Speech therapy data collection in session

By the IEP Atlas editorial teamLast updated November 1, 2026

Short answer

To collect speech therapy data in session, decide before the group starts which method each goal uses, how many trials each student gets and what counts as correct. Take a short cold probe at the start, tally teaching trials with simple codes, then turn each tally into a fraction or percent and graph it weekly against the goal.

Good session data answers one question: is this goal working? You do not need a score for every word a student says. You need the same measure, taken the same way, often enough to see a trend before the progress report is due.

This guide walks through the steps for a typical school group of 2 to 4 students, then follows one fictional student from planning to a progress report. Every goal type in the speech therapy IEP goals pillar fits one of the methods below.

What you need before you start

Each student's goals, with the condition and criterion in front of you. The goal tells you the level, the support and the unit.
The baseline, taken the same way you will probe. If you need one, see how to collect baseline data.
One sheet per group, with each student's targets listed. A clipboard beats a tablet for most SLPs, because you can tally without looking away.
A probe list for each goal that changes every time: new words, new directions, new pictures.

Step-by-step: collecting data during a session

Step 1: Match each goal to a method

Pick the method from the goal, not from habit. Most speech goals fit one of five methods.

Goal type	Method	Unit	Data sheet
Sounds in words or sentences, grammar forms, receptive items	Trial-by-trial tally	Correct of total trials	Opportunity trial sheet
Requests, repairs, asking for help in class	Opportunities	Skill used of times it came up	Opportunity trial sheet
Comments, AAC messages, volunteering to talk	Frequency	Count per fixed time	Frequency data sheet
Retell, presentation, conversation quality	Rubric	Points of total	Rubric scoring sheet
Sentence length, sound use in talk	Language or speech sample	MLU, percent correct in a sample	Recording plus a scoring form

The measurement guide explains each method with examples, and percent accuracy vs percent of opportunities covers the two denominators most speech goals use. For goals scored trial by trial, see percent accuracy.

Step 2: Set the number of trials

Decide how many scored trials each student gets per goal. Ten is a good default for a mixed group. Twenty works for a single articulation target in a small group. Fewer than 10 makes each trial worth too much: one error moves the score by more than 10 points.

Example: in a 30-minute group of 3, each student gets 10 scored trials on one goal and unscored practice the rest of the time. That is 30 scored trials, about 6 to 8 minutes of the session.

Step 3: Write down what counts as correct

Write the scoring rule at the top of the sheet. For speech sounds: correct means no substitution, omission or distortion of the target sound; a close but distorted sound is incorrect. For language: correct means the form is used in an obligatory context, or the direction is completed without a repeat.

Decide how to score prompted answers. A common rule: only independent responses count toward the criterion, and prompted ones get their own code.

Step 4: Use simple codes

Keep codes short so you can tally while you teach.

+ correct, independent
− incorrect
M correct after a model
C correct after a cue (visual, verbal or tactile)
SC self-corrected without help
NR no response

At the end, count only the codes the goal names. If the goal says "independently," only + and SC count. Self-corrections are worth tracking on their own, because they often rise before accuracy does.

Step 5: Take a cold probe at the start

Teaching trials are inflated by your models and cues. A cold probe is not. Spend the first 2 minutes per student on 10 unpracticed items with no help, then teach. Use the probe for the graph and the progress report. Use the teaching data to plan the next session.

Example: Ava's cold probe is 6 of 10 /s/ sentences. Her teaching trials, with cues, run 15 of 20. The 6 of 10 goes on the graph.

Step 6: Turn the tally into a number and graph it

Right after the group, convert each tally into a fraction and a percent. Write it on the graph that day, not at the end of the month. Graph the cold probes once a week or every session if you probe every session.

Check the trend against a goal line drawn from the baseline to the target. A common decision rule is to change instruction if 4 consecutive points fall below the goal line.^[1] The goal line entry explains how to draw it.

Step 7: Log the session and keep the record

Note attendance, minutes, the activity and anything that affected performance, such as a missed session or a new cue. The service log keeps minutes and notes in one place. Many districts also need these notes for Medicaid billing, so follow your district's format.^[3] The IEP must say how progress will be measured and when it will be reported, so the record you keep each week is what the report is built from.^[2]

Worked example: one student from plan to progress report

Ava is a fictional third grader. Her goal: given 20 sentences with /s/ in initial and final positions, Ava will say /s/ correctly in 16 of 20 sentences (80%) across 3 consecutive sessions. Her baseline in September was 5 of 20 (25%), the median of three probes. She is in a 30-minute group with two other students twice a week.

Plan. Ava gets 10 cold probe sentences at the start of each session, then 20 to 30 teaching trials. A second student works on two-step directions; the third on past tense -ed. Each has a 10-item probe.

Session data. Over October, Ava's cold probes, out of 10, were 3, 4, 4, 5, 5, 6, 5, 6. Her teaching trials ran 60% to 75% with cues. Her self-corrections rose from 0 to 2 per probe.

Graph and decision. The goal line runs from 25% in September to 80% in May, which is about 7 points a month. Ava's probes went from 30% to 60% in a month, above the line. No change needed. The team keeps the current cue and adds reading aloud of /s/ sentences in November.

Progress report. "Ava is making progress toward her /s/ goal. On 10 unpracticed sentences, she said /s/ correctly in 6 of 10 (60%) in late October, up from 25% at baseline. She now corrects some errors on her own. Next step: practice /s/ while reading aloud in class."

That report states the measure, the starting point, the current level and the next step, in four sentences a parent can follow. For the full format, see how to write an IEP progress report.

Common mistakes

Weak: Scoring every production during a game, cues included, and reporting 85%. Better: Score 10 cold probe items for the graph; keep cued trials as teaching data.

Weak: Using the same 10 picture cards for every probe. Better: Rotate probe lists so the score reflects the sound or form, not memorized cards.

Weak: Writing "good session, 80%" with no denominator. Better: "8 of 10 independent, 1 SC, 1 incorrect."

Weak: Collecting data only in the therapy room for a classroom goal. Better: Add a 10-minute classroom observation every 2 to 4 weeks, scored the same way.

Weak: Filling in the graph the night before progress reports. Better: Graph after each probe, so you can change course while there is time.

Templates and tools

Start with the opportunity trial data sheet for most articulation and language goals. Write the student's targets down the side and use the codes above in the grid. For groups, the group data sheet puts several students on one page.

For goals that a teacher scores in class, such as voice level or volunteering in discussion, give the teacher a frequency card and check a few sessions together. AAC goals need a column for message type; see AAC goals examples. For autism caseloads with peer goals, speech therapy goals for autism shows how to set up opportunity counts at lunch and recess.

If you graph many students at once, IEP Atlas Pro can hold the data and build the progress report from it; see pricing. The IEP data collection guide covers methods across every goal area.

Frequently asked questions

How many trials should I collect per goal in speech therapy?

About 10 scored trials per goal per session works well in a group. Twenty suits a single articulation target in a small group. Fewer than 10 makes one error move the score too much to read a trend.

Should prompted responses count toward a speech goal?

Only if the goal names that prompt in its condition. Most goals measure independent performance, so prompted responses get their own code and count as teaching data, not as progress toward the criterion.

How often should I take data in speech therapy?

Take a short probe on each goal at least every one or two weeks, and teaching data every session. Graph the probes as you go so you can see the trend well before the progress report is due.

Use it in IEP Atlas

Sources

Progress Monitoring: Data & Goal Setting, National Center on Intensive Intervention
34 CFR §300.320: Definition of individualized education program, Legal Information Institute, Cornell Law School
Documentation in Schools, ASHA

Written and checked by the IEP Atlas editorial team under our editorial policy. This guide describes general practice and federal rules. It is not legal advice; your state may add requirements, and the IEP team makes the decisions for each student.