This dissertation will examine the evidence for the claim that witnessing domestic violence causes serious and lasting harm to children. As it would not be feasible to conduct primary research on this topic at the researcher’s current level of training, given the significant ethical issues involved in working with children and families in this context, it will consist of an extensive critical review of the literature on this topic. This body of evidence will be systematically reviewed to establish the current state of knowledge regarding:
– The strength of the link between exposure to domestic violence and children’s
– Trauma symptoms
– Development
– Social functioning
– Internalising (eg. depression)
– Externalising (eg. aggression, disruptive behaviour)
– Academic performance
– The existence of mediating or moderating factors determining the level of damage caused by witnessing domestic violence, including
– Temperament
– Social support
– Genetic factors
– The prevalence of exposure to domestic violence in childhood.
Preliminary review of the literature
The prevalence of childhood exposure to domestic violence
Intimate partner violence is disturbingly common in the UK: an analysis of recent data gathered by the NHS for various purposes found a lifetime prevalence rate of some experience of domestic violence of 13-31% among the general population of British women (Feder et al, 2009). In the US, Dong et al found (2004) that 24% of respondents (n = 2, 081) indicated that they had been exposed to domestic violence while under the age of 18. These figures indicate that a high proportion of children will, at at least some point, witness acts of violence between (most commonly) their parents or caregivers in the home or another family setting. However, the usefulness of lifetime prevalence figures like this in assessing the real impact of domestic violence on children is rather questionable: these results do not distinguish adequately between individuals who witnessed a single incident, or very infrequent ” mild” violence, and those who were repeatedly exposed to serious violence. More detailed data is required to address the question of how common prolonged exposure really is.
A further problem with the analysis of data for the prevalence of children’s exposure to domestic violence is the high level of co-occurrence with other forms of maltreatment. A large US study (3, 777 males and 4, 411 females) found that 12. 3% of men (n = 482) and 15. 9% of women (n = 703; chi square of difference 15. 9, p < 0. 0001) had witnessed " maternal battering;" however, only 3. 7% of men (n = 139) and 3. 8% of women (n = 168) had only witnessed violence and not suffered physical and/or sexual abuse as well (Edwards et al, 2003). This suggests that it may be hard to identify the specific effect of witnessing violence alone; it is intuitively obvious that homes in which there is violence between partners are likely to be ones in which there is an atmosphere conducive to other forms of abuse, and the parents' orientation to the child is highly likely to be problematic.
Witnessing parental aggression: its effects on child development
There is strong evidence that aggression and violence between the child’s parents or caregivers can have serious negative consequences even if the child is ” too young to understand:” in particular, it has been suggested, very reasonably, that domestic violence negatively impacts the quality of maternal care as poor management of emotions and conflicts may transfer from the couple relationship to the mother-infant one (eg. Krisknakumar & Buehler, 2000). Indeed, women who are in violently abusive relationships may even express more negative attributions about their unborn child while pregnant (Theran, Levendosky, Bogat, and Huth-Bocks, 2005), creating the conditions for an emotionally distant parenting style which can lead to a poor attachment between mother and infant. Of four studies of children aged 3-6 reviewed by Wolfe et al (2003), all but one found moderate to strong effects on internalizing and externalizing symptoms as a consequence of witnessing domestic violence; Levendosky et al. (2002) also found a significant level of post-traumatic stress disorder (PTSD) symptoms in a similar population, and Bogat and her colleagues described clinically significant trauma symptoms in one-year-old infants exposed to family violence (2006). Interestingly, however, in a sample of 7865 British children aged 5-16, Meltzer and his colleagues found that ” Witnessing severe domestic violence almost tripled the likelihood of children having conduct disorder but was not independently associated with emotional disorders” (2009: 491). The picture is yet further complicated by the finding that at least some mothers who suffer domestic violence in fact appear to compensate for this in ways which increase their availability to their children, showing ” heightened sensitivity and responsiveness” (Letourneau, Fedick and Willms, 2007: 649).
Domestic violence and adolescent outcomes
Given the complexity of the picture of the effect of witnessing domestic violence (and of having a caregiver who is a victim or perpetrator of it) which has already emerged, it is to be expected that the impact of this form of maltreatment on the eventual outcomes of children who are affected by it will also be far from easy to determine. High levels of conduct disorder and other adjustment and attitudinal problems in the adolescent children of battered women have been extensively described (Fantuzzo et al, 1991; Holden and Ritchie, 1991, and numerous later studies); these conduct problems have, however, bee n found to be amenable to interventions to improve mothers’ own support, and management of their children (eg. Jouriles et al, 2001). McFarlane and her colleagues found, worryingly, that in a sample of 330 children (including black, white and hispanic ethnicities), ” the mean internalizing behavior score for boys 6-11… as well as girls and boys 12-18… of abused mothers were not significantly different from the clinical referral norms” (2003: 202), suggesting that the impact of witnessing serious domestic violence is enough to lead to clinically significant symptoms – including suicidality and self-harming behaviours – in adolescents. This indicates that, although the mechanisms by which it causes such great damage are as yet unclear, witnessing domestic violence which is either serious or prolonged needs to be treated as a major traumatic incident in a child’s life. However, the prevalence of exposure like this is so great that intervening in the vast majority of cases where harm is being caused would be impossible; we are, furthermore, learning ever more about the factors which determine whether or not these experiences take a lasting toll, both biological and social.
Domestic violence and the biology of trauma
While only a small fraction of the children who suffer maltreatment are, as it were, ” fortunate” enough only to witness domestic violence and not be subject to other forms of maltreatment, even in these cases evidence has been found that ” both [hypothalamic-pituitary-adrenal] axis and sympathetic nervous system functioning were found to differ between children exposed to domestic violence and comparison children (Saltzman, Holden and Holahan, 2005), suggesting that exposure to this form of trauma has lasting biological as well as emotional consequences. While the exact effect of the kinds of changes which Saltzman and his colleagues found on later functioning is unknown, findings like this give cause for serious concern that exposure to domestic violence in early life may have consequences which include functional, particularly social, impairments which are difficult (although not impossible) to overcome. They may, too, have long term health effects: the prolonged effects of biological stress responses observed in PTSD sufferers have been linked to a variety of serious chronic illnesses (Boscarino, 2008), suggesting that children who witness violence may be at risk of ill health long after their exposure itself has ended. Future research seems likely to confirm that the hypersensitivity to verbal conflict displayed by the one-year-old (ie. pre-verbal) infants studied by DeJonghe and her colleagues (2005) has a neural basis; this high level of sensitivity may itself predispose individuals who were exposed to domestic violence as young children to displaying high levels of arousal in conflict situations, contributing to the emergence of aggression and conduct problems in later life.
Rationale for undertaking this research
It is clear from the preliminary review of the literature which has been presented above that exposure to domestic violence is a serious child welfare issue: it affects a large number of children, is frequently combined with other forms of maltreatment, and has been shown to have long-term negative effects on both psychosocial functioning and, more tentatively, on physical health. As such there is an obvious rationale for assessing the current state of research into this topic: there is now a large volume of work on this issue, although it has only been explored empirically since the 1980s, and new techniques such as the use of biomarkers and neuroimaging continue to add dramatically to our understanding of the risks and mechanisms of harm associated with witnessing domestic violence. Producing a broad systematic review of the aspects of this topic of greatest relevance to social policy and professional social work practice will help to inform responses to this grave threat to the wellbeing of thousands of children in the UK, and contribute to the formulation of effective responses to the challenges which family violence poses today.
Outline research strategy
As has been mentioned above, the research strategy which will be adopted here is that of a critical review of the literature, based on a structured search of major journal databases. This strategy is the most appropriate one due in part to the challenges of conducting experimental or observational research in families where domestic violence occurs; given the researcher’s lack of training in managing the care and welfare of vulnerable children and adults, a methodology of this kind would not be appropriate. As such, an approach which does not pose these ethical and practical problems has been adopted.
A structured literature search methodology will be used to search the PUBMED, OVID and Web of Science databases; the terms used will be selected in order to identify literature which deals primarily with exposure to violence without the copresence of other forms of maltreatment. Due to the broad scope of this review, a meta-analytic approach would not be appropriate: where appropriate, meta-analyses of studies on this topic will be included, along with discussion of the individual studies included in them. Particular attention will be given to critical analysis of the effectiveness of the studies’ attempts to exclude the effect of confounding variables, including exposure to other forms of maltreatment and verbal aggression in the home, social factors and other issues.