Measurable goal checklist
A goal that cannot be monitored will still be monitored, badly, for a year. Eleven questions, and most of the goals that fail do it on the same three.
This is not a compliance check. It says nothing about whether a goal is legally sufficient or appropriate for a student. Those are judgements for the team that knows the child. What this covers is narrower and entirely practical: can somebody collect data on this, and will two people collecting it get the same answer?
The eleven
| Question | Notes | |
|---|---|---|
| 1. Is the behaviour observable? Could a substitute see it happen, or does it require inferring a state of mind? “Will understand” and “will be more confident” are not observable. | ||
| 2. Is one opportunity defined? Can you say, without hesitating, what makes a moment count as a chance to perform? | ||
| 3. Is meeting criterion defined? Time limit, completeness, quality bar. If the words “appropriately” or “independently” appear, are they defined elsewhere? | ||
| 4. Are the conditions stated? Setting, materials, who is present, what happens first. | ||
| 5. Is the prompting expectation stated? Does a prompted success count? Which prompts? | ||
| 6. Is the measurement method named? Trials, frequency, duration, latency, interval. Not just a number at the end. | ||
| 7. Is the unit unambiguous? Percent of what. Per what window. Out of how many. | ||
| 8. Is there a real baseline? With its denominator, not a recalled impression of “about half the time”. | ||
| 9. Is the target reachable from the baseline? And is the number a judgement, rather than a habit of rounding up to 80%? | ||
| 10. Is the collection schedule realistic? Weekly across six goals is thirty sessions a week. Is that going to happen? | ||
| 11. Is it clear who collects and who reviews? Named, not assumed. |
Where goals usually fail
- Question 2, every time
- The criterion is precise and the opportunity is not. “80% of opportunities” with no definition of an opportunity means each collector invents their own denominator, and the percentage is not comparable across people or across weeks.
- Question 7
- “80% accuracy” is not a unit. Accuracy across what: opportunities, sessions, days, items? The calculator shows how far apart two readings of the same data can land.
- Question 10
- A schedule nobody can keep produces gaps, and gaps produce a report where the honest sentence is “there is not enough data to say”. Better to write a schedule you will actually meet.
- Question 9, quietly
- A baseline of 20% and a target of 80% inside a year is a specific clinical claim. Sometimes it is right. It is worth being a decision rather than a default.
Rewriting one
Fictional, and deliberately ordinary rather than a strawman.
| Before | After |
|---|---|
| The student will follow directions independently 80% of the time. | When given a two-step direction during the arrival routine, the student will begin the first step within 10 seconds with no repetition, gesture or physical prompt, on 80% of opportunities across five consecutive sessions of five opportunities each. |
The second one is longer, and it is the reason two people can collect it and get the same number. It names the opportunity, the criterion, the time limit, the prompting expectation, the setting, the denominator and the sample size. A substitute could run it on their first morning.