2 Changing Behavior Using the Theory

of Planned Behavior
Icek Ajzen and Peter Schmidt

Practical Summary

This chapter describes the theory of planned behavior (TPB) and its use as a framework
for behavior change interventions. According to the TPB, human behavior is guided by
three kinds of considerations: beliefs about the likely consequences of my behavior,
beliefs about what important others think I should do, and beliefs about my ability to
carry out the behavior. These beliefs provide the basis for the formation of an intention
to engage in the behavior, but actual performance of the behavior depends on
behavioral control. Interventions are effective to the extent that they produce
changes in the beliefs that underpin intentions and when they ensure that people
have the skills and resources needed to enact their intentions. Research has provided
extensive empirical support showing that the TPB can be used not only to predict and
explain behavior but also to help design effective behavior change interventions.

Prochaska & DiClemente, 1983, 1992; see Chapter

2.1 Introduction
10, this volume). In this “transtheoretical model of
Human behavior is determined by multiple factors change,” it is assumed that different psychological
and changing well-established behavioral patterns processes are involved when people move from one
is notoriously difficult. From a purely theoretical stage to the next and that, therefore, different inter-
perspective, the problem is easily conceptualized. vention strategies are required at different stages of
To adopt a novel behavior, people must be moti- the change sequence. Although of possible heuristic
vated to do so and they must be capable of acting value, a moment’s reflection reveals that the same
on their motivation. The motivation for change intervention can influence people at different stages
may be self-generated, as when a person decides of change (Fishbein & Ajzen, 2010). For example,
to go on a weight-loss diet, stop smoking, or a message to the effect that taking an aspirin pill
avoid procrastinating. In these instances, behavior every day can prevent a heart attack may prompt
change interventions must focus on ways and people at the precontemplation stage to contemplate
means to help people initiate and maintain the engaging in this behavior, it may motivate people at
desired behavior. In other situations, however, the contemplation stage to make preparations to take
people are not self-motivated and behavior change a daily aspirin pill, it may get people at the prepara-
interventions must first generate the motivation for tion stage to actually start taking a daily aspirin pill,
change, followed by a focus on implementation. and it may encourage people at the action stage to
Behavior change is often said to proceed in stages, maintain this behavior over time. Consistent with
from precontemplation through contemplation, pre-
paration, action, and maintenance (Bamberg, 2013; https://doi.org/10.1017/97811086773180.002

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 18 ICEK AJZEN AND PETER SCHMIDT

this argument, empirical research (Armitage & et al., 2010) showed that interventions based on
Arden, 2002) has shown that the consecutive stages a theoretical framework were more effective in
in the transtheoretical model correspond to increas- changing health-related behavior than nontheory-
ing levels of behavioral intentions. based interventions. The three most frequently
The approach to behavior change described in used theoretical frameworks were the previously
the present chapter is based on the TPB (Ajzen, mentioned transtheoretical model (see Chapter
1991, 2012a). This theory has been used success- 10, this volume), Bandura’s (1977) social cogni-
fully to explain and predict behavior in a multitude tive theory (see Chapter 3, this volume), and the
of behavioral domains, from physical activity to TPB. In a comparison of research with these three
drug use, from recycling to choice of travel theories, interventions based on the TPB had, on
mode, from safer sex to consumer behavior, to average, the strongest impact (effect size r = 0.36)
name just a few (for meta-analytic syntheses of followed by the transtheoretical model (r = 0.20)
this research, see, e.g., Albarracín, Fishbein, and social cognitive theory (r = 0.15).
& Goldestein de Muchinik, 1997; Armitage & This chapter aims to provide an introduction to
Conner, 1999; Hagger, Chatzisarantis, & Biddle, the TPB as a basis for behavior change interven-
2002; McDermott et al., 2015; Riebl et al., 2015; tions. A brief description of the theory and its
Sheeran & Taylor, 1999; Winkelnkemper, Ajzen, & implications for behavior change is followed by
Schmidt, 2019). The TPB is also increasingly being an in-depth discussion of the practical steps
used as a framework for designing and evaluating involved in designing a promising TPB-based
the effects of behavior change interventions (for intervention and evaluating its effectiveness.
meta-analytic syntheses of intervention studies,
see Sheeran et al., 2016; Steinmetz et al., 2016;
2.2 Brief Overview of the Theory of
Tyson, Covey, & Rosenthal, 2014). Instead of
Planned Behavior (TPB)
describing behavior change as moving through
five distinct stages, the TPB focuses on the two As extensive discussions of the TPB and of
basic facets of change mentioned at the beginning empirical evidence in support of the theory can
of this section: motivating people to engage in be found elsewhere (e.g., Ajzen, 2012a; Fishbein
a desired behavior and, to the extent that engaging & Ajzen, 2010), the following section provides
in the behavior poses difficulties, enabling them to only a brief overview while taking into account
realize the behavior. As discussed in greater detail recent developments of the theory. The structural
in the next sections, the first facet involves forming TPB model as currently conceptualized is shown
an intention to perform the behavior of interest, in Figure 2.1.
while the second facet has to do with the relation
between stated intentions and actual behavior.
2.2.1 Definition of the Behavior
Practitioners have long recognized the value of
a well-formulated, empirically validated theory The TPB starts with a clear definition of the beha-
for designing and evaluating the effects of beha- vior of interest in terms of its target, the action
vior change interventions (see Bickman, 1987; involved, the context in which it occurs, and the
Coryn et al., 2011; Michie & Prestwich, 2010; time frame. Each of these elements can be defined
Rossi, Lipsey, & Freeman, 2004). As Kurt Lewin at varying levels of specificity or generality.
(1952) famously observed, “there is nothing more However, once the behavior has been defined, all
practical than a good theory” (p. 169). Consistent other constructs in the theory must correspond to
with this reasoning, a meta-analysis of behavior the behavior in all four elements. This is known as
change interventions delivered online (Webb the principle of compatibility (Ajzen, 1988). For

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 19

Attitude
Behavioral toward the
beliefs behavior

Intention Behavior

Normative Subjective
beliefs norm

Perceived
Control Actual
behavioral
beliefs behavioral
control
control

Figure 2.1 Theory of planned behavior

example, to study entrepreneurship, an investiga- unanticipated events; insufficient time, money, or

tor may define the behavior of interest at a low resources; lack of requisite skills; and a multitude
level of generality, such as “opening (action) of other factors may prevent people from acting on
a restaurant (target) in Chicago (context) in the their intentions. The degree to which people have
next twelve months (time frame).” Alternatively, control over the behavior depends on their ability
the investigator may be interested in entrepreneur- to overcome barriers of this kind and on the pre-
ship at a more general level and define the beha- sence of such facilitating factors as past experience
vior as “starting (action) a business (target) in the and assistance provided by others. In light of these
next twelve months (time).” Note that the target considerations, the TPB postulates that the degree
has been expanded to include any type of business, of behavioral control moderates the effect of inten-
not just a restaurant, and that the context is left tion on behavior: The greater the actor’s control
unspecified. The particular behavioral definition over the behavior, the more likely it is that the
adopted determines how all constructs in the intention will be carried out.
TPB are to be formulated. The TPB thus contends that the performance of
socially significant behaviors is volitional, the
result of an intention, not capricious or performed
2.2.2 Intentions and Behavioral
automatically without conscious awareness. To be
Control
sure, when behavior has become routine as a result
As in other “reasoned action” approaches (e.g., of repeated performance, intentions may be acti-
Bandura, 1997; Fisher & Fisher, 1992; Triandis, vated automatically and remain implicit unless
1972; see also Fishbein & Ajzen, 2010), the effortfully retrieved (Ajzen & Dasgupta, 2015;
immediate antecedent of behavior in the TPB is see also Chapter 13, this volume). Even strong
the intention to perform the behavior in question; habits are not necessarily unintentional, although
the stronger the intention, the more likely it is that habitual behavior may occur spontaneously, with-
the behavior will follow. To return to the example out a conscious intention (see Ajzen & Fishbein,
in Section 2.2.1, the intention to start a business in 2000 and Chapter 41, this volume). Consider, for
the next twelve months could be measured and data example, individuals who have been using their
collected to determine whether participants did or cars to commute to work for many years. As they
did not implement their intentions. However, get ready to leave, they may proceed without

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 20 ICEK AJZEN AND PETER SCHMIDT

forming a conscious intention to drive their cars to (the experience). In their aggregate, behavioral
work. However, if asked, they could easily retrieve beliefs are theorized to produce a positive or
and report their (implicit) intention. negative attitude toward the behavior.

2.2.3.2 Normative Beliefs and Subjective Norm

2.2.3 Determinants of Intentions Two types of normative belief can be distinguished:
Consistent with the notion of reasoned action, injunctive and descriptive (see Cialdini & Trost,
people’s behavioral intentions are assumed to be 1998; Fishbein & Ajzen, 2010). An injunctive nor-
guided by some measure of deliberation, where mative belief is the expectation or subjective prob-
novel behaviors and important decisions receive ability that a given referent individual or group
more thorough contemplation than relatively less (e.g., friends, family, spouse, coworkers, one’s phy-
important or routine behaviors (Ajzen & Sexton, sician or supervisor) approves or disapproves of
1999). According to the TPB, three kinds of con- performing the behavior under consideration.
siderations guide the formation of intentions: Descriptive normative beliefs, on the other hand,
beliefs about the likely consequences and experi- are beliefs as to whether important others them-
ences resulting from performance of the behavior selves perform the behavior. Both types of beliefs
(behavioral beliefs), which, in their aggregate, contribute to the overall perceived social pressure to
result in the formation of an attitude toward the engage in the behavior or subjective norm.
behavior; beliefs about the expectations and 2.2.3.3 Control Beliefs and Perceived
behaviors of significant social referents (norma- Behavioral Control
tive beliefs), which produce perceived social Control beliefs are concerned with the presence of
pressure to engage or not to engage in the beha- factors that can facilitate or impede performance of
vior, or subjective norm; and beliefs about factors the behavior. Control factors include required skills
that may facilitate or impede performance of the and abilities; availability or lack of time, money, and
behavior (control beliefs), which result in per- other resources; cooperation by other people; and so
ceived behavioral control or a sense of self- forth. A control belief is defined as a person’s sub-
efficacy (Bandura, 1997). It is assumed that the jective probability that a given facilitating or inhibit-
behavioral, normative, and control beliefs that are ing factor will be present in the situation of interest.
readily accessible in memory are the prevailing In their aggregate, readily accessible control beliefs
determinants, respectively, of attitude, subjective produce the prevailing perceived behavioral control.
norm, and perceived behavioral control and that Empirical research has identified two subfactors of
these latter variables influence the behavioral control: capacity and autonomy (see Fishbein &
intention (see Figure 2.1). These processes are Ajzen, 2010). Capacity refers to the perceived abil-
described in greater detail in Section 2.2.4, ity to perform the behavior, as determined by the
which deals with the structural model. assumed availability of requisite skills and
resources, whereas autonomy is the extent to
2.2.3.1 Behavioral Beliefs and Attitudes which people believe that the decision to perform
A behavioral belief is the person’s subjective the behavior is entirely up to them.
probability that performing a behavior of interest
will lead to a certain outcome (instrumental beha- 2.2.4 The Structural Model
vioral beliefs) or involve a certain experience
(experiential behavioral beliefs), for example In most applications of the TPB, the three pre-
the belief that exercising (the behavior) improves dictors of intention (attitude, subjective norm,
physical fitness (the outcome) or is invigorating and perceived behavioral control) have been

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 21

treated as additive factors, although, in the origi- more proximal antecedents of behavior specified in
nal formulation of the theory, Ajzen (1985) dis- the theory. A behavior change intervention can be
cussed the possibility of an interaction between viewed as such a background factor and, like other
attitude and subjective norm with perceived beha- background factors, its impact is expected to be
vioral control. In its current formulation, favor- mediated by the proximal determinants of intentions
able attitudes and supportive subjective norms and behavior. This conception is discussed in greater
motivate people to perform the behavior, but detail in the next section.
this motivation leads them to form an intention
to engage in the behavior only to the extent that
2.3 Using the TPB to Change
they believe that they are capable of performing
Behavior
the behavior in question. This implies that per-
ceived behavioral control moderates the effects of As noted, an intervention can be designed to moti-
attitude and of subjective norm on intention (see vate people to engage in a desired behavior or, if
Figure 2.1). A number of studies provide empiri- they already intend to do so, to help them carry out
cal evidence in support of the proposed interac- their intentions successfully. A motivational inter-
tion effects (see Hukkelberg, Hagtvet, & Kovac, vention based on the TPB can achieve its aim in at
2014; La Barbera & Ajzen, 2018; Yzer & van den least three ways. First, it can influence behavioral or
Putte, 2014). normative beliefs leading, respectively, to a more
It can be seen that behavioral control plays favorable attitude toward the behavior or a more
a central role in the theory’s current formulation. supportive subjective norm, thereby raising motiva-
First, the degree of actual control moderates the tion to engage in the behavior. Many studies have
effect of intention on behavior and, second, per- provided support for this proposition (e.g.,
ceived behavioral control moderates the effects of Bamberg, 2006; Bamberg & Schmidt, 2001).
attitude and of subjective norm on intention. Second, the intervention can increase perceived
Moreover, because it is usually difficult to know behavioral control and thus elevate the effects of
how much control people actually have in a given attitude and/or subjective norm on intention to per-
situation, perceived behavioral control is also form the behavior. Research has indeed demon-
used as a proxy for actual control. Of course, strated that interventions can raise perceived
perceived behavioral control can usefully serve behavioral control and thereby influence intention
as a proxy for actual control only to the extent that and behavior (e.g., Kelley & Abraham, 2004;
it is veridical, that is, to the extent that it reflects Stecker, McGovern, & Herr, 2012), but the mod-
actual control reasonably well. erating effect of perceived behavioral control
has not been studied directly. Third, an interven-
tion can change the relative importance of atti-
2.2.5 Intervention as a Background
tude and subjective norm as determinants of
Factor
intention. Thus, for example, if the behavior of
Many factors not included in the TPB may influence interest was largely under attitudinal control, an
intentions and behavior, including demographic intervention targeted at normative beliefs may
characteristics (age, gender, race, education, raise awareness of social norms and increase
income, etc.), personality traits, life values, political their influence on intention. To the best of our
ideology, mood and emotions, and so forth. In the knowledge, this proposition has not been tested.
TPB, these kinds of variables are considered back- Note that changing one or two behavioral, nor-
ground factors that have no direct effects on beha- mative, or control beliefs may not be sufficient to
vior but can influence it indirectly by way of the bring about a significant change in the overall

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 22 ICEK AJZEN AND PETER SCHMIDT

attitude, subjective norm, or perceived behavioral following section outlines means to evaluate its
control. Moreover, a change in one belief may be effectiveness.
accompanied by an unexpected countervailing As noted, to be effective, interventions must
change in another belief, thereby neutralizing the influence the beliefs that are readily accessible at
impact of the intervention. Only if the intervention the time that behavioral performance is contem-
is effective in changing the total composite of plated. However, although the TPB can help iden-
behavioral, normative, or control beliefs can tify the most likely targets for an intervention, it
change be expected in attitude, subjective norm, does not prescribe how the desired changes in
or perceived behavioral control, respectively. beliefs are to be brought about. Investigators have
An intervention designed to enable people to to rely on auxiliary hypotheses or other considera-
realize existing favorable intentions can achieve its tions to decide on the particular strategy that may be
aim in two ways: (1) it can improve actual control most useful for their purposes (see Trafimow,
over the behavior, allowing people to act on their 2012). Michie, van Stralen, and West (2011) sug-
intentions, and (2) it can raise perceived behavioral gested a behavior change wheel that delineates
control, which may lead people to persevere in the possible methods of influence: education (increas-
face of difficulties. For example, individuals who ing knowledge or understanding), incentivization
are motivated and intend to engage in a weight-loss (creating expectation of reward), coercion (creating
diet can be provided with a list of recommended expectation of punishment or cost), training
food items and taught how to prepare them. An (imparting skills), restriction (using rules or laws
increase in actual control resulting from the instruc- to regulate behavior), environmental restructuring
tions on how to prepare the food items may help (changing the physical or social context), and
them successfully implement the weight-loss diet. enablement (increasing resources or removing bar-
In addition, an increase in perceived behavioral riers). However, perhaps the most frequently used
control may help them persevere if their initial strategy in the behavior change wheel is persuasive
attempts are less than successful. Greater perceived communication, a topic of long-standing interest in
behavioral control can exert this effect by increasing social psychology (Ajzen, 2012b; Hovland, Janis,
the effect of attitude or subjective norm on intention, & Kelley, 1953; for in-depth discussions, see
thus strengthening the intention to adhere to the Dillard & Shen, 2012). Persuasive messages impart
weight-loss diet (see Figure 2.1). new information in the context of one-on-one
encounters, in group discussions and workshops,
or in mass media campaigns, such as public service
2.4 TPB-Based Interventions:
announcements. No matter which strategy is
Practical Considerations
employed, a feature common to all is that they
Some investigators have used the TPB to help expose people to new information designed to
design their intervention as well as to evaluate its effect a change in behavior.
effects (e.g., Norman, et al., 2018; Stecker,
McGovern, & Herr, 2012), while others have
2.4.1 Preliminary Considerations
implemented an intervention that was not based
on the TPB but then used the TPB as a basis for At the risk of belaboring the obvious, the first step
evaluating the intervention by examining its effects before designing an intervention is to ascertain
on some or all of the theory’s constructs (e.g., Macy how many people in the target population already
et al., 2012). This section outlines the practical steps perform the behavior of interest. To take an exam-
involved in using the TPB as a framework for ple, the behavior may be defined as eating a low-
designing behavior change interventions, and the calorie diet, a criterion that specifies the action

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 23

(eating) and target (low-calorie diet) but gener- TPB makes it clear that considerable formative
alizes across context and time. If such a diet is research is required to ensure an effective inter-
recommended after a heart attack, it may be found vention. This formative research involves sev-
that, in a target population of individuals who eral consecutive steps (see Fishbein & Ajzen,
have experienced a heart attack, virtually every- 2010, appendix).
body has already adopted a low-calorie diet and
2.4.2.1 Belief Elicitation
no intervention is required. Alternatively, it may
be found that a subset of individuals has not yet The first step is the elicitation of readily accessible
adopted such a diet and an intervention could be behavioral, normative, and control beliefs in
targeted at these individuals. a representative sample from the target population.
If it is found that a sufficient proportion of This is typically done in a small-group session but
individuals in the target population fail to perform can also be done online (see Meitinger & Behr,
the behavior of interest, a second important con- 2016). Sample participants are given a few minutes
sideration has to do with the reason for the failure. to list (1) the outcomes and experiences they associ-
One possibility is that they are not sufficiently ate with the behavior of interest; (2) the individuals
motivated and currently do not intend to perform or groups who approve or disapprove of performing
the behavior. This situation would require an inter- the behavior and who themselves perform or refuse
vention directed at the motivational component. to perform the behavior; and (3) the factors that
Another possibility is that people are motivated may facilitate or interfere with performance of the
but find it difficult or impossible to carry out their behavior. Responses are content-analyzed and a list
favorable intentions, a situation that requires of the most frequently mentioned behavioral, nor-
a different intervention, one designed to increase mative, and control beliefs is created. A suitable
control over the behavior. The need to examine criterion might be the requirement that a belief
which of these two possibilities explains failure to selected for inclusion must have been listed by at
perform the desired behavior may appear self- least 25–30 percent of the participants in the elicita-
evident but it is not always implemented. For tion study or that the set of beliefs selected contains
example, Sniehotta (2009) used TPB-based inter- at least 50–60 percent of the total number of beliefs
ventions to motivate students to take advantage of listed by all participants.
underused university sports facilities, with little 2.4.2.2 Administration of a Pilot
success. Examination of the findings suggests Questionnaire
that most students already held quite favorable
Items that can serve as direct measures (reflective
attitudes and supportive subjective norms with
indicators) of attitude, subjective norm, perceived
respect to using the sports facilities, that is, they
behavioral control, and intention are formulated.
were generally motivated to engage in this beha-
When behavior is a self-report, behavioral items
vior. The reason they did not do so had mostly to
must also be devised. These items, together with
do with relatively low perceived behavioral con-
the identified behavioral, normative, and control
trol. Although strengthening motivation further
beliefs, are used to construct a pilot questionnaire
could have had some effect, a more promising
that is administered to a new sample from the
approach would have focused on issues of control.
research population.

2.4.2.3 Finalizing the TPB Instrument

2.4.2 Formative Research
As is discussed in greater detail in Sidebars 2.1 and
All too often, interventions are attempted with- 2.2, the data from the pilot questionnaire are
out sufficient preparatory work. Applying the analyzed to establish the reliability, validity, and

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 24 ICEK AJZEN AND PETER SCHMIDT

Sidebar 2.1 Testing the measurement model of the theory of planned behavior

Past research with the TPB has typically settled for reporting the internal consistency
among items designed to reflect the attitude, subjective norm, perceived behavioral
control, and intention constructs directly, usually by means of Cronbach’s α coefficient.
However, beyond internal consistency, it is also important for researchers to examine
the discriminant validity of the items that were used to assess the TPB’s constructs (see
Campbell & Fiske, 1959). The preferred method for evaluating the measurement
model of the TPB is confirmatory factor analysis (CFA). This analysis can confirm that
each item included in the questionnaire contributes adequately to the TPB construct it
is meant to assess (known as convergent validity) and that it does not relate to
measures of the theory’s other constructs (known as discriminant validity). In studies
adopting a true experimental design, it is not necessary to test the invariance of factor
loadings in the experimental and control groups because the random assignment
should account for any potential variations in the equivalence of measures. In the case
of quasi-experimental designs, researchers should test the model using CFA in both
experimental and control groups and formally compare their structures (known as
multigroup analysis) in order to confirm that the psychometric properties of the
measures of theory constructs are consistent across experimental and control groups.
These analyses can also be used to support the validity of the TPB measures across
different populations or behaviors or across two or more time points (Brown, 2015).

dimensionality of the reflective indicators; to test intervention. Several general criteria can be
the relations between behavioral beliefs and atti- suggested.
tudes, between normative beliefs and subjective
1. Examination of mean levels of attitude, subjective
norms, and between control beliefs and perceived norm, and perceived behavioral control measures
behavioral control; and to test the structural TPB in the population of interest can reveal which of
model (Figure 2.1). The results are used (1) to refine these factors is already favorable and which may be
the direct measures of the theory’s constructs; (2) to amendable to strengthening. For example, it may
validate the behavioral, normative, and control emerge that attitudes are highly favorable and that
beliefs as predictors of attitude, subjective norm, subjective norms are also supportive but that per-
and perceived behavioral control, respectively; and ceived behavioral control is quite low. In this case,
(3) to make sure that the theory holds up in terms of the intervention would best be directed at control
predicting intentions and behavior (e.g., de Leeuw beliefs in an effort to raise people’s perceived
et al., 2015). Assuming the results are satisfactory, behavioral control (see Elliott & Armitage, 2009).
2. Testing the structural model (see Sidebar 2.2)
a final TPB questionnaire is formulated (for
will provide information about the relative
a sample questionnaire, see Ajzen, 2019; Fishbein
importance of attitude and subjective norm, and
& Ajzen, 2010, appendix). of the moderating effect of perceived behavioral
control, in the prediction of intentions. All
2.4.2.4 Selecting Target Beliefs
else equal, the intervention should target the
The next preparatory step is to use the pilot data more important of these components. Thus, if
to select the behavioral, normative, and/or con- subjective norms are found to be more important
trol beliefs that are to be targeted in the in predicting intentions than attitudes, the

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 25

Sidebar 2.2 Testing the structural model of the theory of planned behavior

In past research, the TPB model has often been tested by means of multiple regression
analyses in which intentions were regressed on attitude, subjective norm, and perceived
behavioral control and behavior was regressed on intention and perceived behavioral
control. With the widespread adoption of structural equation modeling, it has become
possible to obtain a structural test of the overall fit between the model and the data and
to simultaneously test hypotheses derived from the theory, such as mediation and
moderation, while taking into account errors in measurement (Rodgers, 2010). One way
to test the structural model in interventions or experiments involves including the
intervention as a predictor in the model, as a “dummy-coded” variable (e.g., intervention
group coded as “1” and the control or comparison group coded “0”). The intervention is
then modeled to influence directly the behavioral, normative, and control beliefs and to
have no direct effect on attitude, subjective norm, perceived behavioral control,
intention, and behavior (for illustrations, see Bamberg, 2006; Stark, Berger, & Hössinger,
2018). A second approach relies on multiple group structural equation modeling to
compare theory effects across the experimental and control groups by testing the same
model over the two groups and also, for example, the equality of certain cross-group
constraints, such as equal regression coefficients. This method also allows testing for
moderation effects (Hancock, 2004). A poor fit may be an indication of methodological
issues, such as inadequate operationalization of the theory’s constructs or low variance in
intentions and/or behavior, but it may also indicate lack of full support for the
hypothesized mediation and moderation effects specified in the TPB. At a more general
level, it may raise questions about the TPB’s applicability for the population or behavior of
interest.

intervention could attempt to make subjective employed in which participants are randomly
norms more supportive of the behavior. assigned into experimental and control groups,
3. Examination of the individual behavioral, nor- the intervention is administered in the experimen-
mative, and control beliefs can offer insight into tal group, and its impact is evaluated by compar-
the particular beliefs that need to be changed or ing this group to a control group that is not
new beliefs that could be introduced.
exposed to the intervention (see Shadish, Cook,
2.4.2.5 Pilot-Testing the Intervention & Campbell, 2002; Chapter 20, this volume).
Finally, pilot work is required to make sure that A questionnaire assessing the TPB variables is
the intervention to be used is capable of changing administered in each group following the inter-
the targeted beliefs and that it does not adversely vention and, if long-term effects are to be exam-
affect beliefs not directly targeted. ined, it can be administered again at one or more
later points in time. The impact of the interven-
tion is analyzed by comparing means of the TPB
2.5 Evaluating an Intervention’s variables in the two groups (see Green &
Effectiveness Thompson, 2015; Hancock, 2004).
In the preferred method for examining an inter- The present section describes how, using an
vention’s impact, an experimental design is experimental design, intervention effectiveness is

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 26 ICEK AJZEN AND PETER SCHMIDT

evaluated in the context of the TPB. Before this, of effects to reach the final destination of a change
however, it should be noted that interventions can in behavior. Our analysis also indicates that it is
also rely on quasi-experimental designs without not sufficient to demonstrate the effect of an
random assignment into experimental and control intervention on intentions because many factors
groups, such as the regression discontinuity design can weaken the link between intention and beha-
(Angrist & Pischke, 2015; Shadish et al., 2002). vior or prevent people from carrying out their
Such designs can also allow strong inferences to intentions (for discussions of the intention-
be made. For example, using a quasi-experimental behavior gap, see Fishbein & Ajzen, 2010; and
design, an intervention to encourage the use of Chapter 6, this volume).
public transportation (Bamberg & Schmidt,
1999) assessed the intervention’s outcome repeat-
2.5.2 Empirical Support
edly, showing that the initially observed increase
in the use of public transportation remained stable It is beyond the scope of this chapter to review the
over a three-year period. sizable body of research that has applied the TPB in
efforts to modify behavior. A meta-analytic synth-
esis of 123 behavior change interventions across
2.5.1 Testing the Impact of a TPB
a variety of behavioral domains (Steinmetz et al.,
Intervention
2016) confirms that these efforts are generally quite
According to the TPB, an intervention’s impact successful. As can be seen in Table 2.1, with the
on behavior is contingent on a sequence of effects exception of behavioral beliefs, the interventions
linking changes in beliefs to changes in the target had, on average, statistically significant effects on
behavior (see Figure 2.1). This view has impor- the theoretical antecedents of intentions. As a result,
tant implications for the amount of change in they also had a considerable impact on intentions
behavior that can be expected. Even under the and behavior, with effect sizes of 0.34 and 0.50,
best of circumstances, each link in the chain of respectively. Note, however, that the amount of
hypothesized effects is far from perfect. It follows change varied widely across interventions, as is
that a given amount of change in behavioral, shown by the significant heterogeneity in effect
normative, or control beliefs is likely to produce sizes across studies (see Cochran’s Q coefficients
a smaller change in attitude, subjective norm, or in Table 2.1). This suggests that a significant pro-
perceived behavioral control. Similarly, any portion of the variance in effect sizes is attributable
changes in attitude, subjective norm, or perceived to extraneous factors that moderate the impact of
behavioral control are likely to produce smaller the interventions. Indeed, the meta-analysis estab-
changes in intention, and any change in intention lished that group-based interventions were more
will have diminishing returns in regard to beha- effective than interventions administered in indivi-
vior. This latter phenomenon was confirmed in dual settings and that public interventions had
a meta-analysis of intervention studies (Webb & a stronger impact than private interventions.
Sheeran, 2006), which found that successful Additional research is needed to identify other
interventions produced large changes in inten- potential moderators that may influence an inter-
tions (mean effect size d = 0.66) but much smaller vention’s effectiveness.
changes (d = 0.36) in behavior. It follows that, to A study by Norman et al. (2018) will serve to
have a significant impact on behavior, the inter- illustrate the successful application of the TPB to
vention must be powerful enough to produce design and evaluate a behavior change intervention.
substantial changes in the targeted beliefs, such The investigators tried to discourage excessive con-
that its influence can reverberate along the chain sumption of alcohol (binge drinking) among

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 27

Table 2.1 Main effects of theory of planned behavior interventions (from Steinmetz
et al., 2016)

TPB variables k (ES) N d Q

Behavioral beliefs 11 (14) 2,902 0.39 120.76*

Normative beliefs 12 (17) 3,962 0.54* 130.64*
Control beliefs 6 (6) 1,053 0.68* 59.43*
Attitude 70 (101) 21,374 0.24* 397.80*
Subjective norm 61 (85) 15,711 0.14* 157.77*
Perceived behavioral control 80 (113) 24,897 0.26* 623.12*
Intention 72 (108) 22,030 0.34* 650.21*
Behavior 40 (51) 9,395 0.50* 445.80*

Note. k = number of studies; ES = number of effect sizes; N = sample size; d = weighted

mean effect size; Q = test of heterogeneity. *p < 0.05.

incoming university students. Prior research had not perceived agency), and lower intentions to
identified three prominent beliefs related to binge binge drink. The intervention also reduced the
drinking: that it is fun, that it has a negative impact amount of alcohol consumed and, although the
on studies, and that binge drinking by friends effect on behavior diminished over time, it was
encourages this behavior. Based on extensive for- still significant at the six-month follow-up. Finally,
mative research, a message addressing these three consistent with the TPB, a mediation analysis
beliefs was formulated and delivered online. The showed that the effect of the intervention on beha-
message explained that binge drinking has vior was fully mediated by the TPB variables.
a negative impact on academic outcomes, that it is
possible to have fun in other ways, and that most
2.6 Summary and Conclusions
students do not binge drink on a regular basis. The
message also mentioned the financial cost of heavy It can be difficult to bring about new patterns of
drinking. Three weeks prior to entering the univer- behavior, and interventions intended to do so can
sity, participants reported their alcohol consump- be costly and time-consuming; their design and
tion, they were or were not exposed to the implementation should not be left to intuition.
persuasive message, and they then completed Reliance on the TPB can provide information
a TPB questionnaire with regard to binge drinking. about the factors that motivate a behavior of inter-
Alcohol consumption was assessed again at one est and about the resources and barriers that can
week, one month, and six months after starting influence the adoption of a new course of action.
university and the TPB measures were reassessed According to the TPB, an intervention must first
at one and six months. and foremost be designed to influence behavioral,
Comparison of the intervention condition to the normative, and/or control beliefs regarding the
no-message control group showed that the inter- advocated behavior. To the extent that the inter-
vention had its intended effects on the TPB vari- vention manages to produce substantial changes in
ables: it influenced the targeted beliefs in the these beliefs, corresponding changes in attitudes
advocated direction and it produced less favorable and subjective norms can be expected, changes
experiential and instrumental attitudes toward that raise motivation to perform the behavior. In
binge drinking, lower injunctive and descriptive turn, increased motivation will produce an inten-
subjective norms, lower perceived capacity (but tion to engage in the behavior to the extent that

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 28 ICEK AJZEN AND PETER SCHMIDT

perceived behavioral control is sufficiently strong. Ajzen, I. (2012b). Attitudes and persuasion. In K. Deaux
Finally, the favorable intention will be implemen- & M. Snyder (Eds.), The Oxford Handbook of
ted to the extent that the actor has volitional Personality and Social Psychology (pp. 367–393).
control over the behavior in question. New York: Oxford University Press. https://doi.org
/10.1093/oxfordhb/9780195398991.001.0001
The TPB provides a general framework for plan-
Ajzen, I. (2019). Sample TPB Questionnaire. https://
ning a behavior change intervention, but there is
people.umass.edu/aizen/pdf/tpb.questionnaire.pdf
plenty of room for the investigator’s imagination
Ajzen, I., & Dasgupta, N. (2015). Explicit and
and creativity. If the intervention is to increase implicit beliefs, attitudes, and intentions: The
motivation to perform a recommended behavior, role of conscious and unconscious processes in
the TPB can help to identify the behavioral, norma- human behavior. In P. Haggard & B. Eitam
tive, and/or control beliefs that should be targeted. (Eds.), The Sense of Agency (pp. 115–144).
On the other hand, if people are already motivated New York: Oxford University Press. https://doi
and intend to change their behavior, the theory can .org/10.1093/acprof:oso/9780190267278
direct the design of an intervention that will enable .001.0001
them to carry out their intentions. In either case, it is Ajzen, I., & Fishbein, M. (2000). Attitudes and the
up to the investigator to design an effective inter- attitude-behavior relation: Reasoned and automatic
processes. In W. Stroebe & M. Hewstone (Eds.),
vention strategy that can produce the desired
European Review of Social Psychology, Vol. 11
changes in the targeted variables. These are the
(pp. 1–33). Chichester: Wiley. https://doi.org/10
action hypotheses not contained in the theory itself
.1080/14792779943000116
(Chen, 2014). They can be derived from other Ajzen, I., & Sexton, J. (1999). Depth of processing,
theories, from the results of past interventions, or belief congruence, and attitude-behavior
by consulting practitioners and stakeholders, as correspondence. In S. Chaiken & Y. Trope (Eds.),
proposed by Lewin (1946) and Campbell (1969). Dual-Process Theories in Social Psychology (pp.
Adequate formative research guided by the TPB 117–138). New York: Guilford Press.
that incorporates these action hypotheses can Albarracín, D., Fishbein, M., & Goldestein de
increase the likelihood of a successful behavior Muchinik, E. (1997). Seeking social support in old
change intervention. age as reasoned action: Structural and volitional
determinants in a middle-aged sample of
Argentinean women. Journal of Applied Social
References Psychology, 27, 463–476. https://doi.org/10.1111
Ajzen, I. (1985). From intentions to actions: A theory of /j.1559-1816.1997.tb00642.x
planned behavior. In J. Kuhl & J. Beckman (Eds.), Angrist, D. J., & Pischke, J. S. (2015). Mastering
Action-Control: From Cognition to Behavior Metrics: The Path from Cause to Effect. Princeton:
(pp. 11–39). Berlin: Springer. https://doi.org/10 Princeton University Press.
.1007/978-3-642-69746-3/2 Armitage, C. J., & Arden, M. A. (2002). Exploring
Ajzen, I. (1988). Attitudes, Personality, and Behavior. discontinuity patterns in the transtheoretical model:
Chicago: Dorsey Press. An application of the theory of planned behaviour.
Ajzen, I. (1991). The theory of planned behavior. British Journal of Health Psychology, 7, 89–103.
Organizational Behavior and Human Decision https://doi.org/10.1348/135910702169385
Processes, 50, 179–211. https://doi.org/10.1016/ Armitage, C. J., & Conner, M. (1999). The theory of
0749-5978(91)90020-T planned behaviour: Assessment of predictive
Ajzen, I. (2012a). The theory of planned behavior. In validity and “perceived control.” British Journal
P. A. M. Lange, A. W. Kruglanski, & E. Tory of Social Psychology, 38, 35–54. https://doi.org/10
Higgins (Eds.), Handbook of Theories of Social .1348/014466699164022
Psychology, Vol. 1 (pp. 438–459). London: Bamberg, S. (2006). Is a residential relocation a good
SAGE. https://doi.org/10.4135/9781446249215 opportunity to change people’s travel behavior?

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 29

Results from a theory driven intervention study. de Leeuw, A., Valois, P., Ajzen, I., & Schmidt, P. (2015).
Environment and Behavior, 38, 820–840. https:// Using the theory of planned behavior to identify
doi.org/10.1177/0013916505285091 key beliefs underlying pro-environmental behavior
Bamberg, S. (2013). Applying the stage model of in high-school students: Implications for
self-regulated behavioral change in a car use educational interventions. Journal of
reduction intervention. Journal of Environmental Environmental Psychology, 42, 128–138. https://
Psychology, 33, 68–75. https://doi.org/10.1016/j doi.org/10.1016/j.jenvp.2015.03.005
.jenvp.2012.10.001 Dillard, J., & Shen, L. (2012). The SAGE Handbook of
Bamberg, S., & Schmidt, P. (1999). Regulating Persuasion: Developments in Theory and
transport: Behavioral changes in the field. Journal Practice (2nd ed.). Thousand Oaks, CA: SAGE.
of Consumer Policy, 22, 479–509. https://doi.org https://doi.org/10.4135/9781452218410
/10.1023/a:1006391723214 Elliott, M. A., & Armitage, C. J. (2009). Promoting
Bamberg, S., & Schmidt, P. (2001). Theory driven drivers’ compliance with speed limits: Testing an
subgroup specific evaluation of an intervention to intervention based on the theory of planned
reduce car use. Journal of Applied Social behaviour. British Journal of Psychology, 100,
Psychology, 31, 1300–1329. https://doi.org/10 111–132. https://doi.org/10.1348
.1023/a:1006391723214 /000712608X318626
Bandura, A. (1977). Self-efficacy: Toward a unifying Fishbein, M., & Ajzen, I. (2010). Predicting and
theory of behavioral change. Psychological Changing Behavior: The Reasoned Action
Review, 84, 191–215. https://doi.org/10.1037 Approach. New York: Psychology Press.
//0033-295x.84.2.191 Fisher, J. D., & Fisher, W. A. (1992). Changing
Bandura, A. (1997). Self-Efficacy: The Exercise of AIDS-risk behavior. Psychological Bulletin, 111,
Control. New York: Freeman. 455–474. https://doi.org/10.1037//0033-2909
Bickman, B. (1987). Using Program Theory in .111.3.455
Evaluation. San Francisco: Jossey-Bass. Green, S. B., & Thompson M. S. (2015). A flexible
Brown, T. A. (2015). Confirmatory Factor Analysis for structural equation modeling approach for
Applied Research (2nd ed.). New York: Guilford analyzing means. In R. Hoyle (Ed.), Handbook of
Press. Structural Equation Modeling (pp. 393–416).
Campbell, D.T. (1969). Reforms as experiments. New York: Guilford Press.
American Psychologist, 24, 409–429. https://doi Hagger, M. S., Chatzisarantis, N. L. D., &
.org/10.1037/h0027982 Biddle, S. J. H. (2002). A meta-analytic review of
Campbell, D. T., & Fiske, D. W. (1959). Convergent and the theories of reasoned action and planned
discriminant validation by the multitrait- behavior in physical activity: Predictive validity
multimethod matrix. Psychological Bulletin, 56, and the contribution of additional variables.
81–105. Journal of Sport and Exercise Psychology, 24,
Chen, H. T. (2014). Practical Program Evaluation: 3–32. https://doi.org/10.1123/jsep.24.1.3
Theory Driven Evaluation and the Integrated Hancock, G. (2004). Experimental, quasi-
Evaluation Perspective. London: SAGE. experimental, and nonexperimental design and
Cialdini, R. B., & Trost, M. R. (1998). Social influence: analysis with latent variables. In D. Kaplan (Ed.),
Social norms, conformity and compliance. In The SAGE Handbook of Quantitative
D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The Methodology for the Social Sciences (pp.
Handbook of Social Psychology, Vols. 1–2 (4th 318–335). Thousand Oaks, CA: SAGE. https://doi
ed., pp. 151–192). New York: McGraw-Hill. .org/10.4135/9781412986311.n17
Coryn, C. L. S., Noakes, L. A., Westine, C. D., & Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953).
Schröter, D. C. (2011). A systematic review of Communication and Persuasion: Psychological
theory-driven evaluation practice from 1990 to 2009. Studies of Opinion Change. New Haven, CT:
American Journal of Evaluation, 32, 199–226. Yale University Press. https://doi.org/10.2307
https://doi.org/10.1177/1098214010389321 /2087772

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 30 ICEK AJZEN AND PETER SCHMIDT

Hukkelberg, S. S., Hagtvet, K. A., & Kovac, V. B. new university students: Combining
(2014). Latent interaction effects in the theory of self-affirmation, theory of planned behaviour
planned behaviour applied to quitting smoking. messages, and implementation intentions. British
British Journal of Health Psychology, 19, 83–100. Journal of Health Psychology, 23, 108–127.
https://doi.org/10.1111/bjhp.12034 https://doi.org/10.1111/bjhp.12277
Kelley, K., & Abraham, C. (2004). RCT of a theory-based Prochaska, J. O., & DiClemente, C. C. (1983). Stages
intervention promoting healthy eating and physical and processes of self-change of smoking: Toward
activity amongst out-patients older than 65 years. an integrative model of change. Journal of
Social Science and Medicine, 59, 787–797. https:// Consulting and Clinical Psychology, 51, 390–395.
doi.org/10.1016/j.socscimed.2003.11.036 https://doi.org/10.1037//0022-006x.51.3.390
La Barbera, F., & Ajzen, I. (2018). Control interactions Prochaska, J. O., & DiClemente, C. C. (1992). Stages
in the theory of planned behavior: Re-thinking the of change in the modification of problem
role of subjective norm. Unpublished manuscript. behaviors. In M. Hersen, R. M. Eisler, &
Lewin, K. (1946). Action research and minority P. M. Miller (Eds.), Progress in Behavior
problems. Journal of Social Issues, 2, 34–46. https:// Modification (pp. 184–218). Sycamore, IL:
doi.org/10.1111/j.1540-4560.1946.tb02295.x Sycamore Press.
Lewin, K. (1952). Field Theory in Social Science: Riebl, S. K., Estabrooks, P. A., Dunsmore, J. C. et al.
Selected Theoretical Papers by Kurt Lewin. (2015). A systematic literature review and meta-
London: Tavistock. analysis: The theory of planned behavior’s
Macy, J. T., Middlestadt, S. E., Seo, D.-C., Kolbe, L. J., application to understand and predict nutrition-
& Jay, S. J. (2012). Applying the theory of planned related behaviors in youth. Eating Behaviors, 18,
behavior to explore the relation between 160–178. https://doi.org/10.1016/j.eatbeh.2015
smoke-free air laws and quitting intentions. .05.016
Health Education and Behavior, 39, 27–34. Rodgers, J. L. (2010). The epistemology of
https://doi.org/10.1177/1090198111404702 mathematical and statistical modeling: A quiet
McDermott, M. S., Oliver, M., Svenson, A. et al. methodological revolution. American
(2015). The theory of planned behaviour and Psychologist, 65, 1–12. https://doi.org/10.1037
discrete food choices: a systematic review and /a0018326
meta-analysis. International Journal of Behavioral Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004).
Nutrition and Physical Activity, 12. https://dx Evaluation: A Systematic Approach (7th ed.).
.doi.org/10.1186%2Fs12966-015-0324-z Thousand Oaks, CA: SAGE.
Meitinger, K., & Behr, D. (2016). Comparing cognitive Shaddish, W. R., Cook T. D., & Campbell, D. T. (2002).
interviewing and online probing: Do they find Experimental and Quasi-Experimental Designs
similar results? Field Methods, 28, 363–380. for Generalized Causal Inference. Belmont, CA:
https://doi.org/10.1177/1525822X15625866 SAGE.
Michie, S., & Prestwich A. (2010). Are interventions Sheeran, P., Maki, A., Montanaro, E. et al. (2016). The
theory based? Development of a theory coding impact of changing attitudes, norms, and
scheme. Health Psychology, 29, 1–8. https://doi self-efficacy on health-related intentions and
.org/10.1037/a0016939 behavior: A meta-analysis. Health Psychology,
Michie, S., van Stralen, M. M., & West, R. (2011). The 35, 1178–1188. https://doi.org/10.1037
behaviour change wheel: A new method for /hea0000387
characterising and designing behaviour change Sheeran, P., & Taylor, S. (1999). Predicting intentions
interventions. Implementation Science, 6. https:// to use condoms: A meta-analysis and comparison
doi.org/10.1186/1748-5908-6-42 of the theories of reasoned action and planned
Norman, P., Cameron, D., Epton, T. et al. (2018). behavior. Journal of Applied Social Psychology,
A randomized controlled trial of a brief online 29, 1624–1675. https://doi.org/10.1111/j.1559
intervention to reduce alcohol consumption in -1816.1999.tb02045.x

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002
 Changing Behavior Using the Theory of Planned Behavior 31

Sniehotta, F. (2009). An experimental test of the theory Tyson, M., Covey J., & Rosenthal, H. E. (2014). Theory
of planned behavior. Applied Psychology: Health of planned behavior interventions for reducing
and Well-Being, 1, 257–270. https://doi.org/10 heterosexual risk behaviors: A meta-analysis.
.1111/j.1758-0854.2009.01013.x Health Psychology, 33, 1454–1467. https://doi.org
Stark, J., Berger, W. J., & Hössinger, R. (2018). The /10.1037/hea0000047
effectiveness of an intervention to promote active Webb, T. L., Joseph, J., Yardley, L., & Michie, S.
travel modes in early adolescence. Transportation (2010). Using the internet to promote health
Research, Part F, 55, 389–402. https://doi.org/10 behavior change: A systematic review and
.1016/j.trf.2018.03.017 meta-analysis of the impact of theoretical basis,
Stecker, T., McGovern, M. P., & Herr, B. (2012). An use of behavior change techniques, and mode of
intervention to increase alcohol treatment delivery on efficacy. Journal of Medical Internet
engagement: A pilot trial. Journal of Substance Research, 12, 1–27. https://doi.org/10.2196/jmir
Abuse Treatment, 43, 161–167. https://doi.org/10 .1376
.1016/j.jsat.2011.10.028 Webb, T. L., & Sheeran, P. (2006). Does changing
Steinmetz, H., Knappstein, M., Ajzen, I., Schmidt, P., behavioral intentions engender behavior change?
& Kabst, R. (2016). How effective are behavior A meta-analysis of the experimental evidence.
change interventions based on the theory of Psychological Bulletin, 132, 249–268. https://doi
planned behavior? A three-level meta-analysis. .org/10.1037/0033-2909.132.2.249
Zeitschrift Für Psychologie, 224, 216–233. Winkelnkemper, P., Ajzen, I., & Schmidt, P. (2019).
https://doi.org/10.1027/2151-2604/a000255 The theory of planned behavior: A meta-analysis
Trafimow, D. (2012). The role of auxiliary assumptions using structural equation modeling. Unpublished
for the validity of manipulations and measures. manuscript.
Theory and Psychology, 22, 486–498. https://doi Yzer, M., & van den Putte, B. (2014). Control
.org/10.1177/0959354311429996 perceptions moderate attitudinal and normative
Triandis, H. C. (1972). The Analysis of Subjective effects on intention to quit smoking. Psychology
Culture. New York: Wiley. of Addictive Behaviors, 28, 1153–1161.

Downloaded from https://www.cambridge.org/core. University of Exeter, on 14 Jul 2020 at 16:18:20, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108677318.002