DTT: Discrete Trial Teaching

There are MANY different ways to apply the science of ABA, such as Natural Environment Teaching, Verbal Behavior, and DTT: Discrete Trial Teaching. Many people think DTT is all ABA is; that all ABA therapists do is sit at a table with flashcards and small pieces of candy. That is a common misconception, but ABA is broad and multifaceted. Applied Behavior Analysis is a science that can be applied in a variety of ways to teach new behaviors, and DTT is just one branch on the ABA tree.

DTT is widely used today and probably the best known type of ABA. If your child receives ABA services at a clinic/Autism school, or from an agency, they are probably using DTT techniques for the sessions. It is common that younger children start with a more structured and rote DTT style, and eventually over time or as the child develops they transition to a more naturalistic learning style, such as NET or Incidental Teaching. I love DTT as a teaching methodology because it gives me the ability to remove all the frills and boil my teaching down to exactly what the learner needs to be successful.

When people make complaints or negative statements against ABA, they usually critique ABA for being too robotic, focusing on repetitive drills, giving kids bites of candy all day, etc. Although distorted and inaccurate, there is some truth to these complaints. The DTT method does focus on breaking down skills, repetition, and reinforcement. However when done properly by qualified staff, DTT can bring about amazing and significant gains in language, imitation, attending, self-help, motor, and play skills in children with Autism.  The research is extensive and conclusive: DTT works.

Let’s take a closer look at the DTT method--

  • Teaching Trial: For the DTT method, the following teaching trial is used: Present materials, give SD, prompt incorrect response or reinforce correct response, clear materials. That entire process is called a teaching trial, and it is repeated over and over in a therapy session. It should move at a brisk pace, and the therapist should be collecting trial by trial data after each trial has ended. Each trial is a learning opportunity, and the goal of a DTT session is to give the child hundreds of learning opportunities in order to teach specific skills. Younger children usually do 10 trials of each skill, while older children might do 5 trials of each skill.
  • Breaking down tasks- DTT breaks down complex skills into small, separate tasks to teach. So what is considered a complex skill? Well, pretty much everything. For a child with Autism, learning to write their name could be a 5-10 step process. For example, if the child’s name is “Adam”, first they learn to individually trace “A”, then “d”, and so on. After tracing, they learn to write each letter individually. Finally, the child is given paper and pencil and told “Write your name”, as the therapist provides letter prompting (therapist says as child writes “A…d…a…m”). Over time the therapist would remove her assistance until the child can independently write his name. Sound complicated? It’s really not. The goal when using a DTT approach is to break down the skill as much as is necessary for the child. It might not take that many steps for your client to learn to write his name. Or, your client may need even more steps. Let the child’s progress determine how far to break down the skill. If you have been stuck on a particular skill for weeks, it probably needs to be broken down further to help the child understand.
  • Mass Trial vs. Mixed Trials- Depending on the ability and functioning level of the child, you would teach using “mass” or “mixed” trials. A mass trial simply means you are asking the target multiple times in a repetitive manner. That would look like this: “Trial 1: Touch dog…Trial 2: Touch dog…..Trial 3: Touch dog”. Mass trialing is typically used to teach very young children, children new to ABA, or lower functioning children. Some programs add in generalization at this step by changing the stimuli for each trial. So the 1st trial the child touches a photo of a brown dog, the 2nd trial the child touches a photo of a white dog, and so on. Mixed trialing, also called random rotation, is when you teach targets in a random order and do not use repetition. That would look like this: “Trial 1: Touch dog…Trial 2: Give me red…Trial 3: How old are you?”. Typically you want to transition from mass trialing to mixed trials; although with some children they may need the repetition of a mass trial in order to learn.
  • SD---> Response- The SD is simply the instruction or demand you present to the child. Examples of SD’s include: “Touch your nose/Stand up/What number? /Touch apples”. With DTT, the SD leads to a specific desired response. If I say to a child “Touch your nose” there is a specific response I want the child to do. If the child does the response correctly, I provide reinforcement. If the child doesn’t respond correctly, I provide prompting. The SD must be consistent across time, therapists, and settings. If the morning therapist says “Touch your nose” and the afternoon therapist is saying “Show me your nose” that could be confusing to the child, and could slow down learning. The SD tells the child what you want them to do, and their correct response is how they contact reinforcement. It is the reinforcement that increases the likelihood that the child will respond correctly in the future.
  • Reinforcement- Reinforcement is a critical part of any ABA program, but particularly so with DTT. This is because DTT has so many learning opportunities in a session. Depending on the schedule of reinforcement being given, the child may be reinforced hundreds of time in a therapy session. With a continuous or dense reinforcement schedule, every correct answer contacts reinforcement. With an intermittent or thin reinforcement schedule, only some correct answers contact reinforcement. The schedule of reinforcement you use will depend on the child. Reinforcement must be easily accessible and ready (a bag of unopened chips is a poor choice. The chips should be opened and placed into a bowl or on a plate for easy access), yet out of reach of the child. Reinforcement should be varied, rotated often, and given in small amounts. If the reinforcer is tickles, then don’t tickle the child for 5 minutes after one correct answer. Tickle the child for maybe 4 seconds, and then continue teaching. Many people ask me if DTT sessions have to use food treats. The answer is no, they do not. The reason why so many DTT sessions use food reinforcers is because very young children or children new to ABA often don’t find social interaction, adult attention, or toys/games reinforcing. For those children, the therapists use things like candy, chips, or ice cream because that’s all the child likes. However, the goal is always to transition the child away from external/adult mediated reinforcement as soon as possible, and increase naturally occurring/social reinforcement.
  • Data Collection- Data collection for a DTT trial is done after each trial. As the child is contacting their reinforcement, the therapist is writing down how the child performed at the task. The type of data sheet used can range from very simple to complex. It can be checkmarks, a plus/minus system, graphing, etc. The method of data collection isn’t as important as making sure data is collected. It will be difficult to impossible to determine progress and make programming decisions without solid, accurate data. If you don’t like data collection or want to avoid taking so much data, I’d suggest collecting first trial data or doing a weekly Cold Probe.
  • Generalization- DTT gets a pretty bad rap when it comes to generalization. Many people think that DTT is the most robotic and boring type of ABA out there, and that it teaches skills in a vacuum (the child can answer “pig” when asked “What says oink-oink?”, but says nothing when asked “What does a pig say?”). I have stated this before on my blog, and I’ll say it again: Poor quality ABA providers perform poor quality therapy. The only reason why DTT would be lacking a plan for generalization is if the DTT program was being done improperly.  I already mentioned one way that DTT can plan for generalization (vary the materials). Another method is something called “Maintenance”. This is when known skills are reviewed days, weeks, or even months after the child was taught the skill. So if my client learned to say her phone number last week, I will ask the skill again next month. The purpose of maintenance is to make sure the child retained the skill. Failing to incorporate generalization techniques into a DTT program will only lead to problems in the long run.


DTT Research Articles:

Gresham, FM & MacMillar, DL. (1998). Early intervention project: Can its claims be substantiated and its effects replicated? Journal of Autism and Developmental Disorders,28(1): 5-13

Suzannah Ferraioli, Carrie Hughes and Tristram Smith (2005) : A Model for Problem Solving in Discrete Trial Training for Children With Autism. JEIBI 2 (4), Pg.224- 229

DTT Books:

 Individualized Autism Intervention for Young Children: Blending Discrete Trial and Naturalistic Strategies, Travis Thompson

A Work in Progress: Behavior Management Strategies and a Curriculum for Intensive Behavioral Treatment of Autism, Ron Leaf

Teaching Individuals With Developmental Delays: Basic Intervention Techniques, O. Ivar Lovaas


  1. Love your blog..got lots of great info and will continue to visit..

  2. Hey so I am in training to be a BCBA (ive asked you a question on here before and love to read your blog). My supervisors usually run DTT trials differently and i wanted to get your take on it. Usually i will run the first trial as a probe always.if there is no response/incorrect and the goal was already mastered before i will try to then say the sd again and immediately provide a v/g/m prompt. the only time i use full physical is with some of my young clients (some not yet verbal) when its a completely new goal and it just seems that the client doesnt usually respond to gestural prompts right away. anywho i guess im just wondering what is your take on prompt levels and least to most, etc. another thing i like to do is always use errorless when i can so if i see that the client is going to get the wrong answer i just go ahead and prompt and reinforce but write "-" on the data sheet...

    1. Hi there and thanks for your comment,

      I have been in your shoes, and can definitely relate to being unsure how to score a trial or which types of responses are considered independent. Unfortunately, there is no one answer to your question.
      The way data is collected will vary from one BCBA to the next, and from one case to the next. It all depends on the way the BCBA wants that done, so in your RBT role: you follow what the BCBA has written. After your certification: you design data collection based on the information you need to make accurate treatment decisions.

      You will find that if you ask 5 BCBA's their favorite/go-to way to prompt, you could get 5 different answers. Which is okay, because this stuff isn't always black and white :-)

  3. Hello,

    I just want to say I LOVE referring to your information when it comes to ABA. You always do the best breaking down concepts and making it easier for us to learn.
    I do have a question or two that I have been wondering about.

    First question is when I'm running a DTT trial would this be right:

    Present SD: Match Red
    R: Learner matches card to blue (incorrect)
    Consequence: Uses gestural prompt to learner. Learner matches the correct response.

    Re-present SD: Match Red
    R: Learner matches card to red.

    Would you count that as 1 trial and the second portion is the learning portion?

    I've seen some BCBAs do this.

    SD: Match Red
    R: Learner matches blue.
    Re-present SD: Match Red. (Uses the prompt that the learner needs). That's matching Red.

    But how can we test if the learner knows what you are wanting without allowing them to try it independently again. I hope this makes sense.

    My other question is, today another RBT saw me represent my SD to my learner and she asked, "so why do you repeat the SD?"

    My response was, "it's to show that he can do it independently." (The learner got the wrong answer and I prompted the correct answer.)

    Would you say that's accurate? How would you have responded to that?

    Thank you for taking your time to read this post. I hope it reaches you somehow. I look forward to hearing from you!


    An optimistic RBT!

    1. Hi Optimistic :-),

      The basic DTT teaching method is as follows - (1) The trainer’s presentation or SD (“Point to the red block”), (2) The learner’s response (the child points), and (3) The consequence (correction or reinforcement).

      Beyond those very basic steps, specific BCBAs may customize the way each trial is cycled through depending on the learner or the program. For example, I teach my Tech's to constantly conduct preference assessments during session to ensure they are approaching the client with the most salient and valuable reinforcers at all times. So for them, prior to running a block of trials they would offer a choice of reinforcers (or just ask the child, if the child can speak) and identify the reinforcer that will be used. THEN they would move into teaching.

      So what you are describing is not a matter of wrong or right, it could be a specific teaching variation that works best for that learner/context. I usually have my Techs run a few distractor (maintenance trials) after an error, both to get the learner in contact with reinforcement, and to keep motivation high. So in your example, it sounds like you prompt the correct response and then immediately represent the original demand. I would not say that is wrong, because I don't know your learner. For them, that may work well.

      Ultimately, as long as you are following the teaching procedure outlined by your BCBA and they have strong IOA with you during overlaps, then you are good to go! It isn't so much what another RBT or anyone else thinks, what is important is the person who customized the treatment plan (most likely the case BCBA) states that you are teaching correctly.
      Hope that helps.

  4. It makes me feel better reading what you've provided. I've been going back and forth with myself if I am running the trials correctly. Then I end up being in my head thinking that I am not following procedure and I am not teaching the learner correctly. If I'm understanding this correctly, the scenario I posted was neither wrong nor right, it all depends on how the learner learns AND how the BCBA wants me to run things.

    Thank you so much for your speedy response. You're so awesome! :) By the way, I'm studying to become a BCBA as well. I hope I can become a great BCBA one day.


    An Optimistic RBT!

    1. Correct!
      So glad this was helpful, if you are feeling uncertain about your teaching definitely talk to your BCBA. They are there to help you!
      And know that being passionate about what you do can sometimes mean being hard on yourself. I'm guilty of that as well.

      Also: good luck on your BCBA journey :-)


Copyright T. Meadows 2011. All original content on this blog is protected by copyright. Powered by Blogger.
Back to Top