How to take data in group speech therapy sessions
Four students, four sets of goals, thirty minutes and one of you. Group data collection breaks down when it depends on remembering. Two habits fix most of it: plan the turns so every student gets countable trials, and record each response the moment it happens.
Plan the trials before the group starts
- Know each student's target and trial count for the session. Ten is a realistic minimum for drill-type goals; fewer is fine when each response is long, like a story retell.
- Pick an activity where every turn is a trial for someone. Card games, board games and barrier games work because the turns are built in.
- Let each student work on their own goal within the same turn. With one deck of picture cards, the articulation student names the card using their sound, the language student answers a question about it, and the fluency student describes it using a strategy.
A sample plan
| Student | Goal | Their part of each turn | Trials |
|---|---|---|---|
| Sam | /r/ in words | Names the card | 20 |
| Priya | Answers wh- questions | Answers "who" or "where" about the card | 10 |
| Leo | Uses easy onset | Describes the card in a sentence | 10 |
| Maya | Requests with AAC | Requests the next card ("more," "my turn") | 10 |
Thirty minutes, one deck of cards and four sets of data.
Rotate quickly
Short turns give every student more trials than long ones. Take two or three responses from one student, record them, and move on. A student who waits five minutes for a turn produces no trials in that time.
Give the waiting students a job. Asking them to judge a response ("Thumbs up if you heard the /r/") keeps them listening, and for some goals the listening is practice too.
Record every response as it happens
- Mark it immediately. Totals written at the end of the session from memory are unreliable, and they are usually generous.
- Use a short code:
+and−with a cue letter (I, Min, Mod, Max), and NR for no response. The cue levels guide explains the levels. - Keep the whole group on one sheet, one row per student and one box per trial, so your eyes and your pencil don't travel between pages.
Probe some goals, not all of them, every session
With a big group, or several goals per student, formal data on every goal every session isn't realistic. Some SLPs work on every goal each session but take counted trials on a rotating set: each goal gets a short probe, say 10 trials at the start of a session, at least once a week, and the rest of the time is practice. Whatever the rotation, it should give every goal enough data points to graph before the progress report is due.
Record attendance and minutes with the data
Group sessions also need a service log: who attended, who was absent, which sessions were make-ups, and how many minutes were delivered. Record it during the session on the same sheet. Reconstructing it at the end of the month is slower and less accurate.
Paper or device?
A device beats paper when it does the arithmetic, the graphs and the report wording for you. It loses if scoring a response takes more than one tap, or if you have to look away from the students to find the right button.
- Keep the device flat on the table within reach. A Chromebook folded into tablet mode works well.
- Learn the keyboard shortcuts so you can score without looking at the screen.
- Keep a printed paper data sheet for days when the device is dead or left at another school.