Theory-Based Health Behavior Change: Developing, Testing, and Applying Theories for Evidence-Based Interventions

APPLIED PSYCHOLOGY: AN INTERNATIONAL REVIEW, 2008,

57

(4), 698–716doi: 10.1111/j.1464-0597.2008.00339.x

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ,UK and 350 Main Street, Malden, MA 02148, USA.

Blackwell Publishing LtdOxford, UKAPPSApplied Psychology0269-994X1464-0597© International Association for Applied Psychology, 2008XXXOriginal ArticlesTHEORY-BASED HEALTH BEHAVIOR CHANGELIPPKE AND ZIEGELMANN

Theory-Based Health Behavior Change: Developing, Testing, and Applying Theories

for Evidence-Based Interventions

Sonia Lippke* and Jochen P. Ziegelmann

Freie Universität Berlin, Germany

Theories are needed to explain and predict health behavior, as well as for thedesign and evaluation of interventions. Although there has been a history ofdeveloping, testing, applying, and refining health behavior theories, debatesand limitations in evidence exist: The component of theories which, for example,predicts change should be better elaborated so that we can more easily under-stand what actually drives behavior change. Theories need to be empiricallytestable in two ways. Theories need to specify a set of changeable predictorsto describe, explain, and predict behavior change, and they should enable usto design an effective intervention that produces exactly those changes inbehavior that are predicted by the relevant theory. To make this possible,theories need to be specified in such a way that they can be rigorously testedand falsified. Moreover, for the design of theory-based interventions it mustbe possible to derive change techniques from the theory and to use them togenerate changes in behavior. Based on eight state-of-the-science articles thatmake conceptual and empirical contributions to the current debate on healthbehavior theories, various approaches are discussed to gain further insightsinto explaining and changing health behaviors and the iterative process oftheory development.

On ne peut se passer de théories d’une part pour expliquer et prédire lescomportements relatifs à la santé, et d’autre part pour programmer et évaluerles interventions. En dépit de l’histoire du développement, de la mise àl’épreuve, des applications et du perfectionnement des théories concernant lesconduites liées à la santé, des controverses et des carences dans les confirma-tions persistent: la partie des théories qui, par exemple, prédit les changementsdevrait être plus approfondie de façon à pouvoir mieux appréhender ce quidétermine réellement le changement de conduite. Les théories se doivent d’êtreempiriquement évaluées de deux façons. Elles se doivent de proposer unensemble de prédicteurs variables pour décrire, expliquer et prédire les

* Address for correspondence: Sonia Lippke, Health Psychology, Freie Universität Berlin,Habelschwerdter Allee 45 (PF 10), 14195 Berlin, Germany. Email: [email protected]

We are grateful to Ralf Schwarzer, Susan Michie, and Benjamin Schüz for helpful commentson a previous draft, and Jill Vyse for her editorial assistance on this manuscript.

THEORY-BASED HEALTH BEHAVIOR CHANGE

699

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

changements comportementaux et devraient être à même de nous permettrede programmer une intervention efficace qui produise au détail près lesmodifications comportementales annoncées par la théorie. Pour rendre celapossible, les théories doivent être conçues de telle façon qu’elles puissent êtreméthodiquement éprouvées et falsifiées. En outre, en vue de l’élaborationd’interventions sur des bases théoriques, il doit être possible de déduire de lathéorie des modifications techniques et de les exploiter pour provoquer deschangements de conduite. A partir de huit articles présentant l’état desconnaissances, on apporte une contribution conceptuelle et empirique àl’actuel débat sur les théories concernant le comportement relatif à la santé;on analyse des approches diversifiées pour avancer dans l’explication et lamodification des conduites liées à la santé et dans le lent processus dudéveloppement des théories.

INTRODUCTION

Health behaviors are of key importance in areas such as prevention, treat-ment, and rehabilitation. When investigating health behaviors, theories

1

areimperative to describe and understand processes, gain knowledge, and accu-mulate evidence. It is only from a sound evidence base that effective andinterpretable interventions to promote healthy lifestyles and to reduce riskbehaviors can be developed without reinventing the wheel.

Unfortunately, many publications on health behavior interventions donot provide a theoretical background or rationale. For example, Evers,Prochaska, Driskell, Cummins, Prochaska, and Velicer (2003) evaluated 37online interventions in terms of whether they were designed as theory-based.The majority (76%) did not base their intervention on any theory at all.Equivalent findings were reported by Dombrowski, Sniehotta, Avenel, andCoyne (2007); in a review of behavioral interventions, more than half (54%)of the included studies reported no theoretical background. Even in a meta-analysis of tailored interventions by Noar, Benac, and Harris (2007), 9 percent of the studies had no explicit theoretical basis. In future meta-analyses,however, we also might include those studies with no explicit theoreticalbasis as in this case we can rate the reported behavior change techniquesand thus are able to relate them to specific theories using the taxonomy ofAbraham and Michie (in press).

1. CONTINUUM MODELS AND STAGE MODELS

Health psychology and behavioral medicine have provided numerous theo-ries; rapid advances in theoretical approaches to health behavior change

1

In this article we refer to theories and models as interchangeable. For a detailed discussionof differences in definitions of theories and models and different types of models see Reese andOverton (1970).

700

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

have occurred within the last 20 years. Two groups of theories are typicallycategorised: (1) continuum models and (2) stage models (Weinstein, Roth-man, & Sutton, 1998).

Continuum models

try to identify predictors (such asintentions or attitudes) for behavior or behavior change. These variables aretypically combined into a linear prediction equation that places individualsalong a continuum of behavior likelihood, depending on their level of thevariables considered. If one or more of these determinants are strengthened,the likelihood of behavior or behavior change is also increased. Thisassumption implies that interventions to promote health behavior, based ona continuum model, focus on increasing all the associated variables in allthe individuals (Sutton, 2008).

On the contrary,

stage models

assume that behavior change takes place inseveral discrete stages. Depending on the stage a person belongs to, specificsocial-cognitive variables are more important than others. According tostage models, there is a special equation within every stage that predicts theprogression to the next stage, up to the final stage. Based on these assump-tions—which predictors are imperative at which stage—stage models providedifferent packages for people in different stages. Hence these programsare, according to researchers in favor of stage models, matched to the stage-specific needs of each participant (Weinstein et al., 1998). For stage-matchedinterventions to be successful, it is important that they be based on validstaging algorithms and that the gains of the stage-matched approach out-weigh the costs of tailoring.

2. USE OF THEORIES IN RESEARCH AND HEALTH PROMOTION

In their review of online interventions, Evers et al. (2003) identified 16 percent of the theory-based interventions as being based on stage theories (suchas the Transtheoretical Model, TTM) and 8 per cent on continuum theories(such as the Theory of Planned Behavior, TPB, or the Social CognitiveTheory, SCT). Dombrowski et al. (2007) reported that 21 per cent of thebehavioral interventions published used the TTM and 18 per cent the SCT.

Noar et al. (2007) investigated whether tailoring by theoretical constructs(e.g. only individuals high in intentions receive a planning intervention) wouldmoderate intervention effectiveness. The authors found that those interven-tions tailored only on behavior (e.g. earlier levels of the behavior;

r

+

= .026)were significantly less successful that those tailored on theoretical concepts(e.g. self-efficacy, intention, perceived susceptibility;

r

+

= .065). Also, tailor-ing only 0–3 theoretical concepts was significantly less effective (

r

+

= .062)than tailoring 4–5 theoretical concepts (

r

+

= .093; Noar et al., 2007).From a more applied viewpoint, theories are imperative for planning,

implementing, and evaluating health promotion (Bartholomew, Parcel,

THEORY-BASED HEALTH BEHAVIOR CHANGE

701

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Kok, & Gottlieb, 2006), for example, because they state which variables are(when) important to intervene on and which variables should exhibit aneffect of an intervention (for instance, a planning intervention should pro-duce more precise plans and more behavior but no effect on intention; seeLippke, Ziegelmann, & Schwarzer, 2004). Furthermore, understandingand effectively applying health behavior change theory has direct theoreticaland practical implications: From a more theoretical viewpoint, applying andtesting theories is imperative for theory development and refinement (andthis testing should be in both laboratory-based settings and real-life appliedsettings). This, in turn, is essential for guiding health promotion programs(Rothman, 2004).

In this context it is important to stress that there seem to be at least threeways in which behavior change interventions are developed:

a. strictly based on one theory,b. using several theoretically derived behavior change strategies which

do not necessarily originate from one theory (see Michie, Johnston,Francis, Hardemann, & Eccles, this issue; Bartholomew et al., 2006;Hardeman, Sutton, Griffin, Johnston, White, Wareham, & Kinmonth,2005),

c. not explicitly based on theory.

When the goal is to advance theoretical knowledge, one should aim forstrategy (a), whereas with the goal of maximising the effectiveness of inter-ventions one might be better aiming for strategy (b). However, basing anintervention (and evaluation) on no theory (c) needs to be avoided.

3. THIS SPECIAL ISSUE: CONTENT AND AIMS

This Special Issue gives insights into how to achieve the goal of advancingtheoretical knowledge. It contains state-of-the-science conceptual andempirical articles addressing different theoretical approaches with the aimof explaining and improving health behavior change.

Theory Development

The new theory for physical activity maintenance—namely the

PhysicalActivity Maintenance

(PAM) theory

—

by Nigg, Borrelli, Maddock, andDishman (this issue) is not only innovative because it mainly focuses onbehavior maintenance, but it is also one of the first attempts to include

stress

in behavior prediction. With that, the PAM makes an important con-tribution with regard to stress being an internal barrier (if individual stressresponses are evaluated) or an external barrier (if stress is conceptualised asa stimulus). Explicitly incorporating stress into a theory of health behavior

702

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

is very novel and might direct us toward new pathways for looking at healthbehaviors.

Other very valuable facets of Nigg et al.’s theory (this issue) include thespecial emphasis on stage-specific self-efficacy. Not only is self-efficacyconstructed as an important factor (as in several other models, such as theTPB; see Skår, Sniehotta, Araújo-Soares, & Molloy, this issue; the SCT; seePlotnikoff, Lippke, Courneya, Birkett, & Sigal, this issue; and the TTM;see Prochaska, Wright, & Velicer, this issue), but Nigg et al. also underlinethe stage-specific character. Stage-specific self-efficacy is crucial if one isstudying behavior maintenance and recovery as well as behavior initiation(see Schwarzer, Schüz, Ziegelmann, Lippke, Luszczynska, & Scholz, 2007;Scholz, Sniehotta, & Schwarzer, 2005; Rodgers, Hall, Blanchard, McAuley,& Munroe, 2002).

Criteria for Evaluation of Theories

Theory evaluation can be guided by the criteria presented by Prochaskaet al. (this issue). They propose a

hierarchy

of criteria (based on the recom-mendations of philosophers of science), for assessing the quality of theories:Clarity, Consistency, Parsimony, Testability, Empirical Adequacy, Produc-tivity, Generalisability, Integration, Utility, Practical Usefulness and Impact(see Prochaska et al.’s Table 1).

The authors demonstrate this hierarchy with the example of the

Trans-theoretical Model

(TTM) and review research on the TTM in the light ofthe proposed criteria. However, these criteria can also be employed in thecontext of other theories. Prochaska et al. (this issue) make some very impor-tant points. For example, the theory proposed by the original author mightbe interpreted differently by other scientists (called by Prochaska et al.“someone else’s derivations of the theory”, p. 570). Taking a constructivistview, we would expect that a theory will yield different interpretations bydifferent people (Stoker, 1996). In principle, these different or alternativeinterpretations or derivations from a theory can be beneficial for theprogress of science, but there is also a challenge. For example, the deriva-tions may misrepresent the theory or omit key parts of it. Thus, when com-paring the effectiveness of theories (e.g. in reviews or meta-analyses) orwhen calling for the rejection or refinement of theories, we need to establishwhether or not the theories in question were tested properly (for a detaileddiscussion of state-of-the-science theory testing see Wittmann & Klumb,2006). This appears as a dilemma especially when theories should not beexpected to remain constant over time: Theory refinement starts with aninitially consistent theory. Over time the theory needs to be modified oreven rejected when new data demonstrate the need for refinements (Kuhn,1970). What can be learned from this is that “theory development should

THEORY-BASED HEALTH BEHAVIOR CHANGE

703

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

be an iterative process with theory driving empirical studies and empiricaldata driving theory refinement” (Prochaska et al., this issue, p. 569).

One general component described by Prochaska et al. is

Parsimony

, anaspect often disregarded. For instance, in the future Nigg et al.’s new theoryneeds to be evaluated as to whether or not it meets this criterion. Anotherpotential area where parsimony could be considered is when assessing Skåret al.’s results: The inclusion of additional predictors in multiple regressionanalyses increases the variance explained in the criterion, whenever thepredictor is correlated with the criterion. Judging from the principle ofparsimony, the question is whether the variance explained gained by addingconstructs outweighs the violation of the parsimony principle. Thus, thegains of adding constructs have to be evaluated. We propose a series ofsteps:

First, convergent, discriminant, and predictive validity for the conceptsneed to be demonstrated. Second, the concepts should be theoreticallyembedded; if this recommendation is not met, one may end up with a simplelist of determinants, which will make it difficult to design interventions. Thisleads to the third point: The aim is not only to explain behavior but also tochange it. A recent conceptual addition to health behavior change is theconcept of coping planning, which offers both a better explanation of healthbehaviors and also more effective interventions (Sniehotta, Schwarzer,Scholz, & Schüz, 2005; Ziegelmann, Lippke, & Schwarzer, 2006a).

Testing Theories by Testing Predictions

A further criterion pointed out by Prochaska et al. (this issue) is

Testabilityor Falsifiability

(Popper, 1959). For instance, every theory proposes relation-ships between constructs or sometimes even between specific effect sizes.Both of these can be empirically tested. Studies that demonstrate thisapproach are presented, e.g. by Velicer, Cumming, Fava, Rossi, Prochaska,and Johnson (this issue) or Plotnikoff et al. (this issue).

Velicer et al. (this issue) tested quantitative effect size predictions basedon the TTM. They examined whether the effect sizes found in an empiricalstudy were within the range of the effect sizes derived from theoreticalpredictions. The results are impressive: 11 of the 15 confidence intervalsincluded the predicted value. This was validated with both the 95 per centand the 99 per cent intervals. The authors interpreted their results as overallsupport for the theoretical model.

Velicer et al. (this issue) predicted effect sizes over three stages: Precon-templation, Contemplation, and Preparation (which are only the first threestages theorised by the TTM). However, one could also look at adjacentstages and trends across stages. Discontinuity patterns might emerge, forexample, that variable A increases from stage 1 to stage 2, whereas from

704

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

stage 2 to stage 3 it decreases. By contrast, variable B decreases from stage1 to stage 2 but remains the same in stages 2 and 3 (Sutton, 2000). Theauthors propose that one should “tease out the transitions that might bemasked by the curvilinear nature of some of the variables like the processesacross all five stages” (Velicer et al., this issue, p. 605).

Theory-Based Interventions

Bandura’s

Social Cognitive Theory

(SCT) proposes an interplay of thedeterminants of goals and behavior (Bandura, 2004). One of the first studieswhich explicitly tested this proposed structure was carried out by Plotnikoffet al. (this issue). Invariances between individuals with different types ofdiabetes were investigated and descriptively compared with non-diabeticpopulations. These kinds of analyses are imperative not only for theorytesting and potential theory refinements but also for intervention guidance.Such an intervention was utilised and tested by Luszczynska and Tryburcy(this issue). Their SCT-based intervention (delivered by e-mail) targetedself-efficacy. Thus it makes a contribution to the growing field of eHealthinterventions, which has its unique challenges and opportunities (Ahern,2007). Further, the authors tested whether diabetes/CVD was a moderatorfor intervention effectiveness. Self-efficacy was found to operate as amediator, and diabetes/CVD surfaced as the proposed moderator forintervention success.

Techniques in Effective Behavior Change Interventions

A systematic collection of

techniques

which could be effective componentsof interventions is presented by Michie et al. (this issue). Prochaska et al.(this issue) point out that the field of psychotherapy was fragmented from150 theories in the 1970s to over 300 theories in the 1980s. Such fragmenta-tion can no longer be considered as scientific progress, but is rather harmfulif we want to build upon existing knowledge. Thus we need to reverse thistrend and derive from the abundant literature the core techniques and con-structs. This is one achievement of the article contributed by Michie et al.(this issue), which cataloged 23 techniques judged by expert consensus to beeffective in changing one or more of 11 domains or theoretical behavioraldeterminants.

The article by Michie et al. (this issue) is very innovative in that it is amixture of review, expert brainstorming, and rating. This approach can alsobe seen as an alternative to a Delphi study (see De Vet, Brug, De Nooijer,Dijkstra, & De Vries, 2005). It describes preliminary work for constructinga taxonomy of behavior change techniques to be used for developing theory-based behavior change interventions (p. 665). This sets the first stage for

THEORY-BASED HEALTH BEHAVIOR CHANGE

705

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

further elaboration. The authors identify as forthcoming steps “dealing withissues such as the further identification of techniques, completing techniquedefinitions, and the elimination of overlap between techniques”. Theauthors also question whether a larger sample of experts would deducecomparable results. This remains an empirical question. Different resultsmight emerge when the experts come from other theoretical backgrounds ordisciplines, e.g. stage theories (see Prochaska et al., this issue, or Velicer etal., this issue) or environmental interventions (see Nigg et al., this issue).

Future research should also investigate

which

technique is most appropriate

when

or

for whom

: The question would be whether techniques could betailored/matched to particular target individuals, e.g. based on the stage anindividual is in (see Prochaska et al., this issue, or Velicer et al., this issue).

While this first very helpful attempt by Michie et al. (this issue) can befurther developed, it can already be used to develop and describe inter-ventions which draw from various theoretical backgrounds. For instance, asalready discussed above, when deciding whether or not an intervention istheory-based, we might rate the intervention according to the presence ofthe 11 behavioral determinants that were derived by Michie et al. (this issue)and/or Michie, Johnston, Abraham, Lawton, Parker, and Walker (2005). Inother words, one might detect theoretical aspects in interventions whichappear at first atheoretical or were not theoretically driven. The constructdomains generated by Michie et al. (this issue) are useful when it comesto reporting components of complex interventions, such as the one byMikolajczak, Kok, and Hospers (this issue).

Procedure for Designing Interventions

The article by Mikolajczak et al. (this issue) supplements the other articlesby showing in a detailed and structured way how an intervention can bedesigned: The authors describe the

Intervention Mapping

procedure and thedevelopment of a concrete intervention on this basis. The article demon-strates both “bottom-up” and “top-down” approaches. First, the needs wereassessed; second, the results were organised in a matrix of change objectives;and in a third step the theory-based behavior change strategies werematched to the change objectives. This is where categories of change tech-niques as outlined in the Michie et al. article are of key relevance: Onlywhen we have a pool of validated behavior change strategies can such anapproach be successful. What follows from this is that interventions basedon the Intervention Mapping procedure do not have an exclusive focus onone theory, but rather can select the most promising theoretically derivedstrategies for a given problem.

In addition, this study demonstrated how the

Internet

can be used as anadvantageous setting and how modern technology, such as interaction with

706

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

virtual relational agents and e-mail feedback, can produce motivational andvolitional effects. Here the article by Mikolajczak et al. (this issue) relatesto the intervention by Luszczynska and Tryburcy (this issue). They alsoused an e-mail intervention, with strong effects for physical activity promo-tion in high risk individuals.

For most theories, there is now accumulated evidence about theory-basedchanges of intentions and behavior (e.g. Webb & Sheeran, 2006: Meta-analysiscomparing TPB, PMT, and stage theories). For future meta-analyses itwould be helpful to report the effect sizes for the different stage theories andhybrid models separately. Despite these advances, however, we have onlypreliminary support for strategies to target the variables in the theories.

4. THEORY REFINEMENT AND THEORY INTEGRATION

Findings from studies showing the shortcomings of theories might guidetheory refinement. One example of such a procedure is the development ofthe TPB, which was created by adding PBC to the TRA. There are manyother examples where concepts were added to theories (e.g. in most healthbehavior change theories, the construct implementation intention/plan isnow included, either as an explicit addition or as an add-on, outside of thetheoretical framework, to maximise intervention effectiveness).

Another strategy is to combine stage models with continuum models. Oneapproach is to investigate the means in variables derived from continuumtheories, such as the TPB, across the different stage groups. Another strategymight be to test the architecture of a linear model within each stage groupseparately (see Biddle, Hagger, Chatzisarantis, & Lippke, 2007). To investi-gate stage as a moderator and to examine stage-dependent processes hasbeen shown to be fruitful (Lippke, Nigg, & Maddock, 2007). These pro-cesses are analogous to the assumption of most social-cognitive models withintention formation, action plans, and behavior change. With this strategyone might also test stage-specific prediction patterns by testing whether,depending on the stage, different social-cognitive variables are more or lessinfluential (Weinstein et al., 1998).

Thus one approach in the future might be not necessarily to compare andseparate the theories, but instead to combine theories and behavior changestrategies to make the best use of existing knowledge (Michie et al., thisissue). However, researchers and practitioners have to avoid overloadingtheories and interventions; this can be achieved by matching change tech-niques to the change objectives (Mikolajczak et al., this issue). Theorieshave to be comprehensive but also parsimonious, in other words, clear andsimple (Michie et al., 2005; Prochaska et al., this issue).

One model that aims to achieve this by integrating a linear predictionstructure with qualitatively different stages is the

Health Action Process

THEORY-BASED HEALTH BEHAVIOR CHANGE

707

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Approach

(HAPA; Schwarzer, 2008). Depending on the stage of the targetgroup, different factors are seen as important. In intentional individuals, forexample, plans are crucial for translating intentions into successful behavior(Lippke, Ziegelmann, & Schwarzer, 2005; Ziegelmann et al., 2006a). Riskperception is necessary only for goal-setting (non-intentional stage); but notin the intentional stage (i.e. when planning and initiating the behavior).Other factors are more generic, such as self-efficacy which is imperative inall processes of behavior change: Perceived self-efficacy appears to beinstrumental in the initial process of intention formation, and it continuesto facilitate planning, initiative, maintenance of behavior, as well as recoveryfrom setbacks. Although it is considered to be important across all stagesof health behavior change, it is best conceptualised in a stage-sensitive manner(Nigg et al., this issue; Scholz et al., 2005). To summarise, throughout theprocess of behavior change, perceived self-efficacy facilitates mental andbehavioral components that are inherent in goal-setting and goal pursuit(Schwarzer et al., 2007; Lippke et al., 2005).

In this way the HAPA can be understood as a hybrid model (Biddleet al., 2007; Schwarzer, 2008). The HAPA was tested for different healthbehaviors and proved to be applicable to a variety of behaviors such asphysical activity, nutrition, seat belt use, and dental hygiene (Renner,Spivak, Kwon, & Schwarzer, 2007; Schüz, Sniehotta, & Schwarzer, 2007;Schwarzer et al., 2007). The HAPA proved to be effective not only inexplaining health behavior change, but also in experimentally changing keyconstructs such as action planning and coping planning (Schüz et al., 2007;Sniehotta et al., 2005; Ziegelmann et al., 2006a). Also, it enables us to testinterventions in a stage-specific way (Lippke et al., 2004). The effectivenessof the HAPA when used as the basis for designing intervention still needsto be more systematically evaluated in the future. The work by Luszczynskaand Tryburcy (this issue) as well as by Nigg et al. (this issue) on phase-specific self-efficacy underlines the potential of the HAPA as a theory thatguides health behavior change interventions.

5. CONCEPTUAL OVERLAP OF THEORIES

The articles in this Special Issue answer various questions, such as whichdeterminants and techniques of behavior change are central (e.g. self-efficacy; see Luszczynska & Tryburcy; Michie et al.; Plotnikoff et al.; Skåret al.; and planning; see Michie et al., all this issue). However, at the sametime further questions arise, for example, whether determinants and tech-niques are actually equivalent although they have different labels (Bandura,2004; Schüz, Sniehotta, Mallach, Wiedemann, & Schwarzer, in press;Weinstein, 1993; see also Table 1).

708

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

The correspondence of the TRA and the TPB is obvious. However, thereare also similarities between other theories, for example the TPB and theSCT, due to the relationships between PBC and self-efficacy (Bandura,2004). PBC and self-efficacy may not be clearly separable (Schwarzer, 2008).Another example includes attitudes in the TPB, outcome expectancies in theSCT, and the decisional balance (pros and cons) in the TTM (Williams,Anderson, & Winett, 2005).

Similarly, stage theories share assumptions. It has been argued that thestages can be subsumed under three stages—stage 1, in which individualsdo not yet intend to act, stage 2, in which individuals intend to change buthave not yet changed their behavior, and stage 3, in which individuals actupon their intentions (Weinstein et al., 1998). In a recent study, it has beendemonstrated that predictors derived from common constructs in social-cognitive theories can predict transitions between these shared stages ofchange (Schüz et al., in press).

6. COMPARING THEORIES

A number of open questions remain: Is it the case that theories have morein common than appears at first glance? Do the different labels maskwhat they have in common? Which theories are the better ones? On thebasis of the articles included in this Special Issue very little can be said about

TABLE 1Similarities of Constructs in Different Theories

Social Cognitive determinants of health behaviors

Self-efficacy(perceived behavioral control)

Outcome expectancies

(Behavioral beliefs,subjective norms)

Risk perception

Intention(goals)

Planning (implementation

intentions)

HBM – ✓ ✓ – –TRA – ✓ – ✓ –TPB ✓ ✓ – ✓ –SCT ✓ ✓ – ✓ –PMT ✓ ✓ ✓ ✓ –HAPA ✓ ✓ ✓ ✓ ✓

Note: HBM, Health Belief Model; TRA, Theory of Reasoned Action; TPB, Theory of Planned Behavior;SCT, Social Cognitive Theory; PMT, Protection Motivation Theory; HAPA, Health Action ProcessApproach.✓ the theory in question predicts the determinant as important in the process of behavior change (adaptedfrom Bandura, 2004, S. 147); determinants may be named slightly differently, depending on the theory.

THEORY-BASED HEALTH BEHAVIOR CHANGE

709

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

which theory is the best one. To achieve this would first require that wedetermine the criteria for comparing theories. Maybe one theory is betterfor a special population, another theory is more appropriate for a singlebehavior but not for other behaviors or in changing multiple behaviors.Or is there a theory that is best for explaining different behaviors inseveral different populations? Meta-analyses (e.g. Gollwitzer & Sheeran,2006; Webb & Sheeran, 2006), systematic reviews (e.g. Maes & Karoly,2005; Williams et al., 2005), or comments (e.g. Leventhal & Mora, 2005;Leventhal, Musumeci, & Contrada, 2007) can give insights into andanswers to these questions.

A crucial problem in the comparison of health behavior theories remainsthat theories do not make explicit assumptions about the relative impact ofpredictors on intentions, behavior, or other dependent variables, let alonegive quantifications of these impacts (Weinstein & Rothman, 2005). This leavesmuch room for subjective interpretation of data when comparing theories.

In this line, Prochaska et al. (this issue) argue that ideally they wouldconsider the fit of their model in competition with an alternative theory,rather than in absolute terms (Weinstein, 1993). However, very few studieshave empirically compared different theories, such as continuum models(Garcia & Mann, 2003) or stage models (cf. Lippke, Sniehotta, & Luszczynska,2005). Lippke et al. (2005) underlined the advantage of a more parsimoniousstage model. Garcia and Mann (2003) provided support for the stage 1 ofthe HAPA in comparison with the Theory of Reasoned Action (TRA), theTPB, and the Health Belief Model (HBM). Since we not only aim to explainhealth behavior change (which is expressed in a good fit of a given model),but also aim to change behaviors—comparing effect size in behaviorchange—meta-analyses seem to be the gold standard (e.g. Webb & Sheeran,2006). When making such an analysis we need to take into account variousmoderators (e.g. time interval of follow-up, delivery method, type of controlgroup, type of sample) in order to identify the subsamples, behaviors, orcontexts for which research findings are homogeneous, i.e. for which aparticular theory can be regarded as evidence based.

Skår et al. (this issue) made a point of demonstrating something com-parable: They tested the TPB in contrast to novel developments. They foundthat intention certainty moderates the intention–behavior relationship, andthat the predictive power of the TPB can be enhanced if intention certaintyand congruence are taken into consideration. Testing the effect of pastbehavior, the authors found comparable results to studies which tested stagemodels: Past behavior was the most powerful predictor of subsequentbehavior. Skår et al. attributed this well-known phenomenon to the factthat past behavior might be a function of similar cognitions which indi-viduals developed and maintained previously (cf. Weinstein, 2007). Thesecognitions might stay constant if past behavior was controlled—which

710

LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

could also be done by including

psychological stage

in the predictions (cf.Lippke et al., 2007; Lippke, Ziegelmann, & Schwarzer, 2005).

7. APPLIED PSYCHOLOGY: DIFFERENT DOMAINS CONTRIBUTE DIFFERENT THEORIES

Twenty years ago, when health psychology was still in its infancy, Weinman(1987) described this field as drawing upon theories derived from main-stream psychology. We need to continue taking advantage of the fact thatother psychological disciplines (such as social psychology, clinical psychology,developmental psychology, and personality psychology) have a key interestin self-regulation.

Health psychology

should use this wealth of knowledge to further developthe field, theoretically and in its applications, including behavior change.Theories such as the model of Selection, Optimisation, and Compensation(SOC; Freund & Baltes, 1998), the Lifespan Theory of Control (LTC;Wrosch, Schulz, & Heckhausen, 2005), the Socio-emotional SelectivityTheory (SST; Löckenhoff & Carstensen, 2004), or the Reflective-ImpulsiveModel (RIM; Strack & Deutsch, 2004) have the potential to bring newperspectives for effective health psychology interventions.

In most health behavior interventions we focus on the reflective route tobehavior. However, the RIM approach suggests that we should also bearin mind the impulsive route to behavior when we design interventions. TheSOC theory can guide us, for example, in how we can augment planninginterventions in such a way that individuals plan their behavior change in amore detailed way (Ziegelmann et al., 2006a). In addition, SOC theory, likeSST, guides us in designing age-specific interventions. This is of centralimportance, as health behavior change and the outcomes of health behaviorchange always take place in a life-span context (Wurm, Tesch-Römer, &Tomasik, 2007), and not only chronological age plays a central role but alsofuture time perspective (Ziegelmann, Lippke, & Schwarzer, 2006b), as isalso reflected by temporal self-regulation theory (Hall & Fong, 2007).

8. HEALTHY LIFESTYLES: CHANGING MULTIPLE BEHAVIORS

Prochaska et al.’s criterion for evaluating theories,

Impactfulness

, takes intoconsideration the effectiveness of managing multiple behaviors. This is notonly important from a theoretical point of view. Populations with

multiplebehavior risks

are at greatest risk for chronic disease and premature death,and also account for a disproportionate percentage of health care costs.The theory is expected to generate interventions which change unhealthybehaviors in a heterogeneous population. Prochaska et al. (this issue) assume

THEORY-BASED HEALTH BEHAVIOR CHANGE

711

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

that Impactfulness is the most demanding and most valuable criterion, as itrequires effective interventions for multiple behaviors in entire populations.Such interventions yield the largest effect that theory-based interventionsprovide for prevention, rehabilitation, and disease management and forreducing health care costs (Prochaska et al., this issue).

9. METHODOLOGICAL ISSUES IN EVALUATING THEORY-BASED INTERVENTIONS

In addition to the above-mentioned issues, some data analyses are alsodemonstrated in this Special Issue which are current state-of-the-science.These include statistical moderation (Skår et al., this issue), moderation andmediation analyses (Luszczynska & Tryburcy, this issue) and Multi-groupStructural Equation Modeling (Plotnikoff et al., this issue). A new theoryparticularly for physical activity maintenance was outlined for furtherempirical investigations (see Nigg et al., this issue). Notably, this SpecialIssue covers very different subjects and perspectives. Innovative ideas arepresented and potential future trends are hopefully stimulated.

While there is still a clear need for theory development, this need shouldnot be the only focus of future research activities. The advancement ofresearch methods, for example, allows for more in-depth analyses of indi-vidual change processes within existing theories, which helps elucidate theprocess of behavior change (e.g. Scholz, Nagy, Schüz, & Ziegelmann, inpress); another may be the more frequent use of experimental or moderatedmediation analyses which help disentangle complex mediation relations(Reuter, Ziegelmann, Wiedemann, & Lippke, 2008; Wiedemann, Schüz,Sniehotta, Scholz, & Schwarzer, in press).

Another impact facet emphasised by Prochaska et al. (this issue) is theratio of efficacy rate times and participation rate. On the one hand, if onlyevaluated on the basis of the efficacy criterion, a treatment with a 30 percent adherence rate would be appreciated as 50 per cent better than anintervention that resulted in 20 per cent adherence. On the other hand, ifthe more efficacious intervention reached only every tenth person of a targetpopulation, the impact would only be 3 per cent (30% × 10%). If the lessefficacious intervention reached three out of four individuals in the popula-tion (75%), it would have an impact of 15 per cent. In terms of Impactful-ness, the latter treatment can be seen as being five times better than themore efficacious intervention (Prochaska et al., this issue).

10. CONCLUSIONS

In the future, investigations need to state which theory is the most appro-priate for a particular research question or intervention strategy. However,

712 LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

with the current overlap in constructs it may not be easy to compare thetheories because of difficulties in testing the discriminant validity of theconstructs or how the constructs extend beyond each other. Other and bettermethods to assess the appropriateness of a theory could be to evaluate:

a. whether the theory provides a high match of theory-driven hypothesesand findings (see Lippke et al., 2007; Velicer et al., this issue),

b. how generalisable the theory is to different populations, cultural sub-groups and circumstances (cf. Ziegelmann & Lippke, 2007),

c. how good each theory is in relation to the criteria summarised byProchaska et al. (this issue),

d. its effect size when changing behavior (cf. Webb & Sheeran, 2006) andwhether it is clinically or socially meaningful,

e. whether the findings of correlational tests can be replicated in experi-mental studies, as the magnitude of effects that result from tworesearch designs can differ substantially (Michie, Rothman, & Shee-ran, 2007; Weinstein, 2007),

f. the public health impact, if the theory is tested in a public healthintervention (Nigg & Jordan, 2005), including how capable the theoryis of changing multiple behaviors, and in an entire population.

To summarise, theory-based interventions are imperative for successfulhealth behavior promotion. We will advance our field only when we knowhow to translate theory effectively into practice. In addition, we shouldsystematically use the knowledge gained from the results of theory-basedintervention research to refine our theories in an iterative process. Here it isnot only important that this iterative process involves feedback and forwardprocesses between theory and practice, but we also should be aware of theprinciples of state-of-the-science theory testing (Wittmann & Klumb, 2006).Finally we should not only be concerned with theory refinement but shouldbe aware of new analytical approaches when designing our studies andwhen (re-)analysing our data (Singer & Willett, 2003).

REFERENCES

Abraham, C., & Michie, S. (in press). A taxonomy of behavior change techniquesused in interventions. Health Psychology.

Ahern, D.K. (2007). Challenges and opportunities of eHealth research. AmericanJournal of Preventive Medicine, 32, S75–S82.

Bandura, A. (2004). Health promotion by social-cognitive means. Health Educationand Behavior, 31(2), 143–164.

Bartholomew, L.K., Parcel, G.S., Kok, G., & Gottlieb, N.H. (Eds.) (2006). Planninghealth promotion programs: An intervention mapping approach (2nd edn.). SanFrancisco, CA: Jossey-Bass.

THEORY-BASED HEALTH BEHAVIOR CHANGE 713

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Biddle, S.J.H., Hagger, M.S., Chatzisarantis, N.L.D., & Lippke, S. (2007). Theoret-ical frameworks in exercise psychology. In G. Tenenbaum & R.C. Eklund (Eds.),Handbook of sport psychology (3rd edn., pp. 537–559). New York: Wiley.

De Vet, E., Brug, J., De Nooijer, J., Dijkstra, A., & De Vries, N.K. (2005). Deter-minants of forward stage transitions: A Delphi study. Health Education Research,20(2), 195–205.

Dombrowski, S.U., Sniehotta, F.F., Avenel, A., & Coyne, J.C. (2007). Towards acumulative science of behaviour change: Do current conduct and reporting ofbehavioural interventions fall short of best practice? Psychology and Health, 22,869–874.

Evers, K.E., Prochaska, J.M., Driskell, M.M., Cummins, C.O., Prochaska, J.O., &Velicer, W.F. (2003). Strengths and weaknesses of health behavior change pro-grams on the internet. Journal of Health Psychology, 8, 63–70.

Freund, A.M., & Baltes, P.B. (1998). Selection, optimization, and compensation asstrategies of life management: Correlations with subjective indicators of success-ful aging. Psychology and Aging, 13, 531–543.

Garcia, K., & Mann, T. (2003). From “I wish” to “I will”: Social-cognitive predic-tors of behavioral intentions. Journal of Health Psychology, 8(3), 347–360.

Gollwitzer, P.M., & Sheeran, P. (2006). Implementation intentions and goal achieve-ment: A meta-analysis of effects and processes. Advances in Experimental SocialPsychology, 38, 69–119.

Hall, P.A., & Fong, G.T. (2007). Temporal self-regulation theory: A model forindividual health behavior. Health Psychology Review, 1, 6–52.

Hardeman, W., Sutton, S., Griffin, S., Johnston, M., White, A., Wareham, N.J., &Kinmonth, A.L. (2005). A causal modelling approach to the development oftheory-based behaviour change programmes for trial evaluation. Health EducationResearch, 20, 676–687.

Kuhn, T.S. (1970). The structure of scientific revolutions (rev. edn.). Chicago, IL:University of Chicago Press.

Leventhal, H., & Mora, P.A. (2005). Is there a science of the processes underlyinghealth and illness behaviors? A comment on Maes and Karoly. Applied Psycho-logy: An International Review, 54, 255–266.

Leventhal, H., Musumeci, T.J., & Contrada, R.J. (2007). Theory, translation, andevidence-based practice. Psychology and Health, 22, 381–386.

Lippke, S., Nigg, C.R., & Maddock, J.E. (2007). The Theory of Planned Behaviorwithin the stages of the Transtheoretical Model—Latent structural modelingof stage-specific prediction patterns in physical activity. Structural EquationModeling: A Multidisciplinary Journal, 14(4), 649–670.

Lippke, S., Sniehotta, F.F., & Luszczynska, A. (2005). Social cognitions across thestages of behavior change: A comparison of two stage models. Polish Psycholog-ical Bulletin, 36(1), 43–50.

Lippke, S., Ziegelmann, J.P., & Schwarzer, R. (2004). Initiation and maintenance ofphysical exercise: Stage-specific effects of a planning intervention. Research inSports Medicine, 12, 221–240.

Lippke, S., Ziegelmann, J.P., & Schwarzer, R. (2005). Stage-specific adoption andmaintenance of physical activity: Testing a three-stage model. Psychology ofSport and Exercise, 6, 585–603.

714 LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Löckenhoff, C.E., & Carstensen, L.L. (2004). Socioemotional selectivity theory,aging, and health: The increasingly delicate balance between regulating emotionsand making tough choices. Journal of Personality, 72, 1395–1424.

Luszsczynska, A., & Tryburcy, M. (2008). Effects of a self-efficacy intervention onphysical activity: The moderating role of diabetes and cardiovascular diseases.Applied Psychology: An International Review, 57(4), 644–659.

Maes, S., & Karoly, P. (2005). Self-regulation assessment and intervention in phys-ical health and illness: A review. Applied Psychology: An International Review, 54,267–311.

Michie, S., Johnston, M., Abraham, C., Lawton, R., Parker, D., & Walker, A.(2005). Making psychological theory useful for implementing evidence basedpractice: A consensus approach. Quality & Safety in Health Care, 14, 26–33.

Michie, S., Johnston, M., Francis, J., Hardemann, W., & Eccles, M. (2008). Fromtheory to intervention: Mapping theoretically derived behavioural determinantsto behaviour change techniques. Applied Psychology: An International Review,57(4), 660–680.

Michie, S., Rothman, A., & Sheeran, P. (2007). Current issues and new directionsin psychology and health: Advancing the science of behaviour change. Psycho-logy and Health, 22, 249–253.

Mikolajczak, J., Kok, G., & Hospers, H.J. (2008). Queermasters: Developing a the-ory-and evidence-based Internet HIV-prevention intervention to promote HIVtesting among men who have sex with men (MSM). Applied Psychology: An Inter-national Review, 57(4), 681–697.

Nigg, C.R., Borrelli, B., Maddock, J., & Dishman, R.K. (2008). A theory of physicalactivity maintenance. Applied Psychology: An International Review, 57(4), 544–560.

Nigg, C.R., & Jordan, P.J. (2005). Commentary: It’s a difference of opinion thatmakes a horserace . . . Health Education Research, 20(3), 291–293.

Noar, S.M., Benac, C., & Harris, M. (2007). Does tailoring matter? Meta-analyticreview of tailored print health behavior change interventions. PsychologicalBulletin, 133(4), 673–693.

Plotnikoff, R.C., Lippke, S., Courneya, K.S., Birkett, N., & Sigal, R.J. (2008). Phys-ical activity and Social Cognitive Theory: A test in a population sample of adultswith type 1 or type 2 diabetes. Applied Psychology: An International Review,57(4), 628–643.

Popper, K.R. (1959). The logic of scientific discovery. New York: Routledge.Prochaska, J.O., Wright, J., & 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(4), 561–588.

Reese, H.W., & Overton, W.F. (1970). Models of development and theories of develop-ment. In L.R. Goulet & P.B. Baltes (Eds.), Life-span developmental psychology:Research and theory (pp. 115–145). New York: Academic Press.

Renner, B., Spivak, Y., Kwon, S., & Schwarzer, R. (2007). Does age make a differ-ence? Predicting physical activity of South Koreans. Psychology and Aging, 22,482–493.

Reuter, T., Ziegelmann, J.P., Wiedemann, A.U., & Lippke, S. (2008). Dietary planningas a mediator of the intention–behavior relation: An experimental-causal-chaindesign. Applied Psychology: An International Review, 57(Suppl. 1), 194–207.

THEORY-BASED HEALTH BEHAVIOR CHANGE 715

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Rodgers, W.M., Hall, C.R., Blanchard, C.M., McAuley, E., & Munroe, K.J. (2002).Task and scheduling self-efficacy as predictors of exercise behavior. Psychologyand Health, 17(4), 405–416.

Rothman, A.J. (2004). “Is there nothing more practical than a good theory?”: Whyinnovations and advances in health behavior change will arise if interventions areused to test and refine theory. International Journal of Behavioral Nutrition andPhysical Activity, 1, 11.

Scholz, U., Nagy, G., Schüz, B., & Ziegelmann, J.P. (in press). The role of motiva-tional and volitional factors for self-regulated training behaviour: Associationson the between and within person level. British Journal of Social Psychology.

Scholz, U., Sniehotta, F.F., & Schwarzer, R. (2005). Predicting physical exercise incardiac rehabilitation: The role of phase-specific self-efficacy beliefs. Journal ofSport and Exercise Psychology, 27, 135–151.

Schüz, B., Sniehotta, F.F., Mallach, N., Wiedemann, A.U., & Schwarzer, R. (inpress). Predicting transitions from preintentional, intentional and actional stagesof change: Adherence to oral self-care recommendations. Health Education Research.

Schüz, B., Sniehotta, F.F., & Schwarzer, R. (2007). Stage-specific effects of anaction control intervention on dental flossing. Health Education Research, 22,332–341.

Schwarzer, R. (2008). Modeling health behavior change: How to predict and modifythe adoption and maintenance of health behaviors. Applied Psychology: An Inter-national Review, 57(1), 1–29.

Schwarzer, R., Schüz, B., Ziegelmann, J.P., Lippke, S., Luszczynska, A., & Scholz,U. (2007). Adoption and maintenance of four health behaviors: Theory-guidedlongitudinal studies on dental flossing, seat belt use, dietary behavior, and phys-ical activity. Annals of Behavioral Medicine, 33, 156–166.

Singer, J.D., & Willett, J.B. (2003). Applied longitudinal data analysis: Modelingchange and event occurrence. Oxford: Oxford University Press.

Skår, S., Sniehotta, F.F., Araújo-Soares, V., & Molloy, G.J. (2008). Prediction ofbehaviour vs. prediction of behaviour change: The role of motivational modera-tors in the Theory of Planned Behaviour. Applied Psychology: An InternationalReview, 57(4), 609–627.

Sniehotta, F.F., Schwarzer, R., Scholz, U., & Schüz, B. (2005). Action plans andcoping plans for long-term lifestyle change: Theory and assessment. EuropeanJournal of Social Psychology, 35, 565–576.

Stoker, R. (1996). Constructivism—paradigm, theory, or bundle of techniques? Edu-cational and Child Psychology, 13, 58–63.

Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of socialbehavior. Personality and Social Psychology Review, 8, 220–247.

Sutton, S. (2000). Interpreting cross-sectional data on stages of change. Psychologyand Health, 15, 163–171.

Sutton, S. (2008). How does the health action process approach (HAPA) bridge theintention–behavior gap? An examination of the model’s causal structure. AppliedPsychology: An International Review, 57(1), 84–93.

Velicer, W.F., Cumming, G., Fava, J.L., Rossi, J.S., Prochaska, J.O., & Johnson, J.(2008). Theory testing using quantitative predictions of effect size. Applied Psy-chology: An International Review, 57(4), 589–608.

716 LIPPKE AND ZIEGELMANN

© 2008 The Authors. Journal compilation © 2008 International Association of AppliedPsychology.

Webb, T.L., & Sheeran, P. (2006). Does changing behavioural intentions engenderbehavior change? A meta-analysis of the experimental evidence. PsychologicalBulletin, 132, 249–268.

Weinman, J. (1987). Editorial. Psychology and Health, 1, 1–2.Weinstein, N.D. (1993). Testing four competing theories of health-protective behav-

ior. Health Psychology, 12(4), 324–333.Weinstein, N.D. (2007). Misleading tests of health behavior theories. Annals of

Behavioral Medicine, 33, 1–10.Weinstein, N.D., & Rothman, A.J. (2005). Commentary: Revitalizing research on

health behavior theories. Health Education Research, 20(3), 294–297.Weinstein, N.D., Rothman, A.J., & Sutton, S.R. (1998). Stage theories of health

behavior: Conceptual and methodological issues. Health Psychology, 17, 445–453.

Wiedemann, A.U., Schüz, B., Sniehotta, F.F., Scholz, U., & Schwarzer, R. (inpress). Disentangling the relation between intentions, planning, and behaviour: Amoderated mediation analysis. Psychology and Health.

Williams, D.M., Anderson, E.S., & Winett, R.A. (2005). A review of the outcomeexpectancy construct in physical activity research. Annals of Behavioral Medicine,29(1), 70–79.

Wittmann, W.W., & Klumb, P.L. (2006). How to fool yourself with experiments intesting theories. In R.R. Bootzin & P.E. McKnight (Eds.), Psychologicalresearch: Strengthening research methodology: Psychological measurement andevaluation (pp. 185–211). Washington, DC: APA.

Wrosch, C., Schulz, R., & Heckhausen, J. (2005). Health stresses and depressivesymptomatology in the elderly: A control-process approach. In G.E. Miller &E. Chen (Eds.), Current directions in health psychology (pp. 105–111). UpperSaddle River, NJ: Pearson.

Wurm, S., Tesch-Römer, C., & Tomasik, M.J. (2007). Longitudinal findings on age-related cognitions, control beliefs and health in later life. Journals of Gerontology:Psychological Sciences, 62B, 156–164.

Ziegelmann, J.P., & Lippke, S. (2007). Planning and strategy use in health behaviorchange: A life span view. International Journal of Behavioral Medicine, 14, 30–39.

Ziegelmann, J.P., Lippke, S., & Schwarzer, R. (2006a). Adoption and maintenanceof physical activity: Planning interventions in young, middle-aged, and olderadults. Psychology and Health, 21, 145–163.

Ziegelmann, J.P., Lippke, S., & Schwarzer, R. (2006b). Subjective residual lifeexpectancy in health self-regulation. Journals of Gerontology: PsychologicalSciences, 61B, 195–201.