An Overview of the Risks and Protectors for Children of Separation and Divorce

4. Way in Which Factors Combine to Increase and Decrease Disturbance in Children: Exacerbation and Protection

One of the most important findings to have come out of risk research in the last 20 years is that risks do not operate in isolation.  In section 2, we saw that even though children who experience parental divorce show more disturbance than children who do not experience divorce, most children are not permanently affected by this experience.  What explains why some children show major problems following divorce, while others seem to come through the experience relatively unharmed (Hetherington et al., 1998)?

As we saw in section 3, some of this difference between children can be explained by components of the divorce process (e.g. conflict, economic instability) that contribute to the severity of the stress.  Even when we know about the severity of the divorce components, however, there is still a lot of unexplained variation in children's adjustment.  The reason for this is that divorce occurs in the context of many other events and circumstances (both before and after the divorce) that put children at risk.  Both multiplicative effects of risks on the one hand, and the beneficial effects of protective factors on the other are important for understanding children's adjustment to divorce.

4.1  Exposure to Multiple Risks

Factors that increase the level or rate of disturbance in children are called risk factors.  The main risk factors for children, apart from divorce and conflict, include harsh parenting (Dodge, Bates and Petit, 1990); mental illness in a parent (e.g. depression); parental substance abuse problems (Quinton and Rutter, 1985); being born to a teenage mother (Brooks-Gunn and Chase-Lansdale, 1995); poverty; being raised in a high-crime neighbourhood (Sampson et al., 1997); experiencing many transitions in parental care (Henry et al., 1996); and having a learning disability (Moffitt, 1993).

Rates of disturbance in children have been examined as a function of how many risks children experience in their lives.  This has been done by adding up all the risks to which the child is exposed, regardless of the type of risk.  A child raised by a teenage mother, who had an alcoholic father and whose parents divorced would have a score of three risks on a cumulative risk index.  Rutter (1979) found that children exposed to one risk showed a rate of disturbance comparable to children who did not experience risk.  Children exposed to two risks showed a four-fold increase in their rate of disturbance.  Children exposed to four or more risks showed a ten-fold increase in disturbance.  Jenkins and Keating (1998) examined data from the National Longitudinal Survey of Children and Youth (NLSCY) to investigate the cumulative effects of risk in a Canadian sample of school-aged children.  Their cumulative risk index included exposure to divorce as one of the risks.  They found that among children exposed to two risks in their lives, their rate of disturbance, based on teachers' reports, was 16 percent.  Children exposed to four or more risks showed a 50 percent rate of disturbance.  These data show that risks combine to multiply detrimental effects.  Another way to express this is that risks potentiate one another.

Most children are not exposed to a large number of serious risk factors.  Jenkins and Keating (1998) found that only four percent of children in a nationally representative sample of Canadian children experience four or more risks in their lives.  Nonetheless, risk factors do cluster within families.  Canadian data on factors that predict divorce can illustrate this effect.  O'Connor and Jenkins (2001) looked at the effects of individual, family and community factors in predicting marital breakdown.  Factors such as the number of previous relationships, parental depression and low income all predicted the risk of subsequent marital breakdown.  This has also been found in other population studies in the United Kingdom and the United States (Capaldi and Patterson, 1991; Dunn et al., 1998; O'Connor et al., 1998).  Thus, children experiencing parental divorce are more likely than others to be exposed to multiple serious risks.  The same effect is seen with post-divorce risk factors.  In section 3, we reviewed research showing that lower family income, more negative parent-child relationships and less contact with the non-custodial parent were potential consequences of parental divorce.  Thus, children experiencing parental divorce are often exposed to multiple stressors rather than just one.  When this happens their risk of showing disturbance increases markedly, as the effect of multiple risks potentiate one another.

4.2  Protective Effects

Another way of approaching the issue of how factors operate in conjunction with one another is to examine how positive aspects of children's lives combine with stressful events or risk to reduce the likelihood that children will be affected negatively.  This has been called the study of protective effects.  With respect to the consequences of divorce to children, it has been said that there are "winners, losers and survivors" (Hetherington, 1989).  Let us now consider a resiliency perspective (Hetherington et al., 1998) to understand the factors that protect or buffer children from the stress of divorce.

We must consider an important methodological issue before reviewing the findings on protective factors and divorce.  There are basically two methodologies for examining protective effects (Jenkins and Smith, 1990; Rutter and Pickles, 1987).  The second method we shall describe is much more instructive than the first.  The first involves investigation of a high risk group only.  Investigators examine the factors within the high risk group that explain better adjustment in children.  For instance, in a group of children undergoing parental divorce (high risk group) it might be found that those with easy temperaments show less disturbance than those with more difficult temperaments.  This would lead one to conclude that easy temperament "protects" children undergoing parental divorce from negative outcome.  However, if we had a group of children in low risk circumstances (their parents were not undergoing divorce), we would probably also find children with easy temperaments to be showing better adjusted than children with difficult temperaments.  What sense does it make to say that these children are "protected" by their easy temperaments when they do not need protection?  They do not need protection because they are not at risk.

This criticism led to a second wave of studies on protective effects in which children in high and low risk circumstances were investigated, and protective effects were defined as those factors that reduced disturbance among high risk children, but had little or no effect on adjustment among children at low risk (Rutter and Pickles, 1987).  This type of protective pattern is assessed by examining the statistical interaction between the risk factor and the protective factor.  If the interaction is significant, this means that the protective factor shows a different relationship to children's outcomes when it occurs in high and low risk circumstances.

Because the study of protection is relatively new, only two studies have examined protective effects in divorced families.  Therefore, we have also drawn on studies of family conflict.  As conflict often precedes and follows parental separation, such findings are relevant to children exposed to divorce.

Jenkins and Smith (1990) examined protective factors in children experiencing marital conflict in the general population.  Fifty-seven children in high conflict homes were compared with 62 children in low conflict homes.  The children were between the ages of 9 and 12 years, and children in the two levels of risk were matched for gender, family size, father's social class and geographical area.  Two types of effects were seen.  Factors such as the quality of the mother-child relationship were found to be associated with children's adjustment in low and high conflict homes.  Even if children were not experiencing parental divorce, their adjustment was better if they got along well with their mothers.  The father-child relationship showed the same pattern of effect.

On the other hand, three factors were found to operate differently in low and high risk groups of children.  For instance, a close relationship with an adult outside the family (usually a grandparent) was associated with better adjustment among children in high conflict homes, but made little difference to adjustment among children in low conflict homes.  This acted as a buffer for children in high conflict homes.  Similarly, a child in the high risk group who had a close relationship with a sibling or who engaged in an activity in which he or she gained positive recognition benefited from these protective factors.  This suggests that there are compensatory processes in development.  When children are missing something in their lives because of exposure to stress, other aspects of their environment can compensate.

Formoso, Gonzales and Aiken (2000) also examined protective effects in high conflict families.  The sample included 284 10- to 16-year-old boys and girls in families experiencing high or low conflict.  They examined the protective effects of parent-child attachment, peer attachment and parental monitoring with respect to the development of conduct disorders (stealing, lying, physical violence, etc.) and childhood depression.  Parental attachment and parental monitoring were found to reduce the risk of conduct disorder in girls experiencing high conflict, but no association with disorder was seen among girls experiencing low conflict.  The same effect was not seen for boys.  The results indicate that in the presence of high family conflict, attachment and parental monitoring may protect girls from exhibiting high levels of conduct problems.

Two studies have examined protective effects for children in divorced families.  O'Connor and Jenkins (2001)[1] reported on a representative sample of Canadian families (using NLSCY data), followed up over a two-year period.  Some families had experienced a separation over this period.  As data concerning children's well-being were available before and after their parent's separation, it was possible to examine changes in children's functioning over the period of the separation and divorce.  More importantly, it was possible to examine protective effects.  If children did show an adverse reaction to the divorce, what were the factors that predicted which children showed a less adverse reaction?  Symptoms of anxiety and depression increased as a function of parental divorce, but symptoms of hyperactivity and aggression showed little change.  This was largely because children who would later experience parental divorce already showed more problems with hyperactivity and aggression prior to the parental divorce than children whose parents did not divorce subsequently.  Warmth in the mother-child relationship was identified as the most important factor protecting children from increases in anxiety and depression following their parents' divorce.

Recently, Lengua et al. (2000) looked at the protective effects of temperament on the relationship between parenting and adjustment problems in children of divorced families.  Two aspects of parenting, namely parental rejection and inconsistent discipline, were examined in relation to three aspects of temperament:  positive emotionality (smiling, laughing, sensitivity to positive environmental cues), negative emotionality (fear, frustration, sensitivity to negative environmental cues) and regulation (attention, impulsivity and inhibition).  The researchers were interested in child depression and conduct difficulties as outcome measures.  The sample included 231 mothers and children who had recently experienced divorce (the mean time since the divorce was 1.1 years).  The age of the children ranged from 9 to 12 years and the sample was approximately equal in the number of boys and girls.

Children high in positive emotionality were found to show better adjustment when exposed to parental rejection than children low in positive emotionality.  Positive emotionality was less strongly associated with children's adjustment if they were not exposed to parental rejection.  Illustrating the way that risks can potentiate one another (as described in section 4.1), when children were impulsive in homes characterized by inconsistent discipline, they showed increased problems with depression and conduct disorder.  Impulsivity had a less negative impact on children's outcomes when parents were consistent in their discipline.  In conclusion, these findings indicate that children who are high in impulsivity may be at greater risk for maladjustment, whereas positive emotionality may help buffer children from the effects of negative parenting in divorced families (Lengua et al., 2000).

[1] Human Resources Development Canada supported this work and we are grateful to that department for allowing us to present these results in this report.