The goal says: "from 40 percent to 80 percent accuracy." A new clinician opens the file to find where the 40 percent came from. There is no task, no date, and no cue level. The number is a statement without a source.
A goal without a baseline is hard to measure, because progress is a change from somewhere. This guide covers what a baseline is, what the rules and professional guides say about it, and how to collect one that you can stand behind.
What the IEP has to say
Section 300.320 requires the IEP to include a statement of the child's present levels of academic achievement and functional performance. It requires measurable annual goals. And it requires a description of how progress toward those goals will be measured and when periodic progress reports will be provided, such as quarterly reports or reports issued with report cards.
The word "baseline" does not appear in that list, but the idea is built into it. Progress is a comparison with a starting point, and the starting point is the present level written in a form that can be measured.
What a state technical guide adds
Utah's State Board of Education publishes a technical assistance guide on writing present levels. It says baseline and current performance data are the starting point for writing measurable goals, and that data should be current, specific, measurable, and correspond directly to the annual goals. It tells teams to consider the data that most accurately reflects the student's current level of functioning in each area, and to use data that is already being gathered regularly. It says data covers social and behavior skills, functional and adaptive skills, speech and language, and motor skills, as well as academics, and that past data that no longer applies should be left out.
The same guide says that if a team fails to identify present levels clearly, to establish a baseline for IEP goals, and to ensure parent participation, it could result in a denial of a free appropriate public education. That is Utah's statement of how it sees the issue, and the details of your state's rules may differ, but it shows why a baseline is treated as more than paperwork.
What a good baseline looks like
| Part | What to record | Example |
|---|---|---|
| The measure | The same measure the goal uses | Correct out of total trials |
| The result | A count and a total, not only a percent | 4 of 10 |
| The task | What the student was doing | Picture description, past-tense verbs |
| The conditions | Setting, materials, and cue level | Therapy room, picture cards, no cues |
| The dates | When it was taken | Sept 12 and Sept 19 |
| The source | Who collected it and how | Clinician, probe sheet |
AOTA's decision guide for school OT lists the same ideas: identify the outcome to measure, select a way to measure it, collect baseline data, and document contextual features that may affect performance, such as distractions in the environment and cues provided.
Matching the baseline to the goal
The goal tells you how to take the baseline. If the goal says "accuracy in structured language tasks," a probe in a structured language task is the right measure. If the goal is about how often something happens in a classroom, a tally in the classroom is the right measure.
| If the goal measures | Take a baseline of | Avoid |
|---|---|---|
| Accuracy on trials | Correct and total, with cue level | A single impression such as "weak" |
| How often a behavior occurs | A count over a timed observation | Counts from observations of different lengths |
| How long something lasts | Timed episodes | Estimates |
| Independence | Prompt level for each opportunity | Changing the prompt key between sessions |
| Quality on a scale | A rating against a written description | A scale with no descriptions |
If you already collect data regularly, as Utah's guide suggests, you may have much of this. The key is that the data matches the goal and has a date.
How many samples
No federal rule I found sets a number of data points for a baseline, and the professional sources cited here do not either. Many clinicians take more than one probe, on different days, so that one unusual session does not set the starting point. Whatever your district expects, write down what you did: "Baseline from two probes, Sept 12 and Sept 19, average 4.5 of 10."
A baseline you can paste into a sheet
See progress data sheets for IEP goals for how to carry the baseline into a sheet that tracks sessions.
When there is no baseline
Sometimes you inherit a goal with no starting point. Do not guess and do not reconstruct one from memory.
- Tell your supervisor and the case manager.
- Take a baseline in the first sessions, with the date, task, and conditions.
- Record it in the same form the goal uses.
- Ask whether the team needs to update the present levels, and let the team decide.
Keep conditions the same
A baseline taken in a quiet room with picture cards is not comparable to later data from a noisy classroom with spoken prompts. When you take later data, note when the conditions change. A change in task or cue level is a change in what you are measuring, and the record should show it. See SOAP notes for school SLPs for where those details fit in a session note.
Baseline errors that come back later
- Using impression words. "Poor" is not a baseline.
- Mixing measures. If the goal counts correct trials, do not baseline with a rating.
- Leaving out the cue level. The same score with and without cues is not the same performance.
- Using old data. Past data that no longer applies should not be used.
- Reporting a percent alone. Add the count and the total.
Terms are in the glossary.