Why do protective factors protect? An evolutionary developmental perspective

Highlights

• An evolutionary developmental approach can help us to advance our understanding of protective factors.

• Individual differences in antisocial and prosocial behaviour can be linked to alternative life history strategies.

• What counts as a protective factor depends on specific characteristics of individuals and the environments in which they are embedded

• Ultimately, a better understanding of why protective factors protect will contribution to more effective intervention efforts.

Abstract

A growing body of research has identified a list of protective factors that decrease the likelihood of an individual engaging in criminal and antisocial behaviour. However, despite substantial empirical advances in our understanding of the factors that protect against offending, relatively little theoretical work has been directed at explaining why these factors serve a protective function and the notion of protective factors, like risk factors, is conceptually problematic. In this article I argue that we can advance our understanding of what protective factors are and why they serve a protective function by taking an evolutionary developmental perspective. More specifically, an evolutionary informed framework is provided that outlines two broad models for understanding individual differences in antisocial and prosocial behaviour. First, I suggest that individual differences in antisocial (and prosocial behaviour) are linked to the development of alternative life history strategies in response to different environmental contexts. More specifically, I argue that ‘protective’ factors can often be conceptualised in terms of features of the environment and/or feature of individuals that promote and/or reflect the development of slow life history strategies. Second, drawing from recent research on individual differences in plasticity I highlight how our conceptualisation of protective (and, indeed, risk) factors is contingent on the interaction between individual susceptibility to environmental influence and relevant environmental contexts. In other words, often what counts as a ‘protective’ factor depends on individual differences in plasticity. Inevitably, as our understanding of why protective factors protect improves, there will be advances in the nature and scope of our interventions to improve prosocial outcomes and prevention efforts will be enhanced.

Introduction

Consider the following two young men: Charles and Henry. Charles is impulsive, has a low IQ, was neglected by his parents, grew up in an impoverished neighbourhood, and associates with antisocial peers. Henry, in contrast, has above-average intelligence, low impulsivity, grew up in a warm and nurturing family in a middle-class neighbourhood and associates with non-deviant peers. Which of these two young men is most, and which is least, likely to have a record of violent and antisocial behaviour? The answer is pretty straightforward. Because Charles has the most risk factors for antisocial and criminal behaviour he is more likely to have a record of antisocial behaviour than Henry. Henry has the most protective factors so he is much less likely to have a record of antisocial behaviour compared to Charles. Developmental research on the antecedents of criminal and antisocial behaviour has provided an extensive list of well replicated risk factors for offending (Farrington, 2015). Knowledge of these factors have allowed researchers to predict which individuals are more likely to become serious and persistent offenders (Andrews & Bonta, 2010), and which individuals are at a greater risk for re-offending. A smaller, but growing body of complementary research has also identified a list of protective factors that decrease the likelihood of an individual engaging in criminal and antisocial behaviour (Lösel & Farrington, 2012), and mark an individual as less likely to re-offend (de Vries Robbé, de Vogel, & Douglas, 2013). However, despite substantial empirical advances in our understanding of the risk factors for antisocial behaviour, our understanding of the causal processes that link these factors to offending has not been fully developed (Durrant, 2016, Kraemer et al., 2001, Ward and Fortune, 2016), and many have argued that the very notion of ‘risk factors’ lacks conceptual clarity (Ward, 2016). Similarly, although there are now a number of studies that have identified factors that protect against offending (Lösel & Farrington, 2012), relatively little theoretical work has been directed at explaining why these factors serve a protective function and the notion of protective factors, like risk factors, is conceptually problematic (Serin, Chadwick, & Lloyd, 2016).

In this article I argue that we can advance our understanding of what protective factors are and why they serve a protective function by taking an evolutionary developmental perspective. First, I briefly outline how protective factors have been conceptualised in the research literature and what factors have been identified as serving a protective function. I then provide a broader discussion of how an evolutionary approach can help us understand both antisocial and prosocial behaviour within the normative context of human development. In other words, before we can begin to understand why certain factors promote risk whereas other factors protect we need to have a conceptual understanding of the deeper causal origins of both antisocial and prosocial behaviour. Clearly there are also important individual differences in the likelihood that individuals will become more serious and persistent offenders and in the next section a framework is provided that highlights two idealised ‘models’ for understanding individual differences from an evolutionary perspective. These two models can help us to understand not only why some individuals are more likely to be offenders than others (i.e. why risk factors promote offending), but also why some individuals are less likely to become offenders (i.e. why protective factors protect).

There is now an extensive literature on risk factors for offending. Broadly speaking, two separate (but related) literatures have addressed the characteristics of individuals, families, and environments that make some individuals more likely to offend than others: the first draws from developmental research and focuses on risk factors for offending; the second focuses largely on forensic populations and documents the risk factors for reoffending or recidivism (Durrant, 2016). Similarly, a number of researchers have directed their attention to identifying the developmental factors that protect against offending (Glowacz and Born, 2015, Lösel and Farrington, 2012), while others have focussed on how certain factors might make some individuals less likely to re-offend, and to develop assessment tools that can effectively measure these factors (e.g., de Vries Robbé et al., 2013). Here I focus on the developmental factors that might serve a protective function. First, it is important to be clear just what is meant in the literature by a ‘protective’ factor.

Although there is no widespread agreement on how to best conceptualise protective factors in developmental research, there is an emerging consensus that two different types of protective factor can be identified (Glowacz and Born, 2015, Lösel and Farrington, 2012). First, a direct protective factor is a characteristic of an individual or environment that is associated with a decreased likelihood of specific problem behaviours (aggression, violence, and other forms of antisocial behaviour) occurring. A direct protective factor, thus, can be viewed as a mirror image of a developmental risk factor – a factor that is associated with an increased likelihood of specific problems behaviours occurring. Second, a buffering protective factor is conceptualised as a factor that moderate the impact or influence of risk factors. In other words, in the presence of risk factors a factor that decreases the likelihood of problem behaviours occurring can be viewed as a buffering protective factor (Glowacz and Born, 2015, Lösel and Farrington, 2012). There is clearly much overlap in the use of this term with the widely employed notion of resilience, which is also often employed to capture the idea of largely prosocial development in the presence of multiple risk factors (Glowacz and Born, 2015, Jaffee et al., 2007). A couple of final points are worth noting about the conceptualisation of protective factors. First, as with risk factors, a dose-response relationship is often assumed: the more protective factors (direct or buffering) that an individual has the less likely they are to exhibit antisocial behaviour (e.g., Andershed, Gibson, & Andershed, 2016). Second, although it is possible in some cases to identify a protective factor as simply the opposite end of a continuous variable which might otherwise be labelled as a risk factor (e.g., low IQ is a risk factor, whereas high IQ is a protective factor), it is important to be able to conceptualise protective factors in ways that make them independent from risk factors (Hall, Simon, Mercy, Loeber, Farringotn & Lee, 2012), although in practice this does not always occur.

The evidence base for protective factors is not nearly as extensive as that for risk factors, but a number of longitudinal studies have identified a range of individual, family, school, peer, and neighbourhood factors that protect against offending (Andershed et al., 2016, Dubow et al., 2016, Farrington et al., 2016, Kim et al., 2016, Pardini et al., 2012). A thorough review of this literature is well beyond the scope of this article (see Lösel & Farrington, 2012), but it will be useful to highlight some key findings. It is well recognised that low IQ and low levels of educational attainment are important risk factors for later offending (Farrington & Welsh, 2007). Similarly, above average IQ appears to serve as both a direct and a buffering protective factor against antisocial behaviour. In a meta-analysis of fifteen longitudinal studies that assess the role of intelligence as a protective factor, Ttofi et al. (2016) found that a higher level of intelligence significantly reduced the risk for offending in both low risk (direct protective factor) and high risk (buffering protective factor) groups. Interestingly, the protective effects of intelligence were substantially greater in the high risk group suggesting that a high IQ and related outcomes (e.g., better educational attainment) can play an important role in promoting resilience in the face of risk. Other individual protective factors identified in the literature include negative attitudes towards delinquency (Pardini et al., 2012), anxiety, popularity and religious attendance (Dubow et al., 2016), higher levels of self-control (Farrington et al., 2016), and various biological factors such as high resting heart rate and genetically mediated differences in monoamine oxidase A (MAO-A) activity (Kim-Cohen et al., 2006, Portnoy et al., 2013). We should probably also include gender (female) as a protective factor. Although gender has not often been a specific focus of most research on protective factors, the enduring gender gap in offending, especially for more serious violent offences, suggests that there are characteristics of females that (on average) make them less likely to offend (Campbell, 2013). Family factors that have demonstrated protective effects include having more positive, supportive parents, greater parental supervision, parents less engaged in antisocial behaviour and more positive child-parent relationships (e.g., Dubow et al., 2016, Farrington et al., 2016, Jolliffe et al., 2016). Perhaps unsurprisingly, given the robust relationship between delinquent peers and offending, individuals with more prosocial peers are less likely to offend (e.g., Hall et al., 2012b, Pardini et al., 2012), and greater educational aspirations and stronger attachment to school environments appear to have a protective effect (Dubow et al., 2016, Farrington et al., 2016). Living in a good neighbourhood and higher quality housing also have both direct and buffering protective effects (Jolliffe et al., 2016).

There is, of course, some variation in the protective factors identified in different samples and some factors are protective at some ages or developmental periods but not at others. However, taken together, the research on protective factors highlight the importance of secure, positive, and nurturing family, peer, school, and community environments on the development of prosocial behaviour along with specific psychological characteristics such as high IQ, good self-regulatory capacities, and prosocial attitudes. Despite this accumulating body of research, a number of key empirical and conceptual issues remain. First, although many protective factors are ‘consistent’ with mainstream criminological theories, there is a lack of understanding about the relevant causal processes that link protective factors with positive outcomes. In other words, there is no clear understanding of why protective factors protect. Second, although much of the research described above has focussed on criminal and antisocial behaviour, there is widespread agreement that there are strong relationships between offending and a range of ‘problem’ behaviours including substance abuse, and various kinds of risk-taking, along with other facets of human functioning particularly as they relate to intimate relationships and sexual behaviour (e.g., Copping and Campbell, 2015, Copping et al., 2013). This reminds us that ‘antisocial and criminal’ behaviour may itself be somewhat of an artificial construct that is part of a broader set of individual differences in social functioning. Third, although there is a recognition that there is a dose-response relationship between protective factors and the relative absence of antisocial behaviour, the focus remains on how factors work in isolation rather than considering how they cluster together to form ‘integrative phenotypes’. In sum, although substantial advances have been made in the identification of protective factors, there is scope for further work that helps us to understand why these factors realise their protective effects. Before turning to a discussion of how protective factors can be conceptualised from an evolutionary development perspective, I briefly outline how an evolutionary approach can provide insight, more broadly, into the nature of both prosocial and antisocial behaviour.

There is an enormous body of research, across academic disciplines, devoted to understanding why human engage in aggression, violence, and other forms of antisocial and criminal behaviour. Surprisingly little research, however, has focussed on why most individuals do not commit acts of serious violence or become persistent offenders (e.g., Tyler, 2006). This is an important task, because a better understanding of the causal processes that protect against offending will provide opportunities for improving effective interventions to prevent crime. In order to understand why certain factors might protect against aggressive and antisocial behaviour we need to locate such acts within the broader context of normative human development. In other words, before we can make theoretical progress in understanding individual differences in antisocial behaviour (why some, but not others), and the factors that predict those individual differences (risk and protective), we need to have a general understanding of – for want of a better term – our ‘human nature’ (Durrant & Ward, 2015). As Agnew, 2011, Agnew, 2013 has pointed out, most criminological theories make implicit or explicit assumptions about the kind of animal that we are. Many prominent theories (e.g., control theories) assume that humans are basically selfish and that we are motivated to pursue our self-interests even if this means harming others. The central explanatory task, therefore, becomes one of identifying the features of individuals, their development, and the environments in which they are embedded that promote prosocial or non-criminal behaviour. In other words, there is a natural focus on what factors protect against antisocial behaviour. Other prominent theories appear to take as their starting point the idea that humans are largely cooperative, prosocial animals (e.g., strain theories) and hence the key task becomes one of accounting for deviations from this prosocial nature. In short, the focus turns to risk factors for antisocial behaviour (see Agnew, 2013, Durrant and Ward, 2015 for a more detailed discussion).

From an evolutionary perspective, neither view of human nature – fundamentally selfish, or fundamentally cooperative – provides the complete picture. Rather, natural selection has shaped humans in ways that promote prosocial and cooperative behaviours in many contexts, yet also favours competition among individuals in ways that can lead to aggression, violence, and other forms of antisocial behaviour in other contexts. Space precludes anything like a full discussion of this topic, but a few key points are worth noting. First, there is abundant evidence for humans' prosocial nature. Indeed, the large-scale cooperation that is a feature of human societies is unprecedented among vertebrate species and reflects the outcome of both biological and cultural evolutionary processes (see Crespi, 2014, Henrich, 2016). Evidence for the strong human tendency to cooperate comes from a range of sources including the tendency to adhere to (and punish violators of) social norms (Boehm, 2012, Matthew and Boyd, 2014), and developmental research that highlights the early emergence of prosocial and cooperative behaviour independently of explicit instruction (e.g., Hamlin, 2013, Tomasello and Vaish, 2013). Humans, in a very real sense, are ‘natural born co-operators’.

However, Darwinian theory is fundamentally concerned with relative fitness. Therefore, behaviours that promote differential reproductive success are favoured even if they might involve the infliction of harm on others. The capacity for aggression and violence is almost certainly an evolved feature of human nature with deep phylogenetic roots (see Gomez, Verdu, González-Megías, & Mendez, 2016). Other types of harmful behaviours – cheating, theft, exploitation – may also have been favoured by evolution (Durrant and Ward, 2015, Roach and Pease, 2013). It is crucial to note, however, that such behaviours are highly contingent: given our evolution in small, highly cooperative social groups bounded by strong social norms, the violations of others' rights are only likely to be beneficial (in the evolutionary currency of reproductive fitness) under specific circumstances, for particular individuals. A better understanding these person × environment interactions will contribute to our knowledge about the key causal pathways that underlie the protective effects of specific factors. Finally, what we think of as ‘antisocial’ or ‘criminal’ behaviour may reflect specific outcomes that are part of a broader range of behavioural tendencies that promote reproductive fitness for some individuals in some circumstances. More generally, a good deal of criminal and antisocial behaviour can arguably be conceptualised as either the direct or indirect outcome of intra-sexual competition among males – especially young males – for status and resources (Durrant & Ward, 2015). Thus, various forms of risk-taking, antisocial behaviour, offending, and a tendency to pursue short-term sexual relationships are both highly correlated and are strongly patterned by age and gender: they are more pronounced in males compared to females, and during adolescence and early adulthood compared to other developmental periods (see Durrant & Ward, 2015 for a detailed explanation for why this is the case from an evolutionary point of view). Crucially for the argument developed in more detail in the next section, there are also individual differences in the strategies that individuals employ to realise social status.

There is unlikely to be any widespread agreement on just what constitutes human nature. However, steering a course between Rousseau and Hobbes, an evolutionary perspective highlights how a history of living in small, cooperative groups has favoured the development of prosocial behaviour and a suite of moral emotions (e.g., empathy, guilt), and institutions (e.g., the punishment of norm violations) that promote cooperation under most circumstances. However, various forms of risk-taking and antisocial behaviour have been favoured by males more than females, and during some developmental periods more than others, as the evolutionary benefits of these behaviours outweigh costs for specific individuals at specific points in their developmental trajectories. We need to recognise, then, that both gender (female) and age (any period outside adolescence/young adulthood) are protective factors for offending because antisocial behaviour is likely to reap less (evolutionary) rewards. A final point that is worth making is that although we tend to view prosocial and antisocial behaviour as, respectively, good and bad, for society, from an evolutionary point of view antisocial behaviour may for some individuals, under certain circumstances, advance reproductive success and therefore is an ‘adaptive’ outcome even though it may involve the infliction of harm on others.

What accounts for individual differences in the likelihood of engaging in antisocial and criminal behaviour? More specifically, what features of individuals and their developmental experiences allow us to predict which individuals are more (or less) likely to engage in serious and persistent patterns of offending? These questions are the heart of developmental research on risk and protective factors. However, relatively little consideration has been given to the evolutionary origins of individuals differences. That is, why do individual differences arise at all? Like developmental criminologists, evolutionary psychologists assume that all phenotypic differences emerge through the complex interaction of genetic and environmental interactions across the life course of an individual. An evolutionary approach, however, highlights how these interactions may not simply reflect the outcome of stochastic processes but may, instead, be the result of selection for phenotypic plasticity in ways that promote adaptive outcomes. In other words, although some individual differences may arise as a result of non-adaptive genetic mutations or through the impact of extreme environments on developing individuals, much of the variation in prosocial and antisocial behaviour that we see is likely to reflect adaptive responses to different environment contexts (Durrant and Ward, 2015, Ellis et al., 2012). Moreover, the degree to which individuals may be responsive to features of the environment may also have undergone selection so that some individuals are more or less responsive to factors that promote or protect against antisocial behaviour (Belsky & Pluess, 2009).

In the following Section I discuss how this approach provides two broad models that can help to explain why protective factors serve their protective function. The first model highlights how recurrent features of the developing environment tend to promote slow life history strategies. In other words, this model focuses on adaptive phenotypic plasticity: individuals vary their behavioural strategies in response to cues from the environment in ways that promote beneficial evolutionary outcomes in the environments that they find themselves (Ellis, Figueredo, Brumbach, & Schlomer, 2009). The second model highlights the importance of recognising adaptive individual differences in plasticity: some individuals are more plastic or open to features of the environment than others and hence are more (or less) influenced by the impact of risk and protective factors (Belsky & Pluess, 2009). A consideration of these two models provides the basis for re-thinking just what protective factors are and how they might work to realise their protective outcomes.

Life history theory is centrally concerned with the way that organisms make trade-offs in investment among key life tasks: growth, body maintenance and development, learning, mating, and parenting (Hochberg & Belsky, 2013). Different species allocate investment in different ways to these tasks and can be arrayed along a continuum from slow to fast life histories. Organisms with slow life histories are characterised by delayed growth, later sexual maturation, investment in long-term sexual relationships, and considerable parental investment. Those with fast life histories tend to develop more rapidly, reproduce earlier, produce more offspring and invest less in parenting and offspring (Bjorklund & Ellis, 2014). Evolutionary biologists have also recognised that within species variation in life history strategies are also common: individuals faculatively adjust their life history strategy along the slow-fast continuum in ways that are most likely to promote reproductive success in the environments that they find themselves. The application of life history theory to human development has resulted in a rich body of both theoretical (e.g., Ellis and Del Giudice, 2014, Ellis et al., 2009) and empirical research (e.g., Belsky et al., 2012, Copping and Campbell, 2015, Copping et al., 2013, Simpson et al., 2012) that largely supports the assumption that humans adjust (although not consciously) their life history strategy in response to features of the environment during critical periods in development. More specifically, environments that are harsh (i.e. have cues indicating high levels of violence and morbidity/mortality) or unpredictable (signalled by frequent shifts in residence, family structure and resource availability) are argued to entrain fast life history strategies characterised by accelerated psychosocial development, a greater orientation towards mating rather than parenting (for men), and a tendency to focus on immediate over delayed rewards. Given environmental cues indicating an uncertain future it makes sense from an evolutionary perspective to focus more on immediate rather than delayed reproduction with concomitant changes in cognitions (short time horizon, impulsivity) and sociosexual orientation (more short-term, relatively uncommitted, relationships) (Ellis et al., 2009, Frankenhuis et al., 2016b).

Humans then, like other organisms, are hypothesised to adjust their behaviour in response to specific cues from the environment that signal the most adaptive way to invest in various life history tasks. From this perspective, many of the characteristics that we label as ‘problem behaviour’ – risk taking, uncommitted sexual relationships, antisocial behaviour, and criminal offending, may reflect a relatively ‘fast’ life history orientation characterised by higher levels of aggression, intra-sexual competition, and a focus on immediate gratification (Durrant and Ward, 2015, Ellis et al., 2012). In contrast, individuals who are motivated to develop more enduring sexual relationships, invest in offspring, are relatively risk-averse, and avoid physical aggression are likely to be ‘pursuing’ a slow life history strategy. This suggests that what we view as protective factors are those that either (a) promote or trigger development in ways consistent with a slow life history strategy and/or (b) are downstream consequences of developmental environments that promote slow life history strategies. For example, living in a good neighbourhood and having warm, supportive, prosocial parents provide cues that suggest that future environments are likely to be stable and comparatively free from external risk thus promoting development in ways that encourage delayed reproduction, a long time horizon and the motivation to avoid risk (including crime and antisocial behaviour). Cognitive factors such as good self-regulation and low levels of impulsivity can be viewed, in part, as downstream consequences of these environmental contexts which, in turn, promote more positive relationships with school environments and engagement with prosocial peers. In short, protective factors serve to protect against antisocial behaviour because they are part of a broader orientation towards social and sexual relationships characterised by a slow life history strategy. This approach not only helps us to understand why protective factors protect but also has the advantage of considering offending and antisocial behaviour within the broader context of individual differences in social and sexual behaviour. Clearly more work is needed to flesh out this broad approach. First, there is a need to fully elucidate the underlying causal mechanisms that contribute to shifts in life history strategy, with the stress response system promoted as the most likely candidate (Ellis & Del Giudice, 2014). Second, it will be important to identify the key environmental cues that are important in signalling aspects of future environments and how they affect developing individuals during different periods throughout the lifespan. Currently, early childhood has been the key focus of most life history approaches but other developmental periods may also be important (Del Giudice & Belsky, 2011).

According to life history theory, then, individuals adjust their behavioural strategies in response to salient cues from the environment in ways that promote evolutionarily relevant outcomes. Protective factors, from this perspective, are a mixed bag of environmental cues (features of the neonatal, family and social environment) and outcomes (e.g., increased capacity for self-regulation) that reflect the development of slow life history strategies. A decreased likelihood of engaging in antisocial behaviour is one manifestation of this behavioural strategy. This approach assumes that individuals are behaviourally plastic: they reliably adjust their behaviour in response to enduring features of the environment. However, there is also a widespread recognition among evolutionary biologists that there are also significant individual differences in plasticity that reflect the outcome of selective processes. In other words, individuals vary in the extent that their behaviour is plastic in response to features of the developing environment (Dingemanse and Wolf, 2013, Stamps, 2015).

Two models, developed independently, capture the idea that individuals vary in their susceptibility to environment influences: biological sensitivity to context theory (BSCT) (Boyce & Ellis, 2005), and differential susceptibility theory (DST) (Belsky, 2005, Belsky and Pluess, 2009). Although the original models vary in some respects, they converge on the central idea that some individuals are more susceptible to environmental influence than are others (Ellis, Boyce, Belsky, Bakermans-Kranenburg, & van IJzendoorn, 2011). This susceptibility may arise through genetic and/or environmental factors (and their interaction), but it ultimately leads to differences in how individuals respond to the environment. For some, highly susceptible individuals, features of the environment will have a substantial impact on their developmental trajectory; for others, development is more canalised and they are more impervious to the impact of the environment. Crucial to this approach is the idea that these individual differences in susceptibly mean that some individuals are more affected by both adverse and beneficial environments. In other words, environments are more likely to affect their developmental outcomes for better and for worse (Belsky, Bakermans-Kranenburg, & Van Ijzedoorn, 2007). It should be noted here that the terms ‘better’ and ‘worse’ reflect developmental outcomes from the point of view of mental health or criminal justice outcomes as, from an evolutionary point of view, they are likely to be equally adaptive – in ancestral environments at least (Ellis et al., 2011).

Why might evolution favour individual differences in plasticity? Although facultative adjustments to environmental cues, as noted above, may have been selected for in order to match phenotypes to future environments, ultimately the future is unknowable. Thus, alongside adaptive phenotypic plasticity, selection may have favoured individual differences in plasticity that arise either through exposure to particular environmental cues, or via genetically (or epigenetically) mediated differences (Ellis et al., 2011, Frankenhuis et al., 2016a). The most prominent approach suggests that individual differences in susceptibility arise through a ‘bet-hedging’ process. Parents can best promote their own reproductive success by have offspring who vary in the extent that they are susceptible to environmental influences because it is more likely that at least some offspring will best match likely future conditions thus ‘spreading the risk’.

To get a clearer idea of the potential implications of DST for understanding the role of protective factors it is worth contrasting DST with the more widely employed diathesis-stress model. According to this perspective some individuals possess risk factors (diatheses) that make them more likely to develop adverse outcomes in response to particular kinds of (adverse) environments. In contrast, according to DST, these same individuals may also benefit more from other kinds of (more supportive) environments. From this perspective, certain characteristics may be either risk or protective factors depending on the environment in which they are embedded: they serve as risk factors in stressful, and protective factors in supportive, environments. In other words, characteristics that promote a greater degree of behavioural plasticity may result in the greater likelihood of negative outcomes (antisocial behaviour) in adverse environments, and positive outcomes (prosocial behaviour) in favourable environments. Individuals who have characteristics that make them less responsive to environmental influences, in contrast, are less likely to be influenced by both positive and negative environments.

What does this approach entail for our understanding of protective (and risk) factors? First, what counts as a protective (or risk) factor may itself be dependent on the environment. Individuals who are more behaviourally plastic who find themselves in positive environments are more likely to respond in ways that promote positive outcomes, whereas those who find themselves in negative environments will be more adversely affected. What counts as a direct protective factor, then, depends on it being embedded in a favourable environmental context. In contrast, individuals who are less behaviourally plastic may be less affected by adverse environments: they have characteristics that could be viewed as buffering protective factors and therefore they are more resilient in the face of otherwise developmentally risky contexts. They are also less open to the positive influences of more favourable environments. To complicate matters somewhat, some scholars have also proposed that some individuals may be more responsive only to positive, but not negative, experiences (Pluess & Belsky, 2013). If resilience is viewed as the capacity to be less adversely affected by negative environments, vantage sensitivity captures the idea that positive (but note negative) influences will affect some individuals more than others.

An accumulating body of research supports the underlying assumptions of differential susceptibility theory (see Bakermans-Kranenburg and Van Ijzendoorn, 2015, Van Ijzendoorn and Bakermans-Kranenburg, 2015). Most of the research focus has been directed at genetic variants which are related to functional differences in three different neurotransmitter systems: MAO-A, serotonin, and dopamine. One influential early study found that children with a genotype that coded for low levels of MAOA activity were at a significantly greater risk for engaging in antisocial behaviour, but only when exposed to childhood maltreatment (Caspi et al., 2002) – a result largely supported in a more recent meta-analysis (Byrd & Manuck, 2014). This research clearly suggests that some individuals are at a greater risk for engaging in antisocial behaviour when exposed to certain risk factors compared to others with different genotypes. Research that has examined the role of polymorphisms in serotonin transporter genes indicate that individuals with the 5HTTLPR short allele are influenced more by both negative and positive environments suggesting that 5HTTLPR is a marker for plasticity or differential susceptibility: individuals with the short allele benefit most from positive environments, but are also most adversely affected by negative environments (Baptista et al., 2016, Van Ijzendoorn et al., 2012). Similar results have been found for research on variations in dopamine-system genes with variations in specific dopamine gene polymorphisms related to greater susceptibly to both positive and negative environmental contexts (Bakermans-Kranenburg & Van Ijzendoorn, 2011). In a more recent meta-analysis of randomised controlled experiments that have tested the differential susceptibility hypotheses for these genetic variants, van Ijzendoorn and Bakermans-Kranenburg (2015) found strong evidence to support the idea that individuals with ‘susceptible’ genotypes were influenced to a significantly greater extent by interventions to promote positive outcomes than were individuals without less susceptible genotypes. In other words, some individuals, by virtue of their inherited genotypes, were more ‘amenable’ to the protective effects afforded by the psychosocial interventions than were others. Future works, they suggest, should now examine the mechanisms that underlie these differences and it is clear that this line of research has important implications for intervention efforts (Bakermans-Kranenburg and Van Ijzendoorn, 2011, Bakermans-Kranenburg and Van Ijzendoorn, 2015, Belsky and van Ijzedoorn, 2015).

In Table 1 the various different ways in which protective factors operate to reduce the likelihood of antisocial behaviour are outlined. In many cases, protective factors will promote positive outcomes in response to positive environments because they result in the development of coordinated slow life history strategies that entail less risk taking, aggression, and antisocial behaviour. The first of the evolutionary models, then, suggests that protective factors reflect the outcome of adaptive phenotypic plasticity promoting slow life history strategies. According to the second evolutionary model, protective factors may exert their positive influence on behavioural outcomes due to individual differences in plasticity: some individuals may be (a) less influenced by adverse environments and therefore demonstrate resilience (buffering protective factor); (b) be more influenced by positive environments, thus demonstrating vantage sensitivity (direct protective factor); or (c) be more or less influenced by both positive and negative environments such that high susceptibility acts as a protective factor in positive environments and low susceptibility acts a protective factor in adverse environments.

The approach to understanding protective factors advocated in this section can, I have argued, help us to better understand the causal processes that underlie protection from antisocial and criminal behaviour. An evolutionary developmental perspective has the virtue of integrating multiple levels of analysis, including distal (in terms of evolutionary function and history), developmental (in terms of life history trajectories), and more proximate (in terms of specific psychological and physiological processes) (Durrant & Ward, 2015). It thus promotes a broader and – arguably – more explanatorily complete account of protective factors than those offered by others that focus largely on developmental (Lösel & Farrington, 2012) or proximate psychological (de Vries Robbé et al., 2013) levels of analysis. A more complete understanding of the causal processes underlying protection should, in turn, contribute to improvements in effective interventions to reduce antisocial behaviour.