Understanding Youth Offending

Risk Factor Reserach, Policy and Practice

Stephen Case, Kevin Haines

• Publisher: Willan
• Year: 2013
• Print ISBN: 9781843923411
• eBook ISBN: 9781134028986
• Edition: 1
• Topic: Social Sciences
• Subtopic: Criminology
• Index: Social Sciences

This book aims to provide an understanding of youth offending and policy and practice responses, particularly the risk-focused approaches that have underpinned much recent academic research, youth justice policy and interventions designed to reduce and prevent problem behaviour. There has been growing concern, however, on the part of critical criminologists and others, about the theoretical, epistemological, methodological and ethical bases of risk-focused research with young people. They have pointed particularly to the overly-deterministic and prescriptive nature of the risk factor paradigm.

This book aims to meet the need for an exploration of youth justice and youth offending which takes account of the origins and contemporary manifestations of risk-focused work with young people. It analyses the influence of concepts of risk upon policy development in both England and Wales as well as internationally, highlighting tensions between the proponents of risk factor research and methodological and ethical criticisms of the risk factor paradigm. It will be essential reading for anybody wishing to understand risk factor explanation of crime, contemporary youth justice policy and responses to offending behaviour.

Chapter 1

---

Examining the unresolved
methodological paradoxes of risk
factor research

A large body of research investigating the risk factors associated with youth offending has been developed by criminologists working predominantly in Western Europe, North America and Australasia. The majority of this risk factor research has been positivist in nature and empirically led — typically viewing risks as quantitative, quantifiable, objective, value-fee and scientific ‘facts’ with a consistent, predictable and possibly causal relationship with offending. To this end, risk factor research (RFR) has mainly utilised quantitative methodologies such as structured observation, standardised surveys, psychometric testing, secondary data analysis, random/representative sampling and the statistical control and analysis of variables to reach its ultimate conclusions. These methods have enabled measurement of the prevalence of certain characteristics, circumstances, experiences, behaviours, attitudes, perceptions etc in the lives of young people. This measurement of ‘risk’ has been facilitated by a process of quantification or ‘factorisation’ — attributing a numerical or ordinal value to real world observations to produce ‘variables’ that are amenable to statistical testing. Subsequent statistical analysis has then been conducted on these ‘independent variables’ to ascertain their relationship with the ‘dependent variable’ — offending. Those independent variables (or factors) which are statistically shown to be predictive or causal of the dependent variable (offending) are labelled as ‘risk factors’. In much RFR these risk factors have been identified at an early stage in a child's life (at age ten years or less) and shown to be statistically predictive of offending, usually in the mid-teenage years.

The scientific, empirical, technical and standardised nature of the dominant methodologies of RFR has led to the identification, replication and generalisation of risk factors across different populations of young people, in various countries at different times. Thus, it is claimed, RFR has produced an expanding and robust evidence-base of the psychosocial characteristics of young people which predispose them to offending. In many countries, particularly England and Wales, this research-based knowledge has been transferred into youth justice policy — underpinned by the assumption that if risk factors can be identified then interventions targeted on changing them for the better will reduce the likelihood of offending. Moreover, interventions predicated on this model are often labelled ‘what works’ in reducing youth offending and are offered as prescriptions for ‘effective practice’. Taken together, the identification of risk factors through RFR and their targeting in youth justice policy and practice constitute a ‘risk factor prevention paradigm’. Ostensibly, this paradigm has offered a simplistic, comprehensible and ‘evidence-based’ approach to controlling and managing youth crime and young offenders. Thus, one of the major proponents of RFR, David Farrington, has asserted:

A key advantage of the risk factor prevention paradigm is that it links explanation and prevention, fundamental and applied research, and scholars and practitioners. Importantly, the paradigm is easy to understand and to communicate, and it is readily accepted by policy makers, practitioners, and the general public. Both risk factors and interventions are based on empirical research rather than theories. (Farrington 2000: 7)

However, in reality, the RFR that has underpinned the risk factor prevention paradigm (RFPP) has not been as scientific, value-free, definitive or unequivocal as has often been claimed or depicted. This terrain has, in fact, been quite heavily contested, as has been reflected in the emergence of two very distinct camps and perspectives concerning the generation and application of knowledge through RFR. The first camp is an increasingly large number of researchers across the globe who have pursued quantitative RFR as a valid and productive way of understanding and explaining juvenile delinquency, buttressed by supporters and believers in political and policy-making offices (including many of those responsible for deciding how research funding is spent); the second camp is a smaller, disparate group of academics who have sought to evolve, reorientate, challenge, criticise and oppose the RFR movement and the manner and extent to which this research has been used to inform and develop youth justice policy and practice.

These two camps have proffered disparate views of both the methodological robustness and the practical and policy utility of RFR. Proponents have championed RFR methodology as both scientific (e.g. empirically-based, variable focused and statistically robust) and clinical (e.g. enabling the objective identification of risk as a set of factors that are purportedly predictive or causal of offending and that are amenable to treatment or change). These same proponents have accentuated the empirical basis and sustained validation, replicability and universality of risk factors as robust, conclusive research-based evidence of the utility of RFR for shaping policy and practice, protecting the public and reducing offending and/or re-offending by young people exposed to identified risks. Many opponents, on the other hand, have questioned the methodological, ethical and political bases of RFR and have doubted the validity of the conclusions drawn. Critics have asserted, for example, that the ‘factorisa-tion’ of risk equates to a reductionism that oversimplifies the context and operation of risk, individualises responsibility for subsequent offending and individualises responses to risk; arguing that such pervasive individualisation stigmatises and harms young people rather than helps them (O'Mahony 2008; White 2008). What has emerged has been a polarisation of opinion and a lack of direct engagement between the two camps. RFR has continued to spread and replicate itself, largely without reference to the valid (or otherwise) arguments of the critics. The cries of the critics have, for the most part, gone unheard by risk factor researchers, politicians and policy makers.

Our position does not fit neatly into either camp. We have carried out and published RFR (e.g. Case and Haines 2004, 2007, 2008; Haines and Case 2005), but simultaneously have taken a critical approach to the methodology and conclusions drawn of much pre-existing RFR (e.g. Haines and Case 2008; Case 2006, 2007; Case et al. 2005). In many ways, this book represents the process through which we have developed our own research and the conclusions we have reached about RFR and the uses to which it has been put in the development of youth justice policy and practice.

RFR has potential benefits, not least in affording a practical, real-world methodology for testing theoretical premises, identifying promising targets for intervention and applying knowledge of risk factors to the development of interventions. However, these strengths have been variously overstated and misrepresented by proponents of RFR, too often crudely understood by politicians and policy makers (and others) and frequently clumsily implemented in policies and programmes. On the other hand, critics have tended to proffer generalised critiques of RFR and its uses; criticisms that have often been based on an anti-positivist philosophy, moral or ethical objections and questions concerning the politics of crime control. Much of this critical work provides an important socialising counterbalance to the ‘psychosis of positivists’: the over-riding belief that the complexities of the human social world are reducible to a series of measurable factors which can be shown to exist in statistical relationships to the denial of alternative ways of knowing or understanding. However, rarely have the critics engaged with the substance and actuality of RFR. RFR has, therefore, until now, largely escaped the detailed critical analysis and exposé this book aims to provide.

The RFR movement has emerged from a body of longitudinal studies in the industrialised Western world (discussed in more detail in Chapter 2) that claim to have identified a range of ‘factors’ in childhood that predict an increased statistical likelihood of offending in adolescence. The prediction of adolescent offending from childhood factors has promulgated a very specific (yet almost exclusively implicit) within-individual, developmental understanding of exposure to early-life risk factors as the explanation of adolescent offending and/or antisocial behaviour. This developmental RFR has identified risk factors primarily within domains that have been labelled ‘psychological’ (e.g. relating to cognition, emotion, temperament) and ‘social’, the latter of which has typically been predicated on a restricted definition of the social as relating to the family, school, neighbourhood, peer group and lifestyle, rather than broader socio-structural factors such as gender, class, poverty, societal access routes to opportunities (France and Homel 2006) and the criminalising tendency of interactions with official agencies (McAra and McVie 2005; Armstrong 2004). As a result, a select group of psychological and social (otherwise known as ‘psychosocial’) risk factors for offending have been widely replicated across different groups of young people (e.g. different age groups, genders) across time, place and culture. This sustained and far-reaching replicability has been seen as a major strength of RFR and has been drawn on to impute a robustness and validity to the research findings.

The apparent predictive validity of a widely-replicated, globally-applicable set of psychosocial risk factors generated through RFR has been enormously attractive to politicians, policy makers and practitioners. For many such people, a developmental understanding of the risk-factor-offending relationship has provided clear, evidence-based targets for preventative and ameliorative (early) interventions that have the potential to ‘nip crime in the bud’. Thus, RFR has directed the limited time, resources and finances of policy makers and practitioners towards the most promising targets for intervention and afforded a convenient, commonsense and defensible response to a significant public and political concern.

However, critical criminologists have taken issue with the governmentality and interventionism arising from RFR (see, for example, Armstrong 2004; Pitts 2003a). These same critics have also expressed serious concerns over the methodological robustness and validity of the RFR paradigm; concerns that we intend to evaluate, extrapolate and add to throughout this text. In particular, it has been suggested (see, for example, Haines and Case 2008; O'Mahony 2008; Case 2007; Kemshall 2003; Pitts 2003a) that RFR has:

• relied inordinately on measuring and analysing risk as a broadly-phrased, quantitative factor that is aggregated across groups, thus encouraging a focus on the replication of statistical differences between-groups rather than within-individual changes;

• become dominated by deterministic and probabilistic developmental understandings of predictive, childhood risk factors at the expense of alternative and more holistic and complex explanations;

• lacked coherence and a clear, well-developed understanding of its central concepts, namely the definition of risk factors and the nature of their relationship with offending;

• produced findings that have been applied uncritically and over-simplistically by policy makers more interested in broad headlines than addressing the details and limitations of the research.

Exploring the unresolved methodological paradoxes of RFR

Thinking critically about RFR brings to light a number of inherent methodological issues that can be perceived as paradoxes that beset RFR. These paradoxes are by no means easy or even possible to resolve and (to a greater or lesser degree) reflect the current state of ability and knowledge in the field. Nevertheless, it is essential to illuminate these paradoxes and evaluate their impact on the reliability, validity and practical utility of the conclusions drawn from RFR. These methodological paradoxes are mutually exclusive but they also reinforce each other. They are reflected to a greater or lesser extent across the entire field of RFR as well as manifest variously in individual studies.

• Simplistic over-simplification — RFR has relied on the factorisation of risk in order to produce clear, comprehensible and practical findings. However, the greater the extent of factorisation, the greater the simplification of the lived real-life experiences of young people and the less accurate and representative of real life the data becomes. Moreover, factorisation has been combined with the aggregation of risk and the homogenisation of categories of offending and young people. These methodological and analytical processes have combined to radically over-simplify the lived real-life experiences of young people. Consequently, the data becomes less accurate and less representative of real life. In efforts to try to explain more, RFR is actually able to explain less.

• Definitive indefinity — researchers have confidently disseminated simplistic, clear-cut, evidence-based conclusions which have, in reality, been predicated on indefinite, unspecific and inconsistent definitions and understandings of key concepts. In particular, there has been ambiguity and lack of consensus over how to understand ‘risk factors’, ‘offending’ and the nature and temporal order of the risk-factor-offending relationship (e.g. deterministic or probabilistic, causal or predictive). Thus the attribute of ‘indefinity’ (i.e. a lack of definitiveness) in key concepts is what has enabled and promoted definitive conclusions.

• Risk-dependent protective factors — there has been a lack of consensus about the definition and understanding of so-called ‘protective factors’. The various definitions proffered have, however, explored the concept of protection almost exclusively as a ‘factor’ and in relation to risk (factors) and the prevention of negative outcomes, rather than as powerful elements exerting independent, autonomous and potentially-positive influences. Thus, attempts to explain non-offending, reductions in reoffending or positive outcomes have been limited in scope by notions of risk.

• Replicable incomparability — the development of an impressive range of ‘universal’, globally applicable risk factors has relied on studies collecting and analysing aggregated data. Aggregation has, however, rendered research findings largely unrepresentative of any particular individual differences or contextual differences relating to, inter alia, locality, nation, culture. Thus, the apparent replicability in the risk factors identified has not necessarily implied comparability in the nature or function of these risk factors across different samples.

• Unconstructive constructivism — a constructivist strand of RFR has challenged dominant quantitative, developmental understandings of risk and begun to address the aforementioned methodological paradoxes by exploring young people's ability to construct, negotiate and resist exposure to risk in their interactions with their environment. Constructivism, however, has been simultaneously wedded to dominant developmental definitions of risk factors (at least in part) and intent on exploring qualitative, nuanced, individualised and context-specific understandings of risk and its relationship to offending. Thus, the more constructivism has attempted to challenge and problematise factorised RFR, the less it has actually been able to offer genuine and practical alternatives, findin...

Table of contents

• Cover
• Half Title
• Full Title
• Copyright
• Contents
• Acknowledgements
• List of acronyms
• Introduction: risk factor research
• 1 Examining the unresolved methodological paradoxes of risk factor research
• 2 The origins and development of risk factor research
• 3 Longitudinal risk factor research in England and Wales — achievements, limitations and potential
• 4 Cross-sectional risk factor research in England and Wales — achievements, limitations and potential
• 5 Hunting for the universal risk factor
• 6 Risk assessment in the Youth Justice System: application without understanding?
• 7 Revisiting risk factor research, policy and practice
• Bibliography
• Index