Setting Up Meaningful ABA Programs
The problem I see most often isn't that people don't know what applied behavior analysis is. It's that they write goals that look good on paper and achieve nothing in the room. I've watched RBTs fill out session notes for months using objectives that didn't actually move the needle on a child's daily functioning, and the BCBA on the other end had no idea because the data looked clean. Aba Therapy Goals And Objectives need to be written differently than what most people default to. There's a structural reason for this, and understanding it saves you from months of wasted session time.
What Aba Therapy Goals And Objectives Actually Look Like
A goal in ABA isn't "the child will communicate more." That's a desire, not a behavioral objective. A proper goal has three components that are non-negotiable: the behavior must be observable and measurable, the conditions under which it will be demonstrated need to be specified, and the mastery criteria have to be stated as a percentage with a timeline. Consider the difference between writing "Johnny will request breaks" and writing "Given a preference for an activity, Johnny will independently say 'I need a break' or use a break card within 5 seconds of task demand across 4 out of 5 opportunities for three consecutive sessions." The first one tells you nothing. The second one tells you exactly what to measure, when to measure it, and what success looks like. I once worked with a team that had a goal stating a child would "initiate peer interaction during unstructured time." After six weeks, the data showed the child had initiated approximately zero times. When I asked why this wasn't flagged, the answer was that the behavior definition was too vague to actually measure. You can't count "peer interaction" unless you define what counts as one instance. Does looking at another child count? Standing near another child? The original goal had none of this specified, so the data was meaningless noise.
How to Write Goals That Actually Work in Practice
Start with a baseline. This sounds obvious but it's where most programs fail. I've seen goals written before any data collection happened, which means the target behavior was essentially guessed at. If a child currently mandates zero times per session, setting a goal of 80% independence on manding within four weeks is not reasonable. It needs to account for the actual learning trajectory, which typically follows a curve where early progress is slow and then accelerates. Break goals into teaching steps. Every measurable goal should have a hierarchy of steps leading to it. For a language goal, this means starting with whatever the child can already do reliably and adding one new dimension at a time. I had a case where a child couldn't chain together two-word phrases, so instead of jumping straight to that, we started with reinforcing any vocalization near a preferred item, then required a vocalization within one second of the item being presented, then required a vocalization paired with an adult gaze toward the item, and so on. Each step was its own objective with its own mastery criterion. Specify the reinforcement schedule upfront. This is something I wish more people thought about before writing the goal. A goal that says "the child will..." without specifying how reinforcement will be delivered and faded is incomplete. In my experience, goals that don't account for reinforcement scheduling tend to stall at around 60 to 70 percent accuracy because the behavior is maintained by the adult delivering consequences rather than by natural contingencies.
Get the Full Details
Pitfalls That Derail Programs
The biggest mistake I see is overloading a single goal with multiple behaviors. You might write "the child will follow instructions and make eye contact and wait for turns" as one objective. This conflates three separate skill sets. If the child is at 40 percent on instruction following but 90 percent on waiting, your goal average looks like 70 percent and you have no idea which behavior is actually driving that number. Write separate goals. It takes more documentation but it gives you actual information instead of averaged garbage. Another common failure point is ignoring setting generalization in the goal itself. A child might meet 80 percent mastery in the therapy room with one therapist but perform at chance in the classroom. If the goal doesn't specify generalization conditions, you'll close out the objective on paper while the behavior remains absent from the child's life. I recommend building generalization probes into the goal from the start, not as an afterthought. This means including at least two settings and two people in the mastery criteria, even if the initial teaching phase happens in one setting with one person. There's also the issue of goals that are too narrow. I had a student who could match colors perfectly in a discrete trial format but couldn't identify colors in a natural play setting. The goal had been written around matching behavior, so technically the child met it. But matching color cards and knowing what color something is are not the same skill. The goal needed to include identification in at least one unstructured context to be functional.
When ABA Goal-Setting Doesn't Work
This approach assumes the child has the prerequisite skills for the target behavior. If a child lacks basic attending skills, the ability to sit for two minutes, or any form of symbolic communication, writing advanced social goals is futile. I've seen programs spend months trying to build complex reciprocity goals with children who still hadn't developed the foundational tolerance for structured interaction. The workaround is to step back and assess for prerequisite behaviors before writing higher-level objectives. This isn't a failure of the method. It's a failure of sequencing. Another scenario where traditional ABA goal-setting breaks down is with children who have significant motor impairments that prevent them from demonstrating the target behavior in the standard way. If a child cannot physically produce the response the goal requires, you need to adapt the response mode rather than lower the cognitive demand. The goal should stay the same but the output mechanism changes. I worked with a nonverbal child who couldn't use a break card due to fine motor limitations, so we shifted to a switch-based communication device. The goal of independent break requesting remained unchanged. Only the response modality shifted.
Data Practices That Matter
Track data in a way that lets you see the trend, not just individual session scores. A single session at 70 percent means nothing. You need to see whether the trajectory is upward, flat, or regressing across at least five consecutive data points. I usually recommend graphing every session and reviewing the graph weekly with the supervising BCBA. The visual trend catches problems early that raw numbers hide. Set up maintenance probes immediately after mastery is reached. I've watched teams declare a goal met, stop working on it, and then discover three months later the behavior had completely disappeared because nothing was scheduled to maintain it. Build a maintenance probe into the next goal's protocol from day one. Even one probe per week after mastery is maintained can preserve the skill for six to eight months without additional intensive teaching. Write objectives that allow for measurable regression. This is counterintuitive but important. A goal should specify what happens if the child regresses past a certain threshold. I like to build in a criterion that says if accuracy drops below 60 percent for two consecutive sessions, the goal reverts to the previous teaching step. This prevents the false assumption that a child has mastered a skill when they're actually sliding. It also gives the team a clear decision tree instead of debating what to do when data deteriorates.
