and Showers' foundational research on transfer of training established that the components of PD directly predict whether training transfers to practice:

• Theory and explanation alone: ~5% transfer to classroom practice
• Theory + demonstration: ~10–15% transfer
• Theory + demonstration + practice: ~20–30% transfer
• Theory + demonstration + practice + coaching: ~80–90% transfer

Workshops deliver the first two. The typical "sit-and-get" session — a presenter explains a strategy, demonstrates it, maybe has teachers try it in a low-stakes simulation — lands somewhere around 10–15% implementation when teachers return to their actual classrooms with actual students and actual curriculum demands.

The $18,000-per-teacher figure from TNTP's study didn't produce bad teachers — it produced a system that invested heavily in professional growth and got almost nothing back in changed practice. That's not a criticism of the individuals involved. It's a structural problem with the delivery model.

What Makes Coaching Different

The coaching advantage isn't magic — it's structural. Three specific properties of instructional coaching produce outcomes that workshops can't replicate:

These aren't incremental improvements to the workshop model — they're different mechanisms altogether. A workshop delivers information. Coaching changes behavior. Those are different problems that require different solutions.

The duration effect is particularly important and often underestimated. A single coaching session produces similar transfer rates to a workshop. It's the accumulation of sessions over time — what researchers call coaching dosage — that drives the 60%+ implementation rates. This is why tracking session frequency matters from day one: you can't optimize dosage if you're not measuring it.

The Evidence

The research base on instructional coaching is substantial and consistent. A few key findings worth knowing when making the case internally:

Jim Knight's research at the University of Kansas Coaching Research Project tracked 151 teachers over a school year. Teachers who received instructional coaching changed practice in 60% of cases — compared to 5–10% for workshop-only control groups. The effect size for student achievement outcomes in coached classrooms was 0.4 standard deviations — substantial by any measure of educational intervention.

The Learning Forward standards for professional learning — the framework most districts reference for PD planning — explicitly identify coaching as a key component of effective professional development. Their research synthesis found that effective PD requires a minimum of 50 hours of instruction, practice, and coaching over 6–12 months to produce meaningful change in teacher practice and student outcomes.

More recently, a 2021 meta-analysis by Kraft, Blazar, and Hogan (Harvard Graduate School of Education) reviewed 60 randomized controlled trials of teacher coaching and found an average effect size of 0.49 standard deviations on instructional practice — equivalent to moving teachers from the 50th to the 69th percentile in effectiveness. No comparable meta-analysis of workshop PD produces numbers anywhere near this range.

For the metrics your district should be tracking to demonstrate these effects, see 5 Metrics Every STEM Coaching Program Should Track.

Side-by-Side: Traditional PD vs. Embedded Coaching

The comparison below covers the dimensions that matter most for district decision-making — not just effectiveness, but cost, scalability, and accountability. This is the table worth showing at a board meeting.

The cost-per-teacher comparison deserves attention. At first glance, coaching looks expensive because coach salaries are visible line items. But when you divide a coach's annual reach (20–30 teachers consistently served per coach) into their total compensation, the per-teacher cost is typically lower than the workshop spending it replaces — and it produces demonstrably better outcomes. The ROI calculator models this comparison with your specific district numbers.

Making the Transition

Switching from a workshop-heavy PD calendar to a coaching-forward model doesn't happen in one budget cycle. The districts that do it successfully follow a consistent three-step pattern:

This transition doesn't require eliminating workshops overnight. Most districts run a hybrid model during the transition: workshops for large-group content delivery (new curriculum adoption, policy changes), coaching for the practice and implementation that follows. The key shift is treating workshops as the starting point for learning, not the ending point — and building coaching into the follow-through.