Somewhere between a research journal and your classroom is a gap that no one has fully bridged.

Research tells you what tends to work across hundreds of classrooms, with controlled conditions, carefully selected populations, and months of data collection. Your classroom gives you thirty students, forty-five minutes, Tuesday's specific energy, and about two minutes between class and your next responsibility to reflect on what just happened.

What translates from one to the other is not always obvious. A strategy with strong research support behind it can land completely differently in your specific room, with your specific students, in your specific school context, than it did in the studies. And conversely, something you've developed by instinct over years of practice — something that "works" in a way you can feel but have never articulated — may be doing exactly what the research says effective teaching does, even if you've never named it that way.

We want to help bridge that gap. And we want to do it in a way that honors what you already know — because the most important experiments you'll ever run are the ones that happen in your own room.

We're calling this approach Ground Truth (it can sometimes be called Local Science): the practice of running small, structured, honest experiments on your own teaching, and sharing what you find with other teachers who are doing the same.

– THE PROBLEM

Most teachers encounter new ideas in one of two ways. The first is wholesale adoption: you're convinced enough to implement fully, which means when something doesn't work on day three you're not sure whether to troubleshoot or abandon — because you've already restructured too much to run a fair comparison. The second is indefinite deferral: you're interested but not convinced, so it goes on a mental list of things to try someday, and someday rarely comes.

What's missing from both is the experiment mindset: let me try this deliberately, for a limited time, with a specific question in mind, and see what the data says — in my room, with my students.

This is how teachers actually deepen their practice. Not through overhauls. Through small, specific tests, honestly evaluated, iterated on or set aside based on what you actually observe.

– THE FRAMEWORK / EXPERIMENT

What you're running: A structured one-week classroom experiment on any teaching idea you've been curious about. One question. One change. One week of observation. A decision at the end.

What you need: A teaching idea, ten minutes of pre-experiment design, a brief daily observation note, and five minutes at the end of the week.

How long to run it: Five teaching days — long enough to see a pattern, short enough that nothing is irreversible.

---

Step 1: Write Your Question Before You Start

Every useful experiment has a specific, answerable question at its center. Not "does this work?" — that's too broad. The question is: what specifically do I expect to change, and how will I know if it changed?

Some examples of well-formed experiment questions:

• "If I begin every class with a two-minute silent retrieval prompt, will it take fewer redirects to bring students to focus during direct instruction?"
• "If I stop writing overall comments on drafts and instead write one specific feed-forward note per paper, will students make more substantive changes during revision?"
• "If I use a verbal anchor before moving students off devices, will the first segment after the transition have fewer off-task behaviors than my current baseline?"

Each of these has a specific change, a specific expected outcome, and an implicit measurement method. You need all three before you start. If you can't articulate what you expect to change, spend five more minutes sharpening the question.

Step 2: Design a Fair Test

A fair test means you've isolated the change as well as possible. Teaching is not a laboratory — there are dozens of variables you can't control. But controlling what you can is what determines how much you can learn from the result.

Keep as much constant as possible: run the experiment in the same class, at the same time of day, on the same type of activity across multiple days. If you're testing a transition protocol, test it every time there's a transition in that class — not occasionally.

Document a baseline before you start. Even one day of observing what currently happens — how many redirects you give, how long it takes students to settle, how many engage during the segment you're targeting — gives you something real to compare against. Without a baseline, "it felt better" is your only data, and feelings are not a reliable instrument for evaluating classroom changes.

Pick a narrow window to observe. Not whether the experiment affected everything — just one five-minute window, one type of student response, one specific behavior. One thing. Everything else is noise for this week.

Step 3: Run It and Take One Daily Note

Each day the experiment runs, write one sentence immediately after the relevant moment. Not a paragraph. One sentence.

The simplest form: "[What happened] compared to [what typically happens]."

Examples:

• "Transition took about 90 seconds today — usually runs 3–4 minutes before everyone is with me."
• "Three students revised the specific thing I noted; two revised unrelated things; four didn't revise. Roughly the same as usual."
• "Opening retrieval took exactly 2 minutes; class was quiet and on task for the full first segment without a redirect. Noticeable."

You're not interpreting yet. You're recording. Interpretation comes at the end of the week with all five notes in front of you.

Step 4: Evaluate Honestly at the End of the Week

On Friday, read your five notes. Then answer three questions:

Did I see what I expected to see? Compare your notes to your baseline and your original prediction.

What's the most plausible explanation for what I observed? This is where your professional judgment matters. A classroom experiment has confounders. The week might have been unusual. The class you tested might be atypical. Be honest about what you can and can't attribute to the experiment.

What do I do now — extend, adapt, or set aside?

Extend: You saw a signal worth exploring further. Run it another week, or try it in a different class.

Adapt: You saw something interesting but it didn't quite work as designed. Adjust one element and run it again. A result that surprises you isn't a failed experiment — it's information about what needs to change.

Set aside: You saw nothing, or a net negative. Write down what you tried and what you found. This is still useful knowledge — about your classroom, your students, what doesn't work in your specific context. That knowledge is professionally valuable even when the strategy didn't pan out.

---

What to Do With Your Results

Here's where Ground Truth becomes something more than a personal exercise.

When you run an experiment and find something — even something small, even something that surprised you — that result has value beyond your own room. Other teachers in similar contexts, with similar students, dealing with similar challenges, would benefit from knowing what you tried, what happened, and what you'd adjust.

We want this platform to be the place where that sharing happens. Drop your experiment question, your baseline, your daily notes, and your conclusion in the comments. Not because your findings are generalizable in the way research findings are — they're not, and that's fine — but because your specific, local, honest result is exactly the kind of knowledge that fills the gap between a research journal and a real classroom.

The accumulated Ground Truth of teachers across different schools and contexts is something no study has produced. We want to build it here, together.

– THE HONEST PART

Classroom experiments have real limits. You're not a researcher with controlled conditions and a statistical model. A strategy that fails in your experiment might work well for a colleague with different students. A strategy that works for you might not replicate elsewhere. Your sample is your own room, and one week is a short window.

This is fine. The purpose of Ground Truth is not to produce generalizable findings. It's to produce specific, honest knowledge about your practice — the kind no research study can produce, because no research study has ever studied your room. That knowledge matters. The accumulation of small, honest tests over a career is a significant part of what produces mastery. And sharing that knowledge with a community of teachers doing the same thing is how a profession learns faster than any individual teacher can alone.

– TRY IT OUT

Run the Experiment
Day 1	Write your question. Make sure it's specific. Document one baseline observation.
Days 2–5	Run the experiment. Write one sentence of observation each day. Don't interpret yet.
End of week	Read all five notes. Answer the three evaluation questions. Decide: extend, adapt, or set aside.
Share it	Drop your experiment and result in the comments. What did you try? What did you find? What surprised you?

Before you start, ask yourself: If this works, what will I see that I'm not seeing now — specifically? If you can't answer that, refine the question until you can. That's the question that makes the experiment worth running.

Did you run a Ground Truth? Add your experience in the comments to help more teachers just like you!

---

This is part of Accingo's Learning Lab — where teachers run Ground Truth: small, structured experiments that turn your own room into a source of professional knowledge.