Coaching Letter #233

May 10, 2026

Scaling, part the third

Friends, I hope this finds you well. Thank you for subscribing to the Coaching Letter—you rock. I was out of town last week, so lots of time to write while sitting on airplanes and waiting for things to happen. This work on scale has taken on a life of its own, so here is the next layer. If you want to read the first two in this series, CL #231 is about what I labeled the two logics. Logic Alpha is a top-down, plan-driven approach that prioritizes the measurable parts of scaling up (number of computers purchased, number of teachers trained, number of classrooms visited, number of tests taken, number of students enrolled…) without being clear or precise enough about the practice that is actually making a difference, nor taking into account the unpredictable opportunities that present themselves when implementing. Logic Phi, by contrast, attempts to grow the work by taking advantage of those opportunities, which makes it seem very organic, but it is much more precise (aka “tight” or “constrained”) about what is being scaled. And CL #232 is about the work of John List and his list of ways in which scaling fails—these are what ChatGPT calls failure modes; I don’t know where that terminology comes from, but I like it and have taken to using it.

And one of the many great things about my job is that I get to talk to so many people, and those conversations frequently enable me to make connections between ideas that I hadn’t previously seen. So this CL is an outgrowth of one of those conversations. But also, when these connections occur to me during a conversation that is ostensibly about something entirely different, I write it on a post-it and then often don’t know when the conversation happened or who I was talking to. So if you’re the person I was talking to when I had this brainwave, thanks for the inspiration, and sorry for forgetting.

This Coaching Letter continues the theme of scaling, and focuses on the relationship between the work of Cynthia Coburn on the one hand, and the very common activity that districts engage in of creating a theory of action for their strategic plan.

Let’s start with the idea of a theory of action. This is something that is frequently included in district requirements for school improvement plans, and state requirements for districts that are under some sort of additional oversight because of low student achievement. The basic idea is that districts should have a very clear rationale for the actions they intend to take according to the plan. In Instructional Rounds, the authors list the requirements for a theory of action as follows:

1. “It must begin with a statement of a causal relationship between what I do—in my role as superintendent, principal, teacher, coach, etc.—and what constitutes a good result in the classroom.
2. It must be empirically falsifiable; that is, I must be able to disqualify all or parts of the theory as a useful guide to action that is based on evidence of what occurs as a consequence of my actions.
3. It must be open ended; that is, it must prompt me to further revise and specify the causal relationships I initially identified as I learn more about the consequences of my actions.”

I think this is absolutely true and you should absolutely do this—but very seldom do theories of action actually get tested in the way that the authors of Instructional Rounds intended. For example, I think most educators would struggle to explain what it means to have an open-ended theory of action. And despite the exhortation that the theory of action should be falsifiable, very few organizations go about systematically doing that. Their measurement systems tend to be focused on implementation rather than impact. For example, districts are very tempted to make the very quick leap from talking about what high quality instruction ought to look like to developing a walkthrough tool. It is not uncommon for us to run workshops where district teams will list as their next to do: “create look-fors”.

Why is that not the most appropriate next step? Because it rests on the assumption that replicating what something looks like is the same as replicating what it does.

(I thought about putting that in all caps, but I decided that risked looking a bit hysterical so I settled for bold and putting it in its own paragraph.) This, then, is a conversation about form versus function. A theory of action is supposed to be a theory: if this happens, then we think it’s likely it will cause this, because there is a mechanism at work that means that we can predict results. But this assumes that the form that’s being replicated is a reliable and valid representation of the function. And that is not always the case.

Let’s use BTC as an example. (There are lots of new folks subscribing to The Coaching Letter, so for their benefit: BTC is the acronym for Building Thinking Classrooms, and I’ve written about it several times in CLs #221, #202 and #198). BTC is not revolutionary, in the sense that it is an enactment of research that is decades old—that’s what I wrote about in #221. But it is revolutionary in the sense that a BTC classroom looks very, very different from a traditional classroom—kids are standing the whole time, working in random groups, using dry erase boards and markers to capture their ideas, working on a low-floor high-ceiling grade level task, the best of which can keep kids (and adults) thinking for incredible lengths of time. But if you try to teach using BTC when all you have to go on is watching your colleague down the hall, you would be forgiven for thinking that BTC means taking the curriculum and having kids work on it in groups at non-permanent vertical surfaces—which both is and is not true, and it’s hard to tell the difference if you don’t understand the function that the form is attempting to achieve. The point of form is function, not form for the sake of it.

So it also bears pointing out that the requirements for a theory of action in Instructional Rounds does include the specification of a causal relationship—but mostly in #3, and as we just saw, most districts only get as far as the first half of #2—and it doesn’t really go far enough. I’m fairly certain that most organizations don’t get as far as a conversation about the underlying causation of a particular change that they’re looking for. For example, in addition to the confusion over BTC, we see a lot of schools and districts focused on student discourse, but we don’t hear a lot of conversation about what it is about student discourse that makes student achievement more likely. (We also frequently hear that “the person doing the talking is the person doing the learning”, but that is patently not true—I don’t think you’ll find many studies that show an inverse correlation between teacher talk and student learning, and as Dylan Wiliam has pointed out, there are plenty of countries that score higher than us on PISA where there is more teacher talk than there is in the US.) And unless you know the “why” behind student discourse, it is easy to replicate the form and not the function.

(If you are now thinking, “yikes, I don’t think I could clearly articulate the relationship between student discourse and student achievement based on research and theory”, then this is a good time to subscribe to my colleague Tom’s Substack, and look up his posts #13, #14, and #15. The last one is most directly about student discourse but they function as a set.)

OK, so to tie all that together… A theory of action is making a claim about function. The examples I have given are intended to show the same error: confusing the visible form with the underlying function. If you want a really strong theory of action that will drive your strategy, it needs to be clear what the connection is between the proposed action and the rationale behind it. A theory of action, in other words, needs a because, and everyone needs to know that what comes after the because is actually the driving force behind the strategy—the actual mechanism that causes improved student learning.

And—and this is where Instructional Rounds does not serve us all that well (the book, not the practice)—the because needs real research-based teeth. The definition of a theory-of-action in Instructional Rounds is agnostic when it comes to what strategy you should adopt. It implies: Start where you think you have a good theory and test it out. And my strong opinion is that this is magical thinking—mostly because districts have neither the time, the capacity, nor the patience to start from a loose premise and engage in multiple rounds of research to figure out whether something works. If they did, we would no longer be talking about differentiation the same way, nor student discourse, nor learning styles (still comes up!), nor grouping, nor learning targets, nor… the list is endless. And besides, that’s just really inefficient. So please don’t start there—choose what the research tells us is most likely to lead to improved student learning—you can start by reading the Coaching Letter on OTL, which will lead you to focus on the amount of time devoted to grade level instruction (as opposed to review), the quality and qualities of the tasks students work on, and the kind of instruction that sustains cognitive demand. The definition of a good theory-of-action is not that it meets the 3 criteria, it is that it has a high likelihood of producing increased student learning.

OK, so now about Coburn’s work—because her work helps us see what it would mean for the function described in a theory of action to actually scale. Just to be clear, in the coaching letter about Logic Alpha and Logic Phi, I was already drawing on her work. And my notes are from her 2003 article in Educational Researcher, “Rethinking Scale: Moving Beyond Numbers to Deep and Lasting Change”, but I know she’s written more about the topic since then. If you ask ChatGPT for a one-sentence summary, you’ll get something like: The field has mis-specified the problem of scale by treating it as a quantitative expansion problem, when in fact it is a multidimensional problem of change, learning, and system transformation. She provides clear language to talk about what is being scaled, and that’s more than just form. But as she points out, form is easier to count, so we usually just focus on that. So there you have it. That’s the claim, and she goes on to generate a more nuanced definition of scale that has four dimensions: depth, sustainability, spread, and shift in reform ownership. So here’s what she means by those things, but first, here’s what she says about when scaling is defined by uptake measured solely in numbers:

“This definition is attractive in its simplicity, its intuitiveness, and its measurability. But what does it really mean to say that a reform program is scaled up in these terms? It says nothing about the nature of the change envisioned or enacted or the degree to which it is sustained, or the degree to which schools and teachers have the knowledge and authority to continue to grow the reform over time. By focusing on numbers alone, traditional definitions of scale often neglect these and other qualitative measures that may be fundamental to the ability of schools to engage with a reform effort in ways that make a difference for teaching and learning.”

Here are the four dimensions:

1. Depth. This is the point that “is it implemented?” is the wrong question. What we really want to know is “Has classroom practice changed in deep and consequential” ways, specifically: “I am referring to teachers’ underlying assumptions about how students learn, the nature of subject matter, expectations for students, or what constitutes effective instruction. Many external reform initiatives promote a view of teaching and learning that challenges conventional beliefs about one or more of these dimensions. The question is: Do teachers’ encounters with reform cause them to rethink and reconstruct their beliefs? Or do they alter reforms in ways that reinforce or reify pre-existing assumptions?” Fans of Richard Elmore will recognize similarity to the book Restructuring in the Classroom, which shows how reforms may or may not meaningfully change classroom practice, but an even better case is Cohen’s Mrs Oublier, who had curriculum, materials, and training, but her teaching managed not to shift, regardless of those factors. She didn’t “resist”, and she thought she was doing what she was supposed to be doing, but she was missing the mark without knowing it.
2. Sustainability. How long does a change have to last in order for it to be judged as successfully scaled? Coburn argues that scale is meaningless without sustainability, that most studies don’t actually last long enough to measure sustainability, and that reforms often appear successful at first and then decay over time. I think we would all recognize the phenomenon whereby an initiative is “put in place”, it is easy to assume that the change has happened (like a train being switched from the tracks to one destination to another) and therefore you don’t need to pay attention to it any more. From the practitioner perspective (i.e. the view from the classroom), the the thing that you’ve been talking about all last year suddenly is never mentioned any more, so the understandable inference is that the district has moved on to something else, so you can go back to what you were doing before—and that has happened often enough that it’s the explanation that makes the most sense.
3. Spread. This sounds like it is replication across classrooms and/or schools, but Coburn seems to be talking about diffusion of norms, beliefs, and principles. Which is, of course, the hardest to do. And this connects with Donella Meadows’ work on places to intervene in a system—I know I harp on about how useful that article is, but it really is great.
4. Shift in reform ownership. If a change requires constant oversight and endless conversations about accountability, then the change cannot be meaningfully said to have scaled, because it is dependent on someone higher up in the chain of command to keep it in place. That’s the blunt force version of scaling. But there’s also the issue of a change being dependent on outside expertise to keep it going. I think about this all the time when it comes to our work in instructional improvement. If what we are promulgating is so onerous, technical, or niche that a district cannot be expected to sustain on its own without our ongoing support, we have engaged in maintaining our own job security but haven’t done the client any favors. In either case, what needs to happen is that everyone in the organization understands the form and the function, and is working towards a more perfect expression of both.

What should be clear by this point is I’m arguing that what most districts are trying to scale is really a theory of form rather than a theory of action. I’m trying very hard to land this, so I hope what is intended as emphasis doesn’t read as repetition. When you scale the form of a practice without understanding function—and that’s what you measure and “hold people accountable” for—you get surface consistency but actually a great deal of variation that can be quite hard to see. Second, you get little impact on student learning, because the places where form and function are applied appropriately are canceled out by the places where there is form without function. Therefore, a superintendent, were they so inclined, could legitimately claim both that they had scaled a practice and that they had tried it and it hadn’t worked. Third, you risk abandoning a practice that could have been really successful, which is just painful to witness.

Ironically, Instructional Rounds itself (the practice, not the book) is one of the places you’ll see form implemented over function most often. Years and years ago I talked to a superintendent who has long since retired about her district’s theory of action and their deployment of instructional rounds. The theory of action was in what I would call the “ludicrously weak” category—it was something like, “if teachers collaborate then scores will go up”, which is probably a slight oversimplification, but only slight. And every rounds visit basically showed the same thing: that in general (there is always variation) the tasks students were being given were not challenging, and the instruction was pretty lame. I asked her about the fact that rounds visits showed this repeatedly, and she said, I’m not kidding, “sooner or later they’ll figure it out.” The “they” in question were the principals. I guess she thought that rounds itself was the intervention to improve instruction? Who knows. Anyway, here was a potentially powerful practice being implemented in such a way that the form was in place but the function was woefully misunderstood and therefore no one should have had any expectation that it would actually change instruction for the better.

Just for grins, I asked ChatGPT for a theory of action for Instructional Rounds, and I am copying what it said at the very bottom of this post. Nota bene the following points: it’s not about visiting classrooms, it’s about organizational learning; it rests on developing a shared understanding of high quality instruction; it is attempting to shift mental models… The superintendents who understand this—and there are many—use rounds to ask questions about their strategy and use their findings to inform their tactics—they are, in other words, porting the function of rounds into their own setting. Those who don’t, don’t, and are merely mimicking the form.

Last little provocative remark? Because of Simon Sinek, we have taken to talking a lot about “know your why”, which has come to mean identifying the core purpose, cause, or belief that drives your actions and decisions, rather than focusing only on “what” you do. Supposedly, understanding this underlying motivation provides clarity, boosts motivation during difficult times, and ensures your actions are aligned with your deeper purpose. OK, great, not going to argue with that. But knowing why, in the sense of knowing the mechanism underlying your strategy, and therefore what your thinking about scale should be based on, is much more important for the organization than that every person in that organization “know their why”.

Finally, my organization is in many ways the home of Instructional Rounds, and recently we have lost two of its greatest proponents—who also happened to be great supporters and friends of mine—Andrew Lachman, who wrote the Forward to Instructional Rounds (the book), and Jerry Belair. They were both smart and honorable and dedicated—my dad would have called them two of “the good ones”. Please reach out if we can help you with Rounds, with scaling, with strategic planning—or with anything else. Best, Isobel

Hey, can you get underneath all the guff about instructional rounds and figure out what exactly the theory of action is that connects rounds to instructional improvement.

Short answer: there is a theory of action behind instructional rounds, but it’s usually buried under ritual and language. When you strip it down, it’s a learning system for adults that is supposed to change the instructional core by changing how people see, talk about, and act on instruction.

Let me reconstruct it cleanly.

The implicit theory of action

At its most defensible, rounds rests on something like this:

1. If educators develop a shared, evidence-based understanding of high-quality instruction (through disciplined observation of the instructional core),
2. and if they anchor that understanding in fine-grained, non-evaluative description rather than judgment,
3. then they will improve their ability to notice what actually matters for student learning (i.e., task, student thinking, teacher moves),
4. which will shift their mental models about instruction,
5. which will lead to more coherent and aligned decisions about teaching, coaching, and leadership practice,
6. which, over time, improves the quality of instruction across classrooms.

That’s the clean version. Notice what’s doing the work: perception → shared meaning → decision-making → practice.

What’s actually distinctive about rounds (in theory)

Rounds is not, fundamentally, about visiting classrooms. That’s the surface feature. The mechanism is:

• Disciplined seeing (description before interpretation)
• Collective sensemaking (norming what counts as evidence of learning)
• Deprivatization of practice (making instruction a public, discussable object)
• Network-level learning (not just individual teacher improvement)

In other words, it’s trying to solve a knowledge and coordination problem, not a skill deficit problem.

Where it quietly aligns with your world

If I translate this into your language, the theory of action is trying to operate on the cultural–discursive architecture:

• Build shared language about instruction
• Reduce idiosyncratic interpretations of “good teaching”
• Create a common referent for what’s happening in classrooms

And secondarily, the social–political architecture:

• Shift norms from evaluation → inquiry
• Legitimize talking about practice without defensiveness

But—and this matters—it is weak on the material–economic side. It doesn’t, on its own, change:

• Tasks
• Curriculum
• Time structures
• Team routines
• Coaching cycles

Which is one reason it so often floats above practice.

The real hinge point (and where it often fails)

The entire model depends on a fragile assumption:

Better shared understanding of instruction will translate into better instructional decisions in practice.

That’s a big leap.

It assumes:

• Noticing → knowing what to do
• Knowing → doing
• Doing → sustained system change

You and I both know that chain breaks constantly.

The “guff” you’re reacting to

Most implementations get stuck in one of three failure modes:

1. Descriptive compliance
   People learn to describe without judgment, but it never connects to action. It becomes a kind of observational etiquette.
2. Ambient coherence illusion
   Everyone uses the same language after a while, so it feels like alignment, but practice doesn’t shift.
3. Decoupling from improvement mechanisms
   Rounds sits alongside coaching, team meetings, and PD rather than structuring them.

At that point, rounds becomes what Meyer & Rowan would call a ceremonial practice—it signals seriousness about instruction without necessarily altering it.

The version that might actually work

If you force the theory of action to tighten, it starts to look much closer to what you’re already doing:

• Observation is not an endpoint; it is input into team-based improvement cycles
• The unit of learning is not the network; it is teacher teams working on shared problems of practice
• The output is not “next level of work” in the abstract; it is tested instructional routines (your recipes)
• The goal is not just shared understanding; it is reduction of variation in enacted practice

So a stronger theory of action would read more like:

• If we use disciplined observation to build shared understanding and
• immediately embed that understanding into iterative, team-based enactment cycles focused on task and student thinking,
• then we change not just what people see, but what they do repeatedly,
• which is what actually moves instruction.

Bottom line

Rounds is built on a serious insight:
You can’t improve what you can’t see, and you can’t scale what you can’t name collectively.

But it overestimates how far seeing and naming will take you without:

• structured enactment
• repetition
• feedback loops tied to practice

Left alone, it’s a perceptual intervention.
To drive improvement, it has to become part of a practice-production system.

 

Isobel Stevenson, PhD PCC
Author of The Coaching Letter