Definition

The theory of planned behavior is a theory used to understand and predict behaviors, which posits that behaviors are immediately determined by behavioral intentions and under certain circumstances, perceived behavioral control. Behavioral intentions are determined by a combination of three factors: attitudes toward the behavior, subjective norms, and perceived behavioral control.

Introduction

An extension of the theory of reasoned action (TRA; Fishbein and Ajzen 1975; Ajzen and Fishbein 1980), the theory of planned behavior (TPB) was developed by Icek Ajzen (1985, 1991) as a general model to predict and explain behavior across a wide range of different types of behaviors. A key assumption in the TRA is that behaviors are under one’s volitional control. However, this assumption is likely to be unrealistic in some contexts, as one can imagine that the volitional control of behaviors may vary across different situations. Recognizing that not all behaviors are under volitional control, Ajzen (1985, 1991) proposed the addition of the construct of perceived behavioral control (i.e., one’s view of their degree of control over a behavior) to the TRA. Because the primary distinction between the TPB and the TRA is the addition of the perceived behavioral control component, most of the key assumptions underlying the TRA also apply to the TPB. Thus, a review of the TRA provides a useful foundation for the discussion of the TPB.

The Theory of Reasoned Action

The creators of the TRA, Martin Fishbein and Icek Ajzen (1975; Ajzen and Fishbein 1980), designed the TRA to predict and understand people’s overt behaviors that are under their volitional control. The TRA is based on the assumption that humans are generally rational beings who make systematic use of the information available to them. Thus, the theory presumes that people’s behaviors are not thoughtless, as they consider the implications of their actions before they decide to engage or not engage in a given behavior. When explaining the TRA, it is useful to begin at the final outcome (i.e., behavior) and work backwards along the causal chain (see Fig. 1 for the TRA pathway).

Fig. 1.
figure 1

The theory of reasoned action

Full size image

Behavior

In the context of the TRA, behavior is defined as an overt action or set of actions that is performed by an individual. Behavior can be construed broadly or narrowly and is a function of the researcher’s theoretical and/or applied objectives (i.e., a researcher chooses the behavior he or she wishes to examine). However, according to the theory, any behavior can be conceptualized in terms of four components: action (i.e., the specific act performed by an individual), target (i.e., who or what the behavior is directed toward), context (i.e., in what situation does the behavior occur), and time (i.e., when the behavior occurs). An example of a broad behavior would be eating (action only), whereas a more specific behavior would be eating (action) chips (target) in the lunchroom (context) at noon (time). Each of the four components of behavior can range substantially, as actions can span from a single action to behavioral categories comprised of multiple actions, targets from a single target to multiple targets, context from a single context to multiple contexts, and time from a single point in time to a broader period of time.

Advocates of the TRA argue that any measurement of behavior must specify the four components of behavior if behavior is to be accurately predicted and understood. Indeed, TRA theorists have noted that researchers sometimes mistakenly measure outcomes believing that they are measuring behaviors, which may lead to inaccurate results. Thus, it is important to distinguish behaviors from outcomes, as outcomes are the possible result of behaviors – not the behaviors themselves. For instance, exercising is a behavior and weight loss is a possible outcome of exercising. However, the outcome of weight loss could also be influenced by other factors such as dieting. Thus, ensuring that the behavior is measured, and not the outcome, increases the accuracy of the prediction of behavior. Behaviors can be measured using behavioral indices generated via observations or self-reports of behaviors, which can be advantageous in terms of ease and resources if researchers are confident in the reports’ accuracy (see Ajzen and Fishbein 1980).

Behavioral Intention

According to the TRA, the most proximal determinant of behavior is behavioral intention, which is an individual’s perceived likelihood that he or she will perform a behavior. This specification of intention as the most immediate determinant of behavior has several implications. First, it suggests that intention is the best predictor of behavior. Second, it posits that any other potential predictors of behavior (e.g., attitudes) do not directly influence behavior. Instead, the relationships of these other factors to behavior must be mediated by intention. Third, it proposes that all behaviors are deliberate, as intentions predicate all behaviors.

The accuracy of intentions in predicting behaviors is a function of two factors. First, prediction accuracy can be increased by ensuring that the four components of behavior (i.e., action, target, context, and time) correspond with the four components of intention. For instance, if the behavior is to eat chips in the lunchroom at noon, then asking an individual’s intention to eat chips in the lunchroom at noon will maximize prediction accuracy. Second, prediction accuracy is a function of the stability of intentions. In general, the shorter the duration between the measurement of an intention and the measurement of its corresponding behavior, the more accurate the prediction of behavior from intention will be. Thus, intentions should be measured as close as possible to behaviors in order to maximize prediction accuracy.

Attitude Toward the Behavior

According to the TRA, there are two immediate determinants of behavioral intention. The first determinant is attitude toward the behavior, which is an individual’s evaluation (i.e., favorability or unfavorability) of performing the behavior in question as defined by the behavior’s four components. There are two key points to note here. First, the TRA’s definition of attitudes is directed toward the behavior – not the object associated with behavior. For example, researchers should measure an individual’s evaluation of him or her eating chips, rather than the individual’s evaluation of the chips themselves. As such, this conceptualization of attitudes is different from the definition of attitudes as it has often been conceptualized in the attitude-behavior consistency literature, which is the positive or negative evaluation of the object itself. Second, this definition of attitudes refers specifically to an individual’s own performance of the behavior, rather than the performance of the behavior in general.

When measuring one’s attitude toward a behavior, the attitude measure must be framed at the specificity of the behavior in order to maximize the prediction accuracy of the attitude on intention. Attitudes toward a behavior can be measured using traditional attitude scales (e.g., Osgood et al. 1957), as long as they are framed at the appropriate specificity of the behavior. However, attitudes toward a behavior are often assessed by measuring one’s behavioral beliefs, as the TRA postulates that one’s attitude toward a behavior is a function of one’s beliefs about the behavior. Attitudes toward a behavior are an aggregate of the beliefs that the behavior leads to certain outcomes and the evaluations of those outcomes. Thus, an attitude toward a behavior is a function of the extent to which each outcome is viewed as positive or negative and the likelihood that each outcome is seen as a result of the behavior. Extremely positive attitudes occur when each outcome is viewed as extremely positive and extremely likely. Likewise, negative attitudes result from sets of beliefs regarding outcomes that are seen as extremely negative and extremely likely. Although one can have many behavioral beliefs, only the salient behavioral beliefs will be the immediate determinants of one’s attitude toward a behavior. Furthermore, like the other direct relations in the TRA, behavioral beliefs and attitudes toward a behavior should be measured on the same four behavioral elements in order to maximize the beliefs-attitudes correlation.

Subjective Norms

The second determinant of behavioral intention is subjective norms, which is a person’s perception that most people who are important to him (or her) think he (or she) should or should not perform the behavior in question. However, this perception may or may not reflect what the important others actually think. Subjective norms are determined by normative beliefs and the motivation to comply with specific referents. Normative beliefs are an individual’s beliefs that a referent thinks he (or she) should (or should not) perform the behavior. Not all referents are equally important; thus, individuals are more likely to comply with more salient referents than less salient referents. Subjective norms can be predicted using an index that multiplies an individual’s normative beliefs regarding a specific referent by the motivations to comply with the same referent, and subsequently, sums the products for all referents. Finally, once again, the four behavioral components of normative beliefs must be matched with those of subjective norms.

Relative Importance of Attitude toward the Behavior and Subjective Norms

Although the TRA assumes that attitudes toward the behavior and subjective norms regarding the behavior both influence intention, the theory does not assume that these constructs will always be in agreement or that the magnitude of their individual contributions to intention will be invariant across different behaviors. When these two determinants are in disagreement, the determinant of intention with the greater weight, as determined by action, target, context, time, and/or individual differences, will be more influential on intention. Thus, when attitudes toward a behavior are viewed as more important than subjective norms for a certain behavior, then attitudes toward the behavior will be more influential on intention than subjective norms. Conversely, when subjective norms are perceived as more important than attitudes, then subjective norms will be more influential than attitudes toward the behavior.

External Variables

As previously noted, the TRA’s core assumption is that behavior can only be influenced by intentions, which in turn are influenced by attitudes toward the behavior and subjective norms. Thus, one somewhat controversial assumption of the TRA is that external variables can only influence behavior indirectly through the influence on behavioral beliefs (which influence attitudes toward the behavior), normative beliefs (which influence subjective norms), and the relative importance of attitudes toward the behavior and subjective norms. There is no consistent pattern of influence for external variables on behavior, as their influence on the determinants of behavioral intention varies across different behaviors. External variables include demographic variables (e.g., age, sex, socioeconomic status, etc.), attitudes toward targets (e.g., people, institutions, etc.), and personality traits (e.g., extraversion, neuroticism, agreeableness, etc.).

The Theory of Planned Behavior

While there has been strong evidence in support of the TRA, many of the studies previously reviewed assumed high levels of volitional control (i.e., that a person can willingly decide to perform or not perform the behavior; see Sheppard et al. 1988; Albarracín et al. 2001; Hagger et al. 2002). In recognition of the fact that volitional control can differ across different behaviors, Ajzen (1985, 1991) proposed the TPB. The TPB retains all the key constructs comprising the TRA, but proposes the addition of a third determinant of intentions: perceived behavioral control. The performance of a behavior depends at least to some degree on one’s actual behavioral control (i.e., the ability to perform a behavior). As the required opportunities and resources to perform a behavior (e.g., time, money, skills, cooperation of others, etc.) increase for an individual, the chances of the individual performing the behavior increases accordingly. Perceived behavioral control refers to people’s perception of the ease or difficulty of performing the behavior of interest.

Perceived behavioral control can influence actual behavior in two ways (see Fig. 2 for the full TPB pathway). First, perceived behavioral control may influence intentions. Individuals who believe they have neither the resources nor opportunities to perform a certain behavior are unlikely to have high intentions to perform the behavior, even if they hold a favorable attitude towards the behavior or if there are favorable subjective norms for the behavior. Therefore, perceived behavioral control should influence intention independent of attitude toward the behavior and subjective norms. Second, perceived behavioral control can potentially influence behavior directly. Under situations where perceived behavioral control is an accurate indication of actual behavioral control, perceived behavioral control can serve as a proxy of actual behavioral control. However, in situations where one’s perceived behavioral control is inaccurate in predicting actual behavior control, then perceived behavioral control would be unlikely to predict actual behavior directly. Perceived behavioral control becomes an inaccurate prediction of actual behavioral control when the individual has little information about the behavior, when requirements or available resources have changed, or when new and unfamiliar elements have entered into the situation. This relationship between perceived behavioral control and actual behavior is indicated by the broken arrow in Fig. 2, as this relationship is only expected to emerge when there is high correspondence between perceived behavioral control and actual behavioral control.

Fig. 2.
figure 2

The theory of planned behavior

Full size image

Similar to attitudes and subjective norms, perceived behavioral control is also determined by its beliefs. Control beliefs are beliefs about the presence or absence of factors that facilitate or hinder performance of a behavior. These beliefs may be predicated on past experiences with the behavior or by observation of others engaging in the behavior. Generally, if an individual believes that he or she has the required resources and opportunities that facilitate, as well as the lack of obstacles that impede the performance of the behavior, he or she should have greater levels of perceived behavioral control. Additionally, like other types of beliefs in the model, external variables can sometimes influence control beliefs depending on the type of behavior.

Evidence for the Theory of Planned Behavior

A number of meta-analyses have provided strong support for the TPB. These reviews examined the TPB across various domains of behavior (Armitage and Conner 2001), including specific areas such as health-related behaviors (Godin and Kok 1996), physical activity (Hagger et al. 2002), and condom use (Albarracín et al. 2001). These meta-analyses found that perceived behavioral control predicted intention above and beyond attitude toward the behavior and subjective norms. Additionally, some of these meta-analyses indicated a direct influence of perceived behavioral control on behavior (e.g., Armitage and Conner 2001; Hagger et al. 2002; McEachan et al. 2011). Furthermore, control beliefs have been found to be predictive of perceived behavioral control (e.g. Ajzen and Madden 1986; Armitage and Conner 1999).

Criticisms of the Theory of Planned Behavior

Although the TPB has received substantial empirical support, the TPB and the TRA have also been criticized. Two criticisms in particular have attracted significant attention in the literature (see Eagly and Chaiken 1993, for a review). The first criticism was one regarding the TRA, which applies to the TPB as well. Some researchers have challenged the TRA’s assumption that the two determinants of intention in the TRA (i.e., attitude toward the behavior and subjective norms) are sufficient to fully account for intention. Though Ajzen (1985, 1991) recognized that perceived behavioral control was another determinant of intention, other researchers have proposed a number of other determinants of intention. Among these proposed determinants are past behaviors/habits (Triandis 1977; Bentler and Speckart 1979; Ouellette and Wood 1998), perceived moral obligation (Schwartz and Tessler 1972), self-identity (Biddle et al. 1987), and affective beliefs (Triandis 1977). Although many of these other determinants have received support (see Conner and Armitage 1998), they have not gained much acceptance with the exception of past behaviors/habits. Strong past behaviors/habits have been found to sometimes outperform intentions in predicting behavior (see Ouellette and Wood 1998; Webb and Sheeran 2006). Additionally, while it may be true that the TPB’s three determinants of intention are strong predictors of intentions for behaviors in general, other determinants may be just as strong or even stronger predictors for certain behavioral domains.

The second criticism of the TPB is that it only applies to comparatively deliberate behaviors because it assumes that behaviors are predominantly a result of rational thought (see Eagly and Chaiken 1993). Thus, the TPB and the TRA disregard spontaneous attitude-to-behavior processes. Recognizing this deficiency in the TPB and TRA, Fazio (1990; Fazio and Towles-Schwen 1999) proposed a model that incorporates both deliberate and spontaneous attitude-to-behavior processes: MODE (M otivation and O pportunity as DE terminants) model. According to this model, deliberate and spontaneous attitude-to-behavior processes are determined by motivation and opportunity (i.e., time and resources). When an individual has both the motivation and opportunity to engage in effortful analysis of a given behavior, he or she will likely undergo a deliberate attitude-to-behavior process. These deliberate processes are presumed to be similar to those described in the TRA and the TPB (see Fazio and Roskos-Ewoldsen 1994; Fazio and Olson 2014).

Alternatively, when an individual lacks either the motivation or the opportunity to engage in effortful analysis of an attitude, they will likely undergo a spontaneous attitude-to-behavior process. A model that maps onto this spontaneous process is Fazio’s (1986) attitude-to-behavior process model. There is no behavioral intention in this model; instead, the activated attitudes (toward the object) and subjective norms guide behavior through the definition of the event. It is this definition of the event that determines the direction and nature of the behavior, as approach behaviors typically follow a positive definition of the event and avoidance behaviors typically follow a negative definition of the event (Fazio 1990). Additionally, the MODE model postulates that these two processes can operate simultaneously. The MODE model has been supported in the literature by a number of studies in domains such as prejudice, relationship outcomes, and self-control behaviors (see Fazio and Olson 2014). Some of these studies have indicated that automatically activated attitudes are better predictors of behavioral outcomes than self-reported attitudes.

Other theorists have also argued for more spontaneous and less deliberative models of behavior. For example, Wood and colleagues (see Wood and Neal 2007; Wood and Neal 2009; Wood et al. 2014) have argued that behavioral processes are often strongly influenced by habits. In their model, contextual cues (e.g., physical locations, other people, internal states, preceding actions in a sequence, etc.) activate an individual’s mental representation of habitual response, which, in turn, influences the habitual response (i.e., behavior). Habits develop slowly through context–response learning, which can be strengthened through frequent context–response pairings, as these frequent pairings lead to a stronger associative link between the memory representation of the contextual cues and that of the response. Frequent pairings of a context with a particular response may also weaken associations between the same context and other responses.

Although habitual response is an automatic process and does not depend on goals for performance (other than when individuals are learning context–response associations; Neal et al. 2006), it can be influenced by controlled processes (see Quinn et al. 2010). When an individual has an intention that is incompatible with a habit, he or she could inhibit the habitual response via vigilant monitoring, which involves heightening attentional focus on a response to ensure that it is not performed. Furthermore, an individual could inhibit the habitual response through stimulus control (i.e., removing oneself from the situation) and/or distraction (i.e., thinking about something else).

More commonly, an individual has an intention that is in agreement with a habit, because many habits originate from past goal pursuits. In these situations, habits may thoughtlessly perform the intended behavior that initially prompted the individual to respond repetitively. Ouellette and Wood (1998) supported this notion, as they found that the frequency of past performance (i.e., habit strength) positively correlated with behavioral intentions for many behaviors (r = 0.43). Moreover, habits have the potential to guide intention, as individuals who lack access to the motivations and cognitions guiding behavior draw inferences about their actions from past behavior (Wood and Neal 2007; Neal et al. 2012). For example, people may reason, “I do this often, so I must like to do it.”

Conclusion

In sum, the theory of planned behavior is a theory used to predict and understand behaviors. It posits that behaviors are immediately determined by behavioral intentions, which in turn are determined by a combination of three factors: attitude toward the behavior, subjective norms, and perceived behavioral control. Moreover, under some circumstances, perceived behavioral control can directly predict behaviors. Though proponents of the theory of planned behavior have proposed that the theory predicts all behaviors, opponents of the theory have argued that the theory only predicts deliberate – not automatic – behaviors.

Cross-References