Play-Based ABA Goals for Preschoolers: Replacing Table Time With Real Skills

If your treatment plan for a 4-year-old still says the kid "will remain seated" for most of the session, it is time to rewrite that goal. In her CEU on early childhood programming, BCBA Kelly Brzak literally crosses out the word "seated" on the slide. Turn-taking, FCT trials run by parents, and DTT embedded inside NET are the real targets at this age, and they do not require a table.

Why "seated" is the wrong word in a 2-5 year old goal#

Most preschoolers in ABA already spent the morning being told to sit down at daycare or pre-K. If your session is the third or fourth block of sit-down work they hit before dinner, do not be surprised when they elope, refuse, or melt down. Kelly says it plainly in the recording:

I crossed out the seated because I think it's really unrealistic... if you do find that you're writing a goal and you're putting the word seated, maybe ask yourself why.

This is not anti-structure. It is anti-default. The word "seated" tends to sneak into goals because the BCBA before you wrote it that way, or because seated data feels easier to graph. Neither reason serves the child. Ask whether the skill you actually care about requires a chair. Attention to a task, joint engagement, instruction following, and tact training all happen on the floor, in the kitchen, or chasing a wagon down the hallway.

What play-based programming actually looks like in session#

Play-based does not mean unstructured. It means the antecedent looks like play to the child while the trial structure stays intact for the BCBA. The technician shows up with a bag of items the kid has never seen, follows the child's eyes and hands, and runs trials inside whatever the kid picks up. If the kid grabs a toy car, the tech says "car. beep beep. car." That is three trials, not a missed teaching moment.

Kelly's rule of thumb is to use repetition of single words instead of bombarding the child with questions. Instead of asking "what color is the car, how many wheels, what does it say," the tech narrates from the child's perspective and gives the same target two or three times in a row. The child is more likely to echo, tact, or orient. The data still gets collected. Nothing about that requires a chair.

Sample goal: engagement across 10 activities for 30 to 60 seconds each#

Here is a play-based replacement for the standard "remain seated" goal that almost every preschool treatment plan starts with:

Client will engage in 10 different activities for 30 to 60 seconds each, across five consecutive sessions.

The strikethrough on "seated" matters. You are still building the same prerequisite skill (sustained attention to a task), but the chair is gone and the menu of activities is wide. Shape duration by starting at 15 seconds of touching the item, then 30, then 45, then 60. Two and three year olds can attend to one thing for about a minute when the thing is interesting and the adult is not crowding them. If you aim for 60 seconds and you get five, you celebrate the five. Then you run the next activity.

This goal also forces the team to bring 10 different things into the session. That keeps the tech from running the same puzzle into the ground and gives the data a way to show generalization across stimuli without writing a separate generalization goal.

Turn-taking and sharing goals broken into prerequisite steps#

Sharing is not a single skill. It is a stack. Kelly walks the audience through the breakdown on screen, and her main turn-taking target is small on purpose:

Client will take turns a minimum of two times per session across five consecutive sessions.

Two trials. Not ten. If the kid is not ready for turn-taking yet, you back up one rung and write the prerequisite. The handing goal looks like this:

Client will hand items to others a minimum of three times per session across five consecutive sessions.

Start with a neutral or non-preferred item. A tissue works. A used crayon works. The kid does not need to love the object. The kid needs to release it on cue and have the act reinforced. Once that is fluent, you flip the direction and write a receiving goal: client will take three out of three items from another person across five consecutive sessions. That receiving step is what makes instruction following possible later. Without it, every transition turns into a tug of war.

FCT trials with two family members: how to script them#

The line that catches most BCBAs off guard in this CEU is how aggressively Kelly recommends putting FCT trials into the parents' hands. The target reads:

Client will complete a minimum of five FCT trials with two family members across five sessions.

Two family members, not one. The point is to bake in generalization from the day the goal is written, and to stop the kid from learning that mands only work on Tuesdays with the lead tech. Write the script for the parent. Hand them three high-probability trials they can run in 60 seconds. A puzzle piece request, a snack request, and an "all done" mand are usually enough. Tell them exactly what to say, what to wait for, and what to deliver. If the second caregiver is dad on the weekends, write the same three trials for him on the family-facing summary. Probability of success is the whole point. You are not training the parent to be a tech. You are training the kid that the mand works on multiple humans.

DTT embedded in NET: same components, different setting#

This is the line that should reset every "but we have to do DTT" debate on your team:

Really good DTT is an NET... NET is DTT embedded in the natural environment.

The components are the same. Antecedent, response, consequence, intertrial interval, data. Move the components off the table and into the kitchen, and the trial is still a trial. The recording shows a tech sheet Kelly uses to code technician language during play, because the failure mode is not that NET cannot collect data. The failure mode is that the tech reverts to asking five WH questions per object instead of running clean, repeatable trials. Coding the language fixes that without dragging the kid back to a chair.

A practical tell that you are doing NET well: a stranger walking into the room would not be able to tell which item is the target. The kid is laughing, the tech is following, and the trial is happening anyway.

A second tell: the data sheet at the end of the session has the same number of trials it would have had at the table, or more. Play-based does not mean fewer opportunities. It usually means more, because the kid is not avoiding the setup. When teams complain that NET produces thin data, the usual cause is a tech who is narrating instead of running trials, not the format itself. Coding tech language for a session or two fixes that and the data catches back up.

When a child genuinely needs table time and when they don't#

There are kids who need a table. School-readiness goals for a 5-year-old who starts kindergarten in eight months are a real case for shaping seated tolerance. Feeding therapy with another provider may require seated work. Some discrete imitation goals are easier to capture at a small table for 10 minutes of the session.

The test is whether the table is the cheapest way to get the skill, not the default way. If the answer is yes, write a goal that says exactly that. Specify minutes, specify the activity, and specify the criterion. If the answer is no, take the table out of the goal and let the play do the work.

FAQ#

Is play-based ABA still ABA? Yes. The trial structure does not change. You still have an antecedent, a response, a consequence, and data on every opportunity. What changes is the antecedent and the setting. The components Kelly references are the same components in any DTT manual.

How do I run trials during play without losing data? Use a running record sheet, code by trial type, and tally on a clipboard or a phone. Two and three of two and three out of three criteria are friendlier to play-based data collection than 8 out of 10, because the kid does not have to sit for 50 trials of the same target.

What if the BCBA before me wrote all seated goals? Rewrite them at the next treatment plan update. Keep one seated goal if there is a real school-readiness reason. Convert the rest into engagement, turn-taking, mand, and FCT goals. Document the rationale in the plan so the next reviewer understands the shift.

How do I get a tech to follow the child's lead instead of correcting? Run a side-by-side session and code the tech's language. Count the WH questions and the demands per minute, then model the "repetition of single words" approach. Most techs over-correct because they were trained to. Showing them works better than telling them.

Can I bill insurance for play-based sessions? Yes, when the goals are written as measurable skills and the data shows responding. Insurance reimburses skill acquisition, not the chair. Make the goals specific, keep the data clean, and tie every play activity to a targeted response on the plan. If you get a clawback question, the operative defense is the trial structure on your data sheet, not the furniture in the room.

Watch the full CEU#

Kelly's full session walks through the assessment tools she uses, the caregiver pairing inventory, and the technician care pack she sends into homes with little ones. It is one CEU credit, free, and runs about an hour.

Watch the full CEU on openceu.com