Why 'train and hope' never worked — and what actually changes practice.

The central finding of Fixsen et al.'s (2005) synthesis is blunt and uncomfortable: professional development, in isolation, doesn't work. Not because training is poorly designed. Not because teachers aren't paying attention. But because the gap between learning something in a workshop and doing it automatically during instruction is not a knowledge gap — it's an infrastructure gap.

Fixsen reviewed the evidence from experimental studies across multiple fields: health, mental health, education, juvenile justice, social services. In every domain, the pattern was the same. When practitioners received training without accompanying organizational support, implementation rates hovered between 0 and 20%. When organizational support structures were in place, implementation rates approached 80–90%.

For education specifically — and for MTSS in particular — this means that the majority of districts are spending significant resources on training that, by the weight of the evidence, produces minimal lasting change in classroom behavior. The training is real. The investment is real. The failure is also real, and it's predictable.

The phrase "train and hope" — originally coined by Stokes and Baer in 1977 — describes the dominant model in professional development: provide training, assume transfer, discover failure, provide more training. Fixsen's synthesis found no evidence that this cycle produces different results with repetition. The model itself is the problem.

Fixsen's most important contribution is not simply documenting that training fails — it's explaining what successful implementation actually requires. His framework identifies three categories of organizational drivers, each essential, none sufficient alone:

Competency Drivers build the skills of practitioners and their supervisors. They include staff selection (who you hire matters as much as how you train them), pre-service and in-service training, and coaching that bridges workshop knowledge and daily practice.

Organization Drivers change the structures, systems, and data flows that practitioners work within. They include systems interventions (tools and workflows that make new practices easier), facilitative administration (leadership that removes barriers rather than adds requirements), and decision support data (feedback mechanisms that make implementation visible).

Leadership Drivers sustain implementation over time through both technical leadership (solving known problems with known solutions) and adaptive leadership (navigating complex change with no clear playbook).

Fixsen's framework reveals a critical misalignment in how most districts approach MTSS: they invest almost exclusively in competency drivers (training, coaching) while neglecting the organizational driver tier entirely. The result is exactly what the research predicts — capable teachers who know what to do but lack the systems infrastructure to do it consistently.

Of the three driver types, the organizational driver is the most consistently underfunded and least understood in educational contexts. Fixsen describes it as the infrastructure layer that makes trained behaviors feel natural and automatic rather than forced and effortful. He draws an instructive analogy from Toyota's production system:

Toyota didn't improve quality by training workers to care more about quality. They redesigned every workflow, every physical workspace, every communication channel so that the quality-producing behavior was the path of least resistance. The new way became easier than the old way.

For MTSS, the implication is direct: until the intervention-documenting behavior is easier than the alternative (not documenting), it will be inconsistently performed regardless of training quality. The 30-minute end-of-day logging that most MTSS systems require isn't just inconvenient — it structurally guarantees inconsistent implementation, because it competes with every other demand on a teacher's time and energy at the end of a school day.

Fixsen also identifies decision support data as an essential organizational driver: a regular flow of reliable information about performance, acted upon by practitioners, managers, and the organization. This feedback loop is what distinguishes organizations that improve from those that stagnate. Without visible data on implementation, there is no basis for coaching, no basis for celebration, and no basis for improvement.

Fixsen's analysis of Joyce and Showers' coaching research (2002) provides the single most compelling data point in the entire implementation literature. When you measure the relationship between training type and implementation rates, the gap between coaching and no coaching is extraordinary:

Coaching works because it provides what training cannot: feedback in real time, support in the actual performance context, and accountability across multiple instructional cycles rather than a single workshop experience. A coach can observe implementation, name what they see, suggest specific adjustments, and return to see whether those adjustments stuck.

The problem is simple arithmetic. A school with 40 teachers and one instructional coach cannot provide meaningful coaching across 40 classrooms with sufficient frequency to drive the 80–90% implementation rates the research identifies. The human bandwidth doesn't exist.

One of Fixsen's most operationally useful insights is deceptively simple: for a new practice to replace an old one, it must be made easier — not just better or more evidence-based. Educators are not immune to the behavioral economics of cognitive load. When a new practice requires significantly more time or mental effort than the existing practice, the existing practice wins regardless of its relative ineffectiveness.

Fixsen identifies several common implementation barriers in school settings: difficulty securing release time for training, the compounding effects of absences and staff turnover on training participation, and the particular challenge of requiring teachers to demonstrate new skills in artificial "role play" settings rather than with real students in real instructional contexts.

Each of these barriers points to the same structural problem: most implementation efforts place the burden of change on the practitioner rather than redesigning the environment to make change easier. When burden falls on practitioners, those with the least available capacity — often those teaching the most challenging students — are the least likely to successfully implement new practices.

Fixsen's analysis of successful large-scale implementation programs — including Multisystemic Therapy (MST) — reveals a consistent feature: they all have built-in feedback mechanisms that continuously monitor whether the practice is actually being used as intended. This is not the same as a principal observation or an annual program evaluation. It is ongoing, granular data about practitioner behavior in the actual performance environment.

Notice what this describes: a digital system that captures what practitioners are actually doing, makes that data available to supervisors, and feeds it into ongoing consultation. This is not a once-a-year audit. It is a continuous loop that makes implementation visible at both the individual and organizational level.

For MTSS, the equivalent would be a system that captures what Tier 1 interventions teachers are actually using, with which students, and how often — automatically, without requiring separate self-reporting. This kind of data would allow instructional coaches to see implementation patterns without being physically present for every lesson. It would allow PLC teams to compare intervention effectiveness across classrooms. It would allow administrators to distinguish between high-fidelity and low-fidelity implementation before waiting for state assessment results.

One of Fixsen's most pointed observations concerns the measurement problem at the heart of educational implementation. Districts have robust systems for tracking whether training occurred — sign-in sheets, PD hours, certification records — but almost no systems for measuring whether trained practices are being used in classrooms. The thing that gets measured is the easier thing to measure, not the thing that actually matters.

This creates a persistent blind spot. A district can demonstrate with confidence that 95% of its teachers attended MTSS professional development this year. It cannot demonstrate — with any comparable confidence — what percentage of those teachers are implementing MTSS Tier 1 strategies during daily instruction. The gap between these two numbers is the gap that Fixsen's research predicts, and that most districts have no way to observe, measure, or address.

Every core finding from Fixsen et al.'s 2005 synthesis maps directly to a design decision in Trellis. This is not coincidence — it is the architecture.