Home / RBT Study Guide / Behavior Acquisition / Discrete-trial teaching (DTT): the parts of a trial

Domain C · Behavior Acquisition · Task C.3

Discrete-trial teaching (DTT): the parts of a trial

Discrete-trial teaching breaks a skill into short, repeated teaching units. Each trial has the same parts: an instruction (the SD), a prompt if needed, the learner's response, a consequence, and a brief pause before the next trial.

Last reviewed November 10, 2026 against the BACB RBT Test Content Outline (3rd ed.)

What discrete-trial teaching is

Discrete-trial teaching (DTT) is a structured way to teach in which a skill is practiced in short units called trials. "Discrete" means each trial has a clear beginning and a clear end. Because trials are short, a learner can get many practice opportunities in a few minutes, each one followed by an immediate consequence.

Task C.3 on the RBT outline is to implement discrete-trial teaching procedures. The supervisor chooses the targets, the prompts, the reinforcers and the error correction. You run the trials as written and record what happens on each one.

The five parts of a trial

  1. The discriminative stimulus (SD). The instruction or cue that starts the trial, such as "touch nose" or a picture held up with "What is it?". Get the learner's attention first, give it once, and use the wording in the plan.
  2. A prompt, if needed. Extra help that makes the correct response more likely, given with or right after the SD. The plan says which prompt and how to fade it. See prompting and prompt fading.
  3. The response. What the learner does: correct, incorrect or no response within the time the plan allows.
  4. The consequence. Reinforcement for a correct response, delivered immediately. For an incorrect response or no response, no reinforcement and the plan's error correction.
  5. The intertrial interval. A brief pause, often a few seconds, that separates one trial from the next. Use it to record data and reset materials.

The SD, response and consequence are the same antecedent, behavior, consequence sequence you see everywhere in ABA. The prompt and the pause are what turn it into a teaching trial.

Common mix-up

The intertrial interval is the pause between trials. A time delay is a pause inside a trial, between the SD and the prompt. Latency is a measurement: the time from the SD to the start of the response.

Massed and mixed trials

How trials are ordered is part of the plan.

Massed trialsMixed (interspersed) trials
What it looks likeThe same target is presented several times in a rowThe target is rotated with other targets or with mastered skills
Example"Touch red" five times in a row"Touch red", "clap hands", "touch blue", "touch red"
Often usedEarly, when a target is brand newAs the learner starts responding correctly, to check real discrimination
Watch forCorrect answers that come from repeating the last responseMore errors at first, which is useful information

A learner who is correct only in massed trials may be repeating what worked a moment ago. Mixing targets shows whether the learner is actually responding to the instruction, which is the point of discrimination training.

Error correction and data

There is no single error correction procedure. One plan may have you repeat the SD with a prompt and then retest. Another may have you prompt, insert an easy mastered trial, then return to the target. What matters at the RBT level is that you run the one written in your plan, the same way every time, and that you do not reinforce the error.

Data are recorded on every trial, usually during the intertrial interval. A common code set is a plus for a correct independent response, a P for a prompted response, a minus for an error and NR for no response. Your plan defines the codes and which responses count toward the mastery criterion. In many programs only independent correct responses count, so eight correct responses out of ten trials with two of them prompted would be reported as 60 percent independent.

On the exam

A common question format lists what happened in a trial and asks which part is missing or was done wrong: a late reinforcer, an instruction repeated three times, a skipped intertrial interval, or an error correction that does not match the plan.

A worked example: three trials

Priya is learning to identify body parts. Her plan uses mixed trials, a model prompt that is faded, and a token for independent correct responses. Error correction: repeat the SD with a model prompt, give brief praise (no token) for the prompted response, and move on.

In session

Trial 1. You say "touch head" (SD). No prompt. Priya touches her head within two seconds (response). You say "You got it" and place a token on her board at once (consequence). You mark a plus and wait three seconds (intertrial interval).

Trial 2. You say "touch ears" (SD). Priya touches her nose (incorrect response). You give no token or praise. Error correction as written: you say "touch ears" while touching your own ears (model prompt), and Priya touches her ears. You say "That's ears" with brief praise and no token. You mark the trial as a minus, as her plan directs, and pause.

Trial 3. You say "clap hands" (a mastered skill, mixed in). Priya claps. You praise, place a token, mark a plus and pause.

Three trials took well under a minute. Each one had a clear start, one instruction, one response, an immediate consequence and a data point.

Where DTT fits

DTT gives a lot of practice quickly and makes progress easy to measure. Its limit is that skills learned at a table with one person and one set of cards do not always show up elsewhere. That is why most programs pair it with naturalistic teaching and plan for generalization and maintenance. If a learner is correct at the table but not anywhere else, report it to your supervisor.

Practice questions

Original questions in the style of the exam. Pick an answer before you open the explanation.

1. An RBT says 'touch nose'. Sam touches his nose. The RBT praises him, hands him a token, records the trial, waits three seconds and then gives the next instruction. What is the three-second wait called?

  1. The response latency
  2. A constant time delay
  3. The intertrial interval
  4. The interresponse time
Show answer

C. The intertrial interval

The short pause between the consequence of one trial and the instruction of the next is the intertrial interval. A time delay is a pause between the instruction and a prompt within a trial, not between trials.

2. Mia's plan says that after an error the RBT gives the instruction again with a model prompt. Instead, the RBT says 'no, try again' several times until Mia gets it right. What is the problem?

  1. The error correction in the plan was not followed
  2. The intertrial interval was too long for the learner
  3. The trials should have been massed, not interspersed
  4. The reinforcer was too large for a prompted response
Show answer

A. The error correction in the plan was not followed

Error correction is run exactly as the plan describes. Repeating 'try again' is a different procedure, lets Mia practice errors, and is a procedural fidelity problem to raise with the supervisor.

3. An RBT presents the instruction 'touch red' ten times in a row, with no other targets in between. Which arrangement of trials is this?

  1. Incidental teaching
  2. Total task chaining
  3. Mixed trial teaching
  4. Massed trial teaching
Show answer

D. Massed trial teaching

Repeating the same target on back-to-back trials is massed trial teaching. Mixed (interspersed) trials would rotate 'touch red' with other targets, such as 'touch blue' or mastered skills.

Common questions

What are the parts of a discrete trial?

A discrete trial has an instruction or cue (the discriminative stimulus, SD), a prompt if one is needed, the learner's response, a consequence (reinforcement or error correction), and a brief intertrial interval before the next trial begins.

What is the difference between DTT and NET?

Discrete-trial teaching is led by the instructor, usually at a table, with planned instructions and reinforcers that may be unrelated to the task. Natural environment teaching follows the learner's interest during play or routines and uses reinforcers that are a natural part of the activity. Most programs use both.

What does an RBT do when a learner makes an error in DTT?

The RBT does not reinforce the error and runs the error correction procedure written in the plan, exactly as written. The trial is recorded as the plan defines. If errors keep happening, the RBT reports it to the supervisor.