Can we have a behavioral science of anxiety?

Note: This article originally appeared in the July 2015 issue of Behavior Analysis Quarterly, Vol. 1(2).

By Amber Crane

St. Cloud State University

Recent reports indicate that the prevalence of anxiety disorders in the United States and their respective economic costs are a major societal issue. For instance, a report conducted by the Center for Disease Control estimates that these disorders are as prevalent as depression and reports reviewed estimate the diagnosis at 14% of all adults (Reeves, W. C., et al., 2013). In addition, economic costs nationally, linked to anxiety disorders has been estimated at over $48.72 billion annually, with the direct medical costs totaling $33.71 billion (Shirneshan, E. 2014). So far, the field of Behavior Analysis has been somewhat quiet on the subject in comparison to research for other societal concerns (e.g., autism) in part because of the difficulty in measuring private events and the inadequacies of a working definition of anxiety (Dymond & Roche, 2009; Friman, Hayes, & Wilson, 1998).

The identifying markers available on this subject rely principally on metaphorical examples (Friman et al., 1998), the object or event that the person’s anxiety is orientated towards, physiological events (5th ed., DSM- 5; American Psychiatric Association, 2013), or cognitive “disruptions” (Dugas, et al., 2007). Each of these lacks precision, fails to define anxiety directly, or provides challenges for measuring data. For example, with the orientation of the “phobia”, anxiety is diagnosed through the identification of the object or event to which the individual is anxious or fearful about but this may not provide measureable behaviors and may not be an indication of the controlling variables. In addition, these terms are developed through the individual’s reports of responses to private occurrences and the influences of verbal behavior. It is possible that one can circumvent a report of physiological events through measurements but this may not be practical or possible during treatment. Additionally, Friman, et al. (1998) pointed to the commonalities of physiological events that different emotions share. For instance, anticipation and anxiety may share an increase in heart rate and a measurement will not be reliable. Lastly, these measures are not beneficial in identifying the initial cause of the physiological event and do not measure the individual’s influence in reporting response to the stimulation (Skinner, 1953).

For these reasons, it is ideal to identify the controlling variables of a verbal account of emotions and the overt behaviors that denote anxiety. Given the current literature, it is not yet clear how the overt and covert emotional effects influence each other (Friman et al., 1998; Skinner, 1953). Keeping with a Skinnerian account of emotions, however, it is possible to investigate the controlling variables of all responses separately (Skinner, 1953). The total effects of an emotion are independent responses.

With this in mind, Skinner (1984) talked about the importance of the functional analysis of psychological terms. In his paper he used the term operational but the importance of the methodology is still clear and needed. The controlling variables with which verbal behavior of a private event develops may include, for example, the observations of the self, the methods to which one makes these observations, or neither of these occurrences. The meaning of a word is not found in the response itself but the contextual stimuli of which the response was produced. Leigland (2002) argues conducting a functional analysis of psychological terms would provide a pathway for the field of behavior analysis to follow. It is not the intention to replace the terms but provide direction and clarify a research direction (Leigland, 2002; Skinner, 1984).

An advancement of a science of human behavior requires an ability for the field to predict and influence behavior (Skinner, 1953). When discussing emotional responding, the ability for a behavior analyst to predict and therefore influence the behavior currently faces challenges. The responding is often a private event that cannot be observed and verbal behavior given by the behaving individual may not be complete. (Skinner, 1957).

Total Effects of Emotions

Although the current paper will focus on verbal and other overt behavior that denote anxiety, it is helpful to first address the total effects of emotions to provide some orientation. In other words, the outline of emotion and its effects will provide a point of reference for the other responses of anxiety and to provide evidence that the effects can be analyzed separately.

Skinner (1990), used an analogy of feeling the fabric on a jacket to describe the response to the physiological events termed emotions. Skinner (1957) went on to clarify that these are still responses and they are behaviors themselves. It is not, however, enough to look at just the stimulation and response. When an organism is presented with an aversive stimulus, there may be a physiological event that occurs in the smooth muscles and glands. The term anxiety is used to describe the response to the stimulation—or at least, it is one effect of the aversive stimuli in the environment. Simultaneously or in close temporal location, the organism may also engage in overt behaviors including weakening of behaviors or the organism may engage in behaviors to avoid or escape the aversive stimulus. Lastly, a person might engage in verbal behavior related to the emotion. The response that occurs covertly, however, is not the controlling variable. A person does not run because he is responding to the physiological event related to fear but because historically, this behavior resulted in the removal of aversive stimuli. Often the layperson discusses an overt emotional response as a result of an inner feeling. For example, the statement, “I ran because I was scared” is a common and acceptable statement. The controlling variable, however, is not the covert emotion but the environmental factor that may have elicited an emotional response—a response to stimulation of the smooth muscles—and other behaviors described next.

Looking further in depth at the potential controlling variables of the different effects of emotions will provide a foundation for analysis. It is possible that the aversive stimulus serves biological functions or develops through conditioning in an organism’s life. The conditioning occurs when the aversive stimulus is paired with other—to begin with—neutral stimuli and the now conditioned aversive stimulus elicits an emotional response. The person may engage in either covert or overt behaviors that are elicited. For example, he may scream, jump, or crouch in “fear”.

In addition to the elicited emotional response, the organism may engage in behavior maintained by negative reinforcement of the conditioned aversive stimulus or escape from the aversive stimulus. For example, a person is confronted with a strange figure in the alley and escapes the aversive stimulus by running away. Also, a person with self-defense training takes a stance to defend herself. The stance historically results in the removal of an aversive stimulus.

Identifying behaviors as either overt or covert, as a stimulus-response or maintained by historical consequences does not define them as the controlling variable. Again, it is possible that overt and covert behaviors influence each other but the initial environmental controlling variable must also be identified (Skinner, 1953). Although it is possible that emotional responding at times functions as motivating operations, it should be treated as a behavior with its own controlling variables first and its control on other responding should be considered later. Skinner (1957) cautioned an analysis of motivating operations—absence of friends and familiar places, in his example—as the cause for emotional responding. Nostalgia, an emotional condition, may result in the weakening of other behaviors in a state that may be described as depressed. The behaviors related to development, however, are not “specifically restrained” (Skinner, 1957, p. 166). In other words, the deprivation resulted in two responses. One is any responses that historically have been met with social contact and the other is the emotional response—nostalgia or sadness—that decreased the likelihood of other behaviors. The ways in which emotional dispositions develop is complex. A functional analysis, however, is still possible. Although multiple responses of emotions may influence or relate to each other further, it is beyond the scope of this paper to compare and contrast their potential influences.

Some Current Psychological Interpretations of Anxiety

As stated earlier, other fields of psychology—clinical and cognitive, for example—do provide information related to the definition of anxiety. The Diagnostic and Statistical Manual of Mental Disorders (5th ed., DSM-5; American Psychiatric Association, 2013) defined anxiety—along with separate classifications of the emotion—as being “associated with muscle tension and vigilance in preparation for future danger and cautious or avoidant behaviors” (p. 189). The corresponding additional classifications include: separation anxiety disorder, mutism, specific phobia, social phobia, agoraphobia, and generalized anxiety disorder (5th ed., DSM-5; American Psychiatric Association, 2013). In the cognitive model literature, there are four aspects of anxiety that are used to address and identify anxiety. They include: intolerance of uncertainty, positive beliefs about worry, negative problem orientation, and cognitive avoidance (Dugas, et al., 2007).

Usefulness of worry indicates tendencies that the person views the emotional response as solving problems or altering outcomes, providing motivation, or that the person is kinder for responding with worry. With negative problem solving orientation, the person lacks confidence about an ability to solve problems, and views problems as a threat, which may lead to increased worry (Dugas, et al. 2007).

An article in the journal Depression and Anxiety points to the tendency of literature on generalized anxiety disorder to focus on cognitive avoidance and provides some clarity to the terms related to anxiety. For instance the authors state that worry “functions as a cognitive avoidance response suppressing somatic anxiety and preventing patients from deeper emotional processing of aversive thought and images” (Beeso-Baum, et al., 2012, p. 949).

Excessive checking may be an avoidant behavior for indicators of a future threat—a potential conditioned aversive stimuli. In addition, the term safety behaviors relates to the individual attempting to ensure the safety of the self or others by carefully planning for any potential threats. In the same article, seeking reassurance is discussed as another potential indicator of anxiety. Lastly, patients reporting a loss of control and a weakening of behaviors such as sleep and eating habits (Beeso-Baum, et al., 2012).

Some characteristics of anxiety defined include attentional biases towards threat. One method discussed was to track the individuals’ eye gaze to identify the visual focus on stimuli and to also measure latency orientating towards “threatening” stimuli (Dugas, et al., 2007). The content of the “bias” is individualized given their learning history and the conditioning of objects or events may be specific or generalized. For example, the authors pointed to the diagnosis of social phobia in which the individual’s “illness” is characterized by—in part—a hyper-awareness towards angry faces. A related component discussed in the review is a vigilance-avoidance pattern in which the individual demonstrates shortened latency when orientating towards aversive stimuli and avoidance behavior. The individual exhibits engagement with the stimuli or difficulty removing awareness from the stimuli as well as awareness away from threat stimuli. It is unclear whether the awareness towards and aware from stimuli occur rapidly or in slower succession (Dugas et al., 2007).

A Potential Functional Analysis of Terms

In looking at the literature reviewed, there are a few patterns to point out. First, the categories used in the DSM-5 (5th ed., DSM-5; American Psychiatric Association, 2013) indicate a tendency to act. Many of the other terms imply avoidance or escape-maintained behavior, access to competing contingencies to anxiety, and attention-maintained behavior. The specifics are unique to anxiety and appear varied but categorizing them by function will provide some simplicity and clarity.

Tendencies to Act

Although the objects and events to which the individual is anxious about provides a way to test the occurrence of anxiety, it does not necessarily indicate its controlling variables. For instance, social phobia is anxiety surrounding the topic of social interactions. The individual may be anxious interacting with others and may report worry that others will judge or reject the person. The implication is that the individual’s social interactions contacted reinforcement and the aversive consequence. The individual engages in behaviors that avoid or escape the aversive stimuli. It is also possible, however, that an aversive stimulus was presented while the individual was in a social setting. The aversive stimuli may or may not be related to the social event but due to the conditioning, the social event now elicits a response related to the emotional disposition of anxiety. The conditioning related to development of anxiety may have a comprehensive effect in that even the nuances of the event later elicit a response. For instance, Skinner (1957) used an example of an unexpected death of a friend to demonstrate. The individual is then burdened by a general feeling of anxiety for no particular reason. The reason is difficult to identify because the death was paired with “everyday life.” In addition, the effect of the conditioning may be great due to multiple and “all-encompassing” conditioning. Returning to social phobia provides a good example. When the individual’s brief social interaction is paired with an aversive event, future presentations of a similar interaction now elicit a response. The elicited conditioned response may also be a stimulus that conditions additional responding. Although it is beyond the scope of this paper to address all of the potential effects of emotions especially as it pertains to the influence of responses to stimulation, it is worth pointing out that independence of the two responses and the potential exacerbating effect that the response of anxiety may have on future responding.

Dymond and Roche (2009) pointed to a possibility that direct conditioning does not occur. They point to the fact some individuals affected by anxiety disorders cannot identify when the conditioning occurred nor can they identify the “trauma.” The aversive stimuli may not have the appearance of a traumatic event to generate a response. In addition, the environment that now generates anxiety, may not have at first (Catania, 2013). For instance, after leaving a store, a person is attacked. The store initially did not generate anxiety but with future presentations, it elicits the response. In addition, through induction, similar stores elicit the emotion anxiety and other behaviors related to avoidance. The person may walk into an unfamiliar store and report a feeling of anxiety but is not sure why. The individual does not need to be aware of conditioning for it to occur (Skinner, 1953). Given the complexity of conditioning, it is not the “phobia” that indicates the controlling variable. In other words, it may or may not be aversive stimuli but may be correlated with it. For this reason, the categories used to diagnose anxiety disorders can only be understood as tendencies to act.

Escape-Maintained Behavior

For many of the psychological terms, the potential functional relation appears to be escape-maintained. For instance, with cognitive uncertainty, the individual will more likely avoid or escape from situations that are unfamiliar. This may include subtle examples—a phone call from an unknown caller—to the more obvious example of going on a blind date. Of course, avoidance or escape-maintained behavior is not unique to anxiety. For example, a person runs inside from the rain. The behavior contacts the removal of the aversive stimuli and the person tends to run inside when it begins to rain. Also, conditioning of aversive stimuli may occur and the person will more likely engage in behavior to avoid the conditioned aversive stimuli. The person does not necessarily feel anxious about the occurrence of rain or conditioned stimuli associated with it. It is the total effect of emotional responding that make related avoidant and escape-maintained behaviors categorized as “anxiety” but prediction and influence can still occur through the analysis of the avoidant and escape-maintained behaviors alone.

Access to Approval

With anxiety related to social interactions, the person may engage in behaviors to get approval. This may especially be likely if the person feels more mild anxiety and does not avoid the social interaction altogether. The history of rejection that now functions as a perhaps mild aversive stimuli may also serve as an establishing operation for statements related to favorable reception from others.


Another behavior that may be beneficial to explore further is the vigilance-avoidance pattern discussed earlier. In the article reviewed, the phenomenon was in reference to eye gaze and “awareness” towards and away from aversive stimuli. The effects of anxiety may include a subtle responding similar to this example or it may be a more of a gross response. Skinner (1953) discusses a related response. Algerbraic summation refers to two response to one stimulus that are the same except in opposite direction. For example, an approach and withdrawal response of the same stimuli. When on a smaller scale, the response might be a shaking hand as if in hesitation or the response occurs in a wave-like oscillation.

Reasons for Idiosyncrasies in Emotional Responding

Verbal Community and Its Control

The difficulty of identifying and measuring a covert emotional response, creates a necessity to rely on a person’s verbal behavior of the event. The self-reports may lack precision due to the shaping and maintenance of these responses by the verbal community. In part, they must rely on public accompaniments to reinforce appropriate tacts (Skinner, 1953, 1957). For example, the mother watches her child walk into preschool with wide eyes and slow steps. The mother might say, “It looks like you’re nervous”. Here the child learns to tact the response to a particular stimulation as a form of anxiety. What makes the development of accurate tacts of private events including emotional responding even more difficult is the mother might say—for the benefit of the child—“You’re going to have fun. I’m so excited for you”. With this learning history, the child does not learn to tact the emotional response correctly. The verbal behavior surrounding private events is developed through a long history of perhaps differing contingencies. It may be why, for example, the lay person tacts anxiety when they mean fear or the other way around. Skinner (1957) makes a simple distinction between fear and anxiety. Fear is an emotional response to an aversive stimulus. I am fearful of the man that is running after me. In comparison, a person feels anxious while walking down a dark and unfamiliar alley and the response is akin to I am anxious of what danger historically occurs after the presentation of this stimulus. For purposes of treatment, the distinction is important. Fear is the emotional response that accompanies escape-maintained behavior. The aversive stimuli is present. Whereas anxiety is related to avoidance of the aversive stimulus by escape of the conditioned stimulus and may be referred to as “concern for future events”. The layperson’s verbal behavior may not be a sophisticated or sensitive measure for this distinction.

Several sources point to the use of metaphors in reports of anxiety and the term itself is a derivative of the Latin word anguisse referring to a chocking sensation. Metaphoric extensions, as Skinner (1957) termed them aid in the verbal community shaping a tact of a private event. The wording relates to a public, and therefore, shared experience. The response that wording evokes mirrors the private experience. For example, when a person says, “I have butterflies in the stomach”, the image evokes a response in the listener that is similar to the emotional response of “being” nervous.

In addition, when a person talks about an emotional response, they are often referring to a tendency to act. They are not necessarily tacting the response to stimulation but a statement related to their history of acting in relation to that particular emotional response (Skinner, 1957). For example, a report of anxiety indicates that the speaker will more likely fidget, appear restless, avoid or escape certain situations, or maybe even seem withdrawn depending on the severity of the emotional response. The statement, depending on the learning history of the individual, may function—in part—as a mand for comfort or assistance in the removal of aversive stimuli. The impure tact is controlled by the private event and motivating operations (Skinner, 1953, 1957).

Group Control

A culture may dictate appropriate emotional responding through delivery of reinforcement, punishment, or aversive stimulus. The development of what is generally accepted occurs over time and for a variety of reasons but can be simply defined as what is good for the group. Some cultures engage in overt emotional responding openly whereas others punish those responses and reinforce only emotional responses that are subdued, demonstrated with only immediate family members, or organized verbal behavior related to emotions. The community may be the larger culture of the individual’s country of origin or a smaller community and may be different for individuals depending on their age or standing in the community. For example, different emotional responding is expected and tolerated from a 2 year old then an adult. The control exerted by the group may be in the form of disapproval or shame, a conditioned aversive stimulus. For example, when the customer demands rudely that his plate be taken back because the food is cold, other customers may glare or shake their heads at the response. His date may punish him with silence or withhold future dates. It may also play out that the customer is a powerful member of the community and others identify the angry outburst as an example of that power. Others may say that he is entitled to his demands and the emotional responding is appropriate.

Behavioral Applications for Treatment of Anxiety

Once the functional relation of a behavior is identified, the path for the behavior analyst is significantly clearer. In this section a few examples of how a functional analysis provides viable treatment options. Considering escape-maintained behavior appears frequently in the literature reviewed in this paper, it may be starting point for a behavior analyst. For instance, both respondent and operant conditioning are already an established practice in behavior analysis and other fields of psychology (Dugas, et al. 2009). With extinction, the individual is presented with stimulus that previously elicited emotional responding—anxiety—but the behavior no longer contacts the aversive stimulus. It may be beneficial to lessen the association of the originating stimulus to the presented stimulus by systematically introducing the conditioned aversive stimulus. Perhaps, for example, the person that “fears” social interactions. It would be appropriate, therefore, to present short interactions with more familiar people. Then, once anxiety is not demonstrated or reported, the expectation to contact the stimulus increases.

The aversive stimuli may be unique to the individual. There are several possibilities but some are listed. For “uncertainty to consequences”, the therapist would need to define the term more specifically—lacking a history of clear contingencies. The client would be presented with a situation in which the consequence was unknown. For example, the individual is asked to walk into a room and is told to there will be task but no description of what is expected or what the consequence will be for the behavior is provided.


The complexity of human behavior may generate interest in finding solutions to societal concerns through elaborate methods. It is not always fruitful to ignore the simple intricacies that make up behavior because precision is often lost. In an argument for a science of human behavior, Skinner (1957) uses the beginning student of physics and their view of what a science has to offer a worldview as a starting point. At first the science does not match her experience with it and its lackluster viewpoint is unappealing. As she becomes more astute in her studies, the science demonstrates its power by providing precision and the passion with which she viewed the world is again present. It was never the science that failed her but her ability to grasp the concepts. As the field of behavior analysis pursues its path towards a more comprehensive natural science of human behavior, it must more thoroughly address emotional dispositions and their influence. At first glance, a science will appear to fail in its ability to completely explain the influence of emotions as the layperson experiences it. Once thoroughly developed, however, it will not only examine that experience but provide an explanation when the experience leads to suffering.

The complexity of human behavior should not deter the field from first analyzing simple conditioning. The appearance that the stimulus control of the everyday experience does not match the stimulus control respondent and operant conditioning may or may not be complete. It is the many incidences of conditioning that take place that make up the complexity of human nature.


Email Subscription


  1. Below for your consideration is a simple hypothesis and procedure that demonstrates for the first time how core affective states can be mapped to simple response contingencies.

    The hypothesis is simple. Positively affective states, as mediated by midbrain opioid and dopamine systems, are induced by abstract (how you think) rather than normative (what you think about) information from the environment. Recreate that information as denoted by simple act-outcome contingencies, and you can elicit and sustain positive affect.

    The crux of the argument, as well as the procedure itself, takes only three pages (39-42). As a behavior analyst, the crux of the argument is the procedure itself, and the explanation only justifies its consideration and trial. The explanation has in other words zero shelf life if the procedure it suggests fails.

    The ‘explanation’ is a book on rest and affect that is written for a lay audience, with a scholarly appendix and links to other published supporting journal articles by the author. The work is strongly influenced by the research of the affective neuroscientist Kent Berridge of the University of Michigan, a preeminent expert on the neural processes that underlie affective states, who was kind to vet the book for accuracy and endorse the completed work.

    As an academically trained psychologist and radical behaviorist, but a layman, I would be most grateful for your brief consideration!


    Art Marr
    New Orleans

Leave a comment

Your email address will not be published.