ABC Data Sheets and Behavior Data Systems for Schools
An ABC data sheet helps a school team record what happened before a behavior, what the student did, and what happened immediately after. The value is not the form by itself. The value is a shared way to observe, summarize, and decide what to do next.
If staff are writing long stories after the event, guessing at motives, or collecting data that nobody uses, make the system smaller and clearer. This guide shows school BCBAs how to build an ABC data collection process that fits a classroom.
What ABC data is, and what it is not
ABC means antecedent, behavior, and consequence.
- Antecedent: the observable event or condition immediately before the target behavior.
- Behavior: the observable action, written so another person could identify it.
- Consequence: what happened immediately after the behavior, including what adults, peers, and the student did next.
ABC recording is a descriptive observation method. It records behavior, antecedent conditions, and consequences in temporal sequence in the natural environment. A well-written record can help a team identify patterns and choose what to examine next (Cooper, Heron, & Heward, 2020, p. 56).
ABC data does not prove a behavior function. A pattern such as “task presented, student leaves, task is removed” is a useful observation. It is not, by itself, experimental evidence that escape maintains the behavior. Direct observations can be limited or skewed when they occur in only one time, setting, or interaction. Broaden observations and connect them to a full functional behavior assessment before making a strong function claim (Cooper, Heron, & Heward, 2020, pp. 56 to 57).
For a broader assessment process, use the functional behavior assessment guide.
How to write an ABC observation
Start with one operationally defined target behavior. Do not ask staff to record “defiance,” “attitude,” or “meltdown” unless the team has defined exactly what counts. Record what a person could see or hear. A definition should help two observers identify the same event.
Example: turn a vague note into usable data
Vague note: “He was upset because he did not want to work.”
Usable ABC entry:
| Time | Antecedent | Behavior | Consequence |
|---|---|---|---|
| Time10:14 | AntecedentTeacher placed a math worksheet on the desk and said, “Start problem one.” | BehaviorStudent pushed the worksheet to the floor, stood up, and walked 2 meters toward the door. | ConsequenceTeacher followed, blocked the doorway, and told the student to return. Student stood near the door for 90 seconds. |
The second entry does not claim what the student wanted. It records the instruction, the observable response, the distance, the adult response, and what happened next. It gives the team something to compare across observations.
What to record in each column
Antecedent:
- Activity, location, and people present when they matter.
- The instruction, demand, transition, denied request, peer event, or change in attention immediately before the behavior.
- Relevant setting events only when they are known and useful, such as a schedule change or a missed routine. Label them as context, not as the immediate antecedent.
Behavior:
- The exact action, words, movement, or measurable response.
- Topography and amount when relevant, such as “threw three pencils” or “said ‘no’ twice.”
- The start and end of an episode when duration matters.
Consequence:
- What adults did, what peers did, and what changed in the environment.
- Whether work, attention, an item, movement, or another event followed.
- What the student did next, without claiming that the consequence reinforced the behavior.
Narrative ABC or checklist ABC?
Use narrative ABC when the team is still learning what matters, the behavior is unusual, or the event needs context. Write short, objective phrases in time order.
Use a structured checklist when the team already knows the common options and needs consistent, fast entries. Include an “other” option with a small space for details. A checklist can include common events such as a demand, transition, denied access, peer interaction, free time, adult attention, redirection, break, or task removal. The checklist must match the student, target behavior, and school routine.
Do not force every observation into a checklist. If the available options do not describe the event, use narrative detail and revise the sheet during the team review.
ABC data sheet visual
ABC observation record
Record one episode in time order. Time flows through A, then B, then C.
Student
Target behavior
Observer
Date
Setting
Start time
End time
Time
Before
What happened right before?
Behavior
What did the student do?
After
What happened right after?
Use observable words.
Descriptive data show patterns to examine. They do not prove function by themselves.
Printable sheet text equivalent
Student: ____________________ Target behavior: ____________________
Observer: ____________________ Date: __________ Setting: ____________________
Observation start: __________ Observation end: __________
| Time | A, what happened right before? | B, what did the student do? | C, what happened right after? |
|---|---|---|---|
| Time______ | A, what happened right before?__________________________________ | B, what did the student do?__________________________________ | C, what happened right after?__________________________________ |
| Time______ | A, what happened right before?__________________________________ | B, what did the student do?__________________________________ | C, what happened right after?__________________________________ |
| Time______ | A, what happened right before?__________________________________ | B, what did the student do?__________________________________ | C, what happened right after?__________________________________ |
| Time______ | A, what happened right before?__________________________________ | B, what did the student do?__________________________________ | C, what happened right after?__________________________________ |
Observer reminder: Write what you saw and heard. Use the target behavior definition. Record the adult and peer response. Do not write a motive as if it were an observation.
How to take ABC data in a classroom
- Define the target behavior before the observation. Put the definition at the top of the sheet.
- Pick one routine and one decision. For example: “Does leaving the group happen after a transition, and what support should staff test first?”
- Tell the observer what counts as one episode and what to do if the behavior continues.
- Use a short observation window that can happen without removing the teacher from instruction.
- Record each episode as close to the event as possible. If immediate writing is unsafe or impossible, use a brief agreed-upon code and complete the entry as soon as the routine ends.
- Review entries with the observer. Ask, “What did you see?” before asking, “What do you think it means?”
- Compare observations across relevant times, people, activities, and settings before writing a hypothesis.
Direct observation is useful for selecting and describing behavior, but an observation can miss a pattern if the team looks only in one setting or with one person. Plan observations that sample the situations relevant to the referral question (Cooper, Heron, & Heward, 2020, pp. 56 to 57).
A recommended practice from Rob Spain
It is a good practice to do a brief ABC observation to see whether what you were told matches what actually happens, and to see the precursors for yourself.
How many observations do you need?
There is no universal observation count that makes an ABC dataset sufficient. The right amount depends on the decision, the behavior’s frequency, the safety context, and whether the observations represent the settings where the concern occurs.
Use a decision rule instead of a magic number:
- Continue observing when entries are too sparse to compare.
- Add a setting, person, or routine when the current sample may be masking the behavior.
- Stop collecting descriptive rows when the same pattern is repeating and the team has enough information to choose the next assessment step.
- Collect additional data when team members disagree about the definition, the sequence, or what happened after the behavior.
- For dangerous behavior, follow the school’s safety procedures and do not delay safety action while waiting for a fuller dataset.
Direct and frequent measurement helps practitioners detect change and decide whether to continue, modify, or end an intervention. Measurement starts with identifying the behavior, defining it in observable terms, and choosing an appropriate observation and recording method (Cooper, Heron, & Heward, 2020, pp. 74 to 75).
Summarize ABC data into a working hypothesis
Do not summarize the sheet by counting every word. Summarize the sequence that matters.
- Group entries by target behavior.
- Group antecedents into observable categories that match the actual notes.
- Group consequences by what changed after the behavior.
- Look for repetition across people, activities, times, and settings.
- Check exceptions. A pattern with many exceptions needs a narrower question.
- Write a conditional hypothesis that names the context, behavior, and common outcome.
Example:
When independent written work begins after a transition from a preferred activity, Jordan pushes materials away and leaves the work area. Staff commonly follow Jordan and delay the work. This pattern is consistent with a possible escape-related outcome, but ABC data alone does not establish function.
Then ask what data would test or clarify the hypothesis. Move the result into the team’s FBA to BIP workflow, and use the behavior intervention plan examples page when the team is ready to document prevention, teaching, and response procedures.
Which behavior measure fits?
ABC recording describes context around episodes. It is not the best measure for every progress question. Choose the measure that matches the dimension of behavior the team needs to change.
Which measure fits this behavior?
What do you need to know?
1How many times?
Count or rate
Discrete responses with clear starts and stops.
2How long?
Duration
Behavior that lasts too long, too briefly, or varies in length.
3How soon after an event?
Latency
Time from an instruction or opportunity to the response.
4Was it present during an interval?
Interval recording or momentary time sampling
Ongoing behavior when continuous observation is difficult.
5What remains after the behavior?
Permanent product
A reliable product that can be scored after the event.
6When does it occur during the day?
Scatterplot
Distribution across time periods and activities.
Text version of the measure chooser
- How many times? Count or rate. Discrete responses with clear starts and stops.
- How long? Duration. Behavior that lasts too long, too briefly, or varies in length.
- How soon after an event? Latency. Time from an instruction or opportunity to the response.
- Was it present during an interval? Interval recording or momentary time sampling. Ongoing behavior when continuous observation is difficult.
- What remains after the behavior? Permanent product. A reliable product that can be scored after the event.
- When does it occur during the day? Scatterplot. Distribution across time periods and activities.
Common school measures
| Measure | Use it when you need to know | School example | Watch for |
|---|---|---|---|
| MeasureCount | Use it when you need to knowHow many discrete responses occurred | School exampleNumber of call-outs during a 20-minute lesson | Watch forA count without observation time can mislead; record the observation period. |
| MeasureRate | Use it when you need to knowHow many responses occurred per unit of time | School exampleHand raises per 10 minutes | Watch forUse the same time unit when comparing observations. |
| MeasureDuration | Use it when you need to knowHow long the behavior lasts | School exampleMinutes engaged in work refusal or on-task behavior | Watch forDefine when the episode starts and ends. |
| MeasureLatency | Use it when you need to knowHow long it takes to start after an opportunity or instruction | School exampleSeconds from “open your book” to beginning the first problem | Watch forDefine the opportunity and the response onset. |
| MeasureWhole-interval recording | Use it when you need to knowWhether behavior occurred for the entire interval | School exampleCooperative play for each 30-second interval | Watch forIt can underestimate the total time behavior occurred. |
| MeasurePartial-interval recording | Use it when you need to knowWhether behavior occurred at any point in the interval | School exampleAny disruption during each 1-minute interval | Watch forIt can overestimate total duration and does not show how many responses occurred. |
| MeasureMomentary time sampling | Use it when you need to knowWhether behavior is occurring at the exact check moment | School exampleOn-task behavior at the end of each 2-minute check | Watch forIt can miss behavior between checks. |
| MeasurePermanent product | Use it when you need to knowWhether a reliable product remains to score later | School exampleNumber of completed, correct worksheet items | Watch forThe product must be produced by every target response and by no other response. |
| MeasureScatterplot | Use it when you need to knowWhen behavior is distributed across periods, activities, or locations | School exampleEpisodes by class period across the week | Watch forIt shows distribution and possible patterns, not a demonstrated functional relation. |
Count, duration, latency, and time sampling measure different dimensions. Duration is appropriate when the amount of time matters, latency is the elapsed time between a stimulus and a response, and time sampling uses intervals or specific moments to sample behavior (Cooper, Heron, & Heward, 2020, pp. 78 to 81, 88 to 93).
Whole-interval recording can underestimate occurrence, partial-interval recording can overestimate total duration, and momentary time sampling can miss behavior between checks. Choose one because it fits the question, not because it is the easiest column to add (Cooper, Heron, & Heward, 2020, pp. 88 to 93).
Permanent product is useful only when each target response produces the same product and the product cannot be produced by another response or person. A later score is not automatically a valid measure of what happened during instruction (Cooper, Heron, & Heward, 2020, pp. 93 to 96).
Scatterplots display the relative distribution of behavior across time or conditions. They can help a team decide where to observe next, but the pattern is not proof of function (Cooper, Heron, & Heward, 2020, pp. 138 to 140).
For IEP goal alignment, use the IEP behavior goal examples page. The goal, baseline, data method, and review schedule should describe the same behavior or skill.
For the broader role and systems context, visit BCBA in schools.
Train staff to collect data without making them behavior analysts
Staff training should answer four questions:
- What behavior are we recording?
- What does one example and one non-example look like?
- What is the fastest accurate way to enter the event?
- Who reviews the data, when do they review it, and what decision will follow?
Use behavior skills training when the task matters enough to require consistency: instruction, modeling, rehearsal, and feedback (Cooper, Heron, & Heward, 2020, p. 539). Practice with short classroom examples. Have staff write an entry, compare it with a model, and revise it. Then check a few real entries together and give feedback on the definition, sequence, and level of detail.
Use the shortest form that lets staff produce usable observations during the routine they already support. For a full staff implementation system, see the future staff training for school teams page when it is live.
A pre-observation briefing
- “Today we are watching for this one target behavior: [definition].”
- “Record the event that happened immediately before it.”
- “Write what the student did, not what the student wanted.”
- “Record what adults and peers did next.”
- “If you are unsure, mark the row and bring it to the review. Do not fill the gap with a guess.”
Make behavior data collection doable for teachers
Treat teacher time as part of the measurement system. If a teacher cannot use the form while teaching, the form is not ready.
- Limit each sheet to one decision and a small set of response options.
- Put the target behavior definition at the point of recording.
- Use checkboxes for repeated categories and a short “other” field.
- Keep narrative space for the details that cannot be predicted.
- Set a beginning and end to the observation.
- Decide who enters data, who summarizes it, and when the team will review it.
- Offer large-print, high-contrast, paper, and digital options when staff need them.
- Remove fields that do not change a decision.
Review the data with the people who collected it. Ask whether the sheet matched the routine, whether the definition was clear, and what made an entry hard to complete. Change the system when the school team finds a repeated barrier.
Frequently asked questions
What is ABC data collection?
ABC data collection is a descriptive observation method that records the antecedent, behavior, and consequence in time order. It helps a school team describe patterns around a target behavior. It does not prove the behavior’s function by itself (Cooper, Heron, & Heward, 2020, pp. 56 to 57).
What should be on an ABC data sheet?
Include the student, target behavior definition, observer, date, setting, observation times, and columns for the time, antecedent, behavior, and consequence. Add only fields that help the team answer its decision question.
How do you write an antecedent in ABC data?
Write the observable event immediately before the behavior, such as “teacher placed the worksheet on the desk and said, ‘Start problem one.’” Do not write a motive such as “student did not want to work.”
How many ABC observations should a school team collect?
Collect enough observations to compare the relevant settings and answer the team’s decision question. There is no universal count that fits every behavior. Add observations when the sample is too narrow, the behavior is rare, or observers disagree about what happened.
Does ABC data identify the function of behavior?
No. ABC data can show repeated correlations among events in the observed setting. It does not demonstrate a functional relation by itself. Use it to guide the next assessment step and interpret it with other FBA information (Cooper, Heron, & Heward, 2020, pp. 56 to 57).
What is the difference between ABC data and frequency data?
ABC data records the context around behavior episodes. Frequency or count data records how many responses occurred. Use ABC data when the team needs context and use count or rate when the team needs the number of discrete responses over a defined observation period (Cooper, Heron, & Heward, 2020, pp. 74 to 75, 78).
Should teachers collect ABC data all day?
Usually, no. Choose a short observation window tied to a decision and a routine where the behavior is relevant. A focused observation is more practical than asking a teacher to write a long form while teaching.
What if staff cannot write during the behavior?
Use a brief agreed-upon code, a checkbox, or a simple tally, then complete the sequence as soon as the routine allows. If the event is dangerous, follow safety procedures first.
When should a school team use duration, latency, or interval recording?
Use duration when time engaged matters, latency when the delay from an opportunity to a response matters, and interval recording when the behavior is ongoing or difficult to count continuously. Choose the measure that matches the behavior dimension and decision (Cooper, Heron, & Heward, 2020, pp. 78 to 81, 88 to 93).
What is a scatterplot used for in behavior data collection?
A scatterplot shows how behavior is distributed across times, activities, or settings. It can help the team identify where to observe and what routines to compare. It does not establish function by itself (Cooper, Heron, & Heward, 2020, pp. 138 to 140).
What should happen after ABC data collection?
The team should review the entries, check whether the observations represent relevant contexts, summarize repeated patterns and exceptions, write a cautious working hypothesis, and choose the next assessment or intervention decision. Then document the plan in a behavior intervention plan that staff can implement and measure.
Build a behavior data system your school team can use
Turn ABC observations into a shared process for choosing measures, training staff, reviewing patterns, and taking the next step.
January 14 to February 25, 2027. Six Thursdays, 6 to 8 PM Pacific Time. No session Feb 4, 2027. Applications close Thursday, January 7, 2027.