Discrete Trial Training (DTT): When and How to Use It (Free CEUs)

Discrete trial training (DTT) is one of the most recognizable teaching procedures in applied behavior analysis—and one of the most misunderstood. Done well, it builds a brand-new skill trial by trial with a precision few other methods can match. Done poorly, it produces rote responders who ace the table and fall apart the moment the flashcards disappear.

This guide walks through DTT the way it actually runs in practice—the anatomy of a single trial, prompting and errorless learning, mastery criteria and data-based decisions, a worked example, the errors that quietly sink programs, and the modern, blended, assent-based way clinicians use it today. It’s written for practicing BCBAs and RBTs who want the procedure to click, and it ends with a curated list of free BCBA CEUs on teaching procedures.

What Is Discrete Trial Training?

Discrete trial training is a structured method of teaching in which a skill is broken into small, discrete components, each taught through a series of distinct, repeated learning opportunities called trials. Each trial has a clear beginning and end, a single instruction, one target response, and an immediate consequence—then a brief pause. Because trials are delivered in quick succession, DTT is often described as massed-trial instruction: many repetitions of the same target, packed close together, to build a skill efficiently.

The word “discrete” is doing real work: instruction comes in isolated, tightly controlled units rather than one continuous activity. That control is DTT’s greatest strength—it removes ambiguity, keeps the relevant stimulus in front of the learner, and lets you deliver dense reinforcement and precise error correction—and also its central risk, since stripping too much natural context away can teach a response that only occurs at the table.

DTT is not a synonym for ABA—the single most common confusion in the field, from parents to new supervisors. ABA is the broader science and its full toolbox: assessment, reinforcement, shaping, chaining, naturalistic teaching, function-based intervention, and much more. DTT is one teaching procedure within it. A strong ABA program uses DTT where it fits and reaches for other procedures where it doesn’t.

Where DTT comes from

DTT grew out of the early work of Ole Ivar Lovaas and colleagues at UCLA in the 1960s and 1970s. Lovaas’s landmark 1987 study reported substantial gains for young autistic children who received intensive behavioral intervention, and discrete trial teaching was the procedural engine behind much of it (Lovaas, 1987). The structured, adult-directed table format many people picture when they hear “ABA” traces directly to this lineage. Tristram Smith later helped codify DTT as a defined, teachable methodology distinct from any single treatment package (Smith, 2001), and standard texts formalized its components (Cooper, Heron, & Heward, Applied Behavior Analysis). Just as important, that heavily adult-led format has since been substantially revised—softened, blended with naturalistic teaching, and reworked around learner assent.

The Anatomy of a Discrete Trial

Every discrete trial is built from the same five components, delivered in the same order.

One framing point first. DTT is a direct application of the three-term contingency—antecedent–behavior–consequence (A-B-C)—that underlies all operant learning: the S^D is the antecedent, the response is the behavior, and reinforcement or error correction is the consequence. What a trial adds beyond the three terms is a deliberate prompt and a defined inter-trial interval. The five components below are that contingency, made explicit and repeatable for instruction.

Two habits separate clean trials from sloppy ones: hold the S^D constant until the skill is acquired, and keep the consequence immediate and contingent on the response. Drift on either and the trial loses the stimulus control it exists to build.

DTT vs. Natural Environment Teaching (NET)

DTT rarely stands alone in a well-designed program. Its natural counterpart is natural environment teaching (NET)—instruction embedded in play and daily routines, following the learner’s motivation and using naturally occurring reinforcers. The two aren’t rivals; they’re tools matched to different jobs.

DTT is the better fit for acquisition of discrete, foundational skills: skills that benefit from many clean repetitions and tight stimulus control—early receptive and expressive labels, imitation, matching, foundational listener responses. The massed-trial format delivers the repetition and reinforcement density acquisition often needs.

NET is the better fit for generalization and functional, spontaneous use: getting a skill to occur across people, settings, and materials, and to appear when the learner is actually motivated rather than only on instruction. Because NET runs on the learner’s own motivating operations and natural reinforcers, its skills tend to transfer more readily to real life.

Best practice is not to pick one—it’s to blend them deliberately. A common pattern: establish a new skill with focused discrete trials, then move it into the natural environment to build fluency and generalization. The question is never “DTT or NET?” but “which procedure, for this skill, at this stage?”

Prompting Strategies and Errorless Learning

A prompt is any supplemental stimulus that helps evoke a correct response; the art is choosing the right type, delivering it at the right level, and—above all—removing it on a plan.

Prompt types

Prompts fall into two families. Response prompts act on the learner’s behavior directly—verbal instructions, modeling, gestures, physical guidance. Stimulus prompts act on the materials or the S^D itself—making the correct choice larger, closer, or better positioned. Both work; the difference is whether you’re prompting the person or the array.

Prompt-fading systems

• Most-to-least (MTL): start with the most intrusive prompt that guarantees success (often full physical), then fade as the learner succeeds. Well suited to newer or complex skills and to errorless teaching, because it front-loads correct responding.
• Least-to-most (LTM): start with the least intrusive prompt (or none) and escalate only if the learner doesn’t respond correctly. Gives room for independence, but tolerates more errors.
• Graduated guidance: physical guidance delivered and faded in real time within the response—shadowing the learner’s movement and giving only as much support as the moment requires. Common in motor and self-help chains.
• Time delay: insert a growing pause between the S^D and the prompt (constant or progressive), transferring control to the S^D as the learner responds before the prompt arrives.

Errorless learning and transfer of stimulus control

Many DTT programs use an errorless learning approach: prompt immediately and effectively so the learner rarely errs, then fade the prompt. Minimizing errors reduces frustration, keeps the learner from practicing (and inadvertently reinforcing) wrong responses, and tends to speed acquisition. The mechanism underneath every fading system is transfer of stimulus control—moving control of the response off your prompt and onto the natural S^D, so the learner responds to “Touch red” rather than to your pointing finger.

The failure mode this guards against is prompt dependency: a learner who waits for the prompt and won’t respond without it. The antidote isn’t to withhold help—it’s to help fully, then fade on a plan. A prompt you have no strategy to remove teaches dependence, not independence.

Reinforcement, Error Correction, and Mastery Criteria

The consequence half of the trial deserves as much design attention as the antecedent half.

Reinforcement

Early in acquisition, reinforce every correct response immediately and generously, using reinforcers identified through preference assessment rather than assumption. As the skill strengthens, thin the schedule and shift toward natural reinforcers so the behavior survives outside the teaching setting. Dense reinforcement builds a skill; sparser reinforcement helps it last.

Error correction

When an error occurs, run a planned correction rather than improvising. Common routines share a shape: acknowledge the error neutrally, re-present the S^D with a prompt that ensures a correct response, then re-present the trial without the prompt (a “transfer trial”) to confirm independence. Keep corrections calm—a big reaction can accidentally reinforce the error, especially for attention-motivated learners.

Mastery criteria

A mastery criterion is the standard a learner must meet before a target is considered acquired. A frequently used example is 80% or higher correct, independent responding across a set number of consecutive sessions (say, two or three, sometimes across different instructors or days). Treat that figure as a common convention, not a universal law—the right criterion depends on the skill and learner and should be set clinically. What matters most is that it’s defined in advance and requires independent responding across sessions, so you’re not calling a prompted, single-session fluke “mastered.”

Data Collection and Data-Based Decisions

DTT’s structure makes it unusually easy to measure—which is exactly why data, not clinician impression, should drive every program decision.

Trial-by-trial vs. probe data

Trial-by-trial data records the outcome of every trial (correct, prompted, incorrect). It’s the most granular picture you can get and is invaluable when a learner is struggling, because it shows exactly where responding breaks down—though the scoring effort can pull attention away from teaching.

Probe data samples performance instead—typically the first trial of each target, or a short unprompted probe at the start of a session. Probes are lighter to collect and give a clean read of independent acquisition over time, though they reveal less about within-session dynamics. Many programs use a sensible mix: probe most targets, drop to trial-by-trial recording on any that stalls.

Making data-based decisions

The point of the data is to act on it. Graph it, look at trend and level, and let the numbers trigger decisions: if a target is climbing, keep going; if it’s flat across several sessions, change something—prompt level, reinforcer, task difficulty, or the S^D itself. A DTT program that collects data but never changes course is just generating paperwork.

A Worked Example: Teaching Receptive Identification

Methodology sticks better when you watch it run. The following is an anonymized composite. A 4-year-old learner had no reliable receptive identification of common objects. The team targeted three items (cup, shoe, ball) with DTT, defining the correct response operationally—touching the named object within five seconds—and confirming a reinforcer through preference assessment.

They chose an errorless, most-to-least approach. On early trials the instructor placed the three objects in a field, delivered a clean S^D (“Touch cup”), immediately gestured toward the correct item, and reinforced instantly. They then faded the prompt with a time delay—a growing pause between “Touch cup” and the gesture—so control transferred to the spoken word. On the rare error, the instructor re-presented the S^D with a full prompt, then ran an unprompted transfer trial to confirm independence.

They took first-trial probe data on each target and graphed it. Cup and shoe climbed steadily; ball stalled at chance for four sessions. The team switched ball to trial-by-trial recording and spotted the problem: the learner was scrolling to whatever object sat on the left. Randomizing positions every trial broke the cue, and ball began to climb. Each target was mastered at 80% independent correct across three consecutive sessions—then moved into natural environment teaching: the real cup at snack, the real shoe at the door, the real ball outside, to build generalization the table alone never would.

Common DTT Errors and How to Fix Them

• Prompt dependency. The learner waits for the prompt and won’t respond without it. Fix: fade prompts on a plan from trial one, use time delay to transfer control to the S^D, and reinforce independent responses more richly than prompted ones.
• Rote or scrolled responding. The learner recites answers in a fixed order, or scrolls options, rather than responding to the S^D. Fix: randomize target and stimulus position, intersperse mastered and acquisition targets, and require discrimination rather than sequence.
• Lack of generalization. The skill occurs only at the table, with one instructor and one set of materials. Fix: program generalization deliberately—vary materials, people, and settings, and move acquired skills into NET.
• Running trials too fast. Crushing the ITI or clipping the response window turns teaching into a blur and inflates errors. Fix: keep a real (if brief) ITI and a consistent response window, and let the learner respond.
• Weak S^D control. A wordy or inconsistent instruction means the learner never learns what cue to respond to. Fix: define the S^D, say it once, keep it identical until acquisition, and don’t pad it with chatter.
• Reinforcement that isn’t reinforcing. An assumed reinforcer the learner doesn’t value stalls acquisition. Fix: run preference assessments, rotate reinforcers to prevent satiation, and let the data confirm the consequence is working.

Criticisms, Limitations, and the Modern Blended Approach

DTT has drawn legitimate criticism, and good clinicians take it seriously rather than defending the procedure reflexively. The historical intensive, heavily adult-directed format has been faulted for producing rote, context-bound responding that generalizes poorly; for being repetitive and, at its worst, aversive when run rigidly; and for prioritizing compliance and instructor-chosen goals over the learner’s own motivation. Members of the autistic community and many clinicians have pushed hard, and fairly, on the ethics of rigid, compliance-driven instruction.

The field’s response has been substantive, not cosmetic. Contemporary practice has moved toward a blended, assent-based, more child-led model: loosening DTT toward the naturalistic end—interspersing targets in play, following the learner’s motivation, using natural reinforcers—and pairing it with NET rather than running hours of isolated drills. It treats assent (the learner’s ongoing, moment-to-moment willingness to participate) as something to actively monitor and honor, adjusting rather than pushing through signs of distress. It chooses socially meaningful goals with families and, where possible, the learner. And it keeps what is genuinely valuable in DTT—clean stimulus control, dense reinforcement, precise error correction, measurable progress—while discarding the rigidity that gave it a bad name. Used this way, as one procedure among many and in service of the learner, DTT remains a powerful evidence-based teaching tool.

Free DTT and Teaching Procedures CEU Opportunities

Discrete trial training and related teaching procedures are among the most well-covered topics in continuing education, and a lot of quality coverage is free. Here’s where to find solid, BACB-approved instruction on DTT, prompting, and skill acquisition at no cost:

• University webinar archives — Behavior-analytic training programs regularly post recorded lectures on discrete trial teaching, prompting, and skill acquisition.
• AFIRM (UNC Frank Porter Graham) — The Autism Focused Intervention Resources & Modules team offers a free, evidence-based Discrete Trial Training module with step-by-step implementation guidance and a certificate of completion.
• BACB resources — Task-list breakdowns and practice materials that ground your teaching procedures in the current standards. Verify provider approval at the BACB.
• Conference recordings — Some ABAI and state-chapter presentations on teaching procedures are released free after the event.
• ABA Clubhouse — Unlimited CEUs for a low monthly rate, including courses that walk through DTT, prompting strategies, and errorless learning.
• OpenCEU directory — Browse 475+ free BCBA CEU courses, searchable by topic including skill acquisition and teaching procedures.

One habit worth keeping: before you count any course toward recertification, verify the provider is BACB-approved and the certificate lists ACE provider information. Free is only useful if the hours count.

Frequently Asked Questions

What is the difference between DTT and ABA?

ABA (applied behavior analysis) is the broad science and its full set of assessment and intervention procedures. DTT (discrete trial training) is one specific teaching procedure within ABA—a structured, massed-trial method for teaching skills broken into discrete components. All DTT is ABA, but ABA is far more than DTT; a good program also uses naturalistic teaching, shaping, chaining, and function-based intervention.

Is discrete trial training still used today?

Yes. DTT remains a widely used, evidence-based teaching procedure, especially for acquiring new discrete and foundational skills. What has changed is how it’s used: modern practice blends DTT with natural environment teaching, monitors the learner’s assent, and avoids the rigid, hours-of-isolated-drills format associated with the earliest programs.

What is the difference between DTT and NET?

DTT uses structured, massed trials with tight stimulus control and is best for acquiring discrete, foundational skills. NET (natural environment teaching) embeds instruction in play and daily routines, using the learner’s motivation and natural reinforcers, and is best for generalization and spontaneous use. They aren’t competitors—strong programs blend them, often establishing a skill with DTT and then generalizing it through NET.

How many trials should a DTT session include?

There’s no universal number—it depends on the learner, the skill, the stage of acquisition, and above all the learner’s engagement and assent. DTT delivers many repetitions quickly, but more trials isn’t automatically better; sessions should be paced to keep the learner motivated and adjusted based on data and signs of fatigue rather than a fixed quota.

What is a common mastery criterion in DTT?

A frequently used example is 80% or higher correct, independent responding across a set number of consecutive sessions (often two or three, sometimes across different instructors or days). Treat this as a common convention rather than a universal rule—the appropriate criterion is set clinically for the skill and learner, defined in advance, and should always require independent responding across sessions.

Who can run DTT sessions?

Direct DTT sessions are commonly run by RBTs and other trained behavior technicians. A BCBA is responsible for designing the program—selecting targets, defining responses and mastery criteria, choosing prompting and error-correction procedures—and for supervising implementation, reviewing data, and making data-based decisions. Fidelity training and ongoing supervision keep the procedure effective and ethical.

Key Takeaways

• DTT is a procedure, not the whole of ABA. It’s a structured, massed-trial method for teaching skills broken into discrete components—one tool in a much larger box.
• Every trial has five components: discriminative stimulus, prompt, response, consequence, and inter-trial interval—the three-term contingency made explicit.
• Fade every prompt on a plan. Transfer of stimulus control is the whole game; a prompt with no fading strategy builds dependency, not independence.
• DTT and NET are partners. Use DTT to acquire discrete, foundational skills and NET to generalize them into spontaneous use. Blend, don’t choose.
• Let data drive decisions. Probe and trial-by-trial data exist so you change course when a target stalls—not so you generate paperwork.
• Modern DTT is blended and assent-based. Keep the clean stimulus control and measurable progress; discard the rigidity and honor assent. Quality CEUs to sharpen these skills are free—just confirm the provider is BACB-approved.

Related Resources

• Free BCBA CEUs directory — the full hub of no-cost continuing education
• Verbal Behavior guide — DTT is a common vehicle for teaching verbal operants like mands and tacts
• Autism Intervention Strategies guide — where DTT fits within a broader evidence-based program
• Functional Behavior Assessment guide — the assessment side of behavior-analytic practice
• Behavior Intervention Plan (BIP) guide — teaching procedures like DTT often build the replacement skills a BIP prescribes