The impact of children witnessing domestic violence has become an important focus for investigations over the past few decades (Kitzmann et al. 2003). While the effects on women of living with domestic violence are now well documented (Briere and Jordan 2004; Campbell 2002), the effects on children have been less extensively researched. There is, nevertheless, sufficient evidence to indicate that witnessing domestic violence, particularly if the violence is chronic, poses a threat to the physical, emotional, and psychological well-being of children living in such environments. Recent reviews (Bedi and Goddard 2007; Edleson 1999; Kitzmann et al. 2003; Margolin and Gordis 2000) have confirmed that children, who are exposed to domestic violence, show significantly poorer psychosocial outcomes, and are more at risk for developing a complex range of difficulties than children who are not so exposed.

Reviews have generally reported statistically significant differences between children exposed to domestic violence and their normal population peers in areas of behavioral, social, emotional, and cognitive functioning (e.g., Bedi and Goddard 2007; Margolin and Gordis 2000). Smith et al. (1997), for example, reported that 42% of children three to six years of age who have been exposed to domestic violence exhibited behavioral problems that would warrant clinical intervention. Margolin and Gordis (2000), and more recently Bedi and Goddard (2007), have remarked upon the prevalence of conduct problems, aggression, and other externalizing behaviors in populations of these children. In addition, internalizing problems, such as the development of phobias and fears, somatic complaints, symptoms of depression, and reduced self-worth, are also being reported. Finally, there is emerging evidence that enduring stress in the early years may have an adverse effect upon brain development and organization (e.g., Carrion et al. 2007; Perry et al. 1995), leading to the maintenance of un-adaptive responses to further life stresses.

The recognition that children exposed to a range of stressful and abusive events can develop symptoms of post-traumatic stress (Pfefferbaum 1997), even though the symptoms may not always be expressed according to the established adult criteria (Yule et al. 1999), has led some researchers to redefine many of the behavioral and emotional reactions seen in children affected by domestic violence as trauma responses. Graham-Berman and Levendosky (1998), Lehmann (1997), and Mertin and Mohr (2002), for example, have all confirmed recognizable post-trauma symptoms in children as a consequence of witnessing domestic violence, although the incidence of diagnosable Posttraumatic Stress Disorder (PTSD) (American Psychiatric Association 1994) is variable, ranging from 13% (Graham-Berman and Levendosky 1998) to 56% (Lehmann 1997).

While the above studies provide support for the use of a post-trauma model for understanding the effects on older children of living with domestic violence, there has been little comparable research examining similar effects on preschool-aged children, despite suggestions that relative to the general population, families with documented incidents of domestic violence are likely to have a higher number of children under the age of five living in the home (Fantuzzo et al. 1997). It has been suggested that young children may be at particular risk due to their limited understanding of conflict and their less developed coping strategies (Kitzmann et al. 2003), and it is possible that children as young as three years of age may develop trauma symptoms as a result of witnessing domestic violence (Graham-Berman and Levendosky 1998).

There are few instruments designed to assess symptoms specifically related to childhood trauma, however, and there are surprisingly few trauma-relevant measures available for young children (Briere et al. 2001; Saylor et al. 1999). Scheeringa et al. (1995) first suggested an alternative set of criteria for assessing trauma responses in infants and young children less than four years of age. Although using the DSM-IV (American Psychiatric Association 1994) adult criteria for PTSD as a template, Scheeringa et al. (1995) modified the criteria to be more sensitive to young children’s developmental status and, hence, to more appropriately identify symptoms of re-experiencing, avoidance/numbing, and hyperarousal in young children. Because of less developed language skills in young children, however, rating of the criteria is dependent upon behavioral observation (Scheeringa et al. 1995; Scheeringa et al. 2003).

A search of the literature revealed that, to date, there are only two published studies relevant to this area of inquiry (Bogat et al. 2006; Levendosky et al. 2002). Based on maternal reports, Levendosky et al. (2002) assessed symptoms of posttraumatic stress in 62 children between the ages of three and five years, who had witnessed domestic violence or who were living in households with domestic violence. Two assessment measures were used: one adapted from the Child Behavior Checklist (CBCL) developed by Wolfe et al. (1989), and the other designed for the study (PTSD Symptoms in Preschool Aged Children: PTSD-PAC) based on DSM-IV criteria. No significant sex differences were found for either scale. There were also no significant correlations between the symptoms measured on the scale derived from the CBCL and the PTSD-PAC, and few children met full criteria for PTSD with either measure (one child using the PTSD-PAC, and 24% using the PTSD scale from the CBCL). Children in the study appeared to be most vulnerable to symptoms of re-experiencing and hyperarousal; and they generally had elevated Externalizing scores on the CBCL compared with national norms.

In a similar vein, Bogat et al. (2006) examined trauma symptoms in one-year-olds as a consequence of exposure to domestic violence. Using a sample of 48 mother-child dyads, the infants were assessed using the Toddler Temperament Scale (Fullard et al. 1984) and the Infant Traumatic Stress Questionnaire (ITSQ), an 8-item questionnaire designed for the study measuring “numbing”, “increased arousal” and “fears or aggression”. In addition, in order to examine the relationship between mothers’ and infants’ responses to domestic violence, the mothers’ emotional functioning was assessed using the Beck Depression Inventory (Beck et al. 1961), and the PTSD Scale for Battered Women (Saunders 1994).

Mothers in this study reported that 18 of the 48 infants (37.5%) displayed at least one trauma symptom in a two-week period following an incident of domestic violence. Of these 18 infants, Bogat et al. (2006) found that nine had symptoms of increased arousal, 10 had symptoms of numbing or interfering with development, and 10 had new symptoms, fears, or increased aggression. Results of this study also found that there was a significant relationship between maternal and infant trauma symptoms; that is, infants exhibited trauma when their mothers did, but only when their mothers experienced severe violence. These authors argued that such results were consistent with the theory of relational PTSD (Scheeringa and Zeanah 2001), which posits that the co-occurrence of trauma symptoms in a parent and young child result when the adult’s responses are not well-regulated and, therefore tended to enhance the child’s own responses to stress. However, maternal depressive symptoms and difficult infant temperament did not predict infant trauma symptoms in this study.

Taken together, the above studies suggest that young children show recognizable symptoms of trauma as a consequence of exposure to domestic violence. Based on maternal reports, those symptoms most commonly observed included increased arousal and increased aggression. Although Levendosky et al. (2002) concluded from their results that avoidant symptoms were not as reflective of trauma in their age group (3–5 year olds) as were symptoms of arousal, Bogat et al. (2006) found that 10 infants were reported by their mothers as showing such symptoms. While differences in reported symptoms may have been a function of the different measures used, both studies concluded that the current criteria for PTSD as set out in DSM-IV are unsuitable for use with young children.

The above studies have provided an insight into young children’s responses to environments of stress and threat. Given the importance of brain development and organization in the early years to later emotional and behavioral functioning (Perry et al. 1995), more needs to be discovered about the effects of stress on the developing nervous system, how these effects are manifested through observable symptoms, and of factors in the child’s environment that may either enhance or modify these effects. The present study, therefore, was designed to continue the investigation of young children’s responses to environmental stress. Using domestic violence as the stressor criterion, this study set out to assess the prevalence of trauma symptoms in children aged between 1 ½ to 5 years of age, using Scheeringa et al. (1995) suggested criteria as the template for assessment. In addition, the relationship between maternal stress, as manifested by depression, anxiety and somatization, and child trauma symptoms was also investigated.

Method

Participants

Participants were 46 mothers who had left their homes to escape domestic violence, and who had subsequently received support from a domestic violence crisis service. The mean age of the women at time of interview was 31.7 years (SD = 6.5 years). The mean length of relationship with their spouse/partner before separation was 6.8 years (SD = 4.5 years), while the mean time of separation was 10 months (SD = 12.2 months). A total of 15 (32.6%) women reported that the current separation was the first time they had separated from their partner/spouse. At the time of interview, all women were living away from their spouse/partner either in short term emergency shelter accommodation or in rented accommodation.

Fourteen women in the study had more than one child under the age of 5. Where this was the case, additional data were collected for the second youngest child in the family. Data were, therefore, collected for 60 children (34 female, 26 male) of these 46 mothers. Children were aged between 1.4 years and 5.4 years, with an average age 3.4 years (SD = 1.2 years).

Measures

The following assessment battery was conducted in interview format. The participants were asked each question verbally with a copy of the self report instrument in front of them. Answers were then recorded on the profiles by the researcher.

Adapted Conflicts Tactics Scale (ACTS)

The original 9-item violence subscale of the Conflict Tactics Scale (Straus 1979) was adapted by Mertin (1992) and extended to an 18-item scale to more accurately assess the level of physical, verbal, sexual, and financial abuse experienced by women in domestic violence relationships. The Adapted Conflict Tactics Scale (ACTS) provides an overall indication of the level of abuse experienced by women. The frequency of each item is rated on a Likert-type scale ranging from 0 (never) to 6 (more than 20 times per year), providing a maximum total score of 108. Mertin and Mohr (2000) reported a Cronbach’s alpha of .87, while the alpha for the present study was .89, indicating good internal consistency.

Child Behavior Checklist for Ages 1 ½ -5 (CBCL)

The CBCL is a 100-item scale used to measure children’s behavioral, emotional and social functioning based on parent report (Achenbach and Rescorla 2000). The CBCL was administered to mothers in interview format. Raw scores on the CBCL were transformed to T-scores, with the Internalizing, Externalizing, and Total Problem scales being used as measures of the children’s functioning. The manual recommends T-scores between 60 and 63 as the borderline clinical range. The Checklist is well accepted, with the manual reporting good reliability and validity (Achenbach and Rescorla 2000). In this study Cronbach’s alpha was .95 for the Total Problem scale, .90 for the Internalizing scale and .91 for the externalizing scale.

Parenting Stress Index Short Form (PSI/SF)

The PSI/SF (Abidin 1991) is a shortened version of the full length Parenting Stress Index. The Short Form includes 36 self-report items designed to measure stress in the parent-child relationship. It was designed in response to a need for a valid measure of stress related to parenting that could be administered in a short period of time. This version derives three scales: Parental Distress, Parent-Child Dysfunctional Interaction, and Difficult Child, that can be summed to provide an overall measure of Total Stress. The manual advises that high scores are considered to be scores at or above the 85th percentile. The PSI/SF had a Cronbach’s alpha of .91 in the present sample.

Symptom Checklist-90-R (SCL-90-R)

The SCL-90-R (Derogatis 1994) is a 90-item self-report symptom inventory designed to reflect a set of psychological symptom patterns. Each item is rated on a five-point scale (0–4) ranging from “not at all” to “extremely”. The complete SCL-90-R includes nine primary symptom dimensions. However, only three of the primary symptom dimensions, Anxiety, Depression, and Somatization, were administered as a part of this test battery. Reliability analysis of each symptom dimension revealed a Cronbach’s alpha of .91 for the anxiety scale, .90 for the depression scale and .87 for the somatization scale.

Procedure

Participants were approached to participate in the study initially by a domestic violence shelter worker, who provided information sheets and described the nature and objectives of the study. Participants were then contacted by the first author by telephone and interviewed either in their home/secure shelter accommodation, at the domestic violence service, or at a community health center. Brief demographic data were collected and the assessment battery was then administered to participants in interview format. Each interview took approximately 50–60 min to complete. Where there were families with more than one child under the age of five, the CBCL and the PSI was repeated for the second child. Options for further support or debriefing were offered upon termination of the interview as needed.

Results

Experiences of Domestic Violence

The mean total score on the ACTS was 63 (SD = 23.3) indicating that women who participated had experienced high levels of violence in their relationships. The reported level of violence in the present study is consistent with recent Australian data (Mertin and Mohr 2000) that reported mean ACTS scores of 55 and above. Types of violence endorsed by women as occurring more that 20 times per year included: being called names and/or being put down verbally (67%), being blamed for all of her partner’s problems (76%), unwanted demands for sex (65%) and a high incidence of financial abuse where 63% of women reported that their partner kept all of the money to himself, and 48% of women reported being without enough money to pay for food or the bills. Women also reported a high incidence of physical violence. Sixty three percent of women reported that their partner threw, smashed, hit or kicked something, and 44% reported that they had been pushed, grabbed or shoved. A total of 9 (20%) women reported that their partner had used a knife or a gun either to threaten them or to harm himself in front of her. In addition to the ACTS data, 32 (70%) women reported that they believed they would be killed by their partner at some stage during the relationship.

Further data obtained from the women indicated that 40 (67%) of the children were believed to have witnessed domestic violence in the home on a frequent basis, 18 (30%) children had witnessed domestic violence occasionally, and only 2 (3%) children were reported as not witnessing any violence. Of the 60 children, 48 (80%) had experienced one or more relocations to a different house in their lifetime.

Maternal Functioning and Parenting Stress

Table 1 shows the results of measures of the participants’ functioning, in emotional domains as measured by the Depression, Anxiety, and Somatization scales of the SCL-90-R, and with respect to parenting stress as measured by the PSI/SF. Clinically significant levels of depression, anxiety, and somatization were all recorded on the SCL-R-90 for the majority of participants, with high levels of stress in relation to parenting similarly reported.

Table 1 Maternal functioning and parenting stress
Full size table

Children’s Adjustment

Analysis of the CBCL results revealed that children’s functioning, as indicated by mean scores on the Internalizing, Externalizing, and Total Problem scales, were within clinical ranges in all scales, reaching T-scores of 64 and above. These results are summarized in Table 2.

Table 2 Children’s adjustment as measured by the CBCL (n = 60)
Full size table

Independent samples t-tests were conducted to test for gender differences in CBCL scores. No significant gender differences were found for Internalizing, Externalizing or Total Problem behavior scores. No significant age differences were found for the Externalizing or Total Problem behavior scores. However, children over the age of 3.5 years scored significantly higher on the Internalizing scale of the CBCL (t(58) = -2.58, p < 0.05). Comparisons between witnesses and non-witnesses of domestic violence were not carried out in these analyses because there were only two children in the sample who were reported as not witnessing any violence.

Table 3 shows correlations between child variables, as measured by the CBCL, and mothers’ variables (depression, anxiety, somatization, parenting stress, and experience of violence). Results indicated statistically significant but small correlations between depression, anxiety, and somatization and levels of violence as measured by the ACTS. There were however, larger correlations between levels of parenting stress and total CBCL scores, and between parenting stress and mothers’ reported emotional functioning on anxiety, depression, and somatization.

Table 3 Correlations between mother’s and children’s variables (n = 60)
Full size table

Regression Analysis

Hierarchical Multiple Regression analysis was conducted to determine the best predictors of scores on the CBCL. It must be noted that a total of 14 women in the study had more than one child under the age of five. Regression analyses were conducted initially with the whole sample of children (n = 60), and repeated with the second sample of children (n = 47) where only the eldest child in each family was represented. There were no major differences in the pattern of results for these two samples. Therefore, only regression results based on the whole (n = 60) sample of children are reported. Table 3 shows that violence was not highly correlated to child outcomes, therefore the ACTS was not included in the regression analysis.

Table 4 provides the results of the hierarchical multiple regression analysis on CBCL scores, using measures of mothers’ emotional functioning as predictors. Anxiety, depression and somatization together did not significantly contribute to Total CBCL scores (F[3,56] = 2.14, p > 0.05). However, when Total Parenting Stress was added as a predictor, it explained a significant amount of variance (F[1,55] = 29.95, p < 0.001). The set of predictors together explained 42% of the variance in CBCL scores.

Table 4 Summary of hierarchical regression analysis for variables predicting total CBCL scores (N = 60)
Full size table

As total PSI/SF score was the only significant predictor of CBCL scores, a second hierarchical regression analysis was performed using the subscales of the PSI/SF as predictors of total CBCL scores. As seen in Table 5, each subscale of the PSI/SF explained a significant amount of variance in CBCL scores; the Parental Distress subscale explained 6.2% of the variance, (F[1,58] = 3.86, p < 0.05); the Difficult Child subscale explained 37.2% of the variance, (F[1,57] = 37.41, p < 0.001), and the Parent-Child Dysfunctional Interaction subscale explained 8.8 % of the variance, (F[1,56] = 10.28, p < 0.01). Together the three subscales explained 52.2% of the variance in CBCL scores.

Table 5 Summary of hierarchical regression analysis for variables predicting total CBCL scores (N = 60)
Full size table

Posttraumatic Stress Symptoms in Young Children

In addition to the CBCL data presenting a more global measure of child functioning, the presence of symptoms indicative of post-trauma responses were assessed using items on the CBCL that corresponded to criteria for a proposed model of posttraumatic stress disorder in young children proposed by Scheeringa et al. (1995).

Scheeringa et al. (1995) criteria were matched to corresponding items on the CBCL that were then used as a template for assessing the presence of post-trauma symptoms in the children (Scheeringa et al. 1995, 2003). Following Scheeringa and colleagues, the relevant CBCL items were clustered together on a rational basis to create four categories: Re-experiencing, Numbing of Responsiveness, Increased Arousal, and New Fears and Aggression. The overall alpha for these items was .83 in the present sample. The incidence of post-trauma symptoms in the children, as reported by their mothers, is presented in Table 6. The symptoms were only included if endorsed by the mother on the CBCL as being “very true or often true”, thus ensuring a conservative measure of post-trauma symptoms in this sample.

Table 6 Frequency and percentage of post trauma symptoms in children rated as “very true or often true” on the CBCL
Full size table

As can be seen in Table 6, the most frequently endorsed symptoms (those reported as occurring in 50% or more of the children), were in the Increased Arousal (Can’t sit still, restless or hyperactive), and New Fears and Aggression (Doesn’t want to sleep alone, Clings to adults or is too dependent, and Fears certain animals, situations or places) categories. In addition, a further 48% of children were reported as becoming too upset when separated from parents, and a further 33% of children as having poor concentration. Results also show clear elevations on all items relating to sleep, with 30% of children reported having trouble getting to sleep and crying out in their sleep, 25% of children reported waking in the night, and 28% as experiencing nightmares.

Discussion

Posttraumatic Stress Disorder was first described in the DSM-III (American Psychiatric Association 1980), although the disorder was not initially perceived as relating to children until the publication of DSM-III-R in 1987 (American Psychiatric Association 1987; Davis and Siegel 2000). Subsequent research has confirmed that children can develop PTSD following exposure to traumatic stressors, although the range and type of symptoms displayed may be influenced by developmental factors. Dyregov and Yule (2006), for example, have found that young children show less emotional numbing and have more problems reporting avoidance reactions than do adults. These authors also observed that there was less agreement as to the range and severity of reactions seen in preschool children and that, for this group, their reactions may be determined more by parental reactions to the event.

While a range of post-trauma symptoms have been found in primary school-aged children who have been exposed to domestic violence, comparable research on the effects in preschool-aged children is rare. The present study, therefore, set out to explore the frequency of post-trauma symptoms in young children using a more developmentally appropriate method of assessment. Based on maternal reports of their children’s behavior, results revealed a range of symptoms the most prominent of which were changes in sleeping patterns, becoming more fearful and clingy, increased motor activity, and the development of separation anxiety. In addition, a more global rating of the children’s behavior using the CBCL indicated high levels of general distress, with mean T scores in all three summary scales (Internalizing, Externalizing, and Total Problem) elevated into the clinical range.

These results are in broad agreement with those of Levendosky et al. (2002) who similarly found a range of post-trauma symptoms in their group of preschool-aged children, although these authors reported that symptoms of re-experiencing the trauma and hyperarousal were the most commonly endorsed by the mothers. These authors also found significantly elevated Externalizing T scores on the CBCL with 29% of their sample falling into the clinical range, although unlike the present study, they did not find elevated Internalizing T scores.

Both Levendosky et al. (2002) and the present study relied for their results on maternal reports of their child’s behavior. While the Levendosky et al. (2002) study did not specifically assess for emotional responses in the mothers consequent to the domestic violence, the present study found that, as a group, the mothers were experiencing significant levels of depression, anxiety and somatization, as well as high levels of parenting stress. Dyregov and Yule (2006) remarked upon a probable relationship between a parent’s and young child’s responses to a traumatic event. Also, Bogat et al. (2006) found a significant relationship between maternal and child trauma symptoms in their study of infants exposed to intimate partner violence.

More specifically, Levendosky et al. (2006) examined whether maternal functioning mediated the relationship between domestic violence and infant externalizing behavior. Based on the premise that mental health problems associated with partner abuse are well documented, and that parents suffering from psychological distress are less available and less involved with their children, these authors found that direct and indirect risk factors for infant externalizing behavior included domestic violence, maternal mental health, and maternal parenting. Their results also confirmed the dose-response effect of domestic violence, that is, more domestic violence is related to worse mental health, and this in turn is negatively related to parenting.

Thus, there is emerging evidence that the negative effects of domestic violence on mothers’ mental health mediate emotional and behavioral responses in young children. This evidence is consistent with the relational context of post-traumatic stress disorder proposed by Scheeringa and Zeanah (2001). These authors hypothesized the co-occurrence of post-traumatic symptomatology in an adult caregiver and a young child when the symptomatology of one partner, usually the adult, exacerbates the symptomatology of the other. As examples, these authors proposed that traumatized adults may be significantly less emotionally available to their young children because of their own impairments, or that following exposure of a young child to a trauma, parents may become constrictive and overprotective, or preoccupied by fear that their child may be traumatized again. Scheeringa and Zeanah (2001) concluded, therefore, as with most other forms of psychopathology in young children, post-trauma symptoms are most usefully considered within the context of the parent-child relationship.

The present study found significant relationships between levels of maternal distress (i.e., levels of depression, anxiety, and somatization) as measured by the SCL-90-R, and levels of parenting stress, as measured by the PSI/SF, and by levels of parenting stress and all summary scales (Internalizing, Externalizing, and Total Problem) of the CBCL. Of note is the indication in Table 3 that, while Parental Distress was most highly correlated with levels of depression in the mothers, Difficult Child and Parent-Child Interaction were more highly correlated with the CBCL.

Although these results provide additional general evidence that maternal stress adversely impacts upon the parent-child relationship, the results also raise the suggestion that mothers who are emotionally affected by their own experiences (i.e., domestic violence), and who have diminished reserves of energy and a lowered tolerance for frustration, may perceive their child(ren) as being more difficult for them to manage. This, in turn, may then shape their own responses to their children. Consistent with this idea, Levendosky et al. (2006) suggested that currently experienced domestic violence negatively affected the mother’s ability to respond warmly and sensitively to her infant; and there was the increased risk of hostility and disengagement in the relationship.

The present study found a significant, but modest, relationship between levels of reported violence and levels of emotional distress in the mothers. However, the data showed no relationship between levels of reported violence and parenting stress. This suggests that the experience of violence in a relationship does not impinge directly on the parenting role, but acts by way of elevated levels of emotional distress in the mother as a consequence of the abuse, which in turn adversely affects all aspects of day-to-day functioning. Scheeringa and Zeanah (2001) pointed out that the most powerful change agent for young children’s development and symptomatology is their relationship with the primary caregiver. Further promoting change in a parent’s symptomatology is thus likely to make them better able to respond sensitively to the needs of their children. These authors, therefore, advocate attending first to the needs of the caregiver in order to moderate the traumatic responses in the children.

There are some noteworthy limitations to the current study. First, the presence or absence of post-trauma symptoms in young children was determined by mothers’ reports, and not by direct observations of the children, a limitation which has previously been highlighted in the literature (Sternberg et al. 1993). Second, the template proposed by Scheeringa et al. (1995) used in this study was based on items extracted from the CBCL (Achenbach and Rescorla 2000), a measure that was not specifically designed to assess the presence trauma symptoms in young children. Despite these limitations, and in light of the absence of an agreed assessment tool, the data provide evidence that preschool-aged children can show symptoms of trauma as a result of living in fearful and stressful environments, and that the primary caregiver can be considered the best source of information about young children.

There are also important implications of the study. Our results highlight the need for further research on trauma symptomatology in young children. Despite early attempts to extrapolate the presence of symptoms based on existing measures (Scheeringa et al. 1995), the field has been slow to respond to the need for more developmentally appropriate diagnostic criteria and subsequent standardized tools measuring the impact of domestic violence on young children. There continues to be a gap in accurately defining and assessing infant and toddler trauma symptoms, especially in the age range of 3 years and under. Bogat et al. (2006) surmised that the age of the infants in their study may have had an influence on their results. These authors suggested that the low number of trauma symptoms reported by the mothers may have reflected the possibility that many of the episodes of domestic violence occurred too early in the infant’s life for him or her to respond. Consistent with the idea that young children tend to show more global responses to stressful experiences, mothers in the current study reported significantly less internalizing symptoms in children less than 3.5 years of age when compared to the older children in the sample.

As there is a general acceptance that children do experience adverse effects as a result of witnessing violence (Kitzmann et al. 2003; Margolin and Gordis 2000), it is important that our research efforts begin to focus on younger children, who are undergoing a crucial period of neurological development (Perry et al. 1995). More research is needed to develop more appropriate and standardized ways of assessing trauma symptoms so that clinicians working with preschool-aged children can accurately identify the effects, develop effective interventions, and understand the nature of the child’s difficulties. This is crucial for our field to advance in our ways of supporting women and children.

This study investigated children’s responses to a stressful environment and showed that young children can develop symptoms consistent with post-trauma responses. Based on the results of the present study, and the small empirical evidence available to date, it can be concluded that post-trauma symptoms may be influenced by the children’s own experiences. In addition, more research needs to focus on to what extent the primary caregiver’s own level of distress and possibly diminished coping abilities as a result of violence and abuse, may influence the way they perceive their children. For example, more research is needed to determine to what extent, independent of maternal reports, children’s responses are merely “difficult” or actually clinically significant. Helping to stabilize or reduce the caregiver’s own stress in relation to violence may in turn reduce parenting related stress, which ultimately could assist the children who are living in stressful environments.