A student answers eight of ten sight-word cards correctly in the resource room every day. In the general education classroom, with the same words on a worksheet, she gets three of ten. The goal says "reads 30 sight words." Is it met?
It depends on what the goal says and where the data were taken. This guide covers how to measure whether a skill works in new settings and holds over time, so a mastered goal is a real one.
Two different questions
Generalization asks: does the skill work with different people, in different places, with different materials? Maintenance asks: does it still work next month, after the lessons have moved on? A student can pass one and fail the other.
The PROGRESS Center's module on intensifying instruction treats the first as "transfer." It describes transfer as the application of learned skills to new situations, and says it is crucial because students with disabilities often struggle to generalize across contexts and assessments. It says teaching for transfer helps narrow the gap between performance in an intervention and broader academic success.
What the sources suggest teaching
The IES practice guide on classroom behavior lists strategies that help students apply and maintain new behavioral skills in different environments. They include explaining the appropriate behavior, breaking each skill into concrete steps, modeling it with a variety of examples, offering guided and independent practice and role playing, prompting and cuing, and giving specific feedback. The guide says these are similar to effective academic instructional strategies.
The PROGRESS Center module adds that teachers can explicitly teach how a skill applies in different contexts, give varied practice across settings, use consistent language across instructional settings, add real-world examples, and teach metacognitive strategies.
So a skill that is not generalizing is a teaching question. Does the practice include the other settings and materials? Does everyone use the same words for it?
Back to the sight words
The student read 8 of 10 cards in the resource room and 3 of 10 on a classroom worksheet. The gap is the finding. It says the skill is not yet portable, and it points to what to teach next.
The team adds practice with the classroom worksheet format in the resource room, then moves some of the same practice into the classroom with the general education teacher using the same cue words. Two weeks later, a classroom probe shows 6 of 10. A month later, 8 of 10. The graph now has two lines, one per setting, and the second one is catching up. That is a picture a parent can follow: "She knows the words. We are teaching her to use them in your classroom."
Without the second line, the goal would have been closed on the strength of the resource room alone, and the first time anyone noticed would have been a report card.
Who takes the generalization data
Whoever is in the setting can take it, using the same probe and the same sheet. See data collection in general education classrooms for sharing the task, and paraprofessional data collection for staff who support the student there. The special educator reviews the numbers and decides what they mean. Keep the setting written on every line, so the lines can be separated later.
Measure it on purpose
Taking data in only one setting shows only that setting. To check generalization, plan the same probe in at least one other condition.
| Check | What changes | What stays the same | Record |
|---|---|---|---|
| Different setting | Classroom instead of resource room | Materials, probe form | Count correct out of total |
| Different person | General education teacher instead of special educator | Setting, materials | Count correct out of total |
| Different materials | Worksheet instead of cards | Setting, person | Count correct out of total |
Change one thing at a time. If you change setting, person, and materials together and the score drops, you will not know which one mattered.
State it as a count
Utah's sample present levels does this well. For a seventh grader, it reports that she generalizes shapes to real objects in 12 of 25 opportunities when the real item is paired with a visual item. That one sentence names the skill, the setting (real objects), the support (paired with a visual), and the count with its total.
Use the same pattern on your sheet: "Read 24 of 30 words correctly in the classroom with the general education teacher on Oct 20." The reader can see who, where, and how many.
Maintenance checks
Maintenance is measured by returning to a skill after a gap. NCII says single-skill measures are less useful for monitoring maintenance of skills and progress across the school year, which is one reason it favors general outcome measures for the year's goals. A general probe keeps sampling the earlier skills, so a slip shows up.
If you use single-skill probes, add a maintenance check on purpose. Choose a schedule, such as one week and four weeks after the goal is met, and give the same probe in the same way. Write the dates on the graph. The schedule is practice, not a rule from the sources.
Fade supports without losing data
When a goal is met, NCII's behavior guide says to consider gradually increasing the goal or gradually fading supports while continuing to collect data. Fading is where maintenance can break. Drop one support at a time, such as a visual cue, then an adult prompt, and keep recording. If scores drop after a step, restore the support and try a smaller step. See goal mastered: now what.
Writing it into the goal
A goal can include generalization in its condition. Using NCII's pattern of condition, behavior, and proficiency, an invented example: "When given a list of 30 grade 1 sight words in the resource room and in two general education settings, Maya will read 27 correctly on three consecutive probes in each setting by May." That puts the settings in the goal, so the data have to come from those settings. The baseline should match: take it in all three. See how to set an IEP baseline.
Mistakes that close a goal too early
- Calling a goal met on resource-room data alone. Check the settings that matter.
- Changing several conditions at once. Change one at a time.
- Never rechecking. Schedule a return visit before you close the goal.
- Dropping all supports at once. Fade in steps.
- Treating failure to transfer as the student's fault. Teach in the new setting.
Terms are in the glossary. If a skill will not budge, see goal not progressing: a checklist.
What counts as maintained
The sources describe why transfer matters and what helps. They do not set how many settings to check, how long to wait before a maintenance probe, or what score counts as maintained. Those decisions belong to your team.