GSR Behaviour Change Knowledge Review: Practical Guide: An Overview of Behaviour Change Models and Their Uses

GSR Behaviour Change Knowledge Review
Practical Guide: An overview of behaviour change models and their uses www.gsr.gov.uk
PU547- GSR behaviour change practice guide 2.indd 1
23/7/08 11:17:38
Practical Guide: An overview of behaviour change models and their uses
Andrew Darnton, Centre for Sustainable Development, University of Westminster
23/7/08 11:17:38
Crown copyright 2008 Government Social Research Unit HM Treasury Enquiries: [email protected] For general enquiries about HM Treasury and its work, contact: Correspondence and Enquiry Unit HM Treasury 1 Horse Guards Road London SW1A 2HQ Tel: 020 7270 4558 Fax: 020 7270 4861 E-mail: [email protected] Printed on at least 75% recycled paper. When you have finished with it please recycle it again. ISBN 978-1-84532-488-9 PU547
23/7/08 11:17:38
Foreword
Government Social Research Knowledge Reviews have been established to explore key areas of importance to social researchers throughout government. The overarching aim of the reviews is to improve how government social researchers conceptualise and measure fundamental concepts in order to enhance the use of social research in informing and evaluating policy, and measuring longer-term trends. Behaviour Change was selected as the rst review in recognition of the increasing importance of inuencing behaviour in order to achieve positive policy outcomes. The review was set up and steered throughout by an Advisory Group of Government Social Research members. This report is part of the rst phase of work coming out of the review, which consisted of a synthesis of the evidence on the theory and principles of behaviour change with the aim of cutting through the vast amount of literature in the area and providing a starting point for research analysts in understanding behavioural change models. The Advisory Group are now considering how to take forward additional work in this area, particularly looking at the empirical evidence base and considering how models are applied in the UK public sector context. The rst phase of the review has resulted in two main outputs. This, the Practical Guide is the main report and summarises the key elements of Behaviour Change theory as well as providing information on the use of models. Secondly, a longer Reference Report, has been designed as a resource for research analysts and other interested parties. It presents over 60 social-psychological Behaviour Change models and discusses in more depth issues to consider when using models to bring about change in individuals behaviour. It also contains an electronic bibliography to assist the navigation of literature in this area. Whilst this review is aimed primarily at GSR members we hope it will be a useful addition to knowledge held by other analysts and researchers from other disciplines, and policy colleagues as achieving sustained behaviour change will only be possible through joint working across analytical disciplines and the policy process. More information about the GSR Knowledge Reviews and links to all outputs can be found on the GSR website at www.gsr.gov.uk.
Government Social Research Unit
23/7/08 11:17:38
Acknowledgements
First, thanks are due to the GSR for commissioning this wide-ranging review of the behaviour change literature. Above all, to Anna Richardson (GSRU) for managing the project and nalising the outputs. Also to members of the Advisory Group for their close involvement from designing the specication through to reviewing drafts: Jane Barrett (FSA); Helen Bullock (DfT); Tina Golton and Verity Morgan (MoJ); Rachel Muckle (Defra); Priscilla Netto (WAG); Clare Talbot (DWP); Richard Tonkin (CLG); Rebekah Widdoweld (TSG). Thanks are also due to all those departmental analysts who agreed to be interviewed during the data gathering stage; for some, this turned into ongoing involvement in the study as it developed. All these analysts are listed in Appendix ii) to the Reference Report, but particular thanks go to Kaite Emmerson (DfT); Louis Levy and Vivien Lund (FSA); Karen McNee and David Gordon (TSG); Sara Skodbo (HO). Most importantly, thanks are due to all the external experts who contributed to the data gathering stage; they are all listed in Appendix ii) to the Reference Report. Special thanks should be given to those who found time to discuss the issues and the evidence in detail. Charles Abraham (Sussex); Tracey Bedford (Surrey); Derek Bell (Newcastle); Jake Chapman (Demos); Andy Dobson (Keele); Tim Jackson (Surrey); Dominic McVey (NSMC); Greg Maio (Cardiff); Richard Shepherd (Surrey); Paul Stern (National Research Council). Thanks, as well, to Andy Ross of the Government Economic Service Team for providing invaluable input into the drafting of the Economics and Behavioural Economics sections. A particular debt should also be recorded to the peer reviewers, whose comments on the rst draft were inuential in reshaping the nal ndings. Finally, thanks go to those researchers who happily provided their models to be reproduced in the Reference Report: Stewart Barr (Exeter); Kay Bartholomew (Texas); Birgitta Gatersleben (Surrey); Rick Gibbons (Iowa State); Arianna Haberis (CLG); Martha Monroe (Florida); Otto Scharmer and Janice Spadafore (MIT); Margaret Whitehead (Liverpool).
Andrew Darnton
23/7/08 11:17:38
Contents
1. 2. 3. 4. Introduction Overview of models, theories and frameworks Making use of models Nine principles for developing interventions based on models 9 11 19 23
Appendices i) ii) Tables: Matching behaviours to models Selected references 35 40
GSR Behaviour change knowledge review
23/7/08 11:17:38
GSR behaviour change knowledge review
23/7/08 11:17:38
Introduction
Over the last decade, there has been increasing recognition in the UK of the complexity of managing the performance of public policy. Policy makers now better understand that changing individual and group behaviour is often central to the effective delivery of policy outcomes. There is also a better appreciation of the importance of public preferences, attitudes, and norms in shaping behaviour. In 2002, the Prime Ministers Strategy Unit published a report introducing the concept of public value to the UK policy environment (Kelly, Mulgan and Muers, 2002). Public value refers to the value created by government through services, laws, regulation and other interventions. The report identied three key dimensions of public value outcomes, delivery of services, and trust and highlighted the centrality of public preferences across all three dimensions in determining that value. Crucially, it argued that value only exists if citizens individually or collectively - are willing to give something up in return for it. This might involve monetary sacrice (e.g. taxes and charges), but also granting coercive powers to the State (e.g. in return for security), disclosing information, and giving time or other resources (such as blood). Often then, this involves changing behaviour and this is much easier to achieve if policy makers give greater weight to public preferences for trustworthy government, due process and fair treatment in legitimising this change. This practical guide helps to explain why this is the case, and signposts analysts and policymakers to some of the analytical tools which have already been developed to understand behaviour, building on other syntheses of behaviour change work (see for example Halpern, Bates and Bales, 2003). It provides an introduction to the socialpsychological literature on the factors which underpin individual and group behaviour, and combines it with that on how these can be inuenced to achieve behaviour change. The focus is on social-psychological models (and the economic theory they build upon) because of their relevance to an understanding of public value. Moreover, these models provide the theoretical basis for the behaviour change agenda - nowadays when people in policy circles talk of behaviour change models, they tend to mean social-psychological models. This literature is synthesised more fully in the accompanying Reference Report which provides full referenced detail on over 60 relevant models, theories and frameworks. Much of the supporting evidence for the models is conceptual, but there is also a more limited empirical evidence base. In signposting readers through the theoretical and empirical literature, the Practical Guide provides a framework for developing interventions based on behavioural models. It also provides detailed support for research analysts and policy makers in the task of selecting appropriate models. To that end it includes tables matching specic behaviours to behavioural models (see Appendix i). In particular, the Guide recommends the use of models which consider behavioural inuences that operate at the group or societal levels in conjunction with those that operate at the individual level. It also emphasises that systems thinking approaches are particularly helpful in attempting to address behaviours which have multiple and complex underlying factors. The importance of addressing the wider implications of behaviour change interventions, particularly in terms of equity and ethical issues, is also incorporated in the Guide. The framework highlights the importance of adopting an analytical approach that draws on theory as the basis for assembling relevant empirical evidence. In this way, users of the framework will help build the evidence base
23/7/08 11:17:38
INTRODUCTION
for the effectiveness of models in explaining and changing behaviour. The Guide begins by providing a brief overview of the range of social-psychological models, theories and principles covered in more detail in the Reference Report, followed by a discussion of the key considerations when using behavioural models to design interventions. This sets the scene for the Nine Principles on the use of models in the intervention development process.
10
23/7/08 11:17:38
Overview of models, theories and frameworks

This review of behaviour change theory observes a distinction between models of behaviour and theories of change. Models of behaviour help us to understand specic behaviours, by identifying the underlying factors which inuence them. By contrast, theories of change show how behaviours change over time, and can be changed. The two bodies of theory are complementary; understanding both is necessary in order to develop effective approaches to behaviour change. In order to provide an overview of the types of models, theories and frameworks available to those developing interventions the evidence was broken down further. Four types of evidence reviewed are outlined: Models of behaviour at the individual level Models of behaviour at higher levels of scale Theories of change Applied models and frameworks In addition some of the key features of behavioural models are discussed.
i)
Models of behaviour at the individual level
Models relating to the behaviour of individuals are predominantly drawn from psychology and sociology, the disciplines which are most concerned with understanding the factors inuencing human behaviour. These models build upon standard economic theory which uses the working assumption that individuals tend to behave rationally, with the aim of maximising the benet to themselves (in psychological terms, such models are expected utility models). Economic theory provides the basis for considerations of human behaviour (especially those behaviours featuring a choice based on costs and benets). This body of theory has subsequently been extended to account for the limitations in human decision making. The resulting discipline of behavioural economics presents numerous principles to account for less rational behavioural choices (see eg. Dawnay and Shah 2005). In so doing, standard economic theory draws closer to the perspectives offered by psychology. (The views of human behaviour offered by economic theory and behavioural economics are outlined further in Section 2 of the Reference Report.) Most social-psychological models are consistent with standard economic theory by presenting behaviour as a decision making process. Most of these models are also consequentialist, assuming behaviour to involve planning ahead, based on outcome expectations (see e.g. Loewenstein et al 2001). Models of individual behaviour tend to be linear, or multilinear, in shape and the simplest follow rational choice theory, which assumes that individuals perform cost/benet calculations and act accordingly. Social-psychological models of behaviour go beyond standard economic theory by investigating the origins of behavioural preferences. The resulting models thus incorporate a wider range of inuencing factors. However, at their core these models present behaviour as the product of a deliberative process (based on intention, and expected outcomes). In this way social-psychological approaches to understanding behaviour can be seen as building upon standard economic theory.
11
23/7/08 11:17:38
OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

The essential factor in most social-psychological models is attitudes, which tend to be conceived as the product of a deliberative calculation weighing an individuals beliefs about a behaviour with the value they attach to those characteristics (these are expectancy value models - see eg. Jackson 2005). Over time, the models have developed further (becoming adjusted expectancy value models), building in additional factors to explain behavioural outcomes. These in turn diminish the primacy of attitudes in determining behaviour. Most social-psychological models remain intention-based; the most well-known example is Ajzens Theory of Planned Behaviour (TPB) (1986, in Ajzen 1991 see Figure 1 below). An alternative is offered by Triandis Theory of Interpersonal Behaviour (TIB) (in Triandis 1977); this model allows for less deliberative behaviours by including habit, which can bypass intentions to determine behavioural outcomes directly.
Figure 1: Ajzens Theory of Planned Behaviour (TPB), (1986)
B eliefs about outcomes Attitude towards the behaviour
E valuation of outcomes
R e lativ e importance of attitude and norm
Intention
B ehaviour
B eliefs about what others think
S ubjective norm
P erceiv ed B ehavioural Control
It is important here to give denitions of some of the key factors which feature in models of individual behaviour, if only to provide ready references for the tables supplied in Appendix i. It may be noted that while attitudes, norms and agency are common to most models, habit and emotion only appear in some (eg. the TIB and not the TPB). Looking at Table A1 matching models and factors to specic behaviours, it can be noted that many of the factors highlighted as being important are not featured in the TPB despite its being the most widely-used model in approaches to behaviour change.
Attitudes
As mentioned above, attitudes tend to be conceived as the product of our beliefs about a behaviour (or object), combined with the value we attach to those beliefs. While the origins of attitudes (as preferences) are not addressed in standard economic theory, the attitude
12
23/7/08 11:17:38
formation process in psychology is presented as a deliberative calculation, following rational choice lines. It should be noted however that several other models highlight the role of emotions in generating attitudes (eg. the Risk as Feelings model (Loewenstein et al 2001), and Petty and Cacioppos Elaboration Likelihood Model, in eg. Bagozzi et al 2002). Technically, attitudes are dened as being specic to a behaviour (or object); in contrast, beliefs are more generic, relating to a wider worldview. Values represent a still higher level of innate preference (see eg. Stern at al 1995).
Norms
Social norms appear in the TPB as subjective norms, dened as a persons perception that most people who are important to him think he should or should not perform the behaviour in question (Ajzen and Fishbein 1980, in Jackson 2005). Social norms act as a guide to how we should behave, and how we expect others to behave. Cialdini makes the important distinction between two types of social norms: descriptive norms which specify what is done, based on the observation of the majority of others, and injunctive norms which specify what other people think ought to be done (Cialdini et al 1990). As well as social norms, theory also identies personal norms; these internalised norms are felt as a sense of moral responsibility to help others (eg. Schwartz 1977). Personal norms are particularly useful in accounting for pro-environmental behaviour (eg. Thogersen 2007). Theories of identity relate closely to norms: social identity dening who we are by reference to others (including the concept of in-groups and out-groups see Turner and Tajfel 1979 in e.g. Terry et al 2000), and self identity relating to our concept of who we are, which is subject to ongoing negotiation (see eg. Jackson 2005).
Agency
The concept of agency appears in most social-psychological models, but in a variety of different guises. Agency can be broadly dened as an individuals sense that they can carry out an action successfully, and that that action will help bring about the expected outcome. Self efcacy is the most widely used version of the concept of agency, dened by Bandura as the conviction that one can successfully execute the behaviour required to produce the outcomes (Bandura 1977). Agency appears in the TPB as Perceived Behavioural Control (PBC), a construct which is heavily based on self efcacy. Agency is important in inuencing behaviour as it determines how much effort we will put in, or whether we will attempt the behaviour at all.
Habit
Paul Stern describes habit as an individuals standard operating procedure (Stern 2000). Most frequent behaviours which are undertaken at low levels of consciousness (ibid.) have a large habitual component (for example, turning out the lights in unused rooms). Whereas the TPB holds beliefs as the underlying foundations of behaviour (Ajzen 1991), habit is seen as the primary determinant in the TIB (Triandis 1977). Triandis denes habit as situation-behaviour sequences that are or have become automatic (1980, in Bamberg and Schmidt 2003). It is the automatic element of habit that differentiates it from repeated behaviour.
13
23/7/08 11:17:38
Emotion
Emotion is included in the TIB as the concept of affect (Triandis 1977). Behavioural intentions in the TIB are generated via two paths: cold cognition (the expectancy value construct) and hot evaluation (the affect factor, or emotional response) (see Bamberg and Schmidt 2003). The inclusion of affect as a unique factor in the TIB is relatively unusual as most social-psychological models embed emotions within other components. For instance, the Risk as Feelings model shows emotion to contribute to the process of attitude formation (the affect as information hypothesis see Loewenstein et al 2001). However, in that model, feelings can also follow a direct path to behaviour (side-stepping intentions); fear offers a good example of such an overpowering emotion. Emotions also inuence intentions directly, as anticipatory emotions which inform individuals analysis of the cost/ benets of acting, based on how they would feel if they were to succeed or fail at the behaviour in question (see eg. Bagozzi et al 2002).
Contextual Factors
Contextual factors can be dened as factors beyond an individuals control (Stern 2000), such as access to information or resources (be that money, time or transport, etc). As such, these external factors are usually left off social-psychological models, which only plot inuencing factors which are situated in an individuals psyche. However, most models account for these contextual factors by incorporating them within the agency construct (eg. Perceived Behavioural Control is a measure of individuals perceptions of how much enabling factors are beyond their control see Ajzen 1991). The TIB features contextual factors in the construct facilitating conditions; these are not simply external factors, but include a persons ability to act, their state of arousal (eg. hunger) and their knowledge of the behaviour. Contextual factors such as cost and the availability of information can be important in determining behaviour, but it should be noted these are not simply external, but also depend on how an individual perceives them.
ii)
Models of behaviour at higher levels of scale
While most social-psychological models are concerned with the factors inuencing behaviour from within an individuals own psyche, some models include factors shaping individual behaviour from higher levels of scale. The forces shown in these models include macro-level societal factors, for instance technology and the economy. Thus these may be called societal models. These models are important to those developing policy as often it is necessary to work on the contextual factors limiting behavioural options directly; simply changing a persons perceptions of these material factors (eg. cost) will not be sufcient to enable change.
14
23/7/08 11:17:39

Figure 2: Vlek et als Needs Opportunities Abilities (NOA) Model, (1997)
T ec h n ology
E c onomy
Demography
Ins titutions
Culture
Needs R e lations , de velopment, comfort , ple as ure , work, he alth , privacy , money , s tatus , s afety , nature, freedom , leis ure time , jus tice
Opportunities Availability , adve rtis ement, price s , s hops
Abilities F inancial, temporal, s patial, cognitive, phy s ical
Motiv ation
B ehav ioural c ontrol
Intention
Cons umer behaviour
S ubjec tiv e well-being , environmental quality
Examples of societal models include the Main Determinants of Health model (Dahlgren and Whitehead 1991) which shows the individuals behaviour as one element, beneath four other tiers of inuencing factors. The NOA (Needs Opportunities and Abilities) model of consumer behaviour similarly shows the inuence of societal factors on individual behaviour (1997, in Gatersleben and Vlek 1998 see Figure 2 above). This nested model incorporates a social-psychological model of individual consumer behaviour at its centre, subject to ve macro-level societal factors shaping individual behaviour from above. Notably, the model also shows consumer behaviour inuencing the societal factors, by means of a large feedback loop running from the bottom of the model to the top.
iii)
Theories of change
Theories and approaches derived from work in the social sciences can help us to understand processes of change. These theories are vital to those developing interventions, as it is not enough simply to understand why behaviours happen, when the challenge is to make behaviour happen differently. Types of theories of change identied as part of the review include:
Lewins Change Theory
Social-psychologist Kurt Lewins work on change has provided the basis for many subsequent approaches (Lewin 1951). Lewins Change Theory involves group work to change habitual behaviours, using an unfreezing/refreezing dynamic in which habitual behaviours are lifted up to scrutiny by the group and recongured, before being left to fall back into everyday routines.
15
23/7/08 11:17:39
OVERVIEWS MODEL THEORIES AND FRAMEWORKS Diffusion Models
Models of diffusion can be used to show how a behaviour spreads through a society or network. Rogers Diffusion of Innovations theory is the best known example, although it is most commonly applied to the adoption of technologies, rather than social behaviours (see Rogers 1995).
Staged Models
Staged models of change describe change as a process, broken down into a series of stages through which individuals progress. The best known example is Prochaska and Di Clementes Transtheoretical (Stages of Change) Model, originally developed in the context of smoking cessation (see Prochaska and Velicer 1997). In the light of recent criticism from practitioners, the Model appears more useful for its conceptual lessons than its practical applications.
Learning-based Models
Learning theory going back to the educationalist John Dewey sees learning and change as dual elements in a single process. The theory of double loop learning extends this thinking on learning through doing to allow for both incremental and transformational change (see Argyris and Schon 1996).
Organisational Learning Models
Practical approaches to organisational change build on learning theory to show how transformational change requires the revealing and refashioning of underlying assumptions (see eg. Schein 2004, Scharmer 2007).
Systems Thinking Approaches
Systems thinking emerged from the disciplines of engineering and cybernetics; systems thinking regards behaviour as the product of interactions between the parts of a whole system. As an approach to change, systems thinking is best described as a discipline for seeing wholes (Senge 1990).
iv)
Applied models and frameworks
The review identies a number of process-based models and frameworks which have been applied to policy making for behaviour change. Social marketing, dened by the National Centre for Social Marketing as a process for delivering behaviour change for the public good, is characterised by understanding audiences and interventions through ongoing research (see French and Blair Stevens 2005). It also calls for multiple instruments to be combined in an intervention mix. Social marketing provides step-by-step guidance for developing interventions. This guidance is consistent with more specic frameworks such as Gardner and Sterns Principles for Intervening to Change Environmentally Destructive Behaviour (in the context of the environment - 1996, in Stern 2000) and Intervention Mapping (IM) (for health programmes Bartholomew et al 1998). Such frameworks set out theory- and evidencebased approaches to designing, developing and evaluating interventions.
16
23/7/08 11:17:39
Defras 4Es model builds on social marketing, and offers a checklist for policy makers to help ensure that they use a balanced package of measures to achieve their behaviour change objective (2005, in eg. Defra 2008). The 4Es themselves represent different approaches to policy making for behaviour change (Enable, Encourage, Engage and Exemplify). Against each E, potential policy interventions are specied (for example, Enable includes remove barriers, give information, and provide facilities). The 4Es model has been adapted and used across government; it notably features within a new model of policy making for culture change developed by the Prime Ministers Strategy Unit (Knott et al 2008). That culture change model adopts a nested approach to individual change, reminiscent of societal models such as NOA, discussed above. Finally, Jake Chapman offers a challenge to traditional models of policy making, which he describes as operating along mechanistic lines, based on the principles of command and control. Rather than targets being imposed on service deliverers from the centre, Chapman advocates the application of a systems thinking approach which ultimately aims to create a system of government which can learn for itself, and is thereby able to tackle complex problems, and avoid ongoing policy failure (Chapman 2004).
17
23/7/08 11:17:39
18
23/7/08 11:17:39
Making use of models

This section provides an overview of some of the factors that should be considered when using behavioural models to develop behaviour change interventions.
Limits of behavioural models

While this review advocates that intervention strategies should be grounded in theory, it is important to stress that there are limits to what behavioural models can tell those who are developing interventions. The distinction in this review between behavioural models and theories of change is made primarily to emphasise that adopting models alone is insufcient to bring about behavioural change; an understanding of the process of change must also be applied. The evidence in this review is clear that a behavioural model should not simply be taken up and used as the basis for a behaviour change intervention without rst considering a range of issues (see Reference Report, Section 3, for a more in-depth discussion). These include:
Models are concepts, not representations of behaviour
They show the factors inuencing behaviour, but do not explain the processes for changing behaviour.
Behaviour is complex, but models are deliberately simple
They are concepts to aid understanding they are deliberately simplistic and do not capture all the factors that account for behavioural outcomes.
There is a limit to how far models will stretch
They are developed in the context of a specic behaviour, and tend to work best in that context although some do have wider applicability.
Models dont tend to differentiate between people
They tend to show the behaviour of a statistical everyman and need to be adapted in order to cover different audience groups.
Factors dont always precede behaviour
Most behavioural models present social-psychological factors as preceding behaviour but there are instances where people are compelled to change their behaviour rst, which then leads to change in the social-psychological variables.
19
23/7/08 11:17:39
MAKING USE OF MODELS
Factors are not barriers
Behavioural models can appear to present factors as buttons to be pressed, in the expectation that the behavioural outcome shown will result. This is a misleading interpretation, not only because other factors which are off the model may also need to be accounted for, but because the mechanistic assumptions underpinning such a view are inadequate to explain the nature of change, and to support individuals in that process. The review of theories of change also identies a number of procedural limits to using behavioural models in aiming to bring about behaviour change. At a basic level, theories of change stress that interventions must look beyond the individual to address the context within which they function (their social eld (Lewin 1951), or system of interest (Chapman 2003)). Procedurally, the principles of action research as deployed in diverse methods show that engaging actors in the process is the best way to bring about change amongst them (see eg. Stern 2000, Hobson 2001). In addition, systems thinking approaches have put forward fundamental objections to interventions based on models. Senge (1990), for example in the context of organisational change objects to the adoption of models because their use cannot deliver the transformational change required to remedy complex problems. Systems thinking can best be understood as a discipline for understanding complex problems; the process of analysing the problem itself reveals opportunities for interventions. Systems thinking thus offers an approach based on modelling complex behaviours, rather than advocating the use of existing models. Applying systems thinking to policy problems, Jake Chapman calls for the refashioning of government as a learning organisation (Chapman 2004). He objects to the use of predetermined principles in developing interventions on the grounds that, in appearing to offer solutions, their use blocks further learning among those who adopt them. The same criticism could be applied to the adoption of models at face value.
Using models effectively

This review does not conclude, as systems thinking approaches do, that models should not be used in developing behaviour change interventions. Instead it proposes a process in which the analytical strengths of behavioural models are joined up with approaches to policy development and delivery based on the principles of action research. The development process should involve the active participation of audience groups, and the renement of the intervention through piloting on the ground. The Nine Principles framework resulting from this review is built around using behavioural models. However, it is essential that those models are used appropriately: not as templates for behaviour change policies, but as tools to be used in the design of those interventions. In all instances, analysts and policy makers should use models as aids to thinking, and not seek to impose them on the public uncritically through interventions. In his authoritative review of theories and models of behaviour in relation to sustainable consumption, Tim Jackson identies two purposes for behavioural models (Jackson 2005). One is heuristic, in which models serve to help us understand the factors inuencing behaviour and how they interrelate. The other purpose is empirical, to quantify the relationships between the factors to allow us to predict the behavioural outcomes. Triandis adopts a similar approach, saying that a good model is one which predicts a specic behaviour well (Triandis 1977). However this point does not override the inherent nature
20
23/7/08 11:17:39
MAKING USE OF MODELS
of models as concepts. At the end of presenting the TIB, Triandis (quoting Cronbach 1975), says that models are concepts that will help people use their heads. Predicting outcomes with models advances our understanding, and shows which model is more appropriate for understanding a given behaviour. However models fundamentally do not determine how people behave. While models cant account for all the complexities of behaviour and determine how people behave, they can help to identify some of the factors that inuence those outcomes. As Tim Jackson writes: models reveal factors where policy can work (Jackson 2005). In such an approach, models essentially offer a menu of factors for policy makers to choose from (rather than a recipe for creating behaviour). Behavioural models can be used in the initial design phase to help identify those factors that may be worked upon in the intervention. In turn, the interventions can be evaluated in terms of impact on those target variables, as well as in terms of change in the end behaviour itself. An example of this is the evaluation of the impacts of an HIV prevention programme based on the Information-MotivationBehavioural Skills (IMB) model (Fisher et al 2002). The programmes effectiveness was judged against each of the three factors (ie. knowledge, attitudes, and skills/agency), as well as in terms of the overall impact on safe sex behaviour. An example of a model-based approach to intervention development can be found in work on the FRANK drugs campaign from the Home Ofce, DH and DCSF (see Darnton 2005). Recognising reducing levels of drug use to be a complex problem (and one ultimately beyond the reach of information-based campaigns), the FRANK team commissioned desk research, which identied Gibbons and Gerards Prototype/Willingness Model as having potential for use in campaign strategy development (Gibbons et al 2003). One of the objectives for the campaign is to prevent or delay the onset of drug use among 11 to 14 year olds. The Prototype/Willingness Model offered an ideal basis for campaign development as it shows the process through which young people move from initial risky behaviours (based on behavioural willingness) through to established risky behaviours (based on behavioural intention). The Model also identied an inuential factor on the willingness pathway as risk images (a individuals perceptions of what a typical person undertaking the risky behaviour is like). The risk images factor was highlighted as an appropriate factor for FRANK to work upon, and one which would respond to informationbased instruments. Using audience research, campaign elements were developed to alter young peoples risk images of people who use different illegal drugs, and the campaign was evaluated against measures of change in young peoples risk images. Clearly, this is a pragmatic approach designed to improve the effectiveness of behaviour change interventions. The fuller answer to tackling complex problems like drug use would at the very least involve policy interventions working at all levels of scale (eg. across the four tiers in the Main Determinants of Health model - Dahlgren and Whitehead 1991). It is for this reason that the approach arising from this review recommends the use of behavioural models operating at different levels of scale in conjunction with models of individual behaviour. The method also recommends that systems thinking approaches are used to address the most complex problems, and to build understanding of the role that behaviour change interventions could play in tackling them. But at the centre of the policy planning process, models of behaviour at the individual level can provide policy makers with clues as to where their resources are best deployed, and can help analysts judge how best to evaluate the impacts of those interventions.
21
23/7/08 11:17:39
22
23/7/08 11:17:39
Nine Principles for developing interventions based on models

This review has drawn a distinction between models of behaviour and theories of change. From a conceptual point of view, this distinction can be quite hard to impose, as the two bodies of theory are closely related, and overlapping in places. However, the distinction becomes clearest in the context of practical guidance, where: Models of behaviour help us to understand specic behaviours, by identifying the underlying factors which inuence them. Theories of change identify intervention techniques which can be effective in bringing about change, as well as suggesting broad approaches to intervention design, implementation and evaluation which can underpin effective policy planning and delivery. It is important to note that neither body of theory is alone sufcient to generate effective interventions. Following on from this the Nine Principle framework designed as part of this review, and described below, aims to integrate behavioural models with theoretical understanding of effective approaches to change. The framework provides a starting point for selecting models and developing behaviour change interventions based on this premise. The framework is expressed in Nine Principles. While these unfold in a logical sequence, they should not be regarded as discrete steps, with one being accomplished before moving on to the next. Instead the Principles can be best understood as a cyclical process, with learning from monitoring and evaluation feeding back in to the development of the intervention. Progress through the Principles is also iterative; learning from work on one Principle could require returning to revisit earlier assumptions. Behavioural models sit at the centre of this cycle. The key factors they identify inform both the initial objectives set for the intervention, and shape the measures against it will ultimately be evaluated. The circular shape of the process is thus consistent with an approach based on the action research principles of learning through doing, the intervention being constantly rened as a result of ongoing monitoring and evaluation. Importantly, the cyclical nature of the Nine Principles is also in keeping with existing guidance on policy evaluation, such as the ROAMEF model in the Green Book (HMT 2003), which demonstrates how research can support effective delivery throughout the policy cycle. The Nine Principles are: 1. Identify the audience groups and the target behaviour. If faced with a complex behaviour break it down into its component behaviours and/or adopt a systems thinking approach 2. Identify relevant behavioural models (use both individual- and societallevel models). Draw up a shortlist of inuencing factors 3. Select the key inuencing factors to work on Use these to design objectives in a draft strategy for the intervention
23
23/7/08 11:17:39
NINE PRINCIPLES FOR DEVELOPING INTERVENTIONS BASED ON MODELS

4. Identify effective intervention techniques which have worked in the past on the inuencing factors selected Engage the target audience for the intervention in order to understand the target behaviour and the factors inuencing it from their perspective Develop a prototype intervention based on the learning from working with the actors. Cross-check this against appropriate policy frameworks and assessment tools Pilot the intervention and monitor continuously Evaluate impacts and processes Feedback learning from the evaluation
5.
6.
7. 8. 9.
The Nine Principles resemble existing theory-based guidance for planning interventions, but aim to achieve a synthesis between the different approaches. The key difference between the Nine Principles and other approaches such as social marketing and Gardner and Sterns Principles (in Stern 2000) is the building of behavioural models into the heart of the developing process. The Nine Principles can also be compared to the Intervention Mapping (IM) framework, which similarly centres on behavioural models, but which follows a more programmatic path to intervention development and implementation (Bartholomew et al 1998). The IM approach proceeds through ve steps (following an initial Needs Assessment stage), with each step generating a plan or matrix which becomes the basis of the next step. Despite the somewhat mechanistic method of IM, it represents a problembased approach to using models, starting from the audience and the behaviour in question. Overall, the Nine Principle framework proposed here takes account of the need for exibility in developing interventions. The remainder of this report provides detailed advice on using each of the principles; an emphasis is placed on the earlier principles, in which the use of models and theories is most fundamental. As yet there is limited empirical evidence to underpin guidance on the latter stages of implementation and guidance. However, they are included here as more general principles, in recognition of the theoretical evidence which calls for exible approaches to changing behaviour, with interventions being rened through piloting and monitoring (see eg. Knott et al 2008, Chapman 2004, Plsek 2003).
4.1 Principle 1: Identifying the audience groups and the target behaviour
Before setting out to develop a behaviour change intervention it is necessary to be clear about whose behaviour is to change, and which specic behaviour is to be targeted. Understanding both actors and actions is essential to designing effective interventions; imprecision on either aspect can lead to blanket approaches being adopted. It is standard practice in marketing to differentiate between audience groups, recognising that there is seldom a one size ts all solution. Social marketing goes further by putting the customer at the centre of the campaign development process, and starting from the
24
23/7/08 11:17:39
point of their current behaviour (see eg. French and Blair Stevens 2005). Social marketings preference for developing audience segmentation models is a result of this approach. The Nine Principles framework proposed here takes a slightly different view; by putting behavioural models (not audience segmentations) at the centre of the development process, understanding the behaviour comes before engaging the audience (see Principle 5 below). Nonetheless, both activities are essential to successful interventions, whether a customercentred or model-based method is preferred. Alongside identifying the audience groups, the intervention design process must begin by specifying the behaviour to be changed. This is fundamental to designing effective interventions, whether or not they are based on behavioural models; however, the task of model selection particularly requires clarity about the target behaviour. Existing guidance on developing interventions prioritises this task; see for example Gardner and Stern (1996, in Stern 2000) and McKenzie Mohrs rst step in his four stage process of communitybased social marketing (McKenzie Mohr 2000). This principle may seem self-evident, but it is often the case that policies are unclear what their ultimate outcome should be. In her review of theory for DWP Clare Talbot notes that employment interventions commonly set , measures on the numbers of people in work (the outcome) rather than the numbers seeking work (the target behaviour) (Talbot et al 2007). Such a lack of clarity not only handicaps research analysts in the evaluation phase, but works against effectiveness throughout the lifecycle of interventions. If a policy problem is complex, identifying the specic target behaviour may be more difcult. As discussed above, the most complex problems (or messes in systems terminology) will require systems thinking approaches; they are likely to include numerous behaviours among the interactions between their components. As a discipline for seeing wholes, systems thinking allows those addressing policy problems to stand back and see the patterns at play across the whole (see Senge 2000). The Foresight programmes recent project on Tackling Obesities provides an example of a systems thinking approach to a complex problem (Foresight 2007). Finding no existing model of obesity which covered all the inuencing factors, the project team assembled a group of experts to conduct a systems mapping exercise. The resulting Obesity System Map is a highly complex model of the myriad factors inuencing obesity, based around the energy balance model of weight at its centre. The factors are linked together with feedback loops showing the ow of inuence between them. The model is conceptual, and the relationships between the factors are not quantied. The Map is notably presented as a starting point in the process of developing potential interventions, and the Map itself is to be rened through ongoing research. As a result of systems mapping techniques of this kind, it may be concluded that launching a behaviour change intervention is not the appropriate response. For other complex problems, it may be a more simple matter of identifying the multiple behaviours involved, and choosing which ones to tackle rst. The evidence from theory suggests that different behaviours are driven by different factors and in different combinations, even if the behaviours appear clearly related. For example, the factors inuencing energy efciency behaviours vary according to the specic measures being taken (see Darnton 2007), while those inuencing healthy eating behaviours vary from food to food (see Baker et al 2007). Paul Stern states that each behaviour should be conceptualised differently (Stern 2000).
25
23/7/08 11:17:39
4.2
Principle 2: Selecting models
Using behavioural models can help to develop hypotheses for subsequent testing; however, with so many models available, even experienced researchers nd it hard to know which one to choose (see eg. Bamberg and Schmidt 2003). The aim of the model selection exercise in Principle 2 is to identify relevant inuencing factors. The outcome of this Principle should be a shortlist of factors which can be carried into Principle 3, in which the key factors will be selected to form the basis of objectives for the intervention. A similar process is laid out in the IM framework, combining theoretical models with empirical data from audience research. Model selection, and the identication of inuencing factors, can be presented as a dualpath approach: Figure 3: A Method for Identifying Inuencing Factors
Path 1: From Models to Factors - Search the existing literature to identify relevant models of the target behaviour, and within them, relevant inuencing factors These models may be specic to the target behaviour (ie. developed to describe that behaviour), or they may be general models (eg. the TPB) which may have been applied to the target behaviour (although this is not essential). The identication of which models (and which factors within the models) are the most important should be decided statistically. Models should be supplied with empirical data
26
23/7/08 11:17:39
to show how well they predict the behaviour in question, and which factors are the most effective in doing so. Technically, this involves comparing the level of variance in the end 1 behaviour accounted for by different models (and the factors within them) . It will often be the case that, although empirical data may be supplied to show the relative weights of each of the factors in inuencing the end behaviour, the behaviour or target gruops in the model differ somewhat from that which is being targeted in the intervention. In such cases, analysts may need to content themselves with using comparable data as an approximation for their target behaviour. Such proxy data should still enable analysts to identify the likely relative importance of the inuencing factors, and build a hypothesis about which are the most important to work on. The alternative would be to construct a new model specic to the target behaviour, ideally with data from the audience groups in question. Real limits (of time and money) will often prevent such a step being undertaken, and necessitate the use of proxy data (where it exists). In cases where models are solely conceptual, and not accompanied by empirical data, intervention developers can either undertake audience research of their own to quantify the impact of the variables, or they can look to existing research data to gather evidence which supports the case for an inuencing factor being considered relevant. Path 2: From Research Data to Models and Factors -Search existing audience research data to identify reported barriers and drivers, and compare these with models to identify inuencing factors Research evidence is essential to build understanding of a target behaviour. Research data can work in tandem with theoretical understanding by providing specicity to complement models more generic conceptual strengths. Research data should ideally be specic to the audience groups in question, and to the target behaviour. Such evidence can be both qualitative and quantitative, and can be used to identify barriers and drivers as reported by the audience. However, these barriers and drivers should not be taken at face value. There is often a difference between what audiences say is inuencing their behaviour (especially in response to survey questions), and the factors that underlie those reported barriers and drivers. For this reason, research ndings should be read in combination with models from theory. The research data can conrm which factors on a model are relevant to the audience in question (strengthening the case made by the theoretical evidence), while the models can substantiate and standardise ndings from research. Finally, in many cases research data will identify barriers and drivers that are not featured in theoretical models; often these will be contextual factors (such as access to infrastructure, or cost). It is legitimate to carry these factors forward directly to the shortlist of inuencing factors, if they are shown to have a strong inuence, despite their not being explicitly featured in relevant behavioural models. These factors are most useful if combined with models of behaviour at higher levels of scale, such as the NOA model (Gatersleben and Vlek 1998) or the Main Determinants of Health (Dahlgren and Whitehead 1991). Again, audience research data can ll in gaps in the conceptual models, and together
1(A level of 20% has been shown to be sufcient to bring about change in the end behaviour of a signicant proportion of the target audience see eg. Rosenthal and Rubin 1982, Armitage and Conner 2001. The TPB has been found to account for 40% of the variance in outcomes across a range of behaviours, and this can be taken as strong evidence of its ability to support behaviour change in various contexts see Aiken 2002 in Talbot et al 2007.)
27
23/7/08 11:17:39

identify inuencing factors which should be shortlisted. Above all it should be stressed that the aim of this exercise is not to nd one winning model, but to identify a range of models which can provide a shortlist of factors known to be inuential in determining the behaviour. In order to help analysts and policy makers with the process of model selection, this review includes two tables, mapping behaviours against theories (see Appendix i). It should be stressed these tables do not provide solutions to the problem of selecting models; that process is inherently of value in building understanding. Instead these tables are designed to provide suggestions of likely models for analysts to begin with. Further information about all the models listed can be found in the Reference Report. Table A1 matches behavioural models to specic behaviours, based on references made in the sources in this review and also includes factors identied in the sources as inuencing specic behaviours. Note that these matches are not quantied; a match simply means that a source has identied a link between a behaviour and a model or factor. The Table is thus a starting point in the process of model selection and factor identication (it effectively sits at the start of Path 1 in the method above). Intervention developers should proceed by investigating the supporting evidence for each model or factor, to assess which are the relevant inuencing variables. In some cases, the sources which referenced the match will also supply the empirical evidence for the link; in others, the supporting evidence is not made explicit. Finally, it will be noted that for some behaviours, multiple models and factors are identied; it is up to analysts to determine which of these factors and models are the most relevant for the target behaviour and audiences in question. Table A2 matches behavioural models to the types of behaviour they are effective at describing. For example, Norm Activation Theory (Schwartz 1977) was developed to account for altruistic (helping) behaviours. Table A2 is intended to act as a source of additional models to those in Table A1, and to provide possible starting points where Table A1 does not give a match for the specic behaviour an intervention is targeting. Table A2 effectively sits on Path 2 in the method above. Analysts must rst build a thorough understanding of the target behaviour based on research data, in order to conceptualise it into one of the types of behaviour listed (eg. to see it as altruistic, or habitual etc). Once a specic model has been identied, the strength of its inuencing factors will need to be investigated using the factor identication method described above. As Figure 3 shows, the end point of the Model Selection stage is the identication of a shortlist of inuencing variables, all of which should be supported by empirical evidence. In the IM framework, these factors should be combined into a nal, plausible and empiricallysupported causal model (Bartholomew et al 1998). While the generation of a new causal model may be an ideal situation it is not essential to the subsequent stages of intervention development as it is likely to be adequate simply to evaluate the intervention against the key factors selected, while at the same time measuring change in the target behaviour itself. Importantly, if a single model were generated, it would most likely need to be adapted for use with different audience groups. As has been noted (in Section 3) above, socialpsychological models tend to be undifferentiated, showing the behaviour of an everyman. Such models must be adapted to relate to specic audience groups if they are to be taken up for use in interventions. The Obesity System Map provides a case in point, having already been adapted for specic reference to children (see Foresight 2007). The importance of using differentiated models is underlined by work for Defra on pro-environmental
28
23/7/08 11:17:39
behaviours, in which models were produced for the same behaviour across four different segments of the public (Barr et al 2005). The resulting models show clearly how, even for the same behaviour, different factors are present for different segments, and that they impact at different levels. If a new model is not generated, empirical data, or suitable proxies, will be needed to support each of the key inuencing factors to provide baselines against which to evaluate the interventions impact at the end of the delivery cycle. Again, comparison can be made to the Foresight project on Tackling Obesities: the systems mapping exercise generated a new model, but it was not quantied. Instead, the relative weighting of the inuencing factors was judged qualitatively (on a scale of 1 to 5, with the feedback loops shown using different widths of arrows). Finally, whatever method is used to identify relevant behavioural models, it should be reiterated that these models of behaviour at the individual level will need to be supplemented by models at a higher level of scale. In this way analysts can be sure not to overlook contextual and societal factors inuencing the target behaviour. Selecting these models is a much easier task; the review only features two such models of general relevance (NOA and the Main Determinants of Health), both of which may be useful in identifying key external factors. Note that these models are conceptual in nature; they do not include empirical data (nor would it be obvious how to ll them with data in their current formats). These models should be used alongside existing research data from the audience groups to establish the relative importance of the higher-level factors in inuencing the target behaviour.
4.3
Principle 3: Designing a draft strategy
The process of model selection will have resulted in a shortlist of inuencing factors, each supported by data to show their relative impact on the target behaviour. As part of Principle 3, key factors must be selected from this shortlist to form the basis of the objectives in the draft strategy for the intervention. The process of drafting the strategy is best undertaken by analysts and policy makers together, as the decision on which factors to work on is not simply a statistical question. Often it will not be deemed feasible to target the strongest inuencing factors. For example, external barriers (such as cost) may be beyond the intervention developers control, or it may be that the available intervention instruments (eg. information and communications) do not allow for certain factors to be addressed. The objectives set for an intervention must be appropriate and achievable. The task of selecting key factors is both analytical and pragmatic; the solution is not found in theoretical models themselves. This is a further reason why a more exible framework for intervention development is required.
4.4
Principle 4: Identifying intervention techniques
Having selected the key factors on which to design the intervention strategy, the next step is to develop the elements of the intervention itself.
29
23/7/08 11:17:39

While interventions should be shaped in collaboration with the audience, and in the context in which they will be applied, lessons can be learned from analysis of what has worked in existing interventions. Such an analysis is the topic for a separate study, although a little evidence of this sort is provided in the sources reviewed here. Most obviously, Prochaska and Di Clementes Transtheoretical Model includes stage-matched interventions (see Prochaska and Velicer 1997. Social marketing also includes an element of matching intervention methods to desirable outcomes, as shown in Rothschilds Grid applying education (as information), marketing (as incentives), and law (legislation) tools to the target factors of motivation, opportunity and ability (see eg. Kurani and Turrentine 2002). At a greater level of specicity, Fisher et als work on HIV prevention programmes based on the IMB (Information, Motivation, Behavioural Skills) model also identies techniques known to be effective in changing motivations and attitudes, as well as building self efcacy (Fisher et al 2002). Most helpful of all is Charles Abraham and Susan Michies recent work in developing a typology of Behaviour Change Techniques (BCTs), which are specic intervention methods identied through a meta-analysis of research ndings from 195 healthy eating and physical activity interventions (Abraham and Michie 2007). As well as identifying the BCTs as elements of effective interventions, the study maps each technique against theoretical models which have been used to account for the impact of that technique in research (eg. providing information about others approval is matched to the Theory of Reasoned Action, the TPB and the IMB Model). Abraham and Michie are explicit that the purpose of their taxonomy is to help those delivering interventions to make the link back to theory, in order that they may evaluate the intervention appropriately, and to facilitate the replication of the intervention in another context. In so doing they echo the NICE guidance (to which Abraham contributed) which calls for the retrospective use of models to identify the underlying theories linking actions and outcomes (NICE 2007). While the priority given to behavioural models varies, this review is in agreement that interventions should be informed by theory and developed on the ground, rather than be imposed based simply on the uncritical adoption of a model.
4.5
Principle 5: Engaging the audience as actors
Policy frameworks and sets of practical guidance agree that engaging the audience in the process of policy development is essential to achieving effective interventions. At the same time, audience engagement exercises are seen to build fairness (procedural equity) into the policy process (see eg. Pearce 2007, Knott et al 2008). In staged guidance, engaging the audience usually features upfront, alongside specifying the target behaviour. However, the Nine Principle framework proposed here has recognised that identifying the different audience groups to be engaged is the starting point for designing interventions, at the same time as specifying the target behaviour. These two activities cannot very well be separated, as understanding behaviour involves understanding the population whose behaviour it is. However, in order to allow for the development of an intervention strategy based on theoretical models, and informed by existing research ndings, the task of engaging audience groups is put later on in the Nine Principle process. Effectively this allows a draft strategy to be developed based on theory, before taking it out to test and rene with the audience groups in question.
30
23/7/08 11:17:40
The other model-based framework, Intervention Mapping, adopts a similar method (Bartholomew et al 1998). Identifying the at-risk population is the rst task within the initial Needs Assessment phase. The IM framework also recommends building a linkage system between programme developers, implementers, and adopters (i.e. the audience, as actors). This linkage system is only introduced late on in the development process (at Principle 4), although the author notes it would be benecial to have the links in place at the start of the intervention planning. While theoretical evidence and research data should be considered rst, collaboration with actors should be sought from that point. In the development stages proposed here in the Nine Principles, audience engagement would initially serve the purpose of testing the assumptions made about the target behaviour and the intervention techniques to be adopted. Specically, the initial engagement should include the following activities: i. understanding the target behaviour and the factors inuencing it from the audiences perspective; ii. testing the draft strategy and the key factors it identies;
iii. exploring the potential of the known intervention techniques, and collaboration to work up new intervention elements. The last of these three activities involves genuine collaboration, rather than audience research, and leads into the delivery stages of prototyping and piloting which follow.
4.6
Principle 6: Developing a prototype intervention
Once the draft strategy for the intervention has been agreed with the audience, and potential intervention elements have been identied, these should be turned into prototypes through collaboration with the target audience of actors themselves. These principles are consistent with action research, and can be found in group-based approaches to change such as Theory U (Scharmer 2007). The prototyping process in Theory U effectively involves turning hypotheses into potential projects which can be taken out into the eld and piloted on an experimental basis. The prototyping exercise should also incorporate assessment of the prospective intervention. This may follow standard policy assessment methods, using tools such as the Impact Assessments, but should foreground wider impacts, including ethical issues and equity effects, and unintended consequences (for a further discussion of these issues see the Reference Report). This review has shown that negative impacts on equity and other side effects do not arise as a consequence of which behavioural model (or models) an intervention is based on - no one model is inherently fairer than another. However the way in which a model is applied through the development of an intervention can affect both the fairness of the nal outcomes (distributional equity) and the perceived fairness of the development process (procedural equity see eg. Pearce 2007). Such procedural considerations can ultimately determine the acceptability of the intervention among the public. The policy research evidence also argues for an approach to intervention development based on public engagement. In addition, the prototype assessment phase should not merely be regarded as a means of equity-proong prospective interventions, but an
31
23/7/08 11:17:40

opportunity for building public views into the intervention strategy. More fundamentally, interventions which are regarded as unfair are likely to be ineffective, as the public will be disinclined to comply with them (the Fuel Duty Escalator offers an extreme example see eg. Knott et al 2008). Questions of fairness are inherent in policy making for behaviour change, and consideration should be given to ethical concerns and equity issues from the start of the development cycle (see eg. Lewis 2007). Even where an intervention has been developed in accordance with relevant theoretical and empirical research evidence, it may still be rendered ineffective if it is perceived by the public as unfair. An effort should also be made to review the intervention in terms of its policy consistency: whether the audience groups in question are already subject to other interventions working in opposite directions. Inconsistency is to be avoided as when policies in one place work in an opposite direction to policy in another, negative or at least neutralising effects can result (see eg. Knott et al 2008). At the same, policy research has shown that the negative effects of multiple policies tend to fall disproportionately on those already in vulnerable groups.
4.7
Principles 7 to 9: Piloting, evaluation and feedback
The last three Principles in the Nine Principle framework piloting and monitoring, evaluation and feedback - can perhaps be regarded as more advisory than the rst six. They are certainly less closely related to the central theme of this review, and the resulting advice on how to use behavioural models. Ultimately, the implementation, monitoring and evaluation of interventions should be conducted in accordance with existing guidance on best practice in policy making and evaluation, as presented in the Magenta Book (GCSRO 2003, updated GSRU 2007) and the ROAMEF model in the Green Book (HMT 2003). However these three Principles are included in recognition of the strong case made by the theoretical evidence particularly that relating to learning theory and systems thinking for exible approaches to implementation and evaluation. These calls are echoed in the recent processes and frameworks for policy making for behaviour change included in this review. The Nine Principles framework has already been described as a cyclical process; the last three principles, focusing on monitoring and evaluation, effectively join up the loop of the intervention cycle, by feeding learning back in to the ongoing policy process. This resulting depiction of the intervention process as a cycle is consistent with theories of change, notably those based on action research, with its ongoing cycles of action and reection. While these three research-led principles are presented at the end of the sequence of tasks based on behavioural models, it is imperative that they are also considered at the start of the process. As all good practice guidance in research would suggest (and as the ROAMEF model epitomises) objectives and intended outcomes need to be established together at the outset of the policy planning process. In a reinforcing relationship, each informs the other: what is measurable helps to dene what targets should be set, and vice versa. The implications of theories of change included in this review only underline these principles of good practice.
32
23/7/08 11:17:40

Figure 4: The role of research in the Nine Principles framework
The Nine Principles framework proposes that, once through the assessment stage, the prototype intervention should be implemented in a pilot stage (Principle 7). Processbased theories of change relating to policy call for piloting as the most effective means of developing effective interventions. The pre-test phase is central to Andreasens six stages of social marketing, which then continues in a loop of ongoing monitoring (in Kurani and Turrentine 2002). The PMSUs paper on policy making for culture change calls for the creating of safe spacesto test and trial new ideas and innovations (Knott et al 2008). Piloting is also fundamental to Jake Chapmans approach to policy making, in keeping with systems thinkings endorsement of trial and error methods of innovation (Chapman 2004). Ongoing monitoring should be undertaken in parallel to the pilot activity. As a result of the monitoring the pilot intervention should be adapted, extended or abandoned (see Jowell 2003 for further information on the role of pilots in policy making). Principle 8 involves the evaluation of the nal intervention, the design of which should be considered at the outset. Evaluations should measure change in the target behaviour among the audience groups in question, as well as impacts on the key inuencing factors. Other effects should also be evaluated (potentially using qualitative techniques) in order to build an all-round view of the interventions impact which takes in any unintended consequences. As well as assessing impacts, the evaluation should also address the processes used to design, develop and deliver the intervention. In this way, the models and theories used to underpin the intervention may be reassessed in the context of the target behaviour, and understanding of the behaviour itself may be advanced. Principle 9 represents the closing phase of the cycle in which learning from the evaluation is fed back into the policy process. In keeping with the balance of the evaluation, learning
33
23/7/08 11:17:40

should cover the intervention techniques, their impacts and the process. As part of the process evaluation, the appropriateness of the behavioural models used in the intervention should be assessed. The evaluation ndings should inform both the development of the intervention itself and future interventions in similar policy areas. Since Lewins eld theory, theoretical approaches to change have recommended developing theory through practice. Piloting and evaluation, followed by building learning back in, are the nal stages in a process of intervention development which can be characterised as learning through doing. It is notable that the more programmatic Intervention Mapping framework describes the process of intervention development as iterative, and includes the possibility of returning to previous stages in the process in order to readjust the intervention in the light of subsequent learning. Guidance on policy development needs to be explicit; however, it must also be exible to different contexts. Perhaps fearing that their framework would be adopted too rigidly, IMs authors wrote that one of the potential drawbacks of any policy model is that it will be used as a cookbook (Bartholomew et al 1998). Mindful of these inherent risks, the Nine Principles proposed here are underpinned by evidence from theory which ultimately suggests that behaviour change is best pursued as a craft not a science. To this end it is also strongly recommended that research analysts and policy makers work together throughout the Nine Principles to minimise the risks of developing inexible interventions.
34
23/7/08 11:17:40
Appendices
i) Tables: Matching Behaviours to Models
The two tables below are intended to support analysts and policy makers in the process of model selection. The tables summarise all the references made in the sources under review in which a behaviour (or type of behaviour) is linked to a behavioural model (or factor). These matches are indicative; some are supported by empirical data in the sources referenced, and others are not. In all cases the matches they show should be investigated further before selecting a model (or models) for use in designing an intervention strategy. Table A1: Behaviours matched to models (plus factors), by behaviour domain
[Behaviour] Community Participation Blood Donation Community Participation Community Participation Community Participation Voter Choice Voter Choice Voter Choice Voter Turnout Voter Turnout Consumption Buying Domestic Appliances Consumption Consumption (convenience) NOA Spaargaren and Van Vliets Consumption as Social Practices (self/social identity) (attitudes: automatic) Bagozzi and Warshaws Theory of Trying (social/cultural norms) Taylor and Todds Theory of Composting as Altruism (socio-technical regimes) (contextual factors: information, costs, incentives) Bedfords Environmental Considerations for Food Purchasing Shove 2003 Gatersleben and Vlek 1998 Jackson 2005; Burgess and Nye 2006 Shove 2003 Maio et al 2007 Bagozzi et al 2002 (habit / past behaviour) CLGs Model of Community Empowerment (social/cultural norms) (social capital) Theory of Reasoned Action Clarke et als Valence Politics Model (habit / past behaviour) Theory of Planned Behaviour Whiteley and Seyds General Incentives Model Lewis 2007 CLG 2008 Knott et al 2008 Putnam 2000 Ajzen 1991 Clarke et al 2004 Lewis 2007 Ajzen 1991 Clarke et al 2004 [Model (or Factor)] [References]
Consumption Purchasing Choices Shopping Environment Climate Change Composting Energy Consumption Energy Efciency Behaviours Food Choice
Knott et al 2008 Darnton et al 2006 Shove 2003 Wilson and Dowlatabadi 2007 Bedford 2002
35
23/7/08 11:17:40
APPENDICES
Home Energy Use Pro-Environmental Behaviour (information eg. better billing, smartmeters) (values: altruism) Lewis 2007; Wilson and Dowlatabadi 2007 Thogersen and Olander 2006; Berglund and Matti 2006 Thogersen 2007 Barr et al 2005 Stern 2000; Jackson 2005 Burgess and Nye 2006 Burgess and Nye 2006 Wilson and Dowlatabadi 2007 Barr et al 2005
Pro-Environmental Behaviour Recycling Recycling Recycling Recycling Solar Microgeneration Waste Reduction Health Addiction (eg. Smoking) Alcohol / Drug Use (Giving Up) Alcohol Consumption Alcohol Consumption
(personal norms) Barrs Path Analysis Models of Recycling Behaviour Sterns ABC Model (social norms; personal norms inc neutralisation) (contextual factors: infrastructure) (social norms: descriptive) Barrs Path Analysis Models of Reducing Behaviour PRIME Prochaska and Di Clementes Transtheoretical Model (Stages of Change) (contextual factors: cost) Dahlgren and Whiteheads Main Determinants of Health Model (social norms) Gibbons and Gerrards Prototype/Willingness Model (social norms) Theory of Reasoned Action Theory of Planned Behaviour (self efcacy) Theory of Planned Behaviour Gibbons and Gerrards Prototype/Willingness Model Theory of Planned Behaviour Rosenstocks Health Belief Model (information eg. labelling) (affect: preference/pleasure) (contextual factors: cost) Theory of Planned Behaviour Foresights Obesity System Map
West 2006 Prochaska and Velicer 1997
Lewis 2007 Dahlgren and Whitehead 2007 Rimal et al 2005 Gibbons et al 2003 Schultz et al 2007 Ajzen 1991 Ajzen 1991 Armitage and Conner 2001 Gibbons et al 2003 Gibbons et al 2003 Gibbons et al 2003; Baker et al 2007 Becker et al 1977 Lewis 2007 Conner 2007 Maio et al 2007 Maio et al 2007 Foresight 2007
Alcohol Consumption (esp. Young People) Binge Drinking Binge Drinking Bottle-feeding Babies Condom Use Condom Use Dental Flossing Drug Use (esp. Young People) Eating Vegetables Food Choice Food Choice Food Choice Food Choice Food Choice Obesity
36
23/7/08 11:17:40
APPENDICES
Obesity Dahlgren and Whiteheads Main Determinants of Health Model (social/cultural norms) (personal norms) Dahlgren and Whitehead 2007 Knott et al 2008 Maio et al 2007 Becker et al 1977 Loewenstein et al 2001 Bagozzi et al 2002 Dahlgren and Whitehead 2007 Knott et al 2008
Obesity Positive Health Behaviours
Preventative Health Behaviours Rosenstocks Health Belief Model Public Health Scares (Boycotts) Safe Sex Smoking Loewenstein et als Risk as Feelings Model (affect: anticipatory) Dahlgren and Whiteheads Main Determinants of Health Model (contextual factors: cost) Prochaska and Di Clementes Transtheoretical Model (Stages of Change) Protection Motivation Theory (self efcacy) Social Learning Theory (social proof) (social norms: descriptive) Theory of Planned Behaviour Protection Motivation Theory Prochaska and Di Clementes Transtheoretical Model (Stages of Change)
Smoking Smoking (Giving Up)
Prochaska and Velicer 1997
Smoking (Giving Up) Smoking (Giving Up) Smoking (in Public) Smoking (in Public) Taking Exercise Tooth Brushing Using Contraception
Rogers 1975 Prochaska and Velicer 1997 Dawnay and Shah 2005 Halpern et al 2003 Ajzen 1991 Rogers 1975 Prochaska and Velicer 1997
Using Sunscreen
Prochaska and Di Clementes Shepherd 2006 Stages of Change: decisional balance (social identity: in-group norms) Lane and Potters Car Buying Model Theory of Interpersonal Behaviour Terry et al 2000
Using Sunscreen Transport Car Buying Choice Car Use Car Use Car Use Car Use Car Use Congestion Charging (Paying) Crossing the Street Drink Driving Drink Driving Public Transport Use
Anable et al 2006 Anable et al 2006
Bamberg and Schmidts Model Bamberg and Schmidt 2003 of Car Use (affect) (self identity: symbolic) (habit / convenience) Festingers Cognitive Dissonance Protection Motivation Theory Gibbons and Gerrards Prototype/Willingness Model (affect: anticipatory) (personal norms / values) Anable et al 2006 Anable et al 2006 Lewis 2007 Darnton et al 2006; Knott et al 2007 Rogers 1975 Gibbons et al 2003 Bagozzi et al 2002 Anable et al 2006
37
23/7/08 11:17:40
APPENDICES
Public Transport Use Seatbelt Use Seatbelt Use Seatbelt Use Speeding Speeding Work & Savings Incapacity Benet (Reducing Claims) Pensions / Investments Other Adultery Crime Crime Education Retention Education Retention Littering Playing the Lottery (Gambling) Tax Evasion Gibbons and Gerrards Prototype/Willingness Model Halperns Tripartite Model of Crime (collective efcacy) (contextual factors: incentives eg. EMAs) (social/cultural norms) Cialdinis Focus Theory of Norms (affect: anticipatory) Gibbons and Gerrards Prototype/Willingness Model Gibbons et al 2003 Halpern 2001 Pearce 2007 Knott et al 2008 Knott et al 2008 Cialdini et al 1990 Bagozzi et al 2002 Gibbons et al 2003 Rosenstocks Health Belief Model (inertia) Talbot et al 2007 Talbot et al 2007 (contextual factors: infrastructure) Theory of Planned Behaviour (social norms) (past behaviour / habit) Gibbons and Gerrards Prototype/Willingness Model (social norms: descriptive, injunctive) Lewis 2007 Gibbons et al 2003 Demos/Green Alliance 2003; Dawnay and Shah 2005 Lewis 2007 Gibbons et al 2003 Jackson 2005
38
23/7/08 11:17:40
APPENDICES
Table A2: Types of behaviour matched to models
[Type of Behaviour] Addictive (Early Stages) Gibbons and Gerrards Prototype/Willingness Model Addictive (Established) Altruistic / Helping Coping (involving Risk) Delinquent (Inaction) General - Frequent / Habitual / Low Consciousness General Under Total Volitional Control General Under Less Total Volitional Control Less Appropriate Preventative (inc. Health) Risky / Emotional Visible / Public Wests PRIME Theory Schwartzs Norm Activation Theory Rogers Protection Motivation Theory Sykes and Mazas Norm Neutralization Theory Triandis Theory of Interpersonal Behaviour Gibbons et al 2003 West 2006 [Model] [Reference]
Schwartz 1977; Jackson 2005 Rogers 1975 Burgess and Nye 2006 Triandis 1977; Jackson 2005
Fishbein and Ajzens Theory of Ajzen 1991; Jackson 2005 Reasoned Action Ajzens Theory of Planned Behaviour Gibbons and Gerrards Prototype/Willingness Model Rosenstocks Health Belief Model Loewensteins Risk as Feelings Theory Cialdinis Focus Theory of Norms Ajzen 1991 Gibbons et al 2003 Becker et al 1977 Loewenstein et al 2001 Cialdini et al 1990
39
23/7/08 11:17:40
ii)
Selected References
Abraham, C and S Michie 2007. A taxonomy of behavior change techniques used in interventions [in press]. Ajzen, I 1991. The Theory of Planned Behavior. Organizational Behavior and Human Decision Processes 50, 179-211. Anable, J, B Lane and T Kelay 2006. An Evidence Base Review of Public Attitudes to Climate Change and Transport Behaviour. The UK Energy Research Centre et al for the Department for Transport. Argyris, C and D Schon 1996. Organizational Learning II. Reading, Massachusetts: Addison Wesley. Armitage, C and M Conner 2001. Efcacy of the Theory of Planned Behaviour: a Meta-Analytic Review. British Journal of Social Psychology 40, 471-499. Bagozzi, R, Z Grnao-Canli and J Priester 2002. The Social Psychology of Consumer Behaviour. Buckingham: Open University Press. Baker, A, G Rees, M Stanopoulous, S Bakhshi, A Surujlal-Harry 2007. Overcoming barriers to a healthy food choice in girls from low income and ethnic minority groups. Anna Baker et al at London Metropolitan University for Food Standards Agency. Bamberg, S and P Schmidt 2003. Incentives, Morality, Or Habit? Predicting Students Car Use for University Routes With the Models of Ajzen, Schwartz, and Triandis. Environment & Behavior 35 (2), 264-285. Bandura, A 1977. Self-efcacy: toward a unifying theory of behavioral change. Psychological Review 84, 191215. Barr, S and A Gilg and G Shaw 2005. Promoting Sustainable Lifestyles: A Social Marketing Approach Technical Paper. Stewart Barr et al at the University of Exeter for Defra CAD. Bartholomew, K, G Parcel and G Kok 1998. Intervention Mapping: A Process for Developing Theory and EvidenceBased Health Education Programs. Health Education and Behavior 25, 545-563. Becker, M, D Haefner and L Maiman 1977. The health belief model in the prediction of dietary compliance: a eld experiment. Journal of Health and Social Behaviour 18, 348-366. Bedford, T 2002. Sustainable Lifestyles Report 1. Tracey Bedford at the University of Westminster for the DfT New Horizons Research Programme. Berglund, C and S Matti 2006. Citizen and Consumer: the dual role of individuals in environmental policy. Environmental Politics 15 (4) 550-71. Burgess, J and M Nye 2006. Interim Report II on Sustainable Waste Behaviours of GAP EcoTeams Participants. Jacqui Burgess and Michael Nye at the Centre for Global Risk at the University of East Anglia for Defra Waste R&D Programme. Chapman, J 2004. System Failure (2nd edition). London: Demos. Cialdini, R, R Reno and C Kallgren 1990. A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places. Journal of Personality and Social Psychology 58(6) 1015-1026. Clarke, H, D Sanders, M Stewart, and P Whiteley 2004. Political Choice in Britain. Oxford: OUP . CLG 2008. An Evidence Pack on Community Engagement and Empowerment. The Local and Regional Governance Research Unit at the Department of Communities and Local Government.
23/7/08 11:17:40
Conner, M 2007. Assessing inuences on food choices in children using social cognition models. A presentation to the Food Standards Agency Workshop on Food Choices, March 2007. Dahlgren, G and M Whitehead 2007. European Strategies for Tackling Social Inequities in Health: Levelling up Part 2. World Health Organisation. Copenhagen: WHO. Dahlgren, G and M Whitehead 1991. Policies and Strategies to Promote Social Equity in Health. Institute for Futures Studies, reissued as Working Paper 2007 (14). Darnton, A 2007. 15 Principles for Practice from Behavioural Theory. Andrew Darnton for the Energy Saving Trust. Darnton, A 2005. Understanding Young Peoples Drug Use. Andrew Darnton for FRANK and the COI. Darnton, A, J Elster-Jones, K Lucas and M Brooks 2006. Promoting Pro-Environmental Behaviour: Existing Evidence to Inform Better Policy Making, Chapter 1: Theory. London: Defra. Dawnay, E and H Shah 2005. Behavioural Economics: 7 Principles for Policymakers. The New Economics Foundation. Defra 2008. A Framework for Pro-Environmental Behaviours: Annexes. London: Defra. Demos / Green Alliance 2003. Carrots, sticks and sermons: inuencing public behaviour for environmental goals. A Demos/Green Alliance Report for Defra. Fisher, J, W Fisher, A Bryan and S Misovich 2002. Information-Motivation-Behavioral Skills ModelBased HIV Risk Behavior Change Intervention for Inner-City High School Youth. Health Psychology 21 (2), 177186. Foresight 2007. Tackling Obesities: Future Choices - Project Report (2nd Edition). London: Government Ofce for Science. French, J and C Blair Stevens 2005. The Social Marketing Pocket Guide. Dr Jeff French and Clive Blair Stevens for the National Social Marketing Centre and the Department of Health. Gatersleben, B and C Vlek 1998. Household Consumption, Quality of Life and Environmental Impacts. In Noorman, KJ and AJM Schoot-Uiterkamp (eds.) Green Households? Domestic Consumers, Environment and Sustainability. London: Earthscan, 141-183. Gibbons, R, M Gerrard & G Lane 2003. A Social Reaction Model of Adolescent Health Risk. Social Psychological Foundations of Health and Illness, eds. J Suls and K Wallston. Oxford: Blackwell. Government Chief Social Researchers Ofce 2003. The Magenta Book: Guidance Notes for Policy Evaluation and Analysis. Cabinet Ofce. Halpern, D 2001. Moral values, social trust and inequality: can values explain crime? British Journal of Criminology 41, 236-251. Halpern, D, C Bates and G Beales 2003. Personal Responsibility and Behaviour Change. Prime Ministers Strategy Unit, Cabinet Ofce. Hampshire County Council et al 2005. More from Less: A Stakeholder Perspective. Hampshire County Council Material Resources Strategy. Hobson, K 2001. Sustainable Lifestyles: rethinking barriers and behaviour change, Chapter 11 in Cohen and Murphy (eds) Exploring Sustainable Consumption: environmental policy and the social sciences. Amsterdam: Elsevier Science.
23/7/08 11:17:40
HM Treasury 2003. The Green Book: Appraisal and Evaluation in Central Government. London: TSO Jackson, T 2005. Motivating Sustainable Consumption: A Review Of Evidence On Consumer Behaviour And Behavioural Change. A report to the Sustainable Development Research Network. Jowell, R 2003. Trying It Out. The Role of Pilots in Policy Making. Government Chief Social Researchers Ofce. London: Cabinet Ofce. Knott, D, S Muers and S Aldridge 2008. Achieving Culture Change. The Prime Ministers Strategy Unit. London: Cabinet Ofce. Kollmuss, A and J Agyeman 2002. Mind the Gap. Environmental Education Research 8 (3), 239-260. Kurani, K and T Turrentine 2002. Marketing Clean and Efcient Vehicles: A Review of Social Marketing and Social Science Approaches. The Institute of Transport Studies at the University of California. Lewin, K 1951. Field Theory in Social Science: Selected Theoretical Papers. D Cartwright (ed.). New York, NY: Harper & Row. Lewis, M 2007. States of Reason: Freedom, Responsibility and the Governing of Behaviour Change. London: ippr. Loewenstein, G, E Weber, C Hsee and N Welch 2001. Risk as Feelings. Psychological Bulletin 127 (2) 267-286. Maio, G, B Verplanken, A Manstead, W Stroebe, C Abraham, P Sheeran, M Conner 2007. Social Psychological Factors in Lifestyle Change and Their Relevance to Policy. Journal of Social Issues and Policy Review, 1 (1) 99-137. McKenzie-Mohr, D 2000. Promoting Sustainable Behavior: An Introduction to Community-Based Social Marketing. Journal of Social Issues 56 (3), 543-554. NICE 2007. Behaviour change at population, community and individual levels. NICE Public Health Guidance No. 6. London: NICE. Pearce, N 2007. Rethinking Fairness. Public Policy Research 14 (1), 11-22. Plsek, P 2003. Structures, Processes and Patterns. A presentation to the NHS Trent Improvement Network, http:// www.tin.nhs.uk/index.asp?pgid=1401 Prochaska, J and W Velicer 1997. The Transtheoretical Model of Health Behavior Change. American Journal of Health Promotion 12 (1), 38-48. Putnam, R 2000. Bowling Alone: The collapse and revival of American community. New York, NY: Simon & Schuster. Rimal, R, M Lapinski, R Cook and K Real 2005. Moving Toward a Theory of Normative Inuences: How Perceived Benets and Similarity Moderate the Impact of Descriptive Norms on Behaviors. Journal of Health Communication, 10 (5), 433 450. Rogers, E 1995. Diffusion of Innovation (5th edition). New York: Free Press. Rogers, R 1975. A protection motivation theory of fear appeals and attitude change. Journal of Psychology 91, 93-114. Rosenthal, R and D Rubin 1982. A simple general purpose display of magnitude of experimental effect. Journal of Educational Psychology 74, 166-169.
23/7/08 11:17:40
Scharmer, O 2007. Theory U. Cambridge, Massachusetts: Society for Organisational Learning. Schein, E 2004. Organizational Culture and Leadership (3rd edition) San Francisco, Cal.: Jossey-Bass. Schultz, PW, J Nolan, R Cialdini, N Goldstein and V Griskevicius 2007. The Constructive, Destructive, and Reconstructive Power of Social Norms. Psychological Science 18 (5), 429-434. Schwartz, S 1977. Normative Inuences on Altruism. Advances in Experimental Social Psychology 10, 222-279. Senge, P 1990. The Fifth Discipline. London: Random House. Shepherd, R 2006. The use of the stages of change model with dietary behaviours. In R Shepherd & M Raats (eds) The Psychology of Food Choice, . Wallingford: CABI. Shove, E 2003. Comfort, Cleanliness and Convenience. London: Routledge. Stern, P 2000. Towards a Coherent Theory of Environmentally Signicant Behaviour. Journal of Social Issues 56 (3), 407-424. Stern, P T Dietz and G Guagnano 1995. The New Ecological Paradigm in Social-Psychological Context. , Environment and Behavior 27 (6), 723-743. Talbot, C, E Canavan and K Miles 2007. Understanding Behavioural Change: Theoretical Principles and Guidelines for Policy. Clare Talbot at the DWP Strategy Unit. Terry, D, M Hogg and B McKimmie 2000. Attitude-behaviour relations: The role of in-group norms and mode of behavioural decision-making. British Journal of Social Psychology 39, 337-361. Thgersen, J 2007. The motivational roots of norms for environmentally responsible behaviour. Proceedings of the Nordic Consumer Policy Research Conference 2007, 1-16. Thgersen, J and F lander 2006. To What Degree are Environmentally Benecial Choices Reective of a General Conservation Stance? Environment and Behavior 38 (4), 550-569. Triandis, H 1977. Interpersonal Behaviour. Monterey, CA: Brooks/Cole. Vare, P and W Scott 2007. Learning for a Change: Exploring the Relationship Between Education and Sustainable Development. Journal of Education for Sustainable Development 1 (2), 191198. Wenger, E 1999. Communities of Practice. Cambridge: CUP . West, R 2006. Outline of a Synthetic Theory of Addiction. Available at http://aspsilverbackwebsites.co.uk/prime/ Resources/The%20synthetic%20theory%20of%20addiction2.doc Wilson, C and H Dowlatabadi 2007. Models of Decision Making and Residential Energy Use. Annual Review of Environment and Resources 32, 169203.
23/7/08 11:17:40
www.gsr.gov.uk
Crown copyright 2008. Produced by HMT Publishing Unit 2008.
23/7/08 11:17:40

GSR Behaviour Change Knowledge Review: Practical Guide: An Overview of Behaviour Change Models and Their Uses

Uploaded by

Copyright:

Available Formats

GSR Behaviour Change Knowledge Review: Practical Guide: An Overview of Behaviour Change Models and Their Uses

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GSR Behaviour Change Knowledge Review: Practical Guide: An Overview of Behaviour Change Models and Their Uses

Uploaded by

Copyright:

Available Formats

GSR Behaviour Change Knowledge Review

PU547- GSR behaviour change practice guide 2.indd 1

Practical Guide: An overview of behaviour change models and their uses

Andrew Darnton, Centre for Sustainable Development, University of Westminster

PU547- GSR behaviour change practice guide 2.indd 3

PU547- GSR behaviour change practice guide 2.indd 4

Government Social Research Unit

PU547- GSR behaviour change practice guide 2.indd 5

PU547- GSR behaviour change practice guide 2.indd 6

Appendices i) ii) Tables: Matching behaviours to models Selected references 35 40

GSR Behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 7

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 8

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 9

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 10

Overview of models, theories and frameworks

Models of behaviour at the individual level

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 11

OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

B eliefs about outcomes Attitude towards the behaviour

R e lativ e importance of attitude and norm

B eliefs about what others think

P erceiv ed B ehavioural Control

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 12

OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 13

OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

Models of behaviour at higher levels of scale

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 14

OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

Opportunities Availability , adve rtis ement, price s , s hops

Abilities F inancial, temporal, s patial, cognitive, phy s ical

B ehav ioural c ontrol

Cons umer behaviour

S ubjec tiv e well-being , environmental quality

Lewins Change Theory

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 15

OVERVIEWS MODEL THEORIES AND FRAMEWORKS Diffusion Models

Organisational Learning Models

Systems Thinking Approaches

Applied models and frameworks

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 16

OVERVIEW OF MODELS THEORIES AND FRAMEWORKS

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 17

GSR behaviour change knowledge review

PU547- GSR behaviour change practice guide 2.indd 18

Making use of models

Limits of behavioural models

Models are concepts, not representations of behaviour