Functional Analysis in Clinical Settings

Definition

Functional analysis is a methodology used within the field of Behavioral Analysis to explain, predict, and influence behavior. This is achieved by identifying the functional relationships between behavior, the antecedent stimuli that elicit or evoke behavior, and the consequences of behavior that influence its future occurrence. Within this context, a functional relationship refers to a consistent and observable change in behavior (the “dependent variable”) that results from the systematic manipulation of environmental conditions (the “independent variables”). Functional analyses typically focus on the immediate environmental context, where stimuli functionally related to the behavior are observed. In applied and clinical settings, however, behavior analysts may consider additional variables (e.g., genetic predispositions, social context, learning history) to enhance the accuracy and relevance of their analyses. These variables are usually controlled for or excluded in experimental settings but can play a meaningful role in naturalistic behavior assessment.

1. Introduction

The introduction of the term functional analysis within the field of Behavior Analysis has been attributed to B. F. Skinner [1]. The use of this term and what it represents in Skinner’s thinking reflects the early influence of Ernst Mach’s philosophy, who questioned traditional views on cause-and-effect relationships appealing to the concepts of force or agency [2]. Rather than relying on inner explanations for behavior, Skinner promoted an analysis of behavior in terms of its functional relationships with the environment, redefining cause-and-effect relationships as observable and manipulable contingencies. These ideas were clearly promulgated with Skinner’s publication of Science and Human Behavior in 1953: “A cause becomes a change in an independent variable and an effect a change in a dependent variable” [3] (p. 23). He also wrote: “The external variables of which behavior is a function provide for what may be called a causal or functional analysis. We undertake to predict and control the behavior of the individual organism. This is our “dependent variable”—the effect for which we are to find the cause. Our “independent variables”—the causes of behavior—are the external conditions of which behavior is a function” [3] (p. 35).

Skinner’s concept of functional analysis was soon adopted in laboratory settings, where researchers were able to exert precise control over environmental variables to examine their influence on behavior [4]. The success of the experimental analysis of behavior in such controlled environments encouraged scholars to extend the use of functional analysis into applied settings, such as schools, clinics, and homes [5,6,7]. However, this transition posed significant challenges, particularly due to the lack of environmental control that is relatively easy to achieve in experimental settings but often difficult or impossible outside the laboratory [8]. It also became clear that a different set of variables (e.g., learning history, health conditions, cultural norms) needed to be considered when attempting to understand human behavior in real-life settings. Therefore, these challenges spurred the development of several adaptations of the original methodology to better suit naturalistic environments [9,10]. Additionally, descriptive assessments and indirect methods (e.g., interviews and rating scales) emerged as more practical alternatives when experimental manipulation was not feasible [11]. However, the proliferation of methods also led to terminological confusion, with distinctions often blurred between terms such as functional analysis, which typically refers to the experimental manipulation of antecedents and consequences; functional assessment, a broader term encompassing both experimental and non-experimental approaches; and descriptive analysis, which involves direct observation without manipulation [11]. Arguably, however, the concept of functional analysis is not restricted to any specific type of methods (e.g., experimental, observational) or any specific type of relationship between environmental variables and behavior [1]. It is from this perspective that the term functional analysis is used here.

2. Philosophical Assumptions

The formulation of a functional analysis is guided by a set of premises derived from B. F. Skinner’s philosophy of Radical Behaviorism, which underpins the field of Behavior Analysis. Skinner started to sketch this philosophical position in the early 1930s. However, it was not until his publication of The Operational Analysis of Psychological Terms in 1945 that these ideas were comprehensively discussed and first referred to in print as Radical Behaviorism [12,13].

It is worth noting, for the sake of clarity, that the term “radical behaviorism” had not been uniquely used to refer to Skinner’s philosophy. The first time this term appeared within the psychological literature has been dated to 1921, in an article published in Psychological Review and in reference to John B. Watson’s behaviorism. At the time, Watson’s behaviorism was often viewed as “radical”, in the sense of “extreme” or “iconoclastic”, given its departure from established paradigms (e.g., the study of consciousness) [13]. Such an understanding of the word “radical” (i.e., extreme) has often been mistakenly attributed to the “radical” in Skinner’ philosophy of Radical Behaviorism, wherein the term “radical” is purely linked to its etymology: radix (Latin; i.e., “root”) [14].

The philosophical assumptions outlined by Radical Behaviorism provide an ontological and epistemological basis for the experimental study of behavior and the implementation of interventions seeking to influence behavior in applied and clinical settings. Based on this perspective [2,3,12,15,16], behavior is viewed as the subject matter of psychology, a subject matter in its own right. Defined by Skinner as the interaction of an organism with its environment [3], behavior is understood as a physical phenomenon, measurable across different parameters (e.g., frequency, latency, intensity, duration), and fully intertwined with the context where it takes place; context is defined as the stimulus variability an organism is exposed to. It is in such a context where the variables controlling behavior are to be found, and this is why behavior is said to be “a function of” its context, rather than a phenomenon driven by purposeful or goal-directed action [3,17,18]. Moreover, no distinction is made from an ontological point of view between behavior taking place in the public and private realms (i.e., outside and within the skin); the former encompasses behaviors that can be literally observed by anyone (e.g., playing the violin, drinking tequila), whereas the latter encompasses other phenomena that can only be observed/noticed by the behaving individual (e.g., thinking, remembering, feeling). However, these so-called phenomena are also conceptualized as behavior, given that they are subject to the same lawful learning processes, namely, classical and operant conditioning. Similarly, the context can also comprise stimuli located both outside and within the skin.

Skinner recognized the relevance of both phylogenetic and ontogenetic influences on behavior. The former are manifested in inherited characteristics shaped by natural selection, such as reflex behaviors, which enhance survival and are passed down across generations. The latter refers to changes in an individual’s behavior over her/his lifetime as the result of learning through operant conditioning, a process Skinner referred to as selection by consequences. Further, Skinner proposed that the very capacity to learn through both classical and operant conditioning was itself a product of natural selection, as it confers adaptive advantages and increases an organism’s chances of survival [3].

3. Experimental Analysis of Behavior and Learning Principles

Research within the field of Experimental Analysis of Behavior has established that behavior is not random but follows consistent patterns [19,20,21]. These observations have led to the formulation of the so-called learning principles, which account for the acquisition, maintenance, and elimination of behavior as a function of environmental variables and organisms’ learning history and dispositional characteristics. This section presents a non-exhaustive overview of learning principles typically described under the umbrellas of classical and operant conditioning. Together, they offer the essential components needed to formulate a functional analysis.

3.1. Operant Conditioning

Operant conditioning constitutes a type of learning accounting for how behavior is influenced by its consequences. Such influence is exerted through different processes [22]:

• Positive reinforcement involves the presentation of a stimulus following a response and the subsequent increased probability that behavior will recur in the same or similar contexts
• Negative reinforcement involves the removal of a stimulus and the subsequent increased probability that behavior will recur in the same or similar contexts
• Positive punishment involves the presentation of a stimulus following a response and the subsequent decreased probability that behavior will recur in the same or similar contexts
• Negative punishment involves the removal of a stimulus following a response and the subsequent decreased probability that behavior will recur in the same or similar contexts
• Operant extinction, defined as the discontinuation of the contingency between a response and the stimuli that followed. Operant extinction typically results in a decrease in behavior over time and other phenomena, including extinction bursts (i.e., temporary increases in behavior frequency), and long-term relapse patterns like spontaneous recovery, renewal, and resurgence [23].

A few things are worth highlighting regarding the processes of reinforcement and punishment. First, reinforcement and punishment qualify as such only if they result in an increased or decreased probability, respectively, that behavior will recur in the same or similar contexts. In other words, both reinforcement and punishment are exclusively defined by their effects on behavior, not by the subjective valence of the stimuli that follow behavior (i.e., whether these stimuli are perceived as pleasant or unpleasant). Second, the terms positive and negative denote presentation and removal, respectively, rather than good and bad. Third, the processes of reinforcement, punishment, and operant extinction describe instances of behavior change influenced by the stimuli following behavior. However, behavior can also be influenced by stimuli preceding behavior. For example, when a specific stimulus signals the availability of reinforcement for a given behavior, such stimulus is referred to as a discriminative stimulus. Through repeated exposure to discriminative stimuli, organisms learn to differentially respond in the presence or absence of such stimuli, a process known as stimulus control. This control is established via differential reinforcement (i.e., reinforcing a behavior in the presence of a given stimulus but not others). As a result, a three-term contingency is formed, in which the discriminative stimulus signals the opportunity for reinforcement, the organism emits a response, and a consequence follows—either through the addition of a stimulus (positive reinforcement) or its removal (negative reinforcement) [22].

3.2. Classical Conditioning

Classical conditioning constitutes a type of learning accounting for changes in the conditions under which reflex behavior is elicited [22,24]. Unlike operant conditioning, classical conditioning does not explain the acquisition of new behaviors, but rather the establishment of new functional relationships between reflex behaviors and previously neutral stimuli. Classical conditioning involves a two-term contingency between a conditioned stimulus and an unconditioned stimulus. Here, a neutral stimulus becomes a conditioned stimulus through repeated pairings with an unconditioned stimulus that naturally elicits a reflex response, that is, an unconditioned response. As a result, the newly established conditioned stimulus comes to elicit a conditioned response similar to the original unconditioned response. In higher-order conditioning, a neutral stimulus can also become a conditioned stimulus through pairing with a previously established conditioned stimulus.

For example, salivation in response to food constitutes a reflex behavior. However, salivation can become conditioned to previously neutral stimuli through a contingency if those stimuli reliably precede the presentation of food. This principle explains why Pavlov’s dogs salivated at the sound of a bell, and why a person might experience the same reaction upon entering her/his favorite restaurant. After repeated pairings with food (i.e., an unconditioned stimulus), previously neutral stimuli (e.g., a bell or restaurant) may come to elicit a response that closely resembles the original response triggered by food. Such new learning (i.e., the association between the sound of a bell or restaurant and salivation) can undergo Pavlovian extinction upon repeated presentation of the conditioned stimulus in the absence of the unconditioned stimulus (i.e., the sound of a bell or restaurant in the absence of food).

3.3. Conditions Altering Functional Relationships

The functional relationships between contextual variables and operant behavior can be influenced by establishing operations, that is, by a set of stimulus conditions not functionally related to behavior [25]. Establishing operations may temporarily influence behavior by altering (a) the salience of discriminative stimuli, (b) the reinforcing efficacy of stimuli, and/or (c) the probability of emission of a given response. For example, food deprivation is often used as an establishing operation in experimental studies. Under such stimulus conditions, a rat is typically more likely to engage in certain operant behaviors (e.g., pressing a lever) previously followed by presentation of food, food itself becomes more reinforcing, and certain stimuli (e.g., a lever) and not others (e.g., exercise wheel) gain salience. Establishing operations have received a fair amount of attention in the context of operant conditioning, but not so much in the context of classical conditioning. However, for example, it is expected that the magnitude of an unconditioned response such as salivation in response to the presentation of food to be greatest in the context of food deprivation.

A different set of conditions, mostly relevant to human beings and to life outside the laboratory, can also influence functional relationships. This set of conditions has been referred to as dispositional, organismic, or person variables [10,26], and as in the case of establishing operations, this type of variables are not functionally related to behavior. Unlike establishing operations, however, person or dispositional variables most often connote rater permanent or structural conditions, such as individual characteristics (e.g., age, sex, phenotype, illness, learning history) [9] and environmental and social conditions (e.g., climate, law, culture) [10].

From a functional analytic perspective, dispositional variables may influence behavior in the same manner as establishing operations. Nonetheless, from a pragmatic point of view it may still be useful to distinguish them from establishing operations to highlight the significance of the aforementioned permanent and structural factors, especially for human behavior. Dispositional, organismic, or person variables may constitute an important source of variability in behavior, and therefore researchers in experimental settings take measures to reduce such variability to a minimum, for example, by ensuring that a litter of rats is raised under the same environmental conditions. However, these factors are impossible to control in life outside the laboratory, and therefore, it is important that they are considered by behavior analysts working in applied and clinical settings. All in all, dispositional variables and establishing operations demonstrate that functional relationships are not solely determined by present stimuli but are embedded in a broader context built over time.

3.4. A Word of Caution

First, it is worth noting that classical and operant conditioning are most often presented separately in textbooks on Behavior Analysis, just like they have been presented here, as if they were two independent processes. Independence, however, is rarely the case. For example, a particularly relevant intersection between operant and classical conditioning is observed in avoidance and escape behaviors, which are common in clinical practice. These operant behaviors are typically maintained by negative reinforcement, that is, by the avoidance or removal of an aversive stimulus that may otherwise elicit an aversive conditioned response. Avoidance or escape behaviors, however, may inadvertently facilitate the process of incubation [27], a process accounted for by both classical and operant conditioning. Incubation, however, may interfere with the process of Pavlovian extinction, which would deem avoidance or escape behavior unnecessary. For example, if an individual begins to avoid airplanes following a rough landing, the act of avoidance may immediately reduce her/his fear. This consequence strengthens the avoidance response via negative reinforcement, and the emotional response via incubation. Although such avoidance may be adaptive in the short term, it can become maladaptive over time, restricting the individual’s behavioral repertoire and limiting access to important environments or activities (e.g., work travel, tourism).

Second, while the principles discussed above provide a foundation for understanding classical and operant conditioning, considering additional phenomena related to these learning processes may enhance the scope and precision of hypotheses when formulating functional analyses. These other learning phenomena, not addressed here, include habituation, sensitization, blocking, overshadowing, shaping, chaining, generalization, abstraction, equivalence classes, conditional discrimination, conditioned and generalized reinforcers, generalized operants, rule-governed behavior, extinction induced variability, and arbitrary relational responding [22,28]. Additionally, the understanding of complex behavior may be further elucidated by taking into consideration the potential influence of the different reinforcement schedules, as described by Ferster and Skinner [29].

4. Formulating a Functional Analysis

This section provides a basic overview of the process of formulating a functional analysis in clinical settings. Readers seeking more comprehensive guidelines may refer to Froxán-Parga et al. [10], Sturmey [30], and Farmer & Chapman [26].

As stated above, the field of behavior analysis is underpinned by the assumptions of Radical Behaviorism, which clearly delineate what counts as behavior, the nature of the relationship that may be encountered between a given behavior and environmental variables, and the ways in which behavior may be influenced. However, it is also pertinent to consider further assumptions specifically related to the formulation of a functional analysis in clinical settings:

• Functional analysis can be used to account for any behavior in any context; its formulation is not limited to problem behaviors.
• Functional analysis is present-oriented: while an individual’s learning history is important for generating hypotheses about how behavior was acquired, the primary focus of a functional analysis is on the current variables that maintain behavior.
• Functional analyses can be conducted using different approaches, depending on the level of control and the context in which they are formulated. These include indirect methods (e.g., interviews and questionnaires), descriptive or observational methods (i.e., direct observation of naturally occurring behavior and its context), and experimental analyses, in which antecedents and consequences are systematically manipulated to test specific hypotheses about behavioral function [6,31].
• Functional analyses in clinical settings are meant to be useful, that is, they are meant to provide an idiographic understanding of an individual’s problem behavior, and to inform the implementation of an intervention designed to help the individual achieve her/his goals while taking into consideration the uniqueness of her/his history and surrounding context.

4.1. Functional Hypotheses: Origin and Maintenance

A complete functional analysis in clinical settings ideally results in two complementary hypotheses. The first concerns the origin of the (problem) behavior, that is, the environmental circumstances, establishing operations, and the individual’s dispositional variables that might have influenced the behavior’s first occurrence. This origin hypothesis helps contextualize the development of the behavior over time but does not necessarily explain current behavior. Doing so is accomplished by a second type of hypothesis, that is, the maintenance hypothesis, which identifies the environmental circumstances, establishing operations, and dispositional variables currently sustaining behavior [10].

4.2. Formulating a Functional Analysis—An Overview

The following subsections describe a series of tasks behavior analysists in clinical settings typically follow when formulating a functional analysis. These tasks are not necessarily completed in the order presented here; experienced clinicians may complete some of them simultaneously.

4.2.1. Identifying the Target Behavior and Its Context

The formulation of a functional analysis begins with a process of data gathering, which as described above, may rely on different methods depending on the setting [6,31]. In clinical settings, behavior analysts typically begin this process by means of unstructured interviews, through which clients broadly offer information about their presenting problem(s). Clients may sometimes offer rather vague accounts of their problems (e.g., “I’m depressed”, “I have low self-esteem”), and thus these must be carefully operationalized. The type of data gathered during these initial stages can be useful to inform a descriptive or topographical analysis, that is, a description of problem behaviors based on their observable physical features and the contexts where they emerge. As the assessment process proceeds, clinicians may require more specific information to fine-tune their hypotheses and may resort to other methods of data gathering, such as self-reports, client self-monitoring of problem behavior, client self-observation, or systematic observation.

4.2.2. Determining the Nature of the Problem Behavior

The topography of the behavior under analysis can offer important cues about whether it should be conceptualized as a reflex or operant behavior. For example, physiological responses (e.g., increased heart rate) or emotional reactions (e.g., anxiety) most likely constitute reflex behaviors, whereas motor responses and verbal behavior (including covert verbal behavior) most likely constitute operant behavior. Once this verified, it is possible to gain insight into the nature of the functional relationship between the behavior(s) at hand and the relevant environmental stimuli. Based on the principles of classical conditioning, reflex behaviors (i.e., unconditioned and conditioned responses) are directly elicited by antecedent stimuli (i.e., unconditioned and conditioned stimulus) and are not modifiable through consequences. Reflex behaviors, thus, are embedded within a two-term contingency (i.e., unconditioned/conditioned stimulus—unconditioned/conditioned response). In contrast, based on the principles of operant conditioning, operant behavior is emitted by the individual, evoked by discriminative stimuli, and influenced by reinforcer stimuli following behavior. Thus, operant behaviors are embedded within a three-term contingency (i.e., discriminative stimulus—operant behavior—reinforcer stimulus). Differentiating between these two types of behaviors is essential to correctly interpreting the behavioral function, formulating the appropriate functional analysis, and potentially selecting appropriate intervention strategies.

4.2.3. Identifying Functional Relationships Between Behavior and Contextual Variables

Once problem behaviors are identified, carefully operationalized, and categorized as reflex or operant, clinicians must identify the functional relationships existing between these behaviors and environmental variables. Upon questioning, clients may offer information about the situational variables that precede and follow problematic behaviors or unwanted feelings. However, clinicians must carefully isolate the antecedents and consequences that are truly functionally related to behavior (i.e., unconditioned stimulus, conditioned stimulus, discriminative stimulus, reinforcers), as well as any variables (e.g., establishing operations, dispositional variables) that may influence the functional relationship between the behavior under analysis and relevant environmental variables. It must be noted that, strictly speaking, the terms antecedents and consequences denote temporal relations (i.e., what precedes and follows a behavior), and that not all stimuli preceding or following behavior may be functionally related to behavior.

4.2.4. Identifying the Function of Consequences

In the presence of operant behaviors, behavior analysts must examine the function of consequences. This includes evaluating the temporal contiguity between behavior and subsequent environmental events, as well as analyzing the effect of these events on the future probability of behavior. If stimuli following behavior consistently lead to an increase or decrease in its frequency, they may be functioning as reinforcers or punishers, respectively. In cases where no observable consequences follow the behavior, it is necessary to consider whether the behavior is being maintained under intermittent reinforcement schedules, stimulus control, private reinforcement, or is undergoing operant extinction.

4.2.5. Identifying Variables Modulating Functional Relationships

As mentioned above, establishing operations and dispositional variables can modulate the functional relationships between behavior and environmental conditions. Therefore, these conditions must be identified when relevant and the way they modulate functional relationships of interest must be described.

4.3. Implementing an Intervention

In clinical settings, functional analyses are conducted not only to understand the occurrence of problem behavior, but also to guide the development of effective interventions. Identifying problem behavior is important, as such behavior often interferes with daily functioning, prevents individuals from reaching desired goals, and/or hinders progress toward a better quality of life. However, the key to meaningful behavior change lies not merely in labeling the behavior, but in identifying the environmental stimuli that are functionally related to it. These stimuli—whether antecedents, consequences, or contextual factors—become the primary targets through which interventions exert their effects. That is, it is often by modifying these environmental variables that behavior analysts help clients reduce problematic behaviors and achieve their personal goals.

5. Applications

Research in the field of Experimental Analysis of Behavior had documented from the 1950’s that non-human operant behavior is shaped and maintained by its consequences. Foundational research on learning principles was published at the time in The Journal of the Experimental Analysis of Behavior (JEAB). With the publication of Science and Human Behavior in 1953 [3], however, Skinner went a step further; he argued that human behavior was also governed by the same principles previously identified in studies using rats, pigeons, and other nonhuman animals. This caused considerable controversy. Soon after, nonetheless, behavior analysts began publishing empirical studies supporting Skinner’s assertions, and The Journal of Applied Behavior Analysis (JABA) was launched in the following decade, providing further evidence that human operant behavior is influenced by its consequences.

The application of learning principles to the study of human behavior marked an inflection point in the field of Behavior Analysis, greatly expanding its utility and scope [32,33]. An example of this early research was published in a 1959 case study by C. D. Williams [34], who described using extinction to eliminate tantrums in a 21-month-old child. During the initial nights of treatment, the child’s crying not only increased in intensity but also in duration, a pattern that retrospectively aligns with the features of an extinction burst. These early applications illustrate the ways in which principles developed in basic research were beginning to inform practical interventions. Such behavior-analytical thinking and the transfer of procedures and technology from the laboratory to applied settings, as well as contributions from early behavior therapy (e.g., classical conditioning approaches, systematic desensitization) constitute defining features of the so-called first wave of cognitive behavior therapies [15,35].

The emergence of the second wave of cognitive behavior therapies from around the 1970s [36] marked a significant rupture with the behavior-analytic tradition [15]. The foundational postulates of Radical Behaviorism were gradually displaced by alternative philosophical views. For example, Aaron T. Beck and Albert Ellis, two of the leading figures in the development of Cognitive Behavioral Therapy and Rational Emotive Behavior Therapy, were influenced by Stoicism, a pre-Socratic philosophy emphasizing the rational control of thought and emotion [15,37]. This shift was also accompanied by the resurgence of dualistic thinking, and mental processes were increasingly treated as internal causes of behavior rather than as behavior themselves. Additionally, categorical systems for classifying mental disorders (e.g., Diagnostic and Statistical Manual of Mental Disorders) gained prominence, reinforcing a medicalized view of psychological suffering. As a result, the field of clinical psychology witnessed a widening gap between experimental research and clinical practice, with interventions often losing their grounding in empirical principles. One of the clearest signs of this drift was the diminishing role of functional analysis, which had once served as the cornerstone for understanding clinical problems and guiding intervention strategies. Instead, the second wave of cognitive behavior therapies favored topographically oriented and diagnostically driven approaches.

The third wave of cognitive behavior therapies is marked by a renewed emphasis on functional analytic thinking, reflecting a significant shift back to the foundational principles of radical behaviorism. That is, it is recognized that behavior must be understood in relation to its context, rather than as isolated internal processes. This return to contextual analysis has been accompanied by a resurgence of experimental research aimed at grounding clinical interventions in solid empirical foundations. A prominent example is the revived interest in language and cognition from a Skinnerian perspective [38,39]. Within this framework, functional analysis is once again regarded as the cornerstone for both understanding behavior, providing a systematic method to identify the environmental variables that influence psychological problems and guide effective, individualized treatment strategies [40,41].

6. Limitations

The use of functional analysis in clinical settings is often limited by several methodological and practical challenges rarely encountered in experimental environments. One major limitation is the lack of environmental control inherent to real-world clinical contexts, which makes it difficult to isolate and manipulate variables systematically. While laboratory-based methods (e.g., time-series analyses, ABAB designs) are effective in demonstrating functional relationships between variables, their implementation is often infeasible or impractical in clinical settings. Therefore, there is no absolute certainty that observed reductions in problem behavior result from treatment components rather than extraneous or uncontrolled factors. This limitation is compounded by the reliance on patients’ verbal reports, which are often subjective, inconsistent, or influenced by memory biases, thereby reducing the reliability of the data used to guide functional hypotheses and interventions. Together, these challenges highlight the gap between experimental rigor and clinical practicality in the application of functional analysis.