Este artículo está en inglés. Use el selector EN | ES para navegar el resto del sitio en español.
ClinicalDiscrete Trial Training Data Collection: 10 Tips for Cleaner Trial Data

Ten practical tips for recording discrete trial training (DTT) data: define independent, score every trial, record the prompt level, keep block sizes consistent and report percent independent.
Clean DTT data comes from one shared definition of "independent," a score on every trial, and the prompt level recorded each time. These ten habits make trial data comparable across sessions and across RBTs.
1. Define "independent" in writing
An independent response happens before any prompt, within the target latency. Write the definition once so every RBT scores the same way.
2. Score every trial
Skipping "obvious" trials inflates percentages. Record each trial, including no responses.
3. Record the prompt level, not just correct or incorrect
A prompted correct response is progress, but it is not independence. Capture the prompt used. See ABA prompt hierarchy explained.
4. Keep block sizes consistent
A block of 10 trials one day and 5 the next makes percentages jump. Pick a block size per target and keep it within a phase.
5. Record the instruction, once
Write the discriminative stimulus on the target ("What is this?") so RBTs deliver the same instruction every time.
6. Do not re-score
Change a trial only through a visible undo, not by memory at the end. Real-time scoring beats end-of-session reconstruction.
7. Report percent independent
Percent independent is independent trials divided by scored trials. Show percent correct (including prompted) separately. See mastery criteria examples.
8. Interleave and vary targets on purpose
Mixing mastered and acquisition targets keeps motivation up. Note the order if it matters for the data.
9. Probe in new settings
Run occasional probes with other people and materials. See generalization and maintenance probes.
10. Close the loop in the note
Summarize what was taught, prompt levels and progress in the session note.
In TargetFlo
Trials load in program order with the organization's prompt hierarchy (FP, PP, M, G, V by default). The RBT taps +, − or no response, or a prompt level; independent responses are always scored +, undo is one tap, and the target shows running totals like "70% independent · 90% correct." Mastery for trial targets uses percent independent. See Data Collection and the RBT data collection app.
- discrete trial training data collection
- DTT data sheet
- discrete trial teaching
- ABA trial data
- percent independent
Véalo en TargetFlo
Saquemos la admisión del fax
Reserve una demo de 30 minutos: mapeamos su flujo de referidos al pipeline de 9 etapas, conectamos su fax y mostramos elegibilidad en admisión.
Reservar demoFAQ

