Fix a Vague IEP Goal: Seven Edits That Turn “Improve” Into Something You Can Measure

Some goals fail before the first progress report because the sentence does not tell staff what to observe. The fastest way to repair one is not to force it into a formula; it is to identify the missing decision. Read the draft aloud and ask what a data collector would still need to guess.

Edit one: replace “improve” with a response you can see

Draft: “Kai will improve reading comprehension.” Repair the behavior first: “After reading a grade-level informational passage, Kai will answer inferential questions and cite one supporting detail from the text.” Now the team can decide passage length, number of questions, supports, and criterion. “Understand,” “know,” and “be aware of” create the same problem unless the IEP defines how understanding will be demonstrated.

Edit two: make the baseline and target speak the same language

If present levels say the student writes an average of two complete sentences, but the annual goal is scored as a six-point essay rubric, the growth story is disconnected. Either rescore baseline samples using the rubric or choose a goal metric that matches existing data. The team should be able to draw a line from current performance to the annual target without translating between unrelated measures.

Edit three: expose the support that is doing the work

Draft: “Given support, Elena will complete a three-step routine.” What support? A picture schedule, one reminder, repeated verbal prompts, or hand-over-hand assistance? Each produces a different level of independence. State the permitted support in the condition and, when prompt fading matters, record the highest prompt level used.

Edit four: stop percentages from floating without a denominator

“80% accuracy” is incomplete if nobody knows whether it means 8 of 10 math items, 4 of 5 natural opportunities, 16 points on a 20-point rubric, or 80% of school days. Name the sample. Smaller denominators are also volatile: two successes out of three looks like 67%, and one additional success jumps the score to 100%. Choose a sample size that can show change without taking over instruction.

Edit five: define the mastery window

A single strong performance may be an outlier. If the team wants stable performance, say so: three consecutive weekly probes, three of four monthly writing samples, or at least four of five opportunities across three data weeks. The wording should fit the skill. Do not mechanically add “three consecutive trials” to every goal when natural opportunities are rare or a complex writing product is only sampled monthly.

Edit six: remove accidental compliance goals

Draft: “Avery will stay on task for 90% of class.” Ask what educational skill the team actually wants. Is the need initiating independent work, using a checklist, asking for clarification, sustaining attention for a defined work interval, or completing an assigned chunk? A broad on-task percentage can punish appropriate help-seeking, sensory regulation, or time spent thinking. Measure the functional behavior that improves access.

Edit seven: test the sentence with a cold reader

Hand the goal to a colleague who does not know the student and ask: What task would you give? What counts as correct? What support is allowed? How often would you collect data? When would you call it mastered? Every hesitation identifies an ambiguity worth fixing before the IEP is finalized.

Before-and-after example

Before: “Noah will improve organization with 80% accuracy.”

After: “At the end of the school day, given his posted dismissal checklist, Noah will place the homework folder, charged device, and required class materials in his backpack and mark each item complete without an adult verbal reminder on 4 of 5 school days across four consecutive weeks by the annual review date.” The second version tells staff which routine, which materials, which support, and what independence means.

Check that precision has not become fake precision

A goal can include many numbers and still be poor. “Within 3.5 seconds, 87% of the time” is not defensible if baseline data were casual observations and staff cannot time the behavior reliably. Use only the precision the measurement system can support. A clearly defined “within 10 seconds” may be more useful than an impressive-looking decimal.

If you want a searchable bank of measurable IEP goals organized by subject and grade band, see iepgoalwriter.pro. Treat model goals as examples to interrogate, not sentences to paste unchanged into an IEP.

End with the instructional question

The final draft should answer more than “can we collect a number?” It should help the team decide what to teach next. A decoding goal reveals error patterns; a routine goal reveals prompt dependence; a social response goal reveals where generalization breaks. If the data can only populate a progress report box and cannot inform instruction, the goal probably needs one more edit.

Do one final “data sheet test”

Before approving the goal, sketch the data sheet you would use tomorrow. If you cannot label the columns without inventing new definitions, the goal still contains ambiguity. This ten-minute test catches undefined prompts, missing denominators, unclear materials, and mastery rules that sounded acceptable during the meeting but cannot be scored consistently.