A functional analysis (FA) is an experimental procedure used in applied behavior analysis to identify why a challenging behavior occurs. Rather than guessing at the cause, a functional analysis systematically tests different environmental conditions to determine what specific consequence is reinforcing the behavior. It is considered the gold standard in ABA because it is the only assessment method that can demonstrate a causal relationship between a behavior and its reinforcing consequences.
The method was first described in a landmark 1982 study that used an operant methodology to assess relationships between self-injury and specific environmental events. In six of nine subjects, higher levels of self-injury were consistently associated with a specific stimulus condition, demonstrating that the behavior was driven by distinct features of the social or physical environment. That core logic, testing one condition at a time to see which one produces the most behavior, remains the foundation of every functional analysis conducted today.
The Four Standard Conditions
A functional analysis cycles through four conditions, each designed to test whether the behavior is maintained by a specific type of reinforcement. Three are test conditions and one is a control. The conditions are typically run in alternating, randomized order across multiple sessions.
Attention
The therapist tells the individual they are busy and need to do some work, then moves away and ignores all behavior. If the target behavior occurs, the therapist provides brief attention: a mild reprimand, a statement of concern, or comforting physical contact. Then the therapist goes back to ignoring. If the behavior spikes in this condition, it suggests the behavior is maintained by getting attention from others.
Escape (Demand)
The therapist presents repeated learning trials or tasks that are appropriate to the individual’s skill level but somewhat effortful. If the target behavior occurs, the therapist immediately removes the task and turns away for about 30 seconds before restarting. Elevated behavior here suggests it functions as a way to avoid or escape demands.
Alone
The individual is placed in a relatively bare environment with no social interaction and no engaging materials. There are no programmed consequences for the behavior at all. If the behavior persists even when no one is around, it likely produces its own reinforcement, such as sensory stimulation. This is called automatic reinforcement.
Play (Control)
This condition is the baseline against which the other three are compared. The individual has free access to preferred toys or activities, receives frequent attention, and faces no demands. The behavior should occur at its lowest rate here because everything the person might want is already available. If it does occur, no consequence is delivered.
How the Results Are Interpreted
Data from each session are graphed on a multi-element design, with a separate data path for each condition. The analyst looks for separation between the lines. When behavior in a specific test condition is consistently and visibly higher than behavior in the control (play) condition, that signals a functional relationship. A common quantitative benchmark is a rate at least two times higher than the control condition.
Sometimes behavior is elevated in more than one test condition, indicating multiple functions. A child might hit to escape demands and also to get attention, depending on the situation. Other times, behavior is elevated across all conditions with no clear difference between any of them. That pattern points to automatic reinforcement, since the behavior occurs regardless of what social consequences are available.
How It Differs From a Functional Behavior Assessment
The term “functional behavior assessment” (FBA) is broader and often refers to a combination of indirect methods like interviews and rating scales, plus direct observation in natural settings. These descriptive approaches can identify correlations between behavior and environmental events, but they cannot prove causation. A functional analysis is the experimental component that can.
The correspondence between the two approaches is modest. One comparative study found only 38.5% exact agreement between FBAs conducted with indirect and descriptive methods alone and those that also included a functional analysis. The descriptive assessments were almost twice as likely to identify a function that the functional analysis did not confirm as they were to miss one the analysis did identify. In practical terms, relying solely on interviews and observation carries a real risk of targeting the wrong function when designing an intervention.
The tradeoff is time. Participants in the functional analysis group required an average of about 27 more sessions to complete all assessment and treatment procedures, equating to roughly five additional appointments compared to those assessed without one. That added investment is why functional analyses are most often reserved for severe or persistent behaviors that haven’t responded to less intensive approaches.
Safety During a Functional Analysis
Because a functional analysis involves deliberately evoking challenging behavior, safety planning is essential. The environment is cleared of sharp items, breakable objects, and anything that could be grabbed or thrown. Hard surfaces may be padded. Staff dress without dangling jewelry or loose clothing and tie back long hair. Specialized facilities sometimes use one-way observation mirrors, intercom systems, and electromagnetic door locks.
Several procedural modifications also reduce risk. Sessions can be shortened to as little as five minutes. Reinforcing every instance of the target behavior on a continuous schedule keeps the frequency and intensity lower than it would be under more intermittent reinforcement. Clinicians can also include a surrogate response, such as placing a cushion nearby and instructing the individual to hit the pad if they are upset.
For behaviors too dangerous to allow at all, a precursor functional analysis is an option. Clinicians first identify behaviors that reliably precede the dangerous behavior, like picking up a toy before throwing it. The analysis then programs consequences for those precursors instead. Because precursor behaviors are typically less dangerous, the analysis can proceed safely while still identifying the same maintaining function.
Variations for Natural Settings
The traditional functional analysis requires a controlled therapy room and extended session blocks, which isn’t always practical. A trial-based functional analysis adapts the method for classrooms and other everyday environments. Each trial is broken into short segments: a two-minute control period where the motivating condition is absent, followed by a two-minute test period where it is present and the behavior produces a consequence.
Trials are embedded into the natural flow of the day. Attention and tangible trials happen during free-play periods. Demand trials happen during instructional time. Other than a data sheet and a timer, no special materials are needed beyond what is already in the classroom. A typical trial-based analysis runs 8 to 16 trials per day across 4 to 6 days, with around 20 total trials per condition. The total active assessment time averages about four and a half hours, spread across those days in brief segments rather than requiring long dedicated blocks.
The Interview-Informed Approach
A more recent modification called the Interview-Informed Synthesized Contingency Analysis (IISCA) starts with an open-ended caregiver interview and a 15- to 30-minute clinical observation. Based on what caregivers report, the clinician designs just two conditions: a test condition that presents all suspected motivating factors at once and a control condition that provides continuous access to all suspected reinforcers so that the motivation to engage in problem behavior is minimized.
This approach has a practical appeal because it is faster and may better reflect real-world situations where multiple motivating conditions overlap. The tradeoff is precision. A traditional functional analysis identified single, isolated functions for 50% of participants in one comparison study and multiple distinct functions for another 41.7%. The synthesized approach, by combining all contingencies into one test condition, was less likely to tease apart which specific reinforcers were driving the behavior. That distinction matters when designing targeted interventions, since knowing the exact function allows for a more precise replacement strategy.
Who Conducts a Functional Analysis
A board-certified behavior analyst (BCBA) typically designs and oversees the analysis. However, the people running individual sessions can include teachers, parents, residential staff, and even trained undergraduate students, provided they receive proper training and supervision. The key requirements are that the person conducting sessions can follow the protocol with high accuracy, manage the necessary materials and pre-session setup, and collect and graph data reliably. The supervising analyst is responsible for interpreting the results and translating them into a treatment plan.