Measurement workshop

Every example below is invented. No real student is described. Run the exercise on a fictional case first; bringing a live goal into the first pass turns a definitional argument into an argument about a child.

Why this exists

Two adults collect on the same goal all term. One counts an opportunity every time the direction is given. The other counts one only when the student was attending first. At the end of the term there are sixty data points, the percentage looks fine, and it describes nothing, because it is two different measurements added together.

No amount of careful graphing fixes that afterwards. The only place it can be fixed is here, before the first tally mark.

How to run it

Who

Everybody who will record on this goal. If a related-service provider or a second para collects even occasionally, they belong in the room.

How long

About 40 minutes for the first goal. The second goal takes ten, because most of the arguing was about the general case.

What you need

This sheet printed twice, the goal in front of you, and a pen. Fill it in separately, then compare. Filling it in together defeats the exercise: you will converge on whoever spoke first.

What you end up with

One agreed page, which becomes the handoff sheet and the criterion line on your data sheet.

Part 1 — What is one opportunity?

Write the sentence that would let a substitute count opportunities correctly on their first day.

The five cases that usually split a room

1. The direction is given while the student is mid-task on something else.
2. The direction is given twice because the first one was not heard.
3. A peer answers first.
4. The student starts, then a fire drill interrupts.
5. The opportunity arises naturally rather than being set up.

There is no universally right answer to any of them. There is only a right answer for this goal, and you need it written down.

Part 2 — What counts as meeting criterion?

The most common failure is a criterion that sounds precise and is not. “Independently and appropriately” is two undefined words in a trench coat.

Part 3 — Prompts and context

A prompted response is not automatically wrong, missing, or ineligible. What it means comes from the criterion you just wrote. Decide now, not at reporting time when the answer is convenient.

Part 4 — Zero, and nothing

This is the one that changes results. Five opportunities given and none met is a zero. No opportunities given is not a zero, it is an absence of data. On a sheet they look nearly identical and they mean the opposite.

The one people argue about hardest

A refusal after the direction was given. The opportunity was administered; the student did not meet criterion. Under most criteria that is a scored zero, not missing data. Recoding it as “not observed” because it feels unfair to the student is the single most common way a percentage quietly improves without anything changing in the classroom.

If your team decides refusals are excluded, that is a defensible choice. Write it down, apply it to every session, and say so in the report. What is not defensible is deciding session by session.

Part 5 — Rebuild the summary

Both of you take these four fictional sessions and produce a term summary. Do it separately. Then compare numbers before you compare reasoning.

The answers, and why there are two of them

60%. Pooled: 15 opportunities met out of 25 administered. Every opportunity weighted the same.

52.5%. The mean of 40%, 90%, 80% and 0%. Every session weighted the same.

Both are arithmetically correct and they answer different questions. Which one belongs in the report is decided by how the goal is written, not by which number is higher.

75% is what you get if you drop the zero, and it is wrong. Fifteen points of apparent progress, produced by deleting a day the student was present for.

The correction replaces the original. If Friday appears as both 3 of 5 and 4 of 5, ten opportunities enter the denominator where five were administered.

You can check all of this in the calculator, which shows both summaries side by side and will not let the correction be double-counted.

Part 6 — Write it down

The output of the workshop is one page that goes wherever the data is collected. Not a shared drive nobody opens. The clipboard.