3.3  Risk and resilience

Precise pathways or causal sequences explored in developmental models may have little relevance to the practical choice of intervention since they are so complex that they may not offer much policy guidance on where and how to invest in relation to specific outcomes. The difficulties of determining causation have led to the use of a pragmatic risk and protective factor approach to intervention (see for example US Surgeon General, 2001). This approach is used to identify the factors that indicate that an individual is at increased risk of negative outcomes, and those that are likely to lead to a decreased risk of such an outcome. A caveat in using this approach is that while such factors may be correlated with particular outcomes, they are not necessarily causal, and may simply act as markers. For example, while having a teenage mother may be a risk factor for a number of poor outcomes for children, early motherhood per se is not necessarily the underlying cause of subsequent problems, as discussed below. Risk and protective factors differ for particular outcomes, although common themes emerge.

Risk factors are characteristics of a group of people that are associated with an elevated probability of undesired outcomes (Masten 1994). In contrast, a protective factor decreases the likelihood that a risk factor will be associated with a harmful effect. Risk is a statistical concept and actual outcomes for people who are in a risk group, or exposed to risk, will vary. According to (Huffman, Mehlinger and Kerivan 2000a, p 5):

Risk factors will predict problematic outcomes, but may or may not be causally related to the onset or maintenance of problems … to show a characteristic as a risk factor requires a demonstration that the risk factor temporally precedes the adverse outcome and is correlated with it. A characteristic that cannot be shown to precede the outcome but is related to it is a “correlate”, not a risk factor. A major problem in the past literature is that correlates are presented as risk factors, and sometimes even as causal factors.

Risk factors are not necessarily causes of adverse states of well-being, but indicate the probability that some negative outcome will occur (ie, they include a broader range of variables than strict causal models would indicate). Researchers typically identify risk factors by tracking the development of children and measuring the extent to which particular factors are linked statistically to later outcomes such as violent behaviour.

Risk factors are useful for targeting resources and interventions on particular people. Much research isolates and measures the probabilistic effect of particular variables, such as family composition, income levels, or neighbourhood characteristics, on particular outcomes for targeting.

The interaction of a number of factors can increase the probability of a particular adverse or positive outcome. For children aged 2 to 3 years, the number of risk factors has been found to be a strong predictor of negative outcomes (Landy and Kwan Tam 1998). A meta-analysis concludes that all negative outcomes have several risk factors in common, as shown in Table 6 (Durlak 1998). There are multiple pathways and processes which determine the ultimate outcomes (Landy and Kwan Tam 1998). For example, poor educational attainment is typically found to be strongly associated with all the main negative outcomes in childhood (Bynner n.d., Kiernan 1995). Although these risk factors are common to a number of adverse outcomes, specific risk factors are relevant for particular adverse outcomes.

Source: Durlak (1998); Hema (2000)

Risk factors appear to have a multiplicative or cumulative effect – the likelihood of adverse outcomes grows disproportionately as a child is exposed to more risk factors. Any single risk factor makes a relatively modest contribution to individual risk, and exposure to one or two risk factors, unless they are extreme, is unlikely to have a negative impact on the likelihood of a child’s development proceeding normally (Fergusson 1998). Having four or more risk factors, however, can lead to a ten-fold increase in the probability of poor outcomes irrespective of their causal role (Landy and Kwan Tam 1998, Zimmerman and Arunkumar 1994).

Risk, like causality, is a dynamic process. Children can be exposed to multiple risk factors, can move in and out of risk at various times, can be exposed to risk factors at varying levels over the course of their life, and may be more vulnerable to particular risk factors at different points in their development. Risks can accumulate over time, or can be additive at one particular point in time. Risk factors may operate in causal chains, with one risk factor leading to another, with the ultimate result being a problematic outcome. Within a causal chain there may be “mediator” risk factors (eg, harsh discipline), which explain how an earlier risk factor (eg, low income) works to produce a negative outcome (eg, behaviour problems in school), and “moderator” risk factors (eg, male gender), which help to identify the individuals on whom this additional risk factor is most likely to exert an effect (Huffman et al 2000b).

There are a number of limitations regarding the use of risk factors to identify those who are likely to experience bad outcomes in the future and who thus merit intervention before these outcomes occur (US Surgeon General 2001). No single risk factor or set of risk factors can predict negative well-being with certainty, particularly not at the individual level. Not all individuals who exhibit all the best-established risk factors for violence become violent, for example. Risk factors can be used for reasonably robust prediction in groups with particular characteristics, but are far less useful for individuals. Some risk factors are not amenable to change, and so are not good targets for intervention (Huizinga, Loeber and Thornberry 1995). Even if some risk factors, such as poverty, can be changed, they may not be realistic targets for change, in the short run, at least. Longitudinal studies used to identify risk factors may not identify those factors that are relevant in specific situations or that are relevant to all population sub-groups due to limitations of size and representativeness of samples.

Most children living in disadvantaged circumstances grow up to be well-functioning members of society. The fact that some children are resilient in the face of adversity has led to an interest in identifying the factors and processes that may protect children from outcomes.

Resilience is “a dynamic process encompassing positive adaptation within the context of significant adversity” (Kalil 2001). Resilience occurs only with exposure to risk and adversity – it involves successful engagement with risk, rather than an evasion of risk. Resilience, like risk, alters with circumstances so that the attributes or factors that create resilience in one situation may not produce it in another.

Protective effects may operate by buffering or mitigating the negative effects of a risk factor (eg, assertiveness and high self esteem are protectors against the negative influence of parental conflict on the mental health of teenage girls) or by enhancing the protective effects of variables found to decrease the probability of negative outcomes (eg, cultural identity enhances the effects of self-esteem as a protective factor lowering the risk of alcohol and drug use)(Brook, Brook, Gordon and Whiteman 1990). The presence of protective factors explains why individuals who face the same degree of risk may be affected differently.

Not all protective factors are common to all children or are similarly protective against all risks. Nevertheless there may be common characteristics amongst those children who achieve positive outcomes despite facing a high level of risk and adversity during childhood, as shown in Table 6 (Durlak 1998, Masten and Coatsworth 1998). For example, positive social relationships with siblings and with peers and adults outside the immediate family have been found to play an important role in buffering children against the negative effect of risk factors such as marital conflict and low income.

Developmental research suggests the existence of three types of protective factor: the individual characteristics of the child; a close relationship within the family, not necessarily with a parent; and a certain type of social support outside the family (Bynner 1998a). The internal resources of a child that promote resilience are an easy temperament, at least average intelligence and a sense of self-efficacy. One strong, emotionally close relationship in the family seems to protect children even when the family is otherwise dysfunctional. Social support outside the family that combines a close relationship with practical help can promote resilience. Although general protective factors play a role in the general resilience of children, specific factors are relevant in conferring protection against specific adverse outcomes (Bynner 1998b).