Synonyms

Behavior change; Behavioral change

Definition

Behavior modification means the change of a current behavior by adopting a new behavior or by increasing or decreasing the current behavior.

Understanding, predicting, and evoking behavior modification is a key topic in all areas of psychology. It is a crucial means to reaching personal, organizational, and societal goals across the lifespan, including the major goal to stay healthy and well. With chronic disease on the rise, partly due to an increasing aging population, promoting and stabilizing health and well-being is more topical than ever (cf. Scholz et al. 2015). Behavior modification is a major factor to prevent and manage chronic diseases, such as diabetes and cardiovascular diseases (e.g., Ornish et al. 1998), be it, for example, to normalize blood pressure levels by exercising or eating more healthily or by better adherence to medication regimes; behavior modification and long-term maintenance are crucial at any time during the lifespan, even though the goal of behavior modification may differ depending on age. Whereas increasing health and well-being or preventing disease may be more achievable goals for younger persons, stabilization is often the more realistic goal in older age. Irrespective of this, however, behavior modification is usually a prerequisite. The following sections introduce the key principles of behavior modification. An overview is given of the behavioral determinants, the techniques by which behavior modification can be engendered, and how these can be selected for interventions. The entry ends with concluding remarks and directions for future research.

A Social Cognitive Approach to Behavior Modification

Before the cognitive revolution in the second part of the twentieth century, behavior modification was mostly understood as the learning process termed operant conditioning (learning by reinforcement or punishment), which is associated most notably with the works of behaviorist B.F. Skinner. The processes by which behavior modification occurs were considered a “black box,” as they were deemed unobservable and could therefore not be empirically studied. The cognitive revolution changed this understanding and led to the social cognitive approach to behavior modification.

In contrast to the simplistic stimulus–response view of behaviorists, social cognitive theories of behavior modification assume “that social behaviour is best understood as a function of people’s perceptions of reality, rather than as a function of an objective description of the stimulus environment” (Conner and Norman 2005, p. 5). To illustrate this, imagine elderly persons who have difficulties to walk. The objective description would predict that these difficulties will impair them from getting enough physical activity. The social cognitive approach, in turn, would assume that whether or not these persons find a way to overcome this barrier and engage in physical activity despite their walking impairments (e.g., by doing yoga exercises at home) depends on the persons’ perception of this barrier, for example, on their motivation to exercise and on their belief that they can exercise, even when this is difficult (a behavioral determinant known as self-efficacy, Bandura 1999). Various theories have developed from this social cognitive approach. These theories have identified several behavioral determinants, with the theory of planned behavior (Ajzen 1991) and the health action process approach (Schwarzer 2008) as a classic and a more recent example, respectively.

A further principle of the approach is that the behavioral determinants (or causal processes) are modifiable and that they can be specifically tackled with behavior change techniques (BCTs, Michie et al. 2008). Behavior change techniques are “…an observable, replicable, and irreducible component of an intervention designed to alter or redirect causal processes that regulate behaviour” (Michie et al. 2013, p. 82). There are a great number of BCTs with which researchers and practitioners aim at modifying behavioral determinants and behavior. Recent efforts have been directed at standardizing definitions of BCTs in order to accumulate evidence on their efficacy (Michie et al. 2008, 2013) and to link BCTs to specific behavioral determinants (Abraham 2012). Figure 1 summarizes the elements of theory-based behavior modification.

Behavior Modification, Fig. 1
figure 21figure 21

Schematic display of the principles of behavior modification (Adapted from Michie et al. (2008))

Full size image

Determinants of Behavior Modification

There are essentially two types of social cognitive theories that have been proposed to understand and predict behavior modification: continuum theories and stage theories. Continuum theories assume that persons can be characterized on a continuum from action readiness to actual behavior modification (Abraham 2012). Stage theories, on the other hand, assume that the process of behavior change comprises a discrete number of stages an individual has to pass through to modify a behavior from adoption to long-term behavioral maintenance. A representative of each type of behavior theory is presented next.

A classic continuum model is the theory of planned behavior (Ajzen 1991). At the theory’s core are behavioral intentions. They reflect “… people’s decisions to perform particular actions” (Sheeran 2002, p. 2). The theory of planned behavior predicts that when people form an intention to perform a behavior, they are more likely to carry out this behavior. Behavioral intentions, in turn, are predicted by attitudes toward the behavior (i.e., persons’ valuations of the behavior and its consequences and the expected likelihood that these will occur), the subjective norm (i.e., beliefs about others’ expectations regarding the behavior and a person’s willingness to comply with these), and perceived behavioral control (i.e., beliefs about facilitating and hindering factors and their subjective power to impede behavior performance). The latter is also assumed to have direct effects on behavior (Ajzen 1991). In summary, the theory of planned behavior predicts that persons are most likely to perform or modify a behavior when they are highly motivated (i.e., have strong intentions as determined by their attitudes, subjective norms, and perceived behavioral control) and have high perceived behavioral control. The TPB has been widely researched and proven useful to explain and predict several health behaviors (Conner and Sparks 2005). However, a major criticism is the finding that while intentions are usually well explained by its predictors, behavior is not. This phenomenon is commonly referred to as the intention–behavior gap and indicates that, contrary to many behavior theories’ assumptions, people who have stronger intentions than others are only moderately more likely to perform a specific behavior than others (Sheeran 2002). This has led to fruitful research on self-regulatory (or volitional) factors that may explain how intentions translate into actions (Schwarzer 2008) and how to overcome the gap (Sheeran 2002). Timely health behavior models have incorporated these factors to mediate intention–behavior relations. The health action process approach (HAPA, Schwarzer 2008), for example, includes volitional mediators such as action planning, coping planning, and action control (Sniehotta et al. 2005). According to Schwarzer, the HAPA is a hybrid model, meaning that it can be applied as both a continuum model and a stage theory.

Stage theories have traditionally put more emphasis on factors that can translate intentions into actions than continuum theories. As mentioned above, stage theories assume that the behavior change process can be divided into a fixed sequence of qualitatively distinct stages (or phases). At the core of these theories are transitional variables (e.g., the decision to take action) that “move” persons from one stage into the next (e.g., from the pre-action to the action stage, cf. Schwarzer 2008). Each stage transition is predicted by a specific set of stage determinants that are causally related in some theories (e.g., the HAPA, Schwarzer 2008). Stage theories are commonly considered more comprehensive than continuum theories, but also more complex. One of the earlier and possibly the most prominent stage theory is the transtheoretical model of behavior change (Prochaska and DiClemente 1983). In its most frequently used version, the transtheoretical model proposes five stages of change: precontemplation, contemplation, preparation, action, and maintenance. Stage progression is specified by decisional balance (pros and cons of behavior), self-efficacy (confidence and temptation), and ten processes of change: five cognitive (e.g., consciousness raising) and five behavioral (e.g., stimulus control). While the idea of separating the behavior change process into distinct phases may seem appealing, empirical tests of the transtheoretical model and other stage theories have generally yielded mixed evidence for the distinction of stages (e.g., Sutton 2005). Nevertheless, they remain particularly popular among practitioners, perhaps because of their clear-cut directions for intervention development that is a consequence of the stage assumption: if persons in different stages of behavior change are qualitatively distinct, they would require different BCTs to promote their transitions between the stages. This point will be further touched upon in a subsequent section. But, first, an overview of BCTs is given.

Behavior Change Techniques (BCTs)

There are a vast number of techniques that have been proposed and applied to modify behavior, with the abovementioned operant conditioning being one of the first of what is now termed a BCT. A major challenge in behavior modification research is the fact that the same BCTs are often termed differently by behavior change professionals from different fields. Or the same term is used, but different techniques are meant by it. This consequently limits the potential of intervention research to produce evidence on the effectiveness of specific BCTs. In an effort to address this issue, several research groups have recently focused on creating taxonomies of BCTs with standardized definitions. This work is crucial to building a cumulative science of behavior modification. The most widely accepted, systematic, and comprehensive taxonomy that has emerged from these efforts is the BCT taxonomy v1 by Michie and colleagues (2013). In its first version, the taxonomy comprises 93 distinct BCTs that were collected through extensive reviews of the scientific and applied behavior modification literature from various fields, such as clinical psychology, social psychology, and health psychology. The BCTs are grouped into 16 clusters. These include clusters of social–psychological BCTs, e.g., goals and planning, which comprise BCTs such as problem solving and action planning, or the BCT cluster feedback and monitoring. Other clusters contain BCTs to foster social support, making contextual changes of antecedents (e.g., BCTs restructuring the physical environment and adding objects to the environment) or providing reward and threats.

The taxonomy should potentially be applicable to behavior modification in any field of interest, from clinical to health psychology and pro-environmental behavior modification to changing workplace behavior. There are, however, also behavior-specific taxonomies that may be helpful in providing a subset of these BCTs, as not all of the abovementioned techniques are relevant to all behaviors. This can be especially useful when such taxonomies also include information on the effectiveness of the BCTs for modifying a particular behavior.

The BCT taxonomy v1 was a vital first step toward standardization of BCT research and practice. Still, it can be expected that this taxonomy will further develop in the coming years, as it will be refined by researchers in psychology and other fields as well as practitioners. Another topic that also needs to be addressed is the mapping of BCTs onto specific behavioral determinants, i.e., what BCTs can modify which behavioral determinant. This will help selecting specific BCTs to target behavioral determinants that are particularly important to change a specific behavior.

Selecting BCTs

Social cognitive theories and empirical evidence provide guidance which determinants to focus on to achieve behavior modification. BCT taxonomies and empirical research provide an overview of techniques available to modify behaviors. But behavior change researchers and practitioners also need knowledge about the link of BCTs and behavioral determinants, so BCTs can be specifically selected to tackle the intended determinants. Also, when evaluating behavior change interventions, the mechanisms of the intervention can be ascertained by assessing the behavioral determinants assumed to be modified by the administered BCTs and performing mediation analysis. This not only offers a tool to test social cognitive behavioral theories but can also deliver important information as to why an intervention was successful (what were its active ingredients?) or not (did the intervention fail to enhance the behavioral determinants?). For example, in an intervention study on physical exercise in cardiac rehabilitation patients, two intervention groups were compared to a standard-treatment control group (Scholz et al. 2007). The first intervention group received an action planning intervention, that is, participants were asked to plan when, where, and how to implement their physical exercise. The second intervention group received a combined action plus coping planning intervention. The coping planning part comprised asking participants to think about barriers to their physical exercise and to subsequently form detailed plans, when, where, and how they will overcome these barriers. Two months later, the combined planning group was the most successful in increasing their physical exercise levels. This effect was independent of the age of participants. However, a closer look at the behavioral determinants revealed age-differential effects: older individuals reported the highest levels of coping planning already at the baseline assessment compared to young and middle-aged participants. The latter two age groups increased their coping planning after the intervention whereas the older participants reported relative stability of coping planning across the two months. Self-reported action planning in contrast was not changed by the intervention, nor were there age-differential effects over time (Scholz et al. 2007). Thus, not only analyzing the direct effects of BCTs on behavior modification but also examining the effects on the behavioral determinants provides important information on what the active ingredient of an intervention is and whether or not this applies, for example, to people of all ages.

Unfortunately, with few exceptions (Mosler 2012), behavior theories provide little guidance on which BCTs can modify which behavioral determinants, and empirical research on the BCT–behavioral determinant–behavior link is still rare. There are, however, some expert groups that have proposed links (Michie et al. 2008; Abraham 2012) or are currently working on this. Abraham (2012), for example, provided a menu of 40 BCTs linked to behavioral determinants. In any case, much empirical research is needed to test these proposed links.

When planning an intervention, another question is which behavioral determinant to target. On the one hand, this depends on the goal of the endeavor, e.g., testing a particular theory or the more applied goal of evoking greatest possible behavior modification. It also depends on the theory that the intervention is based on. Continuum theories and stage theories have different implications for selecting behavioral determinants for interventions. As discussed above, continuum theories assume that their behavioral determinants increase the likelihood of people’s behavior performance. This implies that behavior can be modified by changing any of the behavioral determinants of the theory and that this holds for all individuals, wherefore this approach is sometimes termed “one size fits all.” If it is a causal theory, the most distal behavioral determinants should be targeted, as it is assumed that they will work their way through to behavior change by modifying the more proximal behavioral determinants of the theory (Sutton 2008). Alternatively, one could “jump into the causal chain” (Sutton 2008, p. 73) and aim at directly altering proximal determinants. In addition, some approaches, such as intervention mapping (Bartholomew et al. 2011) or the RANAS (Risk, Attitude, Norms, Ability, Self-Regulation) approach (Mosler 2012), suggest procedures to target behavioral determinants that are of particular importance to the target population and context (Mosler 2012; Bartholomew et al. 2011). In the intervention mapping approach, this step is referred to as needs assessment (Bartholomew et al. 2011). It entails a detailed literature review and survey in the target population to carefully adapt the intervention to the context.

A criticism of the one-size-fits-all approach is that individual particularities are not taken into account. Tailored interventions seek to overcome this. They are defined as “…any combination of strategies and information intended to reach one specific person, based on characteristics that are unique to that person, related to the outcome of interest, and derived from an individual assessment” (Kreuter et al. 2000, p. 277). One form of tailoring is stage tailoring. As discussed above, stage theories assume qualitatively distinct behavior change stages. Consequently, different interventions result for persons who are in different stages of change. Following a stage theory approach, the stage of change of each person needs to be assessed beforehand, and everyone receives the intervention that is tailored to their current stage. By stage theories’ rationale, interventions matching individuals’ present stage of change should allow transition to the next stage, whereas mismatched interventions should have nil or possibly adverse effects. The prerequisite to perform stage-tailored intervention is a staging algorithm that can reliably assess individuals’ stage of change prior to the intervention. The transtheoretical model, for example, assumes time-based criteria to determine individuals’ stage of change. This has been frequently criticized, because the time criteria seem arbitrary. More recent theories, such as the HAPA, therefore developed psychological staging algorithms that characterize persons regarding their current intentions and behavior.

An advantage of stage-tailored interventions should be that they take into account the characteristics of the target persons and may therefore be potentially more effective than not-tailored interventions. However, evidence on this is mixed, which may be due to the lack of reliable staging algorithms, and the lack of clear definitions of and evidence for the predictors of each stage transition. In particular, it has been criticized that stage-tailored interventions usually only use few behavioral determinants for assessing stages (Abraham 2008). The menu-based approach, in contrast, considers many social cognitive factors, possibly from a compilation of behavioral theories (Abraham 2008). Each individual’s characteristics are considered, wherefore this approach may lead to a menu of as many BCTs as behavioral determinants that were considered. A disadvantage of this approach is the increased effort and complexity for developing and implementing the great number of interventions required to meet the needs of all participants.

Conclusions and Outlook

In summary, behavior modification is related to behavioral determinants that can be modified by BCTs. In the health behavior modification field, which is of particular importance to the aging population, theory and research on the behavioral determinant–behavior modification link is much advanced. Despite an ongoing discussion whether behavior modification is best understood as a continual process or as divided into discrete stages, the behavioral determinants of importance are now, in principal, understood. A behavior change theory that elegantly incorporates both the continuum and the stages of change approach, and may therefore have gained popularity fast, is the HAPA model (Schwarzer 2008). It specifies causal pathways to behavior change, similar to the theory of planned behavior, but it extends the latter by volitional factors (action planning and coping planning, phase-specific self-efficacy, and action control) in an effort to overcome the intention–behavior gap. Yet, to plan and implement interventions, the HAPA can also be divided into at least three stages: pre-intention/motivation, pre-action/volition, and action.

In contrast to the behavioral determinant–behavior change link, BCTs and their mechanisms to modify behavior require much further research. Important groundwork has been done by producing standardized definitions of the BCTs (Michie et al. 2013) and by hypothesizing their links to behavioral determinants (Abraham 2012). Now, these definitions require large-scale adoption by researchers and practitioners, and empirical investigations need to test the mechanisms by which their BCTs modify behavior.

Regarding environmental factors, the social cognitive approach makes the argument that their influence on behavior is mediated through individuals’ perceptions thereof. However, other fields of research, e.g., environmental psychology, suggest that behavior is best understood as an interaction of person by environment. Following this viewpoint, it could be helpful to consider environmental factors that may hinder or facilitate behavior modification.

Finally, a further line of future behavior change research concerns the mode of delivery of interventions. Whether an intervention is delivered personally (e.g., by a health professional), by mass media (e.g., leaflets or television), and by the Internet or smartphone should make a difference in its efficacy. But little is known on this subject yet. Also, the ideal mode of the delivery could differ for different populations, e.g., for different age groups.

Cross-References