Beyond Values Clarification: Addressing Client Values in Clinical Behavior Analysis

Abstract

Ethical principles of psychology, as exemplified in the American Psychological Association (APA) Code of Ethics (2002), provide impractical advice for addressing client values during psychotherapy. These principles seem to argue that each client's values should be respected and protected at all times, except in cases in which this would result in harm to the client or the general public. Although the code is appropriately designed as a defense against potential and actual abuses of professional power, this general proscription against directly targeting client values for change is based on an understanding of values that grants them special ontological status and has resulted in limited clinical interventions focusing solely on “values clarification.” With its strong foundation in a unified philosophical system, clinical behavior analysis offers a sophisticated alternative approach to values that both defines what they are and identifies when they can be ethically targeted to improve the lives of clients.

There are many definitions of values. Colloquially, a person's values are said to be what is important to him or her. Something (e.g., a specific object, a state of being, or a way of life) is important because it is what is good to pursue or possess. In the psychological literature, values are typically viewed in a similar manner; values are what people perceive, think, judge, or feel to be good (Tjelveit, 1999). Exemplifying this view, one popular definition states that “A value is an enduring belief that a specific mode of conduct or end-state of existence is personally or socially preferable to an opposite or converse mode of conduct or end-state of existence” (Rokeach, 1973, p. 5). Clinical behavior analysts also hold to this notion of values as being related to what is good or important, and it is in this sense that the present paper employs this term. The literature on acceptance and commitment therapy (ACT) has previously provided one specific approach to values (see, e.g., Hayes, Strosahl, & Wilson, 1999; Luoma, Hayes, & Walser, 2007). However, the present paper foregoes an exploration of the ACT approach in favor of presenting the broader, foundational approach to values found in clinical behavior analysis (CBA).1 This is done with the goal of providing clinical behavior analysts with a thorough understanding of what values are, how a behavioral analysis points out problems in a literal application of APA ethical principles during the course of behavior change, and how client values can be ethically addressed during therapy.

CURRENT ETHICAL APPROACHES TO CLIENT VALUES

That therapy can and does affect client values has been noted for many years, and this has led to some ethical quandaries for psychologists. Of particular concern has been the phenomenon of values convergence, in which a client's values change during the course of therapy to more closely reflect the values of his or her therapist. Rosenthal's (1955) seminal study first documented this type of change in client values, and Kelly's (1990) descriptive review of psychological research on values concluded that it does indeed occur on a regular basis. Some of the terms used in the discussion of this topic indicate the seriousness of the ethical violations that can occur when client values change in this manner. For example, Tjelveit (1986) suggested that the term values conversion replace values convergence in the literature because it more accurately reflects the finding that client values alone (as opposed to both client and therapist values) change during therapy. In addition, Meehl (1959, p. 257) described therapists who consistently influence the values of their clients as “crypto-missionaries.” Both of these examples are suggestive of the concern over the abuse of power that occurs when therapists inappropriately proselytize clients.

These ethical concerns regarding the changing of client values in therapy have been addressed in the American Psychological Association (APA) Code of Ethics, which serves as the current statement for the ethical principles that govern the activities of all psychologists (APA, 2002). The APA code has also been adopted by the Association for Behavior Analysis International (Code of Ethics, n.d.), establishing it as a code of conduct to govern clinical behavior analysts. Although the APA code does not make specific statements regarding values (in fact, the specific term is never used), it takes an unambiguous position on how to approach client values. In particular, one of the five main principles outlined in the code, “Respect for People's Rights and Dignity” (Principle E), states that differences between therapists and clients should be identified and respected. The principle also states, “Psychologists respect the dignity and worth of all people, and the rights of individuals to privacy, confidentiality, and self-determination” (APA, 2002, p. 1063). The right to “self-determination” can be interpreted to mean that clients should be allowed to freely generate and hold their own values without the undue influence of the therapist.

When strictly following this interpretation of the APA code, two conclusions may be drawn. First, clients have a right to hold to their own unique values (i.e., views about what is good or important). Second, it is unethical for therapists to target these values directly for change during therapy. This means that psychologists can advise a person on how to reach a goal, but they should not tell that person what goal to seek. The concerns about identifying and respecting client values have led to efforts to minimize the impact of therapist values on the therapeutic process (Slife, Smith, & Burchfield, 2003). The general strategy is to make use of some type of values-clarification procedure with two purposes. The first is to alert the therapist about what the client's values are so the therapist does not inadvertently target them for change. The second goal is to increase the likelihood that the client him- or herself defines values-consistent therapeutic goals.

THE DILEMMA FOR CLINICAL BEHAVIOR ANALYSTS

Adherence to such an approach to client values (i.e., broad proscriptions against influencing client values, sole reliance on values clarification) is problematic for clinical behavior analysts and psychologists who target behavior change. On a practical level, it is unclear how therapy intended to produce behavior change can proceed without affecting client values. It is unrealistic to expect that changes in client behavior during the course of therapy would not be accompanied by changes in values. According to Bem's (1967, 1972) self-perception theory, people's beliefs about themselves (which include their values) reflect their overt behavior; a person describes him- or herself as an outside observer would. If a person begins to behave in new ways, this is likely to be accompanied by changes in stated values. For example, a client who learns social skills and begins to form intimate relationships for the first time will begin to report new values concerning relationships (e.g., “My friends are important to me.”). Behavior-change agendas common in therapy settings will lead to changes in client values. Consequently, complete avoidance of the influence of client values by therapists could prevent therapists from pursuing behavior change during therapy. This would be antithetical to the goals and approach of CBA.

Furthermore, eschewing changes in client values is obviously problematic when clients generate therapeutic goals that are impossible to achieve. For example, a client might have the goal of developing intimate relationships without having to be vulnerable to interpersonal rejection. The latter value generally precludes achieving the former. In other cases, clients have therapeutic agendas that are unethical for the clinician to support. One example is an adult who has the goal of continuing to engage in sexual behavior with children without experiencing any negative consequences (e.g., guilt, social disapproval).2 In both of these cases, therapists cannot support the values of the client (i.e., help the client to achieve those values). At the same time, therapists are to respect the chosen values of clients (i.e., should not challenge those values). As a result, therapy cannot progress during this type of standoff unless some aspect of client values is made a legitimate target for change.

A standard values-clarification approach may also interact with other factors to unduly promote the practice of concentrating on the reduction of suffering to the neglect of emphasizing more functional therapeutic goals. The field of psychotherapy is currently dominated by a medical model (see Follette, Houts, & Hayes, 1992), and this has influenced the manner in which therapists and clients view the therapeutic process. Clients emphasize a desire to feel better, and therapists focus on the elimination of diagnosed syndromes. This makes it easy for the therapist and client to reach a consensus on adopting the reduction of the client's psychological symptoms (e.g., depressed mood, poor performance resulting from anxiety) and the learning of distress-tolerance skills as an overarching goal for therapy.

In contrast, it may be much more difficult to reach agreement on the positive goals of therapy (e.g., success in a particular romantic relationship, improved functioning in one's career setting, expanding the range of experiences that may lead to an increase in contingent reinforcement) without engaging the client's values. This is especially the case when a client's goals are based on values that the therapist does not share. For example, a therapist may be reluctant to support activities that will allow a client to increase monetary wealth at the expense of his or her relationship with his or her family without at least pointing to the possible adverse outcome or the inconsistency between providing for and attending to the needs of the family. Consequently, it may be difficult for therapy to work toward positive, functional goals without ensuring that the client has an adequate repertoire for evaluating the consequences of choices he or she might make.

This is clearly at odds with CBA, which has long adhered to the constructional approach described by Goldiamond (1974), in which new client behavioral repertoires are shaped based on behaviors already present in existing repertoires (see, e.g., Hawkins, 1986). This constructional approach to behavior change highlights the congruence between the person-centered positive psychology movement and CBA (Follette, Linnerooth, & Ruckstuhl, 2001). The focus of therapy is on both the reduction of contact with aversive consequences and increasing access to positive reinforcement. In other words, when a client enters treatment, the therapist may not only attempt the reduction of problematic behaviors but may also improve the functioning of the client. The movement toward a positive goal inevitably engages the values of the client and the therapist. Thus, a strict adherence to the values-clarification approach and the broad proscriptions against targeting client values for change underlying it preclude clinical behavior analysts from effecting positive changes in the lives of clients without at least having clients inspect the consequences of alternative behaviors.

At a philosophical level, this clarification-only stance towards values also conflicts with the underlying assumptions of CBA. The APA code accords values special ontological status in which they are intimately connected to a person's essence. They are assumed to be freely chosen by the individual (i.e., generated by “free will”), and that freedom to choose contributes to the dignity and worth of each person. Western culture, especially since the start of the modern period, has generally promoted freedom in this regard (Slife et al., 2003). Thus, generating and maintaining one's own values are viewed as rights on which no one should infringe. As a result, defending a client's ability to choose values is defending that which is special about each individual. The APA code seemingly presumes that values are immutable in the light of new experiences or that one should be protected from experiences that could indirectly change their values. This is certainly inconsistent with the way behavior analysts analyze values.

It is also contrary to an enormous set of empirical findings in clinical, social, and cognitive psychology. The data showing that clients' values change during therapy make it difficult to alter behavior while still granting values the special status accorded to them in the APA code. Furthermore, an emphasis on values clarification has been ineffective at achieving the intended goal of preventing changes in client values. In spite of an awareness of the issue, values convergence continues to occur during psychotherapy (Kelly, 1990). This happens without the intent or knowledge of the therapist, even when the therapist is taking measures to prevent it, and even when the therapist takes a values-neutral, relativistic stance (Kelly, 1990; Slife et al., 2003). All of this clearly demonstrates that values are not freely chosen; they are subject to influence by an individual's environment.

In summary, broad proscriptions against targeting client values and the resulting reliance on values clarification are inconsistent with CBA on both practical and philosophical levels. In some cases, clinical behavior analysts must directly address client values in order to fulfill their duties as therapists. In other cases, they attempt to direct behavior changes in clients that will result in a change in a client's values. This presents a dilemma: How do clinical behavior analysts behave in an ethical manner when beneficial behavior change frequently requires or leads to change in a client's values? This paper provides one approach to this question. It outlines an understanding of values grounded in behavioral science. That analysis will confirm that values change is an inevitable consequence of behavior change. There are, however, circumstances in which the behavior analyst must remain circumspect in facilitating some kinds of behavior change when consequences are not predictable. It is in these circumstances that the APA's recognition of client autonomy is most applicable.

A BEHAVIOR ANALYSIS OF VALUES

Skinner (1945) provides a philosophical framework for defining terms: Identification of environmental variables that occasion the use of that term can be used to scientifically define the term (i.e., what we mean when we use that term). This framework can be applied to generate a behavior-analytic understanding of the term values (Leigland, 2005). The term is typically used to refer to constructs such as beliefs, preferences, and attitudes about what is “good,” and its use is generally occasioned by the same variables that occasion the use of those terms. From a radical behavioral perspective, these terms cannot refer to the expression of the free will or essence of an organism. Statements about beliefs, preferences, and attitudes can only be made in reference to observed behaviors (e.g., the person who eats vanilla ice cream when both vanilla and chocolate ice cream are available is said to prefer vanilla). The same is true regarding values; they are inferred from the behavior of individuals. (This is the main reason behind claims that statements of facts and statements of values are identical; see Leigland, 2005, for further discussion.) Consequently, a behavioral analysis of values will focus on aspects of the behaviors of individuals that typically occasion the use of the term. It will be useful to make distinctions among the terms valuing, functional values, and statements of values as we try to clarify the distinction between occasions when values are inevitably changed during the process of therapy and occasions when behavior change and corresponding values change should be as completely under the control of the client as possible.

Valuing

In a general sense, values are synonymous with behaviors. How a person behaves demonstrates his or her values: A person who does charity work values doing charity work. Likewise, a man who exercises values exercise, a woman who eats breakfast values eating breakfast, and a child who does homework values doing homework. These values are held regardless of the values a person verbally endorses. Consider a person who claims to value gender equality but does not provide equal pay across genders. This person does not actually value this type of equality. At this level, values are best understood in terms of ongoing activities (i.e., valuing) rather than static entities. As a result, a college student may be said to be valuing talking on the phone one moment and valuing watching television in the next. In most cases, though, people do not talk about values in such a momentary manner. They instead identify values by analyzing patterns of valuing (see Bem's self-perception theory for an account of this phenomenon; Bem, 1967, 1972). For example, a man who is conscious of the extra hours he consistently spends at his office may say that he values working.

Functional Values

A complementary behavioral understanding of values focuses on the variables that maintain behavior (valuing). In this case, what a person values is indicated by what controls his or her behavior: The man who works many hours may be said to actually value the money provided by his paycheck or the praise he gets from his boss. Skinner (1971) specifically highlighted the reinforcing consequences of behavior as being intimately connected with values (Leigland, 2005). From this perspective, values are understood as the actual functional consequences that maintain behavior (i.e., functional values).

Values have many sources. For example, they can be established by a person's phylogenetic history. According to Skinner (1971), “it is part of the genetic endowment called ‘human nature’ to be reinforced in particular ways by particular things” (p. 104). Examples of these consequences include sexual contact and pain; they are part of an individual's functional values. In addition, functional values can be established through conditioning processes. For example, a discriminative stimulus involved in a behavioral chain can function as a conditioned reinforcer of the behavior that produces it (e.g., completion of a job application can be of functional value to an individual even if he or she still needs to turn in the application and attend an interview before being hired).

Consistent accompaniment of an event with previously functioning reinforcers can also lead to that event taking on a reinforcing function of its own (e.g., gold stars are of functional value to a child when they are consistently correlated with prizes and social praise). Conditioned reinforcers can even take on a generalized function. This means that they affect behavior on their own, without needing further direct attachment to other positive reinforcers. A common example of a generalized conditioned reinforcer is money. These examples demonstrate that a wide variety of current and historical conditions influence the functional values of an individual at any given moment. Like the behaviors they influence, functional values can be affected by many variables. For example, a young girl's functional values may be identified in animal crackers when she is hungry and in a stuffed animal when she is not. As a result, Leigland (2005) highlights the importance of the establishing operation (Michael, 1982; see also Laraway, Snycerski, Michael, & Poling, 2003, for the expansion of this concept to the motivative operation) in the analysis of the term values.

There are a wide variety of constructs, principles, and theories stemming from the field of behavior analysis that provide further examples of how functional values may change. (A few of these are mentioned here to orient the reader to the potential complexity involved in functional values.) One traditional example is deprivation: The functional value of food items is increased as nourishment is withheld.3 Similarly, a client may present for therapy when he or she becomes increasingly isolated and lonely following the end of a failed relationship. Herrnstein's (1974) matching law demonstrates a mathematical approach to quantifying changes in functional values. The most important implication of that theory is that adding reinforcers to an environment reduces the relative effectiveness of already present reinforcers. Thus, a person's functional values change as additional reinforcing consequences are made available in his or her environment. For example, a businessperson's functional values related to his or her job may become weakened as he or she enters friendships outside the context of the office. Informed by research on delayed discounting, the Ainslie-Rachlin principle (Ainslie, 1975; Rachlin, 1970, 1974) identifies how functional values change as a function of temporal distance between behaviors and consequences. The consequential function of an event is enhanced as it becomes more temporally proximate (e.g., the functional value of a positive reinforcer increases as its delivery approaches). This is evident in the increased likelihood of an individual accepting an invitation to go out for a drink later the same day as opposed to 2 weeks later (i.e., all else being equal, the more immediate opportunity for social contact is of greater functional value to the individual).

Statements of Values

In addition to identifying the valuing (behaving) and functional values (maintaining variables) of individuals, the term values has been frequently used in reference to verbal behavior of clients related to these two variables. This type of verbal behavior will be referred to as statements of values. These statements can be classified into two formal categories: (a) evaluative statements (i.e., statements that judge goodness or preferences) and (b) prescriptive statements (i.e., statements that identify appropriate courses of action). As noted by Leigland (2005), the analysis of statements of values is extremely complex because the same formal statement can have multiple functions depending on the variables that control its emission (Skinner, 1957). As a result, statements of values must be analyzed according to their function rather than their content. (There are many statements that may be directly related to values in their content. However, these may not meet the present functional definition of statements of values. See Day, 1977, for an account of possible variables that control statements that do not meet the functional definition used here; e.g., they may be intraverbals or the results of audience effects.)

The most basic statements of values can be thought of as simple statements of fact (Day, 1977). In the analysis of verbal behavior, this is described as tacting (Skinner, 1957). According to Skinner, tacting is functionally defined by appropriate discriminative control of the environment on the content of a statement (i.e., the statement reflects actual conditions of the environment) combined with reinforcement for the statement in the form of understanding by the verbal community (i.e., the listener understands what the speaker is saying). In the case of tacting, the verbal community usually reinforces correspondence between environmental context and content of statements.

Individuals may emit tacts that fit into both formal categories of statements of values. Some evaluative statements are tacts of functional values. In this case, an individual verbally identifies the reinforcing functions of stimuli and events. Examples are often found in the form of “— is good.” “Good” is used to describe those things that have served as positive reinforcers, and “bad” is assigned to things that have served as negative reinforcers or punishers (Skinner, 1971, p. 105). Evaluative statements can also be tacts of valuing. Here an individual identifies the frequency of his or her behavior. A common statement in this category is “I enjoy —.” A word such as “like” is ascribed to behaviors in which one frequently engages, and “hate” might be used in reference to an activity one often avoids. Prescriptive statements can also function as tacts. Examples include, “I should —,” and “— is right.” Skinner interprets these statements as tacts that identify contingencies of reinforcement. In this sense, they identify relations between valuing and values.

Although tacts of functional values, tacts of valuing, and tacts of the relation between the two can all be analyzed as separate statements of values, they are interrelated in many ways. Tacts of valuing help individuals to identify their functional values. Tacts of functional values help individuals to identify potential sources of reinforcement. Tacts of relations between valuing and functional values help individuals to identify ways to behave in order to contact reinforcing consequences. Making statements of values in the form of tacts leads to a number of benefits. Tacting one's functional values may allow an individual to efficiently mand for things from the verbal community (i.e., to identify reinforcing consequences that the listener can provide; Skinner, 1957). For example, tacting that one momentarily values not having a headache (i.e., identifying a negative reinforcer) makes it much easier to make a direct request that will lead to that consequence (e.g., “May I have an aspirin?”). In fact, the statement “I have a headache,” although formally a tact, may function as an indirect mand for the verbal community to remove the headache (Skinner, 1957).

Tacting one's valuing and functional values is also important for the generation of rules: statements that identify three-term contingencies that operate in the environment (Skinner, 1969). Rules, as statements identifying behaviors to be emitted, can serve as discriminative stimuli that occasion the emission of behaviors identified in them (Skinner). When a verbal event functions in this fashion for an individual, that person is engaging in rule following or rule-governed behavior (Hayes, 1989). Thus, the utility of statements of values in the form of rules is dependent on the individual's capacity for rule following.

Rule following is widely regarded as operant behavior (see, e.g., Hayes, 1989). It is strengthened when reinforcement (positive or negative) is provided contingent on completion of the behavior specified by it, and it is weakened when punishment (positive or negative) is provided contingent on failure to complete the behavior specified by it (Malott, 1989). Hayes, Zettle, and Rosenfarb (1989) have distinguished different types of rule following, including pliance and tracking. The difference between these is the source of the contingencies that maintain the rule following. In pliance, an individual engages in a behavior specified by a rule because of socially mediated consequences. In tracking, an individual engages in a behavior because of the apparent accuracy of the behavior–consequence relation identified by the rule. Take, for example, a teenage boy who follows the rule, “go to college.” If the teenager's behavior is maintained by contingencies mediated by his parents (e.g., he has been chastised for talking about seeking out an apprenticeship, he has received praise for his school performance), pliance is occurring. If the teenager's behavior is maintained by contingencies mediated by the extrasocial environment (e.g., the reinforcing effects of increased knowledge or the social interactions the school environment provides), tracking is occurring.

Rule following can be beneficial in many ways. It may allow a person to learn more efficiently to behave in an environment by eliminating the need for processes of shaping by contingencies. In many situations certain behaviors are more effective than others, and rules can clearly identify these behaviors (or their qualities) that are essential for success. For example, a person who wants to get a job can be told how to best fill out an application (e.g., “use good grammar,” “give references you know will be supportive”). Rules may also allow individuals to contact the contingencies that are instrumental in shaping fine-tuned behavior. It is difficult to tell an aspiring writer directly how to create interesting books, but he or she can easily be given a general task that can lead to this outcome (e.g., “practice writing and have other writers evaluate your work”). Rules can also be very important in aiding a person to behave with respect to long-term contingencies that would not allow learning via shaping processes (e.g., the relation between diet and life span). (See Malott, 1989, for further discussion on this topic.)

The benefits of tacts that serve as rules are dependent on two main factors. First, the utility of a rule is hampered when the tacts that underlie it are inaccurate. If the rules specify contingencies that are not present in the environment, they are not likely to be helpful in guiding behavior. Second, the utility of a rule is dependent on its relevance; the rule must refer to the actual valuing or functional values of that specific individual in order to be of use. In regard to valuing, rules are useful only if they describe the consequences (reinforcing or punishing) of the actual behavior of the individual. If a rule is relevant only to behaviors an individual never emits or never has the opportunity to emit, it will not be of benefit to the individual. For example, a financially insecure middle-class family will likely benefit from advice about reducing monthly spending but not from advice about high-end real estate investment. With regard to functional values, rules are useful only if they pertain to the current functional values of the individual. If a rule identifies consequences of behavior that have no meaning for the individual, it will have no useful purpose. For example, if a person does not find social contact reinforcing, he or she may not find tips for improving social interactions helpful.

Statements of values may be influenced in a number of ways in regard to rule following. A specific value may be established by aiding a person in identifying the contingencies that are operating in an environment. A person's verbal behavior can be shaped to label a particular relation between behaviors and environmental responses (i.e., to tact a rule). For example, in training a statement of values against stealing, one could highlight the negative impact of theft (e.g., criminal charges, loss of trust). The desired rule can be produced by actual contact with the contingencies, but it is often easier to provide it to the person verbally (i.e., the person is told, “You will go to jail if you are caught stealing.”). This approach is particularly effective when the individual who provide the rule has historically provided other accurate rules (i.e., serves as a discriminative stimulus for rule following; see Malott, 1989). Another way a statement of values can be reinforced is by providing socially mediated contingencies that reflect the contingent relation one desires to have reflected in the rule. To promote a rule about stealing, one might punish acts of thievery. This experience can then be used to formulate a rule, such as “I should not steal.” These examples show that there are a number of paths that could result in the formation of rules that could be defined as statements of values (e.g., “stealing is wrong”).

Statements of values can also be important in other ways. Specific focus has been placed on the role of complex human verbal processes in changing functional values. One construct of interest in this domain is the augmental, a type of verbal rule that changes the way in which stimuli function as consequences of behavior (Hayes et al., 1989). Two types of augmentals have been defined. A motivative augmental alters the degree to which a previously established consequence serves as a reinforcer or punisher (Barnes-Holmes, O'Hora, Roche, Hayes, & Bissett, 2001). For example, making the statement “I have strong faith” may increase the value of access to religious activities. A formative augmental establishes an event as a functional consequence for behavior (Barnes-Holmes et al.). For example, being told by a trusted authority figure (e.g., a parent, a teacher) that one political candidate is better than another can establish values that will maintain behaviors that support the candidate. Thus, statements of values may directly influence values rather than simply mirror them.

Relational frame theory (RFT; Hayes, Barnes-Holmes, & Roche, 2001) provides one framework for understanding augmenting processes. (RFT can also allow many extensions to the present analyses involving human language, e.g., rule following. Although fruitful in other contexts, pursuit of these extensions is beyond the scope of the present paper.) This literature has demonstrated that stimuli can derive functional relations through an operant history (see, e.g., Hayes, Kohlenberg, & Hayes, 1991; Ju & Hayes, 2008; Whelan & Barnes-Holmes, 2004; Whelan, Barnes-Holmes, & Dymond, 2006). Briefly, in the context of the word equal, the word good or bad can change the way that other stimuli function as a consequence of behavior. Theoretically, any other sounds (e.g., bueno or malo in Spanish) with similar verbal histories may influence subsequent events when presented in a similar context (e.g., this sentence). For this reason, stimuli that participate in these emergent relations (a product of consistent verbal operant histories) have been termed arbitrary. Furthermore, these emergent relations allow a transfer of function to unencountered events. This is important in the development of values that concern unexperienced consequences (e.g., the afterlife, world peace).

Summary

A behavior analysis of values examines three different phenomena that are typically thought of as related to values. The first is the actual behavior an individual emits (termed valuing). The second is the functional consequences maintaining the person's behavior (termed functional values). The third is a particular kind of verbal behavior (termed statements of values): tacts of an individual's valuing and values and rules that influence that person's behavior (during pliance, tracking, and augmenting). Although they can be theoretically differentiated, these three phenomena are intimately involved with each other. Changes in functional values lead to changes in valuing. Changes in reinforcing consequences lead to changes in behavior. Changes in functional values or valuing can lead to changes in statements of values. If the verbal community supports accurate tacting, changes in behaviors and variables that influence behavior should result in corresponding changes in their verbal identification. (Again, see Bem's, 1967, 1972, self-perception theory for a more comprehensive behavioral analysis of how changes in statements of values can result from changes in overt behavior.) Likewise, changes in statements of values can, in turn, lead to changes in functional values and valuing. Generation of new rules (by the individual or another person) that highlight particular contingencies can lead to the emission of new behaviors, and augmenting can change the functional consequences that influence behavior.

IMPLICATIONS OF THE BEHAVIORAL ANALYSIS OF VALUES

Within the radical behavioral framework, the variables involved in client values are no different than in other forms of behavior (Leigland, 2005). Valuing is understood in terms of the variables that maintain the behavior emitted by a client. Functional values are understood in terms of the processes that establish the reinforcing and punishing functions of events for a client's behavior. Statements of values are understood in terms of the variables that maintain a client's tacting, rule following, and augmenting. The definition of these terms is not an attempt to diminish the therapist's ethical responsibility. Rather, the behavioral understanding of client values provides conceptual support for the ethicality and even inevitability of directly targeting some client values for change.

Following this analysis, it could be relatively easy to conclude that each of these three components (valuing, functional values, and statements of values) can be ethically targeted by clinical behavior analysts (i.e., that proscriptions against changing values should be completely rejected). The goal of psychotherapy is behavior change; this is identical to changing the valuing of a client. To deny therapists the ability to effect behavioral change is tantamount to denying therapists the ability to engage in the therapeutic process. As a result, it cannot be unethical to target client valuing for change. A similar argument can be made in regard to client functional values. Manipulation of client functional values is requisite for changing client behavior. Thus, denying therapists the ability to manipulate the functional consequences of behavior also unnecessarily constrains the ability of the therapist to conduct therapy. The same is true in the case of statements of values. Therapists must be allowed to aid clients in accurately tacting their own valuing and functional values (as well as contingencies of reinforcement that operate in the environment) so that they can direct their own behavior more efficiently. Moreover, therapists must be able to influence clients' verbal behavior in ways that will result in augmenting that leads to effective actions.

However, such a conclusion is too strong. Just as the broad proscription against targeting values suggested by the APA code has been limited in its ultimate utility, granting a broad license to clinical behavior analysts that allows them to target client values at their whim is also too simplistic a solution to this issue. It would increase the likelihood that actual abuses of power (against which the APA code was developed to protect) will occur during therapy. Thus, it seems most appropriate to provide general guidelines that clinical behavior analysts can use to direct their activities effectively when targeting client values. Identification of these guidelines requires further consideration of the purpose of psychotherapy. CBA has as its purpose the improvement of the functioning of the client. This involves two broad goals that directly inform two principles that can guide the targeting of client values.

Reduce Harm

The first goal of CBA is the appropriate reduction of suffering (Follette et al., 2001). Consequently, clinical behavior analysts should target client values that contribute to unnecessary client suffering. At the level of valuing, this may include changing behaviors that lead to aversive consequences. One example is the behavior involved in chronic drug use, which (although it leads to immediate positive reinforcement) can significantly affect the client's physical health and lead to financial and interpersonal difficulties. At the level of functional values, reduction of suffering involves the elimination of unnecessary and destructive aversive consequences in a client's environment. An example of this is helping a client to leave a spouse who, although providing support to the client, abuses him or her consistently. At the level of statements of values, this can consist of decreasing the control of a dysfunctional rule that exerts aversive control over the behavior of the client. A rule of this type commonly takes the form, “If I don't do —, my life is meaningless.” It must be noted that the principle regarding the reduction of harm is not limited to the client. It is also applicable to the client's relationship with others. It is the professional responsibility of clinical behavior analysts to prevent clients from contributing to the suffering of others. Examples of contexts in which intervention on client values in this regard is appropriate include cases in which clients actively engage in discriminatory practices or threaten the physical well-being of others.

Increase Access to Positive Reinforcement

The second goal of CBA is to increase a client's access to positive reinforcement (Follette et al., 2001). In terms of valuing, this means that clinical behavior analysts should encourage client behaviors that lead to appetitive consequences. For example, when a client desires to expand his or her limited social network, the clinical behavior analyst can shape the client's interpersonal repertoire to be more effective. In regard to functional values, clinical behavior analysts should work to establish additional positive reinforcers for the client's behavior. An example of this could be to tell a client about the benefits of holding a job such that keeping a job begins to reinforce work-related behaviors. Increasing access to positive reinforcement may also involve manipulations of a client's statements of values. For example, a clinical behavior analyst may target a client's verbal behavior to increase his or her behavioral flexibility and variability (i.e., to ensure that client verbal behavior does not prevent him or her from being shaped by environmental contingencies) while encouraging patterns of behavior that lead to consistent and long-term reinforcement (e.g., help the client to tact his or her values with the goal of establishing effective rules). (The ACT approach to values has highlighted these activities of therapists. The reader is again referred to examples from the ACT literature, e.g., Hayes et al., 1999, and Luoma et al., 2007, for further discussion of that approach.)

Leigland (2005) provides a helpful observation that has bearing on the application of these principles: “Scientific analysis may indeed tell us what changes to make with respect to a particular problem, if the analyses have produced reliable, practical, evidence-based results that may be put to use in the solution of the problem” (p. 135). The implication of the elimination of the distinction between facts and values is that scientists can provide judgments regarding the appropriateness of various values (Leigland). Thus, the empirical knowledge clinical behavior analysts have concerning the benefit or harm of particular values (understood in terms of valuing, functional values, and statements of values) can be used in identifying appropriate circumstances for intervening on client values. For example, a therapist's knowledge about the potential negative effects of sexual promiscuity (e.g., sexually transmitted diseases, unplanned pregnancy) can inform an intervention strategy for changing the client's valuing (e.g., encouragement of safe sex practices).

Fortunately, the clinical behavior analyst does not need to be the authority on whom the client relies to define the client's values. Most often clients present for treatment when they are experiencing an excess of aversive consequences or a reported insufficient level of reinforcement. In these instances, a great deal of benefit can occur by training the client to become sensitive to the relation between his or her behaviors and the aversive or reinforcing consequences available in the environment. An idiographic intervention in which the client is encouraged to explore these relations is often sufficient. All that may be necessary is for the therapist to encourage any alternative behavior and then to prompt the client to notice the effect the behavior change produces. Thus the therapist does not have to say, “I think it is good to be active.” The client may say, “I feel depressed when I sit around all day,” to which the therapist may only need to reply, “Why not try something else, and let's just see what happens?” Eventually statements of values will emerge when the client learns how to do a functional analysis of his or her own behavior.

The therapist may also be a source of useful information that can increase the likelihood that a client will contact reinforcing contingencies more readily than if the client made use of only trial-and-error learning. For example, a client may report having a difficult time meeting interesting people or romantic partners. If the therapist learns the client is looking for new acquaintances only in bars, the therapist may reasonably suggest that the client look in places where people with common interests might be found at a higher probability (e.g., concerts, volunteer events, a hiking club). Here the therapist is making some predictions about the likelihood of a client's reinforcers being found in specific circumstances. These predictions may be based on many factors, including one's own experiences, logic, or information provided by others.

Now consider the instance in which a client asks the therapist's opinion on whether he or she should be the first in the family to try to go to college. For the sake of discussion, assume a typical outcome in which the therapist has successfully taught the client to identify what behaviors he or she finds reinforcing. At that point, the therapist can ask the noncommittal question, “Well, what do you think?” The therapist could then ask the client to make predictions about the appetitive and aversive qualities of the steps necessary to achieve that particular academic goal. In a more demanding scenario, the client may ask what it is like to be in college. At this point, the therapist can qualify the answer by saying all he or she can report is his or her own experience (not true, of course, because he or she witnessed many others' experiences while in college), but the response is likely to reflect the therapist's values. The value could be partly based on whether he or she enjoyed the experience, but it would also be influenced by the expectation that the outcome would expand the potential sources of reinforcement for the client.

It is here that we recognize where the APA's cautions become clearly applicable. The reliance on empirical evidence in making treatment decisions can justify the direct targeting of client values, but it highlights the limits of this approach. Clinical behavior analysts should target client values only when they can make reasonable predictions about the outcomes that might result from changes in the client's valuing, functional values, and statements of values. If this type of prediction is not possible, then the proscriptions against targeting client values stressed in the APA code should be followed. An example of a context in which one may not be able to make a reasonable prediction (and thus should not directly target client values) is the hypothetical case of a 60-year-old female client who is considering getting a divorce. She has had limited social contacts for many years, so it is difficult to predict how this client will adjust to life after divorce; she may adapt to single life very well (e.g., form many new relationships), or she may not (e.g., struggle to develop a social network).

In service of making reasonable predictions (or in determining if a reasonable prediction can be made), clinical behavior analysts need to assess the client's environment, or better yet, teach the client to do so. They need to identify values that can and will be maintained in the context in which the client lives. It would not be judicious to try to shape values that contradict values of a person's close social environment (e.g., active members of a particular political party). Conversely, the clinical behavior analyst needs to ensure that he or she is not inadvertently supporting an environment that is exploitative or otherwise promotes harm. Just because a social environment does not support particular values, it does not mean that these values should not be encouraged. A youth in a family of low socioeconomic status who dreams of going to college should not be dissuaded from that goal simply because his family and friends view it as a “foolish idea.” The necessity of being able to make reasonable predictions about the outcomes of the targeting of values also highlights the benefits of the constructional approach's emphasis on building on the client's current repertoire (Goldiamond, 1974; Hawkins, 1986). One can make the college example even more complicated by adding a cultural component in which the family thinks the college decision is not just foolish but disrespectful while at the same time friends are encouraging him to cut ties with his family and go to college. In this instance, the therapist would have little, if any, basis for predicting the consequences of either decision.

The final and most important consideration of the approach outlined here is that clinical behavior analysts need to continually assess the evidence they use to make decisions concerning the targeting of client values. They must ensure that it is based on empirical findings and on careful assessment rather than other sources. Anecdotal evidence is not a reliable basis for a reasonable prediction. Moreover, it is of the utmost importance that clinical behavior analysts do not unnecessarily promote their particular values. Just because the therapist functions well with particular kinds of valuing, functional values, and statements of values, it does not mean that any of these are appropriate for a client. (This is the primary foundation for the calls to protect client values from undue therapist influence.) Lastly, clinical behavior analysts can and should use the progress of each client to assess the benefits of an approach that directly targets values. Interventions that are successful should be continued, and interventions that are not successful should be discontinued and alternative strategies explored.

When making these determinations, clinical behavior analysts should consider one significant finding of the values-convergence literature: Changes in client values seem to be related to improvement as measured by therapist ratings but not by client ratings (Kelly, 1990). It is possible that this disparity between therapist and client evaluations simply reflects a difference in their abilities to identify improvement (i.e., therapists are better at evaluating increases in psychological health). However, it is likely more than this, because the therapist ratings often do not correlate with objective measures of symptoms (Kelly). It seems that therapists tend to consider a client “cured” when his or her verbal behavior is reflective of the verbal behavior of that therapist, regardless of the actual changes in the client's overt behavior (Kelly; Rosenthal, 1955). Therapists' behaviors are reinforced by clients who agree with them, and this means that therapists may not be accurate evaluators of the success of approaches that target client values. Because of this, a therapeutic approach that targets values must base tests of its efficacy (both at a general level and on an individual client basis) on objective empirical data that reflect the actual functioning of each client (e.g., specific behavioral outcomes).

CONCLUDING REMARKS

The present considerations serve as an initial step in the discussion of the role of values in CBA. We argue that client values can be targeted for change in an ethical manner. We also argue that, if the common reading of the APA's (2002) Code of Ethics is accurately interpreted as meaning values are not to be changed in therapy, no effective therapy can adhere to that constraint. The targeting of client values is, at least partially, a necessary pursuit in beneficial psychological treatment. At the same time, taking this approach requires that clinical behavior analysts carefully assess their reasons for selecting treatment targets (e.g., philosophical assumptions, personal biases, empirical data, relevance to a client's functioning).

The guidelines provided here attempt both to reflect the general approach of CBA and to conform to the spirit of the principles in the APA code (e.g., protection of the client from undue therapist influence, working for the client's best interest, prevention of harm). Ultimately, the utility of the position outlined in this paper can be judged based on its effectiveness in generating beneficial client outcomes. The assessment of the specific utility of targeting a client's values and the application of these guidelines are left to individual clinical behavior analysts. At the same time, the more general approach is open to evaluation at the philosophical, theoretical, and technological levels. Continued refinement at any of these levels would be welcomed. For example, a proposal for a methodology to assess the impact of more specifically targeting client values in which the environment supports the type of valuing targeted would help to advance this discussion.