Two teachers watch the same student for ten minutes. One records three "outbursts." The other records none. They saw the same student, and used different meanings of the word.
A definition fixes that. This page shows how to write one that holds up for counting, graphing, and writing into a plan.
The test
Queensland's guide states the standard: define the behavior in specific observable and measurable terms, so that it can be counted or timed, and ask whether someone who does not know the student would immediately recognize it. There should be no room for a subjective opinion about what it looks or sounds like.
The IES practice guide makes the same point from the classroom side. If descriptions of behavior are vague, for example "Jacob is always disruptive," it is difficult to assess the extent of the problem, when and where it most often occurs, and how to intervene. Its concrete examples name an action and a setting: blurting out answers without raising a hand during whole-class instruction, hitting and kicking peers during recess, leaving a seat without permission during small-group instruction.
Labels and descriptions
| Label | Description | Why it works |
|---|---|---|
| Disruptive | Calls out an answer or comment without being called on during whole-group instruction | Names the act and the setting |
| Off-task | Looks away from the work, or has no pencil or hand on the materials, for 5 seconds or more during independent work | Gives a visible boundary and a time |
| Refuses | Says "no" or does not begin the task within 30 seconds of the direction | Gives a response and a time |
| Aggressive | Hits, kicks, or pushes another person with enough force to be felt | Lists acts, not a character |
| Elopes | Leaves the assigned area with both feet outside it, without permission | Gives a clear edge |
Notice that each description names an action, and most set a boundary: a time, a place, or an intensity. A boundary is what lets two people agree.
Examples and non-examples
NCII's guide says to create anchored examples and nonexamples when defining the behavior. A short list on each side shows where the line falls.
For the call-out example, the definition card might say: Marcus says words aloud during whole-group instruction without being called on. Counts: shouting an answer before the question is finished; commenting on another student's answer without raising a hand. Does not count: answering when called on; whispering to a neighbor during partner talk. Now a second adult can score the same minute and land on the same number.
What to leave out
- Motives. "To get attention" belongs in the hypothesis. It cannot be seen.
- Feelings. "Frustrated" and "angry" are inferences. Describe the action instead.
- Judgments. "Rude," "lazy," and "defiant" invite disagreement.
- Other people's behavior. NCII lists a goal that measures an adult's behavior, such as fewer teacher prompts, as a common problem.
- Long lists. One behavior, one definition. If a series of acts always happens together, define the series, or choose the most important one.
Define it so you can measure it
A good definition suggests its own measure. If the behavior is brief and countable, a tally works. If it has a clear start and stop and matters for how long it lasts, time it. If it is ongoing, sample it in intervals. The measures are covered in the data hub, starting with data sheets for BIP progress. The data guides on this site assume you have a definition in hand.
Check agreement before you start
Have two people watch the same ten minutes, or the same video if your district allows it, and count independently. If your counts differ by more than a point or two, find the line you disagreed on and rewrite the definition. Do this once at the start and again if a new person takes over the sheet, such as a paraprofessional or a substitute. It takes ten minutes and prevents a month of data that cannot be trusted.
Three drafts of one definition
Start with the sentence a teacher might write in an email: "Dev gets angry and acts out."
Draft one removes the feeling: "Dev disrupts the class." Better, but "disrupts" still means different things to different adults.
Draft two names an action: "Dev throws materials." Closer. Does a pencil dropped on purpose count? What about a book pushed off a desk?
Draft three sets the edge: "Dev throws or pushes a school object, such as a pencil, book, or chair, so that it leaves the desk or travels more than one foot, during any lesson." With a non-example: "an object dropped by accident." Two observers could now agree.
The three drafts took about five minutes. Each one removed a question that two adults would otherwise have answered differently. That is the whole job of a definition.
One definition for the whole team
A definition only works if everyone uses the same one. Put it at the top of every data sheet, in the plan, and in the handout for substitutes and paraprofessionals. If it changes, replace every copy the same day, and mark the date on the graph so that data before and after are not compared as if they were the same thing.
If you are a paraprofessional or a substitute and the definition is unclear, ask. A question on day one is cheaper than a graph that cannot be read in month two. Teachers can make this easy by saying so at the start: "If you are not sure whether something counts, write it down with a note, and we will sort it out together."
Where definitions go wrong
Illinois's guidance lists vague or subjective definitions of the target behavior as the first of several obstacles to an effective plan, and incomplete measurement criteria as the second. They tend to travel together: if the definition is fuzzy, the data are fuzzy. Other common slips:
- Defining too broadly. "Any inappropriate language" will include things the team does not mean.
- Defining too narrowly. Count only the loudest incidents and the plan will miss the pattern.
- Changing the definition mid-way. A new definition means a new baseline.
- Writing it for the file, not for the people. Staff who count have to be able to read it in a few seconds.
Putting it in the plan
The definition appears twice: in the FBA, where it anchors the data, and in the BIP, where it anchors the goal. Colorado's guide says the hypothesis statement includes a brief operational definition of the target behavior. See hypothesis statement examples for how that looks in a sentence. Terms are in the glossary.
What stays local
Which forms record the definition, and who approves changes to it, are set by your district. If your district has a standard format, use it, and keep the examples and non-examples attached.