Article Review- APA- 2 PAGES- references

Facilitating progress in health behaviour theory development and modification: the reasoned action approach as a case study

Katharine J. Head a

and Seth M. Noar b+

a Department of Communication, University of Kentucky, 124 Grehan Journalism Building, Lexington, KY 40506-0042, USA;

b School of Journalism and Mass Communication and

Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, 363 Carroll Hall (CB 3365), Chapel Hill, NC 27599-3365, USA

(Received 26 July 2012; final version received 18 February 2013)

This paper explores the question: what are barriers to health behaviour theory development and modification, and what potential solutions can be proposed? Using the reasoned action approach (RAA) as a case study, four areas of theory development were examined: (1) the theoretical domain of a theory; (2) tension between generalisability and utility, (3) criteria for adding/removing variables in a theory, and (4) organisational tracking of theoretical developments and formal changes to theory. Based on a discussion of these four issues, recommendations for theory development are presented, including: (1) the theoretical domain for theories such as RAA should be clarified; (2) when there is tension between generalisability and utility, utility should be given preference given the applied nature of the health behaviour field; (3) variables should be formally removed/ amended/added to a theory based on their performance across multiple studies and (4) organisations and researchers with a stake in particular health areas may be best suited for tracking the literature on behaviour-specific theories and making refinements to theory, based on a consensus approach. Overall, enhancing research in this area can provide important insights for more accurately understanding health behaviours and thus producing work that leads to more effective health behaviour change interventions.

Keywords: health behaviour theory; theory development; theory of reasoned action

Over the last four decades, researchers have developed a number of individual-level

health behaviour theories (HBT) to understand and predict health behaviours.

Reviews suggest that the most frequently used theories in the field are the

transtheoretical model (TTM) and stages of change, social cognitive theory (SCT),

the health belief model (HBM), the theory of reasoned action (TRA), and the

theory of planned behaviour (TPB) (Glanz & Bishop, 2010; Painter, Borba, Hynes,

Mays, & Glanz, 2008). The latter two HBTs stem from what has recently been

labelled the reasoned action approach (RAA); this approach includes the TRA,

TPB and the most recent development in this line of research, the integrated

behavioural model (IBM; Fishbein & Ajzen, 2010). In addition to the TTM, SCT

and HBM, these theories that make up the RAA have been widely used for the

*Corresponding author. Email: [email protected]

Health Psychology Review, 2014

Vol. 8, No. 1, 34�52, http://dx.doi.org/10.1080/17437199.2013.778165

# 2013 Taylor & Francis

purposes of explaining and predicting a variety of health behaviours (Albarracı́n,

Johnson, Fishbein, & Muellerleile, 2001; Blue, 1995; Montaño, Selby, Somkin,

Bhat, & Nadel, 2004). Additionally, all the theories mentioned above have been

broadly applied to design health behaviour change interventions (Glanz, Rimer, &

Viswanath, 2008), including the RAA theories (Fishbein, 2000; Hardeman et al.,

2002).

Despite the widespread use of HBTs contained in and beyond the RAA,

researchers are beginning to raise concerns about the trajectory of theory

development in the health behaviour field. In discussing the role of HBTs and

cumulative knowledge, Noar and Zimmerman (2005) argued that ‘[just] because we

are conducting more research on health behaviors does not necessarily mean that

we are adding substantive cumulative knowledge to this area of research’ (p. 275).

This is especially concerning for those who apply this knowledge to the practice of

health programme development; Crosby and Noar (2010) assert that ‘unfortunately,

theory development has not proceeded at a pace commensurate with the evolution of

health promotion practice’ (p. 259). Related to this, Rothman (2004) observes that

‘although theories may fluctuate in their popularity, their properties have remained

strikingly static over time’ (para. 4). He makes the critical point that theory should be

‘treated as a dynamic entity whose value depends on it being not only applied and

tested rigorously, but also refined based on the findings afforded by those tests’

(para. 4). Indeed, a critical aspect of theory testing is that theories are proposed,

empirically tested and then modified based upon the findings of those empirical tests

(Crosby, Kegler, & DiClemente, 2002). However, whether or not researchers are

actually modifying the theories is contestable. Simply because a theory is testable

does not necessarily mean that researchers are putting the said theory to the test. van

Ryn and Heaney (1992) note that ‘the testable nature of theory gives it a practical

advantage over personal belief systems or common sense’ (p. 319). But this

advantage only holds if researchers are critically testing and concurrently modifying

theory. In an examination of this and other issues, Ogden (2003) conducted a systematic

review of theoretical tests of HBTs. She found that:

. . . the majority of studies reported results that were not consistent with the predicted associations between constructs and left much of the variance in the outcome variable unexplained. However, rather than using the data to challenge the models, a range of explanations were offered relating to the wording used, the population studied, the behavior of concern, or the need for additional variables. All data are used to support the models, but it is not clear what data would enable the models to be rejected. (p. 426, emphasis added)

Thus, this review suggested that researchers are not using their data to challenge

existing theory or to critically inform theory development.

In the current article, we explore in some detail one particular line of research � the RAA � as a case study that may help advance an understanding of why HBTs have been relatively static over time. In doing so, we raise and discuss several

significant issues or barriers that are essential to HBT development and modifica-

tion, and we ultimately make recommendations for how to move forward in this

important area of inquiry.

Health Psychology Review 35

Health Behaviour Theory and the Reasoned Action Approach

To ask and potentially answer questions about theory testing and development, we

must first define what a theory is. A broad definition of theory is given by Kerlinger

and Lee (1999): ‘a theory is a set of interrelated (concepts), definitions, and

propositions that present a systematic view of phenomena by specifying relations

among variables, with the purpose of explaining and predicting phenomena’ (p. 11).

Hochbaum, Sorenson, and Lorig (1992) define HBTs specifically as ‘tools to help

health educators better understand what influences health-relevant individual, group,

and institutional behaviors and to thereupon plan effective interventions directed at

health-beneficial results’ (p. 298). In addition, DiClemente, Crosby, and Kegler

(2009) suggest that HBTs should ‘provide a conceptual framework for selecting

key constructs hypothesized to influence health behavior and, as such, provide a

foundation for empirical investigations, intervention development, implementation,

and evaluation’ (p. 11). It is also important to note that there is a distinction between

two general types of HBTs: stage models and continuum models. Stage models, like

the TTM, identify processes of change and identifiable stages that individuals may

progress through during the health behaviour change process. Alternatively,

continuum models consider a number of predictors and their relationship to one

another that ‘reflects the likelihood of action’; the RAA is of this type (Schwarzer,

2008, p. 3) (also see Weinstein, Rothman, & Sutton, 1998).

The RAA is an approach that extends beyond just the health arena, but as will be

demonstrated in the current article, it has been heavily applied in health and,

therefore, is one of the major HBTs (Glanz et al., 2008; Noar & Zimmerman, 2005).

Indeed, the RAA approach is best understood as a line of research with four

chronological phases (see Figure 1 for the RAA with shading to illustrate the

different theories that make up this approach). In the first phase, Fishbein and Ajzen

developed the TRA, a development that can be traced back to 1967 (Fishbein, 1967).

The TRA posits that in addition to a host of indirect influences (e.g., demographic

variables, norms and personality traits), attitudes towards the behaviour and

subjective norms concerning the behaviour are direct antecedents to behavioural

intention. Behavioural intention is then posited as most the direct antecedent to

behaviour (Fishbein & Ajzen, 1975; Montaño & Kasprzyk, 2008). Across a number

of behaviours, the TRA has been shown to have relatively strong predictive utility. In

a meta-analysis of 87 studies, Sheppard, Hartwick, and Warshaw (1988) found an

average correlation for the intention-behaviour relationship to be 0.53; the average

relationship between attitude/subjective norm-intention was 0.66. Despite its strong

performance, Ajzen believed the TRA was deficient in one significant way: it was

insufficient in explaining behaviours that were not under volitional control (Ajzen,

1991). Subsequently, Ajzen developed the TPB by including the concept of perceived

behavioural control in the model. The addition of perceived behaviour control, which can be described as phase two

in the RAA line of research, required two considerations for how it fit into the

previously established TRA. First, Ajzen and Madden (1986) note that in part, ‘the

effect of perceived behavioral control on behavior is completely mediated by

intention’ (p. 458). In addition, they posited that ‘perceived behavioral control can

help predict goal attainment independent of behavioral intention to the extent that it

reflects actual control with degree of accuracy’ (pp. 458�459). The revised model,

36 K.J. Head and S.M. Noar

which included perceived behavioural control, was named the TPB. Ajzen (1991)

claims the TPB ‘provides a useful conceptual framework for dealing with the

complexities of human social behavior’, especially for instances in which a person

feels they do not have complete volitional control over performing the behaviour

(p. 206).

The third phase occurred in 1991 when a group of theorists (Albert Bandura,

Marshall Becker, Martin Fishbein, Frederick Kanfer, and Harry Triandis) convened

for a workshop sponsored by the National Institute of Mental Health (NIMH); the

group was charged with developing a unified model of behaviour. The ultimate result

of the workshop was the development of the IBM, which derives its main

components from overlapping constructs in five health behaviour change theories,

including TRA, TPB, SCT, HBM and the theory of interpersonal behaviour

(Fishbein, 2000, 2009; Fishbein et al., 1992; Montaño & Kasprzyk, 2008). This

model and similar versions of this model have proven useful in understanding and

predicting both HIV prevention behaviours and cancer screening behaviours

(Kasprzyk, Montaño, & Fishbein, 1998; Montaño, et al., 2004; Montaño,

Figure 1. Reasoned action approach theories.

Note: Figure 1 uses colours to illustrate the development of theories within the RAA. Attitude,

Norms, Intention and Behaviour are shaded grey but are present in all four models. The

versions of each theory in this figure come from the best representations of the original theory,

as found in following sources: TRA (Ajzen & Fishbein, 1980), TPB (Ajzen & Madden, 1986),

IBM (Montaño & Kasprzyk, 2008), and RAA (Fishbein, 2008).

Health Psychology Review 37

Thompson, Taylor, & Mahloch, 1997; von Haeften, Fishbein, Kasprzyk, &

Montano, 2001). Despite its promising utility, the IBM in its current form has not

yet been employed extensively by researchers. This may be the result of inadequate

exposure for the model; the workshop occurred in 1991 and while an NIMH report

of the model appeared in 1992 (Fishbein et al., 1992), it wasn’t until 1998 that a

primary research article testing this model appeared in the mainstream research

literature (Kasprzyk et al., 1998). In the fourth phase of RAA, Fishbein and Ajzen began working together again.

They note in their last book (Fishbein & Ajzen, 2010) that their career paths had

diverged in the early 1980s, with Fishbein focusing on his work on HIV prevention

using the TRA and Ajzen devoting his time to developing the TPB. However, after

the NIMH theorists’ workshop when Fishbein proposed the IBM, it did not go

unnoticed that the IBM ‘was almost identical to Ajzen’s TPB . . . [but] incorporated Bandura’s . . . notion of self-efficacy rather than Ajzen’s more recent concept of perceived behavioral control’ (p. 19). This led both men to realise that ‘even though

we were at that time working quite independently, we were moving in similar

directions’ and they began working together again in 2001 ‘when we began to

reconcile the differences between our models’ (pp. 19�20). This reconciliation and the broader programme of research that encompasses all of the theoretical variations of

this approach has been labelled the RAA. In describing the approach, Fishbein

(2008) states that ‘what the reasoned action approach attempts to do is to identify a

relatively small set of variables that can account for a substantial proportion of the variance in any given behavior;’ these variables include ‘intentions, attitude, perceived

norms, self-efficacy or perceived behavioral control, behavioral beliefs (which are

often referred to as cost-benefits or outcome expectancies), normative beliefs, and

control beliefs’ (pp. 834�835). In sum, the RAA is essentially an approach that began with the TRA and is now

composed of a number of similar theories which posit that variables such as attitudes,

norms and perceived behavioural control are important predictors of behavioral

intention and, ultimately, behaviour. A critical examination of the entire RAA,

however, reveals that the approach has changed little since this line of research first

began. First, the external variables or ‘background influence’ variables (see Figure 1)

that are in many ways presented as ‘new’ in the IBM (Fishbein, 2009) were actually

included in the TRA more than 30 years ago (see Ajzen & Fishbein, 1980, Figure 7.1,

p. 84). This set of variables has thus long been hypothesised to affect behaviour

indirectly through RAA variables (see Ajzen & Fishbein, 1980, pp. 82�90). Second, more fine-tuned understandings of some variables (e.g., norms � now represented as injunctive and descriptive norms, and attitudes � now represented as experiential and instrumental attitudes) have added value to the model, but these arguably do not

represent significant theoretical modifications. Third, the addition of self-efficacy to

the model, as indicated above, essentially parallels Ajzen and Madden’s addition of

perceived behavioural control to the TPB, which took place more than 25 years ago

(Ajzen & Madden, 1986). It should be noted that other researchers had suggested

adding self-efficacy to the TRA more than a decade before Fishbein formally added

the variable to the model in the form of the IBM (see de Vries, Dijkstra, & Kuhlman,

1988). Also, as will be described below, several additional variables have been tested

and shown empirical value but have not been added to the model. Fourth, interesting

(and potentially important) feedback loops from attitude to beliefs and behaviour to

38 K.J. Head and S.M. Noar

beliefs, evident in early TRA work (see Fishbein & Ajzen, 1975, pp. 15�16, Figures 1.1, 1.2), have apparently been dropped from newer scholarship. In this manner, one

could argue that theories such as IBM are less complex and realistic than the earliest

work with the TRA and have changed little for the better across several decades.

Finally, if we consider the fact that other ‘new’ IBM variables such as environmental

constraints and skills and abilities are listed as part of the IBM but only treated in a

very cursory manner (see Fishbein, 2000, 2009), we are perhaps left wondering why

so little has changed over four decades of research (i.e., the IBM of today looks quite

similar to the original TRA of 1975). Scholars who conduct health behaviour change

research should critically consider if HBTs like those in the RAA have been

adequately tested and refined. If not, what is obstructing progress in theory

development and how can we overcome barriers in the future?

Important theory development considerations

Theoretical domain

If we are to first consider why HBTs such as the RAA have not advanced more fully

than they have, we must first address what we are trying to develop in the first place.

That is, what is the purview of an HBT? As an example, if the point is to develop a

theory that is focused only on beliefs and attitudes, then we should only consider

those types of variables as possible additions to the RAA. If the point is to develop a

more comprehensive theory of health behaviour, then we should consider a much

broader possible set of influences for the theory.

Advocates of the RAA approach, such as Montaño and Kasprzyk (2008), state

that the ‘TRA was developed to better understand relationships between attitudes,

intentions and behaviors’ (p. 68). This suggests that the purview of the theory is

relegated to the social psychological realm of attitudes and beliefs. Ajzen (1991)

seems to take a broader view, stating that the TPB is ‘a theory designed to predict

and explain human behavior’ (p. 181). While clearly the earliest work with the TRA

was focused on attitudes and beliefs (Fishbein & Ajzen, 1975), Fishbein’s more recent

IBM does include some variables that go beyond that realm (e.g., environmental

constraints, skills and abilities). However, as indicated above, these variables seem to

be added in a more cursory fashion and the focus appears to be the extent to which

they constrain or advance the ability of attitudes and intention to affect behaviour

(Fishbein & Ajzen, 2010). The point here is that HBT researchers should engage in a

dialogue about what type of theory is most valuable to ‘build’. Without a roadmap

that directs the path, we are sure to get lost along the way.

In addition, the purview of an HBT has to do with whether its ultimate goal

is prediction, intervention or both. Writings on the RAA appear to suggest that

this theoretical approach serves both purposes � behavioural prediction and intervention � with perhaps prediction being the primary goal and intervention being a secondary use (or application) of the theory (Fishbein, 2009). The

requirements for such a theory are likely to be different than one being developed

solely for prediction or intervention, and to date, the RAA may be more precise for

the former than the latter. For example, Hardeman et al. (2002) examined how the

TPB was applied in behaviour change interventions and unfortunately discovered

that scholars were not clear in how they used � and tested � the TPB. First, they

Health Psychology Review 39

found that the targeted components of the TPB were poorly identified or not

identified at all, meaning it was difficult to assess how the TPB was actually used in

the intervention. Second, despite finding positive changes in behavioural intention

and behaviour in these studies, it was unclear how the TPB was used to design the

interventions and, therefore, the findings could not be attributed to the TPB’s role.

They argue ‘to allow judgment of the effectiveness of using the TPB to develop

interventions . . . studies would need to apply the TPB more comprehensively and be more explicit about how it has been applied’ (p. 148). They conclude that ‘at present

there thus is insufficient evidence to judge whether TPB components mediate

changes in intention and behavior within evaluated interventions’ (p. 149). Related to this, a study conducted by Cooke and French (2008) examined the

TRA and the TPB to predict intentions and attendance at screening programmes.

They claim ‘the TRA/TPB was an effective framework for predicting screening

intentions and attendance. The next step is to perform experimental research that

builds on these findings to improve screening attendance’ (p. 763). Several scholars

have recently advocated such an approach (Noar & Mehrotra, 2011; Sniehotta, 2009;

Weinstein, 2007), as there is no guarantee that factors found to be associated with

intentions and behaviours can be applied in interventions as causal factors that will

result in behavioural changes. To date, however, the field has almost entirely operated

under a survey research paradigm, using (mostly cross-sectional) survey research

studies to test HBTs and then applying those HBTs as bases for health behaviour

interventions. This reasoning essentially involves a leap of faith that those factors

found to be associated with behaviours in tests of HBTs will act as causal

mechanisms in the context of interventions. While the paradigm in this area may

begin to shift towards more experimental research, currently most theory testing is

conducted using survey research. Thus, the status quo is such that we have much

more evidence that HBTs contain factors that are associated with and may predict

behaviour and much less evidence that changes in those factors in the context of

interventions will lead to health behaviour changes. While there is some support for

the notion that theory-based interventions are more efficacious than those that lack

a theoretical basis (Glanz & Bishop, 2010), the difficulties in separating out the

contribution of theory to intervention efficacy, including the lack of mediation

analyses in many published interventions, has left open questions about the precise

role of theory in intervention efficacy (Noar & Mehrotra, 2011).

Tension between generalisability and utility

A key aspect of theory is that it is generalisable, or it is ‘robust and therefore may be

applicable across diverse venues, populations, and social environments’ (DiClemente,

Crosby, & Kegler, 2002, p. 3). Thus, if a theory performs similarly well across diverse

behavioural areas, then this contributes to evidence of its generalisability. Another

aspect of theory is that it must have utility, or the degree to which the theory is ‘useful

and helpful’ in the field (Prochaska, Wright, & Velicer, 2008, p. 577). In many ways,

these two dimensions could at times be in conflict with one another, particularly in

the HBT area. As health behaviour change researchers, we should ask ourselves � is it better to have a broad theory that predicts across behaviours but is not very precise

(in other words, one that is generalisable), or a more specific theory that predicts

40 K.J. Head and S.M. Noar

more precisely (and has utility to practitioners/researchers) but has differences across

behaviours or with regard to other factors?

When we examine the data in the HBT area, it is fairly clear that the relationship

between theoretical constructs in the RAA varies depends on the behaviour studied. For example, Godin and Kok (1996) reviewed 56 health behaviour studies reporting

87 applications of the TPB and found that the average correlations between

theoretical constructs in the TPB varied according to the health behaviour category.

The average correlation between attitude and intention for addictive behaviours was

r � 0.53, while the average correlation between attitude and intention for healthy eating was r � 0.34. Moreover, behaviour-specific reviews of TRA/TPB have often thoughtfully suggested adding variables to the theories that are specific to a

particular domain (e.g., Sheeran, Abraham, & Orbell, 1999), but to our knowledge no formalised behaviour-specific TRA or TPB exists. The literature does suggest,

however, that researchers should strongly consider having behaviour-specific versions

of theories like the TPB in order to (1) better understand particular behaviours and

(2) provide more relevant theoretical guidance for designing interventions for specific

behaviours.

Moreover, it is apparent in practice that many researchers are already using the

behaviour-specific (or utility) approach. Painter et al.’s (2008) examination of the use

of HBTs found that researchers use theory along a continuum, from studies being merely informed by theory to the opposite end where theory is being built and

created. Applying an entire theory for theory testing or intervention was found to

rarely be the case. One category, testing theory, was used by only 7.2% of the studies

examined and even then, a study only had to measure and explicitly test half of the

theoretical constructs for the theory used. They concluded that to advance the use of

theory in health behaviour research, ‘theory should be used more thoroughly . . . this can be done by measuring and testing the full set of key constructs in a theory’ (p.

362). Despite this call, it is more common for researchers to pick and choose theoretical and other constructs based upon what support is found in the empirical

literature concerning the particular behaviour under study, rather than attempting to

measure and test the entire theory (Glanz & Bishop, 2010).

Hochbaum et al. (1992) provide a good explanation for this phenomenon. They

state that:

although academics may wish to test whether (or demonstrate that) some given theories contribute to a project’s success and look for opportunities to do so, practitioners search for ways to assure success . . . .[and] they must search for and utilise anything and everything that will help them plan and conduct programs to assure success. (pp. 303, emphasis added)

In sum, the tension between generalisability and utility can in some ways be

understood as a tension between academics, who desire theoretical fidelity in

research, and practitioners, who desire theories that can guide programme development for specific contexts.

Criteria for adding/removing variables

One of the most visible ways a theory can develop is through the addition or removal

of variables that can help better explain and predict behaviour. This is also one of the

Health Psychology Review 41

most important ways that HBT develops because ‘improvements in both HBT and

intervention methods depend on each other’ (Rothman, 2004, p. 2). Glanz et al.

(2008) argue ‘the best theory is informed by practice; the best practice should be

grounded in theory’ (p. 24). However, there is little agreement and even less guidance

for how theorists and/or researchers should go about adding and removing variables.

In other words, it is unclear how researchers should use data from tests of theory in

the field to inform theory development and add or remove variables. Indeed, the

existing evidence suggests that when researchers observe null findings for particular

variables in their theory-testing studies, they often ‘explain away’ this phenomenon

by pointing to measurement problems or other issues rather than possible problems

with the theory itself (Ogden, 2003). While we cannot prove the null hypothesis, and

thus there may indeed be methodological issues that play into this problem, we still

need criteria that can help guide what decisions to make about adding and removing

variables from a theory. This may not be as important in the context of a single study,

but it is critical when, over time (e.g., meta-analytic review), it becomes evident that

evidence is gathering that a particular variable should be added or removed from a

theory.

Dubin (1978) posits that an ‘unconstrained willingness to admit all possible units

into a scientific model provides the widest range of opportunities for theory building’

(p. 58). He outlines numerous ways in which theorists may add (or invent) variables

(or units) for a theory. These include invention by extension of an existing unit,

invention by subdivision of an existing variable, invention through disproving the

null hypothesis (i.e. a relationship or significant difference does exist), invention

through factor analysis, invention through scale analysis and invention of an

intervening variable. Despite these many suggestions, Dubin is unclear about the

specific standards for adding a new variable. For instance, many argue that a new

variable must explain variance in the outcome variable. Ajzen (1991), when

discussing additions to the TPB, says, ‘the theory of planned behavior is, in

principle, open to the inclusion of additional predictors if it can be shown that they

capture a significant proportion of the variance in intention or behavior after the

theory’s current variables have been taken into account’ (p. 199). Furthermore,

Dubin discusses the removal of variables from a theory. He states that ‘when it is

possible to postulate no interaction between units, we may exclude one or both from

a model’ (p. 86). In other words, if a predictor variable does not explain variance in

an outcome variable, or it is not meaningfully related to other model variables in a

mediating or moderating role, it should be considered for removal from the theory.

The RAA has undergone some changes in this category since the original TRA

was developed. First, one can see that a few new variables have been added. For

example, as previously mentioned, the added construct of perceived behavioural

control has been shown to be a significant predictor of behaviour, especially in

behaviours not completely under volitional control (Ajzen, 1991). Second, some

variables have been divided into more precise constructs; for example, norms in the

IBM is conceptualised as both injunctive norms and descriptive norms, rather than

solely injunctive norms (originally called subjective norms in the TRA; Fishbein &

Ajzen, 1975). These changes appear to have been made based upon empirical

evidence demonstrating the contribution of these variables to the theory, and that

progress is to be commended.

42 K.J. Head and S.M. Noar

However, as indicated earlier in this paper, the RAA has made what we would

describe as only minor changes after more than four decades of research. In addition,

there are likely to be several important variables that could be strong candidates to

add to the RAA approach, given that there is room for improvement in the theory. In

fact, research has demonstrated a host of variables that meet Ajzen’s criteria listed

above, and yet these variables have not been formally added to the theory. While

RAA theories have fairly good prediction/association for a theory that attempts to predict behaviour, meta-analyses also suggest that much of the variance is left

unexplained (Conner & Armitage, 1998; Sandberg & Conner, 2008). For example,

Armitage and Connor (2001) conducted a meta-analysis (k � 185) of studies testing the TPB and found that the average multiple correlation of attitude, subjective norm

and perceived behavioural control with behavioural intention was R � 0.63, explaining 39% of the variance (R

2 � 0.39). Additionally, the average multiple correlation of behavioural intention and perceived behavioural control with

behaviour was R � 0.52, explaining 27% of the variance (R2 � 0.27). While these data represent fairly good prediction in the context of these studies, we need to

recognise that there are several limitations of such theoretical tests. These include

errors and bias in self-report data (e.g., social desirability), the overreliance on cross-

sectional data which may exaggerate a theory’s true effects (Weinstein, 2007), the lack

of controlling for past behaviour in analyses and the limitations of predicting

intention as it relates to the intention�behaviour gap (Sheeran, 2002). Thus, these theories may not actually be performing as well as the data suggest that they do, and

we should continually seek to improve the prediction of our theories where possible. To further inform this discussion, Table 1 presents data from several meta-

analyses of the TPB, which have considered the influence of additional variables that

are not formally a part of the theory. While each meta-analysis tends to confirm the

association of the formal TPB variables with behavioural intention (and in one case,

with behaviour), these meta-analyses reveal that other variables � such as anticipated regret, moral norms and self-identity � exhibit associations with intention that are equal to or greater than the traditional TPB variables. Moreover, each meta-analysis

demonstrates that the novel variables add unique variance in the prediction of

intention over and above the traditional TPB predictor variables, and thus they do

not appear to be redundant with current TPB variables. Thus, these variables appear

to meet Ajzen’s criteria for adding variables to the TPB. While the first meta-analysis

lends support for a change that was made to the IBM (adding descriptive norms to

the theory), none of the other novel variables in these meta-analyses have been

formally added to any RAA theories. Nor have other mediating or moderating

variables been added to the theory, despite compelling research on such factors (e.g.,

Gollwitzer & Sheeran, 2006; Sheppard et al., 1988). In fact, one of the most compelling areas for extension of the RAA is with regard to the intention�behaviour gap, which refers to the phenomenon that many intenders do not engage in the

intended behaviour, in contrast to the clear prediction from RAA theories (Sheeran,

2002). Recent meta-analytic (experimental) research demonstrates that a large

increase in intentions produces only a small increase in behaviour, further illustrating

this point (Webb & Sheeran, 2006). While much research suggests a variety of

variables that may help us better understand and ‘close’ this gap � such as implementation intentions (Gollwitzer & Sheeran, 2006; Orbell, Hodgkins, &

Sheeran, 1997), preparatory behaviours (Abraham et al., 1999; Bryan, Fisher, &

Health Psychology Review 43

Table 1. Some examples of meta-analyses that have empirically demonstrated the value of

additional variables in the context of the theory of planned behaviour.

Study Variable r-I r-PB r-FB Additional findings

Rivis and Sheeran (2003) � 18 studies conducted

across various behaviours

Descriptive

norm

0.46 � � Descriptive norm was significantly (p B 0.001)

associated with intention

when controlling for all TPB

predictor variables.

Attitude 0.58 � � Subjective

norm

0.44 � �

Perceived

behavioural

control

0.21 � �

Sandberg and Conner

(2008) � 20 studies conducted across various

behaviours

Anticipated

regret

0.47 0.34 0.28 Anticipated regret was

significantly (p B 0.001)

associated with intention

and future behaviour when

controlling for all TPB

predictor variables. When

past behaviour was added to

the models, anticipated

regret remained significant

(p B 0.001) in the intention

model but was reduced to

non-significance in the

future behaviour model.

Attitude 0.44 0.30 0.27

Subjective

norm

0.43 0.18 0.21

Perceived

behavioural

control

0.30 0.31 0.11

Rise, Sheeran, and

Hukkelberg (2010) � 33 studies conducted across

various behaviours

Self-identity 0.47 � � Self-identity was significantly (p B 0.001)

associated with intention

when controlling for all TPB

predictor variables; when

past behaviour was added to

the model, it remained

significant (p B 0.001).

Attitude 0.50 � � Subjective

norm

0.39 � �

Perceived

behavioral

control

0.35 � �

44 K.J. Head and S.M. Noar

Fisher, 2002) and strategic or action planning (Schwarzer, 2008; Sniehotta, Scholz, &

Schwarzer, 2005) � RAA theories have done virtually nothing to integrate such work into its approach. Moreover, the notion of intention itself has been challenged and

questioned, with studies showing that in at least some cases, constructs such as

behavioural willingness (Gibbons, Gerrard, Blanton, & Russell, 1998) or suscept-

ibility (Pierce et al., 1996) may be more appropriate to understanding individuals

who may engage in the behaviour. To date, the RAA approach has not integrated any

of this work.

It is important to note that Fishbein and Ajzen (2010) have at least considered

some of these variables as possible additions to their approach. One key reason given

for not adding such variables is that some of these variables are seen as only applying

to particular behaviours and not more broadly across numerous classes of behaviour

(Fishbein & Ajzen, 2010, pp. 282�284). Thus, whereas generality appears to be preferred by the developers of the RAA, practitioners/interventionists are more likely

to prefer specificity and better prediction of a particular behaviour.

Power to change a theory

Rothman (2004) states that ‘the development and specification of theories of human

behavior depend upon an iterative series of research activities in which theoretical

principles initially formulated by basic behavioral scientists are tested and evaluated

Table 1 (Continued )

Study Variable r-I r-PB r-FB Additional findings

Rivis, Sheeran, and

Armitage (2009) � 27 studies conducted across

various behaviours

Moral norms 0.47 � � Moral norms was significantly (p B 0.001)

associated with intention

when controlling for all TPB

predictor variables; when

anticipated affect was added

to the model, it remained

significant (p B 0.001).

Anticipated

affect

0.42 � � Anticipated affect was significantly (p B 0.001)

associated with intention

when controlling for all TPB

predictor variables; when

moral norms was added to

the model, it remained

significant (p B 0.001).

Attitude � � � Subjective

norm

� � �

Perceived

behavioural

control

� � �

Note: All r’s are weighed correlations from meta-analysis; r-I � correlation with intention; r-PB � correlation with past behaviour; r-FB � correlation with future behaviour.

Health Psychology Review 45

by applied behavior scientists’ (p. 3). He goes on to say, ‘these tests provide critical

information that enables basic scientists to revise, refine, or reject their initial

principles’ (p. 3). Despite evidence that changes should be made to the RAA, few

changes have been made over the years. In addition to the three previous sections discussing this issue, one final question remains � who has the authority to change a theory? Only the original theorist of that particular theory? Any researcher in the

field? This is an important issue that has been given only scant attention in the

literature.

One approach would dictate that any researcher who finds evidence of a needed

change and publishes that information in an academic journal has suggested a

change in the theory. However, the published literature is full of studies in which

empirical evidence suggests changes that could or should be made to the RAA, and these changes are often not embraced by the theorists themselves (Fishbein & Ajzen,

2010) or other researchers in the field. If we look at the history of RAA theories, we

see that the TRA and TPB were developed primarily by two researchers who

presented these theories at conferences and published work in academic journals and

books. However, with the massive amount of health-related research and so many

researchers undertaking such work, an important conference presentation or journal

publication might get lost in the milieu. Indeed, simply keeping up with the large

amount of research in particular theoretical domains can be difficult given how much research is published, particularly in the health behaviour field. Also, the theorists

themselves may not be as open to adding new variables to their theory as compared

to particular research communities. Indeed, we have already made the case that

agendas differ � while the theorists wish to understand the smallest set of variables that predict the largest numbers of behaviours, applied researchers are more

interested in the most precise understanding of a particular behaviour. These theory

development goals are quite different from one another.

Instead, organisations that sponsor work in theory development in specific behavioural domains and provide avenues for dissemination of new or modified

theories may be a better way to organise theoretical developments. In fact, it is worth

pointing out that the catalyst for the development of the IBM was a theorist’s

workshop organised by the NIMH that brought several scholars together with the

goal of developing a unified theory of behavioural prediction focused on HIV/AIDS-

related behaviours. If this effort had not occurred, the RAA approach may have

evolved even less than it has to date. Also, if this effort had been followed up with an

organisational focus on HIV/AIDS-related theories and models that were supported by the latest research, more developments might have come from that workshop

effort than solely the IBM. Such products could have been disseminated to relevant

researchers, using several mechanisms at the disposal of organisations such as the

National Institutes of Health (e.g., special journal issue, website, Funding

Opportunity Announcement, etc.).

Recommendations for theory development in the RAA

The previous section proposed four important barriers to consider in theory

development, modification and dissemination. While these were presented within

the context of a case study of the RAA, they can and should be considered with

regard to the development of other HBTs. With that in mind, suggestions on the use

46 K.J. Head and S.M. Noar

and modification of RAA theories and other HBTS are discussed below and

summarised in Table 2. First, the theoretical domain for theories such as RAA should be clarified.

Theories like the TRA, TPB and IBM were initially developed as social

psychological theories, but it remains unclear as to what theoretical territory they

seek to cover now and in the future. Are the theories open to variables that reside

outside the social psychological domain, or are they instead relegated solely to that

domain? This issue is also related to the second issue of intervention development, as

to date, virtually all of the RAA variables are social psychological and thus amenable

to change in psychologically oriented interventions. An exception to this was the

addition of environmental constraints to the IBM, which has the potential to move

the theory beyond the social psychological domain. To date, the addition of that

variable appears to be largely cursory in nature; it may also be related to calls to

separate actual control (i.e., environmental constraints) from perceived behavioural

control in the TPB (Godin & Kok, 1996).

We have noted throughout this paper that HBTs are continually used as bases for

health behaviour interventions. However, we have also noted that the bulk of

research on these theories is correlational, and the primary function of HBTs has

been to explain and predict behaviours. Thus, we should be careful when using

theories in designing interventions, as the extent to which the predictor variables

represent causal mechanisms in behaviour change is not known. In addition to

clarifying how theoretical variables are translated within particular interventions

(Hardeman et al., 2002), we also need increased experimental research, using the

Table 2. Recommendations for advancing health behaviour theory development and

modification.

Issue Recommendations

Theoretical domain Clarify theoretical territory of RAA theories; advance

discussion of prediction vs. intervention applications of

RAA theories; conduct new experimental research on

RAA theories

Generalisability vs. utility Recognise differing agendas of basic versus applied

researchers; consider two lines of advancement for RAA

theories � a general theory that applies to the most behaviours and behaviour-specific theories in key health

areas such as diet, exercise, safer sex, etc.

Criteria for changing theory Advance conversation on criteria for adding/amending/

removing variables from theory; make formal changes to

general and behaviour-specific RAA theories based on

the research literature (in particular using data generated

from meta-analysis)

Organisational tracking of

theoretical developments

Discuss new ways to track theoretical developments in

HBT; consider a consensus approach that takes theory

modification decisions out of the hands of the few; move

forward with either an expert panel approach or a wiki-

platform approach to theoretical tracking and

modification

Health Psychology Review 47

RAA approach (Noar & Mehrotra, 2011; Weinstein, 2007). Indeed, it is somewhat

remarkable that it took more than two decades from the development of the theory

for the first experimental test of the TPB to be conducted, published quite recently

(Sniehotta, 2009). Rather than conducting a test of the theory, using the typical

survey research approach, this study conducted a factorial experiment to examine the

impact of interventions based on particular TPB factors (e.g., behavioural beliefs and

normative beliefs). While results indicated some support of the theory in terms of changing some TPB factors and intentions, results with regard to behaviour change

were inconsistent with TPB predictions. While increased experimental research has

the potential to greatly advance our understanding of HBT and behaviour change

mechanisms, to date it has only seldom been applied in testing HBTs (Noar &

Mehrotra, 2011; Sniehotta, 2009). More experimental research with HBTs is greatly

needed.

Second, when there is tension between generalisability and utility, utility should

be given preference, given the applied nature of the health field. We have already

demonstrated that applied researchers are apparently giving preference to utility over

generalisability (Glanz & Bishop, 2010; Painter, et al., 2008). Additionally, and

perhaps more importantly, by giving preference to utility we create a research

environment in which we are following conceptual thinking but also empirical data

to where they lead. Moreover, the health field is largely divided into areas where

researchers study different diseases and behaviours, and to the extent that different

behaviours can be best predicted by variations on particular theories, we should work to understand this and formalise such theories. However, given that Fishbein and

Ajzen’s (2010) stated goal is to have a general theory that applies broadly across

behaviours, such a general theory will likely always exist. Further, in the HBT

domain, understanding what factors are common to behaviour and behaviour

change across theories is certainly of interest, especially in the context of multiple

behaviour change interventions (Noar, Chabot, & Zimmerman, 2008). However, in

an applied context, it is clear that behaviour-specific versions of theories such as TPB

will be most precise in terms of behavioural prediction, and they are also most likely

to be instructive for intervention development. As one example of this, research in

the realm of safer sex has suggested a whole set of factors that could fruitfully inform

a safer sex TRA/TPB, from partner norms to condom communication to

preparatory behaviours (Abraham et al., 1999; Bryan et al., 2002; Sheeran et al.,

1999). Despite this, no behaviour-specific TRA/TPB formally exists, even though

such a development would likely better build the cumulative theoretical knowledge in

the safer sex and other arenas.

Third, and related to utility, we must be open to changing our theories when particular elements within those theories do not work when empirically tested in the

real world. More specifically, if variables are not performing in a particular area, then

they should be seriously considered for removal (or amendment) from the theory in

that particular domain; similarly, variables demonstrating important (theoretical and

empirical) contributions should be seriously considered for addition to a given

theory. Importantly, this draws attention to the fact that there can be vastly different

characteristics for different health behaviours. In fact, previous research has shown

that for individual behaviours, the RAA may work in different ways and additional

variables may be needed to explain and predict particular behaviours, while other

variables may need to be removed or amended with regard to other behaviours (Blue,

48 K.J. Head and S.M. Noar

1995; Godin & Kok, 1996). This is entirely consistent with a behaviour-specific

approach.

Fourth, the process for tracking and disseminating findings on theory develop-

ment can be greatly improved. In an age, where a plethora of information exists and a

large number of researchers and practitioners use theory, it is easy for potential

developments to get ‘lost in the shuffle.’ Organisations with a stake in particular

health areas may be best suited for tracking the literature on particular behaviour-

specific theories and, over time, making refinements to the theories. Organizations

such as the Centers for Disease Control and Prevention and the National Cancer

Institute have major stakes in the accuracy and completeness of HBTs in particular

health domains, and they also have considerable influence and dissemination

capabilities. Such organisations would thus be well suited to the task of tracking

theoretical developments in high-priority behavioural domains, and putting out their

own versions and suggested modifications of HBTs in the form of publications, on

websites, and in relation to funding announcements. Moreover, the advantages of

putting a theory online are such that a ‘living’ version of the theory could be posted

and modified over time as additional empirical evidence is generated and evaluated

by an expert group charged with this task.

However, to ensure that changes to the theory are made thoughtfully, an advisory

committee could consult on how empirical data would be used to modify the theory

and would recommend theoretical changes at specified intervals, based on the

empirical evidence in the literature. Alternatively, this process could use a ‘wiki’

model where the broad community of researchers has direct control over modifica-

tions to the theory online, and changes are made by anyone within the research

community. Under this model, which has been very successful in the case of websites

such as Wikipedia, changes made that are inaccurate or not agreed upon by most of

the community are amended by a member of that community. Such a project could

be hosted by the National Cancer Institute’s grid-enabled measures website, which

already uses a wiki-based platform for health behaviour (and other) constructs,

definitions, measures, datasets and other items (see http://cancercontrol.cancer.gov/

brp/gem.html).

Conclusion

The purpose of this paper was to explore the issues related to barriers to theoretical

development in HBTs. Using the RAA as a case study, we explored, in-depth, four

important considerations for theory development and testing and then provided

recommendations for stimulating progress in these areas. If those of us who research

health behaviour and test and apply these theories begin to demand more of our

theories, the result will ultimately be more advanced and precise theories than those

that exist today. We will not be guilty of continuing the trend of a large theory-testing

literature that seems to have relatively little impact on the actual make-up of our

current HBTs (Crosby & Noar, 2010; Rothman, 2004). Instead, we can develop

theories that are informed by data and are more effective at explaining and

predicting health behaviours as well as improving the ability of our interventions to

change health behaviour.

Health Psychology Review 49

References

Abraham, C., Sheeran, P., Norman, P., Conner, M., de Vries, N., & Otten, W. (1999). When good intentions are not enough: Modeling postdecisional cognitive correlates of condom use. Journal of Applied Social Psychology, 29(12), 2591�2612. doi:10.1111/j.1559-1816.1999. tb00127.x

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior & Human Decision Processes, 50(2), 179�211. doi:10.1016/0749-5978(91)90020-T

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22(5), 453�474. doi:10.1016/0022-1031(86)90045-4

Albarracı́n, D., Johnson, B. T., Fishbein, M., & Muellerleile, P. A. (2001). Theories of reasoned action and planned behavior as models of condom use: A meta-analysis. Psychological Bulletin, 127, 142�161. doi:10.1037/0033-2909.127.1.142

Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta- analytic review. British Journal of Social Psychology, 40(4), 471�499. doi:10.1348/0144 66601164939

Blue, C. L. (1995). The predictive capacity of the theory of reasoned action and the theory of planned behavior in exercise research: An integrated literature review. Research in Nursing & Health, 18, 105�121. doi:10.1002/nur.4770180205

Bryan, A., Fisher, J. D., & Fisher, W. A. (2002). Tests of the mediational role of preparatory safer sexual behavior in the context of the theory of planned behavior. Health Psychology, 21(1), 71�80. doi:10.1037/0278-6133.21.1.71

Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A review and avenues for further research. Journal of Applied Social Psychology, 28, 1429�1464. doi:10.1111/j.1559-1816.1998.tb01685.x

Cooke, R., & French, D. P. (2008). How well do the theory of reasoned action and theory of planned behaviour predict intentions and attendance at screening programmes? A meta- analysis. Psychology & Health, 23, 745�765. doi:10.1080/08870440701544437

Crosby, R. A., Kegler, M. C., & DiClemente, R. J. (2002). Understanding and applying theory in health promotion practice and research. In R. J. DiClemente, R. A. Crosby, & M. C. Kegler (Eds.), Emerging theories in health promotion practice and research: Strategies for improving public health (1st ed, pp. 1�15). San Francisco, CA: Jossey-Bass.

Crosby, R., & Noar, S. M. (2010). Theory development in health promotion: Are we there yet? Journal of Behavioral Medicine, 33, 259�263. doi:10.1007/s10865-010-9260-1

de Vries, H., Dijkstra, M., & Kuhlman, P. (1988). Self-efficacy: The third factor besides attitude and subjective norm as a predictor of behavioural intentions. Health Education Research, 3(3), 273�282. doi:10.1093/her/3.3.273

DiClemente, R. J., Crosby, R. A., & Kegler, M. C. (2002). Emerging theories in health promotion practice and research. San Francisco, CA: Jossey-Bass.

DiClemente, R. J., Crosby, R. A., & Kegler, M. C. (2009). Emerging theories in health promotion practice and research (2nd ed). San Francisco, CA: Jossey-Bass.

Dubin, R. (1978). Theory building. New York, NY: The Free Press. Fishbein, M. (1967). Attitude and the prediction of behavior. In M. Fishbein (Ed.), Readings

in attitude theory and measurement pp. 4477�4492). New York, NY: John Wiley. Fishbein, M. (2000). The role of theory in HIV prevention. AIDS Care, 12(3), 273�278.

doi:10.1080/09540120050042918 Fishbein, M. (2008). A reasoned action approach to health promotion. Medical Decision

Making, 28, 834�844. doi:10.1177/0272989x08326092 Fishbein, M. (2009). An integrative model for behavioral prediction and its application to

health promotion. In R. J. DiClemente, R. A. Crosby, & M. C. Kegler (Eds.), Emerging theories in health promotion practice and research (2nd ed., pp. 215�234). San Francisco, CA: Jossey-Bass.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.

50 K.J. Head and S.M. Noar

Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: The reasoned action approach. New York, NY: Psychology Press.

Fishbein, M., Bandura, A., Triandis, H. C., Kanfer, F. H., Becker, M. H., & Middlestadt, S. E. (1992). Factors influencing behavior and behavior change. Final report, theorists workshop. Bethesda, MD: National Institute of Mental Health.

Gibbons, F. X., Gerrard, M., Blanton, H., & Russell, D. W. (1998). Reasoned action and social reaction: Willingness and intention as independent predictors of health risk. Journal of Personality and Social Psychology, 74(5), 1164�1180. doi:10.1037/0022-3514.74.5.1164

Glanz, K., & Bishop, D. B. (2010). The role of behavioral science theory in development and implementation of public health interventions. Annual Review of Public Health, 31(1), 399�418. doi:10.1146/annurev.publhealth.012809.103604

Glanz, K., Rimer, B. K., & Viswanath, K. (2008). Theory, research, and practice in health behavior and health education. In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 23�40). San Francisco, CA: Jossey-Bass.

Godin, G., & Kok, G. (1996). The theory of planned behavior: A review of its applications to health-related behaviors. American Journal of Health Promotion, 11(2), 87�98. doi:10.4278/ 0890-1171-11.2.87

Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 38, pp. 69�119). San Diego, CA: Elsevier Academic Press.

Hardeman, W., Johnston, M., Johnston, D. W., Bonetti, D., Wareham, N. J., & Kinmonth, A. L. (2002). Application of the theory of planned behaviour in behaviour change interven- tions: A systematic review. Psychology & Health, 17, 123�158. doi:10.1080/088704402900 13644

Hochbaum, G. M., Sorenson, J. R., & Lorig, K. (1992). Theory in health education practice. Health Education & Behavior, 19, 295�313. doi:10.1177/109019819201900303

Kasprzyk, D., Montaño, D. E., & Fishbein, M. (1998). Application of an integrated behavioral model to predict condom use: A prospective study among high HIV risk groups. Journal of Applied Social Psychology, 28, 1557�1583. doi:10.1111/j.1559-1816.1998.tb01690.x

Kerlinger, F. N., & Lee, H. B. (1999). Foundations of behavioral research. Belmont, CA: Wadsworth.

Montaño, D. E., & Kasprzyk, D. (2008). Theory of reasoned action, theory of planned behavior, and the integrated behavioral model. In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 67�96). San Francisco, CA: Jossey-Bass.

Montaño, D. E., Selby, J. V., Somkin, C. P., Bhat, A., & Nadel, M. (2004). Acceptance of flexible sigmoidoscopy screening for colorectal cancer. Cancer Detection and Prevention, 28(1), 43�51. doi:10.1016/j.cdp.2003.11.005

Montaño, D. E., Thompson, B., Taylor, V. M., & Mahloch, J. (1997). Understanding mammography intention and utilization among women in an inner city public hospital clinic. Preventive Medicine, 26, 817�824. doi:10.1006/pmed.1997.0215

Noar, S. M., Chabot, M., & Zimmerman, R. S. (2008). Applying health behavior theory to multiple behavior change: Considerations and approaches. Preventive Medicine, 46(3), 275�280. doi:10.1016/j.ypmed.2007.08.001

Noar, S. M., & Mehrotra, P. (2011). Toward a new methodological paradigm for testing theories of health behavior and health behavior change. Patient Education and Counseling, 82, 468�474. http://dx.doi.org/10.1016/j.pec.2010.11.016.

Noar, S. M., & Zimmerman, R. S. (2005). Health behavior theory and cumulative knowledge regarding health behaviors: Are we moving in the right direction? Health Education Research, 20, 275�290. doi:10.1093/her/cyg113

Ogden, J. (2003). Some problems with social cognition models: A pragmatic and conceptual analysis. Health Psychology, 22(4), 424�428. doi:10.1037/0278-6133.22.4.424

Orbell, S., Hodgkins, S., & Sheeran, P. (1997). Implementation intentions and the theory of planned behavior. Personality and Social Psychology Bulletin, 23(9), 945�954. doi:10.1177/ 0146167297239004

Health Psychology Review 51

Painter, J. E., Borba, C. P. C., Hynes, M., Mays, D., & Glanz, K. (2008). The use of theory in health behavior research from 2000 to 2005: A systematic review. Annals of Behavioral Medicine, 35, 358�362. doi:10.1007/s12160-008-9042-y

Pierce, J. P., Choi, W. S., Gilpin, E. A., Farkas, A. J., & Merritt, R. K. (1996). Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychology, 15(5), 355�361. doi:10.1037/0278-6133.15.5.355

Prochaska, J. O., Wright, J. A., & Velicer, W. F. (2008). Evaluating theories of health behavior change: A hierarchy of criteria applied to the transtheoretical model. Applied Psychology: An International Review, 57, 561�588. doi:10.1111/j.1464-0597.2008.00345.x

Rise, J., Sheeran, P., & Hukkelberg, S. (2010). The role of self-identity in the theory of planned behavior: A meta-analysis. Journal of Applied Social Psychology, 40, 1085�1105. doi:10.1111/j.1559-1816.2010.00611.x

Rivis, A., & Sheeran, P. (2003). Descriptive norms as an additional predictor in the theory of planned behaviour: A meta-analysis. Current Psychology, 22(3), 218�233. doi:10.1007/ s12144-003-1018-2

Rivis, A., Sheeran, P., & Armitage, C. J. (2009). Expanding the affective and normative components of the theory of planned behavior: A meta-analysis of anticipated affect and moral norms. Journal of Applied Social Psychology, 39(12), 2985�3019. doi:10.1111/j.1559- 1816.2009.00558.x

Rothman, A. (2004). ‘‘Is there nothing more practical than a good theory?’’: Why innovations and advances in health behavior change will arise if interventions are used to test and refine theory. International Journal of Behavioral Nutrition and Physical Activity, 1(1), 1�7. doi:10.1186/1479-5868-1-11

Sandberg, T., & Conner, M. (2008). Anticipated regret as an additional predictor in the theory of planned behaviour: A meta-analysis. British Journal of Social Psychology, 47, 589�606. doi:10.1348/014466607x258704

Schwarzer, R. (2008). Modeling health behavior change: How to predict and modify the adoption and maintenance of health behaviors. Applied Psychology: An International Review, 57(1), 1�29. doi:10.1111/j.1464-0597.2007.00325.x

Sheeran, P. (2002). Intention*behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12(1), 1�36. doi:10.1080/14792772143000003

Sheeran, P., Abraham, C., & Orbell, S. (1999). Psychosocial correlates of heterosexual condom use: A meta-analysis. Psychological Bulletin, 125(1), 90�132. doi:10.1037/0033-2909.125.1. 90

Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research, 15(3), 325�343. doi:10.1086/209170

Sniehotta, F. (2009). An experimental test of the theory of planned behavior. Applied Psychology: Health & Well-Being, 1(2), 257�270. doi:10.1111/j.1758-0854.2009.01013.x

Sniehotta, F. F., Scholz, U., & Schwarzer, R. (2005). Bridging the intention�behaviour gap: Planning, self-efficacy, and action control in the adoption and maintenance of physical exercise. Psychology & Health, 20(2), 143�160. doi:10.1080/08870440512331317670

van Ryn, M., & Heaney, C. A. (1992). What’s the use of theory? Health Education Quarterly, 19, 315�330. doi:10.1177/109019819201900304

von Haeften, I., Fishbein, M., Kasprzyk, D., & Montano, D. (2001). Analyzing data to obtain information to design targeted interventions. Psychology. Health & Medicine, 6, 151�164. doi:10.1080/13548500120035409

Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132(2), 249�268. doi:10.1037/0033-2909.132.2.249

Weinstein, N. D. (2007). Misleading tests of health behavior theories. Annals of Behavioral Medicine, 33(1), 1�10. doi:10.1207/s15324796abm3301_1

Weinstein, N. D., Rothman, A. J., & Sutton, S. R. (1998). Stage theories of health behavior: Conceptual and methodological issues. Health Psychology, 17(3), 290�299. doi:10.1037/ 0278-6133.17.3.290

52 K.J. Head and S.M. Noar

Copyright of Health Psychology Review is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.