Picture a single line on a graph. The horizontal axis is dates, one mark per session, running from September to December. The vertical axis is the percentage of opportunities where a child asked for help instead of pushing the task away.
For six weeks, the line sits low and jagged, somewhere between ten and thirty percent. Then there's a vertical dotted line on the graph where the team changed something, and after it the points climb: forty, fifty-five, fifty, seventy, eighty.
That graph is what all the tally marks add up to. It is also, in my experience, the most reassuring thing a parent can be shown, because progress with a young child rarely feels like progress while you're living it. The week to week is noisy. The line is not.
This post is about what your child's team is recording during sessions, why they picked those particular measures, how they manage to record anything while keeping a three-year-old engaged, and how to read the graphs when they land in front of you.
What Is Data Collection in ABA Therapy?
Data collection in ABA therapy is the ongoing recording of how a child performs on each goal in their treatment plan, session by session. Depending on the goal, a therapist might count how often something happens, time how long it lasts, record how many opportunities the child responded to correctly, or write down what happened before and after a difficult moment. Those numbers are graphed and reviewed by the supervising BCBA, who uses them to decide whether to keep a teaching strategy, change it, or move to a new goal.
If you only read this far:
- Data starts with a precise definition of the behavior, because vague targets can't be measured reliably.
- Different goals need different measures, and the table below shows which is used when.
- ABC data is the tool for understanding difficult behavior, and it needs several occurrences before a pattern is trustworthy.
- Recording happens in short bursts and samples, not by testing your child constantly.
- The graphs drive real decisions, including when a strategy gets dropped.
- You can ask to see the data, and there are specific things worth looking for.
It Starts With Defining the Behavior
Before anything can be counted, it has to be pinned down.
The IRIS Center at Vanderbilt University's Peabody College, which produces training resources for educators, puts it this way in its guidance on behavior assessment: a behavior needs an operational definition, meaning it is defined clearly enough to be measurable and to be identified the same way by two or more observers, across time and in different settings.
That's why treatment plans read the way they do. "Improve communication" can't be measured. "Independently requests a preferred item using a two-word phrase within 10 seconds of the item being visible" can. Two different therapists watching the same moment should agree on whether it counted. If they wouldn't, the definition needs work, and any numbers collected against it aren't worth much.
The same goes for behaviors a family wants reduced. "Aggression" is not a definition. "Hitting an adult with an open or closed hand" is.
What Gets Measured, and Why That Measure
Not every goal is measured the same way, and the choice tells you something about the goal.
| Measure | What it records | When it's the right fit | Example |
|---|---|---|---|
| Frequency or count | How many times something happened | Behaviors with a clear start and stop that don't run constantly | Number of spontaneous requests in a session |
| Rate | Count divided by time | When session lengths vary, and you need comparable numbers | Requests per hour |
| Duration | How long the behavior lasted | Behaviors with a distinct beginning and end, or very high-rate ones | Minutes of crying, or minutes of independent play |
| Latency | Time between an instruction and the start of the response | Goals about responding to a direction or cue | Seconds between "time for shoes" and starting to put them on |
| Percentage of opportunities | Correct responses divided by chances given | Skills taught in repeated trials or across the day | Responds to name in 8 of 10 chances |
| Level of prompting | How much help was needed | Skills being faded from hand-over-hand toward independence | Handwashing completed with a gesture prompt instead of physical guidance |
| Task analysis steps | Which steps of a routine the child did alone | Multi-step daily living skills | 5 of 7 steps of the bedtime routine independent |
| ABC narrative | What happened before and after an incident | Understanding why a challenging behavior is occurring | Written record of a meltdown at transition time |
Duration and latency are worth understanding as a pair. IRIS describes duration recording as appropriate for behaviors with a distinct beginning and end or that occur at very high rates, and latency recording as measuring the time that passes between an antecedent, such as an adult's directive, and the start of the behavior. A child who eventually does everything asked but takes four minutes to start has a latency problem, not a compliance problem, and the plan should reflect that.
ABC Data and Difficult Behavior
When a behavior is unsafe or disruptive, counting it isn't enough. The team needs to know what it accomplishes for the child.
That's what ABC data is for: recording the antecedent (what happened right before), the behavior itself, and the consequence (what happened right after). IRIS notes that the ABC model is used to identify the antecedents that set the stage for the behavior and the consequences that appear to be maintaining it, and that an observer may need data across several sessions before a clear pattern emerges, usually around eight to ten occurrences, other than in cases of extreme behavior. The observer should also note the setting, the time of day, and who was present.
This is why a BCBA may seem slow to give you an answer about why something is happening. Two incidents are an anecdote. Ten incidents with a consistent pattern (say, the behavior nearly always follows a request to stop a preferred activity, and nearly always results in the activity continuing) point to a specific function and a specific replacement skill to teach.
How Anyone Records This While Actually Running a Session
A fair question, and one parents ask often: if the therapist is writing, who is teaching?
Good teams solve this in a few ways. Some data goes down in natural pauses, while a child is engaged with a toy or between activities. Some goals use a first-trial probe, where only the first opportunity of the day is scored, because that's the cleanest read on whether the skill is independent. Others use sampling, recording a set number of opportunities rather than every one. Momentary time sampling, where the therapist notes whether a behavior is happening at the end of each interval, tracks something like engagement without a stopwatch running nonstop.
Most teams now use an app on a phone or tablet, which is faster than paper and graphs automatically. Some still use paper, which is fine. What matters is that the method is consistent and that the therapist isn't so busy recording that your child is left waiting.
Two quality checks worth knowing about. The first is reliability: when two people score the same session independently and compare, do their numbers match? The second is honesty about tradeoffs. Sampling methods are easier to run but less precise than counting every instance. A well-designed plan accepts that tradeoff deliberately rather than pretending it doesn't exist.
What Happens to the Numbers
Data that gets collected and never looked at is just paperwork. The professional standard is explicit on this point. The BACB Ethics Code for Behavior Analysts requires behavior analysts to ensure that data collection procedures are selected and implemented correctly, to display the data graphically, and to use the data to decide whether to continue, modify, or end services (Standard 2.17). Standard 2.18 requires continual monitoring, and when data show the desired outcomes are not being achieved, the behavior analyst is required to assess the situation and take corrective action.
Broader practice standards say the same at the program level. The ABA Practice Guidelines from the Council of Autism Service Providers describe standards of care for planning, implementing, and evaluating ABA services, with ongoing measurement of progress central to that evaluation.
In practice, that means your BCBA is looking for a few things in the graph: whether the line is moving in the right direction, how fast, and whether it changed after the team changed something. A flat line for several weeks is information. It usually prompts a question like whether the prompt fading moved too quickly, whether the reinforcement still motivates your child, or whether the goal was too big a jump from where they were.
What This Looks Like in Practice
The example below is a composite of children we've worked with. Details have been changed, and it doesn't describe one specific child.
A five-year-old's aggression data looked random to everyone, including his parents. Some days there were no incidents, some days there were six, and nobody could see a pattern. His team ran ABC data for three weeks and noted the time of day and who was present along with each incident.
The pattern wasn't the demands everyone assumed were the trigger. The incidents clustered in the forty minutes after his older sister got home from school, when the house went from quiet to loud, and his one-on-one attention dropped off. The behavior reliably produced an adult coming straight over.
In our sessions, we've seen data change the conversation from "why is he doing this" to "what is he getting, and how else can he get it." In this case, the plan shifted to a scheduled check-in during that window, a way for him to ask for attention that didn't require hitting anyone, and a small transition routine when his sister arrived. The line came down over about six weeks, with a noticeable spike the week the family went away, and the routine dropped. That spike was useful too. It told everyone the routine was doing real work.
How to Read Your Child's Graphs
Ask to see them. Then look for:
- The two axes. Time along the bottom, the measure up the side. Check what the measure actually is, because "percent of opportunities" and "number of incidents" tell very different stories.
- The vertical phase lines. Those dotted lines mark where the team changed something. Compare the points before and after.
- Direction and steadiness. For a skill you want up, is the trend rising? For a behavior you want down, is it falling? Are the points getting less scattered?
- The mastery criterion. Most goals have one written into the plan, such as 80 percent across three consecutive sessions with two different people. Ask what your child's is.
- The gaps. Illness, holidays, and staff changes show up in the data, and they explain a lot of dips.
If a graph is confusing, say so. Explaining it plainly is part of the job, and a team that runs data review with families rather than around them will welcome the question.
Numbers That Turn Into Decisions
Tally marks and timers can feel clinical when what you want is for your child to be happy and understood. But careful measurement is what separates a program that adjusts when something isn't working from one that repeats the same approach for months.
At Centerbrite, our BCBAs write clear, measurable goals for each child, track them session by session in your home, review the graphs regularly, and sit down with families to explain what the numbers mean and what's changing as a result. We provide in-home ABA therapy for autistic children across Bergen, Essex, and Morris counties in New Jersey. If you'd like to see how progress would be measured and shared for your child, contact us today to get started.
Frequently Asked Questions
1. Is my child being tested the whole session?
No. Most data is recorded during ordinary teaching and play, often in short samples or on the first opportunity of the day. Your child should experience a session as activities with an adult they like, not as a series of quizzes.
2. Doesn't all this recording get in the way of the therapy?
It can, if it's done badly. That's why teams use sampling and probes rather than scoring every single moment, and why most have moved to apps that take a second per entry. If it ever looks like your therapist is writing more than teaching, raise it.
3. Can I see my child's data, and can I get copies?
Yes. Ask your BCBA for the graphs and the progress report, and ask them to walk you through what each one measures. Families are entitled to understand how progress is being tracked, and questions about documentation and reporting are covered on our FAQ page.
4. What if the data shows no progress?
That's exactly what the data is for. A flat or worsening trend should trigger a change in approach, not more of the same. Ask what the team plans to adjust and when they'll check whether the adjustment worked.
5. Should I be collecting data at home?
Usually not in a formal way. What helps more is telling your BCBA what you notice: which times of day are hard, what happened right before a rough moment, and where a skill does or doesn't show up. A note on your phone is plenty.
Sources:
- https://iris.peabody.vanderbilt.edu/activity/behavior-assessment-duration-and-latency-recording/
- https://bacb.com/wp-content/ethics-code-for-behavior-analysts/
- https://www.casproviders.org/asd-guidelines/
- https://www.autismspeaks.org/expert-opinion/autism-aggression
