The 4 Functions of Behavior in ABA: SEAT, Examples and the Exam

Two children throw their worksheets on the floor. One does it because the adult always comes over and talks to them. The other does it because the worksheet gets taken away. Same behavior, different reason, and if you respond to both the same way, you'll make one of them worse.

That reason is what applied behavior analysis calls the function of behavior, and the four functions of behavior are the shorthand the field uses for it. Most pages on this topic are written by therapy providers for parents. This one is written for people learning to do the work: RBT candidates, students in an ABA program, and anyone heading toward the BCBA exam. It covers what each function means, where the four-function model came from, how a BCBA actually identifies a function, and how the topic is tested.

What Are the 4 Functions of Behavior?

The four functions of behavior are sensory, escape, attention and tangible. Each one names the consequence that keeps a behavior going. In behavior-analytic terms, a function is the reinforcer that maintains the behavior, so "what's the function?" really means "what does this behavior get the person, or get them out of?"

The common mnemonic is SEAT: Sensory, Escape, Attention, Tangible. Some trainers order it as EATS, and our ABA glossary uses that version. The letters are the same and so is the content. Neither acronym comes from the Behavior Analyst Certification Board (BACB). They're teaching aids, and the exam won't ask you to spell one.

The key idea is that function is about why, not what. What the behavior looks like (hitting, crying, eloping, rocking) is its topography, and topography tells you very little. Hitting can be escape for one child and attention for another, and the same child can hit for different reasons in different settings.

The Four Functions, One at a Time

Sensory (Automatic Reinforcement)

A behavior with a sensory function produces its own reinforcement. Nobody has to do anything: the rocking, humming, hand-flapping, skin-picking or spinning feels good, relieves something unpleasant, or just provides stimulation. That's why behavior analysts usually call this automatic reinforcement, and it's the term you'll see on the BCBA exam.

The clue is that the behavior happens when the person is alone, at similar rates no matter who's in the room or what anyone does. If it drops when you leave, it probably isn't automatic.

Escape (Avoidance)

An escape-maintained behavior ends or postpones something the person doesn't want: a demand, a task, a noisy room, a transition, a social interaction. When it works, the unpleasant thing goes away, and that removal is what reinforces it. This is negative reinforcement. "Negative" means something was taken away, not that anything bad happened.

The clue is timing. Escape behavior clusters around demands and tends to start right after an instruction. If a tantrum reliably ends with the work being put away, even for a minute, that's worth noting.

Attention

An attention-maintained behavior produces a reaction from another person. The reaction doesn't have to be pleasant. A reprimand, a sigh, eye contact, "stop that," a long explanation of why we don't throw blocks: all of it is attention, and all of it can reinforce the behavior it's meant to stop.

The clue is that the behavior rises when adults are busy, on the phone, or attending to someone else, and that the person often checks for a reaction afterward.

Tangible (Access)

A tangible-maintained behavior gets the person an item or activity: a toy, a snack, a tablet, more time outside. Like attention, this is positive reinforcement, because something is added. The difference is that what's added is a thing or activity rather than a person's reaction.

The clue is that the behavior happens when a preferred item is taken away or denied, and stops once the item is back.

Four Functions, or Two? The Model Behind SEAT

SEAT is a simplification, and knowing what it simplifies is the difference between memorizing the list and understanding it. Underneath the four labels are two questions: does someone else deliver the consequence, and is something added or removed?

Laid out this way, SEAT puts three labels in the socially mediated row and squeezes the whole automatic row into "sensory." That's why some trainers teach a fifth function: automatic negative reinforcement, where a behavior removes an unpleasant internal state on its own, like scratching an itch or head-hitting during an ear infection. It's not a different theory, just the bottom-right cell made explicit.

This framing is what the BCBA exam expects. Task B.6 of the BCBA Test Content Outline (6th ed.) asks candidates to "identify and distinguish between automatic and socially mediated contingencies," and B.4 covers positive and negative reinforcement. Learn the grid and SEAT falls out of it.

Where the Four Functions Came From

The model wasn't invented as a teaching aid. It comes from an experimental method. In 1982, Brian Iwata and colleagues published "Toward a Functional Analysis of Self-Injury," reprinted in the Journal of Applied Behavior Analysis in 1994. They observed the self-injury of nine people with developmental disabilities across brief, repeated analogue conditions that varied three things: whether play materials were present, whether an experimenter placed demands, and whether social attention was absent, noncontingent or contingent on the behavior. In six of the nine, self-injury was consistently higher in one specific condition.

That's the core of a functional analysis: arrange conditions that test each possible consequence, and see which one the behavior tracks. The conditions map straight onto the functions. An attention condition tests attention, a demand condition tests escape, an alone condition tests automatic reinforcement, and later work added a tangible condition.

Twelve years later, the same group pooled 11 years of cases. Iwata, Pace, Dorsey and colleagues (1994) summarized 152 single-subject functional analyses of self-injurious behavior, run over about 4,000 sessions, and reported which function each one found.

Escape was the most common function, at 38.1% of cases. Attention or access to items (the study counts these as one category, social positive reinforcement) came to 26.3%, and automatic reinforcement to 25.7%. Eight cases (5.3%) had more than one controlling variable, and seven (4.6%) were uninterpretable.

Read those numbers for what they are. This is one behavior (self-injury), in people with developmental disabilities who were referred for treatment, three decades ago. It's not a rate for "all problem behavior," and no one should tell you escape is the answer 38% of the time in your caseload. What the study shows is that every function is common, so you can't guess. That's the argument for assessing instead of assuming.

Functional analysis became standard practice from there. A 2003 review by Hanley, Iwata and McCord describes pretreatment functional analysis as "a hallmark of contemporary approaches to behavioral assessment."

How a BCBA Identifies the Function

Finding the function is called a functional behavior assessment (FBA). It usually moves through three levels, from cheapest and least certain to most rigorous.

• Indirect assessment. Interviews and rating scales completed by people who know the client, such as the Functional Analysis Screening Tool (FAST), the Questions About Behavioral Function (QABF) or the Motivation Assessment Scale (MAS). They're quick and they point somewhere, but they're opinions about behavior, not observations of it.
• Descriptive assessment. Direct observation in the natural setting, most often ABC data: the antecedent (what happened just before), the behavior, and the consequence (what happened just after). Patterns in ABC data suggest a function. They can't prove one, because nothing is controlled.
• Functional analysis. The experimental method above: conditions arranged to test each consequence, with the behavior measured in each. It's the only one of the three that demonstrates a function rather than inferring it.

Designing these is BCBA work. The 6th edition outline lists "design and evaluate descriptive assessments" (F.5) and "design and evaluate functional analyses" (F.6). An RBT's part is narrower but real: the RBT Test Content Outline (3rd ed.) has RBTs "participate in components of functional assessment procedures (e.g., descriptive assessment, functional analysis)." In practice that often means collecting the ABC data or running sessions in a functional analysis the BCBA designed.

Why the Function Decides the Treatment

This is the reason the whole idea matters. An intervention that ignores function can reinforce the very behavior it targets. The classic example is time-out for a behavior maintained by escape: the child hits to get out of work, the adult removes the child from the work, and the hitting has just paid off.

Function-based treatment usually has two parts: stop the behavior from working, and teach a better way to get the same thing. The second part is what the 6th edition outline means by "select socially valid alternative behavior" (H.3).

Several of those procedures appear by name in the RBT outline's Behavior Reduction domain: antecedent interventions such as noncontingent reinforcement (NCR), high-probability request sequences and demand fading (D.2); differential reinforcement procedures including DRO, DRA, DRI, DRL and functional communication training (FCT) (D.3); and extinction (D.4). Extinction itself only works when it matches the function. Ignoring a behavior is extinction for attention, but it does nothing for escape, where extinction means the demand stays in place.

An RBT implements these plans. The BCBA chooses them, based on the assessment, and writes them up as a behavior intervention plan. That division is in the 6th edition outline too: H.2 is to "identify and recommend interventions based on assessment results."

Functions of Behavior on the RBT and BCBA Exams

On the RBT exam, the topic is task D.1: "Identify common functions of behavior," the first task in the Behavior Reduction domain. That domain carries 14 of the 75 scored questions (19%), and task B.3 (participating in functional assessment) sits in Behavior Assessment, which carries 8 more. Expect short scenarios: a behavior, what happened before and after, and four functions to choose from.

On the BCBA exam, the same idea runs through several domains: automatic versus socially mediated contingencies (B.6), descriptive assessment and functional analysis (F.5, F.6), and choosing interventions from assessment results (H.2, H.3). The questions go past naming the function to interpreting functional analysis data and picking the treatment that matches.

Try these. Each is the kind of scenario an RBT exam question builds on.

Common Mistakes With the Four Functions

• Assigning a function from what the behavior looks like. Crying isn't an "attention behavior." Function comes from the antecedents and consequences, never the topography.
• Treating "attention-seeking" as a criticism. It's a technical description of a reinforcer, not a judgment of the person. Everyone's behavior is maintained by attention some of the time.
• Assuming one behavior has one function. In the 1994 data, about one case in twenty had more than one controlling variable. A behavior can be escape at school and tangible at home.
• Stopping at the label. Naming the function is the start. The point is to choose a treatment that removes the payoff and teaches a better route to the same outcome.

For the bigger picture of where this fits in the job, see what an RBT does, and for related terms, browse the ABA glossary, including differential reinforcement and preference assessments.

Frequently Asked Questions

Sources

1. BACB, RBT Test Content Outline (3rd ed.) - tasks B.3 and D.1 to D.6, scored questions per domain
2. BACB, BCBA Test Content Outline (6th ed.) - tasks B.4, B.6, F.5, F.6, H.2, H.3
3. Iwata, Dorsey, Slifer, Bauman and Richman, "Toward a Functional Analysis of Self-Injury," Journal of Applied Behavior Analysis 27 (1994), 197-209, reprinted from 1982
4. Iwata, Pace, Dorsey et al., "The Functions of Self-Injurious Behavior: An Experimental-Epidemiological Analysis," Journal of Applied Behavior Analysis 27 (1994), 215-240
5. Hanley, Iwata and McCord, "Functional Analysis of Problem Behavior: A Review," Journal of Applied Behavior Analysis 36 (2003), 147-185