Post-Trauma Coping

The Australasian Journal of Disaster
and Trauma Studies
ISSN:  1174-4707
Volume : 1999-1

Prediction of Children's Coping Following a Natural Disaster - the Mount Ruapehu Eruptions:
A prospective study

Caryl A Huzziff,
Kevin R Ronan

School of Psychology,
Massey University
New Zealand

Abstract

The present study used a conceptual model to examined predictors of coping ability in 187 elementary school-aged children after the Mount Ruapehu eruption. The model included the examination of several factors: exposure, demographics, home factors, negative cognitive and emotional style (depression, affect, state & trait anxiety), PTSD symptoms and coping. Five primary factors assessed one-month post-disaster (exposure, child demographics, home factors, negative cognitive and emotional style, and PTSD symptomatology) accounted for 44% of the variance in children's initial level of coping ability. Three months post-disaster, 47% of the variance in children's coping was accounted for by these five primary factors plus one additional factor, initial level of coping ability. Additionally, home-based factors accounted for 11% of the total variance initially and 10% prospectively. The findings are discussed in the context of how to organize intervention in the wake of natural disasters.

Prediction of Children's Coping Following a Natural Disaster - the Mount Ruapehu Eruptions:
A prospective study

Eighteenth of September, 1995 Mount Ruapehu (New Zealand) began erupting. Eruptions continued to mid 1996. Early eruptions resulted in lahars down several rivers. Later eruptions resulted in dry sustained ash fallout within 250 km of the volcano. Agricultural land and some 20 communities were effected along with air and road transportation and national power supplies (Johnston, Houghton, Neall, Ronan & Paton, 1998). Ash covered grass, roads and polluted local waterways and supplies. A general alert was issued by the Ministry of Civil Defense to people living within a 100km radius of the volcano. Many people evacuated while others prepared to do so. Late September, one town reported stores to have no tinned food left and garages sold out of petrol (Keys & Williams, 1996, cited in Johnston et al., 1998).

The impact of natural disasters goes beyond physical effects, often resulting in a range of stresses, both at social and psychological levels. Stress may be particularly detrimental to ones health. Stress has often been associated with negative outcomes including ill health, loss of motivation, and depression (Coddington, 1972; Vogel & Vernberg, 1993). It has been suggested that up to eighty percent, or more, of all visits to health care professionals result from excessive stress, and it's physical and psychological manifestations (Cummings & Vanden Bos, 1981, cited in Everly & Lating, 1995). While stress and coping with stress in adults has received much research, little is known about children in this area. Knowledge about how stress impacts on children and what factors are associated with effective coping with increased stress is essential if long-term negative outcomes are to be prevented.

Studies investigating the impact of natural disasters on children and adolescents have found negative emotional and behavioral consequences, to varying degrees. While diagnosable psychopathology may not be present, many children do report significant levels of emotional distress and frequent occurrence of sub-clinical levels of post traumatic stress disorder (PTSD) (Belter & Shannon, 1993; McFarlane, 1987). Recent research by La Greca, Silverman, Vernberg and Prinstein, (1996) and Vernberg, La Greca, Silverman and Prinstein, (1996) have used integrative conceptual models to investigate children's responding to traumatic stress (i.e., Hurricane Andrew). Their findings indicated that exposure to a traumatic event, individual child characteristics, access to social support, life events and the child's coping style may all impact on children's stress responding. Using a similar conceptual model the current study aims to move beyond what is known to impact on the development of PTSD symptomatology in children, and investigate how similar factors impact on children's coping ability following a natural disaster.

Children vulnerable to the effects of disaster may benefit from efforts aimed at mobilizing adaptive coping strategies (La Greca et al., 1996; Terr, 1989). Coping refers to overt and covert cognitions and behaviors adopted to manage a problem or emotionally arousing situation. These efforts allow for the management and alteration of the person-environment relationship to reduce negative emotions or solve stress-related problems (Titchener, 1988). If coping is successful it is thought to act as a buffer against stress and increased vulnerability. Consequently, adaptive coping is viewed as perhaps the most important non-environmental factor for dealing with traumatic events and related stress (Compas, Forsythe & Wagner, 1988; Joseph, Williams & Yule, 1995; Vernberg et al., 1996).

Children are believed to be exceptionally vulnerable to the effects of trauma due to their underdeveloped cognitive coping mechanisms (Gibbs, 1989) and their reliance on adults (Atkins, 1991). In terms of perceived coping ability, adults have the advantage of incorporating skills attained through life experience and have more control over their external environment (Ryan-Wagner, 1992). In support of this developmental progression, Atkins (1991) has shown that increased age is positively related to more active modes of coping in the face of traumatic stressors. In addition, children have less control over environmental contingencies found helpful for reducing stress compared to adults (Atkins, 1991). As a consequence of these individual and ecological factors, children are often more limited than adults in the flexibility afforded them for coping with their environment. However, while coping style has been examined, the extent to which such factors impact upon a child's ability to cope after disaster strikes has not (Compas & Epping, 1993, cited in Vernberg et al., 1996).

Previous research has shown that children may have a continued preoccupation with a traumatic event(s) that may last for periods up to one year or more (McFarlane, 1987; Vernberg et al., 1996). Theory has additionally suggested that in the absence of effective coping, the meaning and impact of traumatic events may continue to play a role in the personality and psychological development of children (Hodgkinson & Stewart, 1991).

A review of the literature regarding children's post-disaster reactions (Vogel & Vernberg; 1993) concluded that the selection of the variables for investigation within this area were somewhat unsystematic and random. Consequently, knowledge of the individual factors related to children's reactions to trauma is unclear. To better understand children's coping ability the current research aimed to use an integrative conceptual model guided by previous theoretically based research (e.g., Vernberg et al., 1996; La Greca et al., 1996). The current model was designed to assess contributing factors to children's ability to cope following a disaster _ the New Zealand, 1995 Mount Ruapehu eruptions. The current study's model assumes that the factors impacting post-disaster coping ability in children are multiple and complex, and include characteristics of the stressor, the child, and the post- disaster environment.

The predictor factors for children's initial coping ability are: exposure to traumatic events, preexisting child characteristics, home factors, and multiple cognitive and emotional factors including PTSD symptoms. To predict coping ability at the three-month post-disaster mark, children's initial level of coping ability will be added as an additional factor in the conceptual model. Each of these factors have previously been linked with children's responses to natural disaster (e.g., Belter & Shannon, 1993; La Greca et al., 1996; Shannon et al., 1994; Vernberg et al., 1996). The primary, global hypothesis of the current study is that each of the seven factors within the current conceptual model would account uniquely for variance in children's coping, when entered at the appropriate point in the model. Each factor is believed to make unique contributions to coping in children following a disaster.

In the following sections, a brief summary of the applicable background literature that directed the selection of the factors in the current model is presented. As mentioned successful coping is thought to act as a buffer against stress, reducing ones vulnerability to experience stress related symptoms (Compas et al., 1988; Williams & Yule, 1995; Vernberg et al., 1996) such as negative emotional and cognitive symptoms. The following sections refer to increased levels of negative cognitive and emotional symptoms such as PTSD symptoms as indicative of increased inability to cope.

Exposure to the Traumatic Events
Exposure to the traumatic event was selected as the first factor in the conceptual model. Most models of trauma exposure have considered exposure to be the most salient and critical factor for the development of post-traumatic stress symptoms. Exposure to a traumatic event triggers the initial need for coping. Level of exposure has been noted to affect the level of PTSD symptomatology reported by children. High levels of direct exposure to trauma have been found to result in lingering stress and PTSD symptoms. The closer the proximation to the disaster, the greater the potential of psychological impact (La Greca et al., 1996; Shannon Lonigan, Finch & Taylor, 1994; Vernberg et al., 1996; Yule, 1994). Besides location, events that are perceived as life-threatening during a disaster have often been linked to increased distress in children (Vogel & Vernberg, 1993) indicating reduced coping ability; however, self-talk related to the relationships between life-threat and coping ability is unclear and in need of investigation.

Preexisting Child Characteristics: Age, Gender, Ethnicity
Age. Literature reviews by Green, Korol, Grace, et al., (1991) and Lonigan, Shannon, Taylor, Finch and Sallee (1994) have shown that younger children are more likely to develop PTSD symptoms following a disaster. Pre-school aged children are reported to show lower levels of global psychological distress than older children, but a higher incidence of specific behavioral disturbance (e.g., instances of acting out). Younger children apparently are also more impacted by caretakers' reactions to the traumatic event. School-aged children typically begin to show the emergence of traditional PTSD symptoms: trauma-specific fears and anxieties, somatic concerns, sleep disturbance and school problems, very possibly as a result of increased ability to understand the traumatic event (Green et al., 1991; Yule, 1994). That is age appears to impact on children's ability to cope and deal with stress following trauma.

Gender. In general, girls tend to report more PTSD symptoms than boys (Green et al., 1991; Yule, 1994). However, recent studies (La Greca et al., 1996; Vernberg et al., 1996) have reported mixed findings concerning the relationship between gender and traumatic stress. Consequently, how gender predicts PTSD symptom development and the ability to cope is unclear and needs further investigation.

Ethnic Group. Although research is limited, ethnic differences in children's reporting of PTSD symptoms have been reported (La Greca et al., 1996). The current research investigated possible ethnic differences.

Preexisting child characteristics were entered as the second group of major factors due to the influence that one's existing characteristics have been found to have on coping (Atkins, 1991; Gibbs, 1989) and PTSD symptoms (Belter & Shannon, 1993; Green et al., 1991). It is also appropriate to place preexisting child characteristics second in the model as they could influence the use of coping strategies (the reverse is less likely, Vernberg et al., 1996).

Home Factors
The third block of factors included within the current conceptual model is home factors. These factors are believed to exert effects on PTSD symptoms after the initial shock of the disaster, impacting on coping ability (Schreurs & de Ridder, 1997; Vogel & Vernberg, 1993). Home-based social support systems are particularly salient for children. Parents serve as models for coping behavior and provide comfort and the feeling of physical safety (Comps & Epping, 1993; Pynoos & Nader, 1988; Vernberg & Vogel, 1993). Family atmosphere contributes to the prediction of increased PTSD symptoms. For example, Ronan (1997) found a relationship between perceived parental upset and increased levels of posttraumatic stress in asthmatic children following volcanic eruptions.

Negative Cognitive and Emotional Style (NCES)
Suls and David's (1996) review of recent literature suggested that specific coping behaviors have cognitive and emotional underpinnings. Research has shown that there are distinguishing characteristics between positive illusions, and denial and avoidance on coping (Taylor & Armour, 1996). For example, optimistic individuals seem to be more likely than their pessimistic counterparts to make active efforts at coping. In the light of such research, the current study sought to investigate what cognitive and emotional factors were associated with coping with traumatic events, including increased anxiety, depression, and negative self-talk. The current research considered cognitive and emotional factors fourth in the conceptual model as it was thought that these factors would be impacted by the prior four factors in the model.

Posttraumatic Stress Disorder and Coping
As a specific form of cognitive and emotional functioning, PTSD symptoms are last in the conceptual model for prediction of coping ability at Time 1. This factor is entered alone rather than as part of general negative cognitive and emotional functioning to assess whether PTSD symptoms, a known effect of trauma on children, has predictive value alone for children's ability to cope. For prediction of coping at Time 2 (i.e., at three months), the initial level of coping is the first factor in the model following the assumption that past behavior is the best predictor of future behavior.

Method

Participants
Participants were 187 children (89 boys, 95 girls, 3 not reported) at three New Zealand elementary schools (year/grade 2 _ year/grade 8) located within the surrounding communities of Mount Ruapehu (Raetihi, 68 children, 11km from the base of the volcano _ leeward, Ohakune, 56 children, at the base of the volcano -- leeward; Waiouru, 62 children, at the base of the volcano _ windward). The mean age was 10 years and 4 months (SD 1 year and 7 months)(Table 1). The ethnic composition of the sample was: 113 Pakeha/European, 20 Maori, 37 both (European and Maori), and 17 'other'.

At Time 1 (approximately 1-month post-disaster), 187 children were evaluated. At time 2 (approximately 3-months post-disaster), 127 children participated. The attrition rate was 32% (n=60) due to the unavailability of some classrooms (i.e., end of year camping trips).

Measures
Exposure. Exposure was measured in two ways, location and life threat. Location served as measure of direct exposure to the disaster with subjects naturally divided into exposure groups according to the distance from the volcanic activity (Johnston, Ronan & Houghton, 1996). For a measure of the extent to which children felt their lives were threatened, children were asked to endorse the frequency of occurrence of the self-statement “I thought my world was coming to an end” following the eruptions. The response scale ranged from (1) not at all, to (5) all the time).

Demographics. A basic demographic measure was used to obtain data concerning children's age, gender, grade, ethnicity, and school location.

Home Factors. A measure of children's perception of how much their parent's are upset by the volcano eruptions and how upsetting they found parental talk regarding the volcano. Two items were developed to create the measure: 'how much have your parents been upset by the volcano?' And, 'when you hear talk about the volcano at home, do you feel scared, afraid, or upset? The response scale consisted of none-, little-, some-, much-, or most of the time. The items showed moderate internal consistency (alpha = .65). Children's home factor scores reflected parental upset. Cronbach's alpha reliability between home factors and parents scores on the Post-traumatic Stress Disorder Reaction Index (RI) was .78.

The Children's Depression Inventory (CDI)(Kovacs, 1981), is a 27-item self-report questionnaire designed to assess areas of cognitive, affective, and behavioral signs of depression in children and adolescents aged 8 to 17 years. The measure uses a three-point scale with children endorsing one of three sentences that best describes their feelings and ideas during the past two weeks. Reliability for this measure is relatively high. Internal reliability is reported at the .87 level with a sample of Toronto public school children (n=860). Nine-week test-retest data on a sub-sample of school children showed a reliability level of .84 (Kramer & Conoley, 1992). The internal reliability with the current sample was .90. The scale also correlates in the expected directions with measures of related constructs such as self-esteem, hopelessness, and negative cognitive attributions (Kramer & Conoley, 1992: Kendall, 1994).

The Negative Affect Self-Statement Questionnaire (NASSQ) is a self-report questionnaire designed to assess self-statements related to negative affect in children aged 7 to 15 years (Ronan, Kendall, & Rowe, 1994). The questionnaire is divided into two separate age defined questionnaires. The NASSQ for children 7-10 years contains 14 items; the NASSQ for children 10-15 years contains 39 items. Children endorse self-statements on a scale ranging from (1) 'not at all', to (5) 'all the time'. Incorporated within the measures are items specific to anxiety and depression, and questions common to both. Items are designed to assess thoughts over the past week. For example, 'I am very nervous', 'I felt like crying' and 'I want to stay in my room forever'. This measure has demonstrated both reliability and validity. The Spearman Brown alpha reliabilities for the NASSQ: 7-10 and NASSQ: 11-15 were .87 and .94 respectively. Alpha reliabilities were shown to be .89 and .96 respectively (Ronan et al., 1994). Alpha reliability using the current sample was found to be .88. Concurrent validity has been established (Ronan et al., 1994). The measure has also shown specific sensitivity to treatment effects (Kendall, 1994; Ronan et al., 1994; Treadwell & Kendall, 1996).

The State-Trait Anxiety Inventory for Children (STAIC)(Spielberger, 1973), is a self-report measure which includes two separate 20-item scales. These scales measure levels of perceived situationally based anxiety (A-State) and the enduring tendencies to experience anxiety (A-Trait). Alpha reliabilities are acceptable (e.g., .82 and .78 for males, and .87 and .81 for females for A-state, and A-trait respectively). Test-retest coefficients range between .65 and .71 and .31 to .47 for A-trait and A-state respectively. The relatively low A-state test-retest coefficients are to be expected due to the transitory nature of state anxiety. Concurrent and construct validity have also been demonstrated (Buros, 1978; Ronan, 1996).

The Post-traumatic Stress Disorder Reaction Index for Children (RI) is a 20-item self-report measure of PTSD symptoms. One particular advantage of this measure is that items were designed to be made situation specific. In respect to the current research the questionnaire was made 'volcano' specific. For example, 'is the volcano something that would upset, or bother, most children your age?' 'Do thoughts about the volcano come back to you even when you don't want them to?' Children endorse self-statements on a scale ranging from (1) 'none of the time' to (5) 'most of the time'. Reliability and validity data has been reported for the self-report format of the RI scale across disasters (Frederick, Pynoos & Nader, 1992, cited in La Greca et al., 1996; Vernberg et al., 1996). Correlations with PTSD cases have been good (.91)(Frederick, 1985, cited in La Greca et al., 1996), while internal consistency has also been acceptable (alpha = .83, Vernberg et al., 1996). The alpha reliability using the current sample was found to be .88.

The Coping Questionnaire - Child (CQ-C), is a three-item questionnaire used to assess children's perceived ability to cope with specific anxiety provoking circumstances. The CQ-C provides a global score. Like the RI, the CQ-C is designed to be situation specific (Kendall et al., 1992). Thus, in the current study children rated their ability to cope with stimuli related to the volcanic eruptions. The questions were 'when you are thinking about the volcano are you able to help yourself feel less upset?,' 'when you are hearing talk about the volcano at home are you able to help yourself feel less upset?' and 'when you are seeing information about the volcano on TV or in the newspaper are you able to help yourself feel less upset?' Children rate their ability to cope with these stimuli on a 7- point scale ranging from (1) not at all able to help myself, to (7) completely able to help myself feel comfortable. In the current sample, CQ-C alpha reliability was found to be .71 (Ronan, 1997). The CQ-C has also been shown to be sensitive to the effects of treatment (e.g., Kendall, 1994).

Procedure

Children were evaluated in October, 1995 (1-month post-volcanic activity, Time 1) and in December, 1995 (3-months post-volcanic activity, Time 2). Informed consent was obtained from parents or guardians. Children with parental/guardian consent had the study explained, the consent procedure reviewed, queries answered, and then participation from the children themselves requested, and questionnaires administered.

Administration was carried out within classrooms. The administrator of the measures was present for the entire session to read each item, ensure that the measures were completed correctly and answer any questions. Each session started with a statement explaining that children with parental/guardian consent may decide whether or not to participate, and that there are no right and wrong answers to the questions. Those who agreed to participate then had the administrator of the measures read each item aloud, while children followed along and marked their answers. Time for completion was between 30 to 45 minutes.

At time one, questionnaires containing the full set of measures were read aloud to at each of the three schools by the same trained child researcher in an effort to avoid any confounding factors that may have been associated with administration (i.e., reading difficulties). At Time 2, questionnaires were administered again following the same procedure as at Time 1. However, at Time 2, and for the purposes of this study, children were only required to answer questions on the CQ-C section of the questionnaire. The variable of main interest at this time was the continued level of children's coping ability (the primary criterion variable).

Results

Analyses using total CQ-C scores as the dependent variable were carried out using hierarchical multiple regression. This allowed for the assessment of the effects of shared and unique variance of the predictive factors (Cohen & Cohen, 1983) within the current conceptual model. Contributions of five of the primary factors specified in the conceptual model (exposure, preexisting child demographics, home factors, NCES, Posttraumatic stress) on total coping at Time 1 were tested using this method. To predict coping at Time 2, coping at Time 1 was included as an additional predictor.

When a primary factor was comprised of two or more variables (e.g., child demographics was represented by age, sex and ethnic group), the variables were entered as a block. Blocking groups of variables allowed for a test of effects of both shared and unique variance among a group of variables, and avoided misleading conclusions resulting from multicollinearity between independent variables (Cohen & Cohen, 1983).

The standardized beta coefficients for each variable within the blocks are reported. By examining the beta coefficients at each step it is possible to observe the effects of individual variables on the dependent variable within each step, and the extent to which subsequent steps alter these effects. {Beta weights are reported as opposed to semi-partial correlation coefficients as they follow previous research protocol (e.g., La Greca et al., 1996; Vernberg et al., 1996). The beta weights allow standardized comparisons results to be comparable with previous and future research findings.} The total variance explained by each step of the equation is provided (R2 and adjusted R2) along with the added variance explained by each block of variables while controlling for previous blocks (R2 change).

Findings are organized into three sections. The first presents descriptive statistics. The second, analysis of children's coping with posttraumatic stress approximately one-month post-disaster. The third, analysis of the prediction of children's coping approximately three months following the volcanic eruptions.

Descriptive Statistics
To aid in the interpretation of results means and standard deviations for relevant variables, and bivariate correlations for all variables are presented in Table 1, and Table 2 respectively. In general, the negative impact of the Mount Ruapehu eruptions on children was moderate to low. Children's mean scores generally reflected moderate to low levels of stress and moderate levels of coping, both initially and over time (Table 1). All variables correlated at a statistically significant level with total CQ-C scores at time one with the exception of the exposure variable, location and two child demographic variables - sex and ethnicity (Table 2). All variables correlated at a statistically significant level with total CQ-C scores at time two with the exception of the exposure variable, location and the child demographic variable ethnicity (Table 2). All NCES factors (including ratings on the RI) significantly correlated with each other.

Table 1.  Means and Standard Deviations for Life Threat, Age, Home Factors, Cognitive and Emotional Style, Reaction Index and Coping Ability

Note n = 184;   *n = 127

Table 2.     Bivariate Relations between Coping Ability, Exposure, Child Demographics, Home Factors, Cognitive & Emotional Style and Posttraumatic Stress.

*p<0.05,  **p<0.01,   ***p<0.001
Note: n = 127 for bivariate correlations between variables and Coping at time two;
n = 186 for all other bivariate correlations.

Exposure. Exposure (location and life threat) explained 21% of the variance in coping, F(2,181)=25.25, p<.001. Perception of life threat was the most important exposure factor with life threat showing a beta contribution of -.46 (p<.001), while location only showed a beta contribution of .08 (ns)(Table 3). Life threat continued to make a significant contribution to the model at step six of the analysis, once all the primary variables were included (beta = -.16, p<.05)(Table 3). Decreased perception of life threat was associated with increases in coping.

Preexisting child characteristics. Inspection of Table 3 indicates that child demographics (age, sex, ethnic group) explained a significant unique change in the variance in coping of 5% when entered after exposure, F(5,178)=12.79, p<.001. Age was the only individual factor within this group to have a significant impact on coping. Increases in age were associated with increases in coping (beta = .21, p<.01). Age continued to be significant when home factors (step 3) were entered (beta = .13, p<.05). However, once NCES factors were entered (step 4) age no longer made a significant contribution to the model.

Home factors. Home factors accounted for a unique change of 11% in coping, F(7,176)=14.88, p,.001. Beta coefficients indicated that upsetting talk at home about the volcanic activity had the more significant effect on coping (beta = -.24, p<.01), followed by children's increased perception of how much their parents were specifically upset as a result of the eruptions (beta = -.21, p<.01)(Table 3). Once all the primary variables where included neither of the home factor variables remained significant as individual contributors within the model.

Negative cognitive and emotional style. NCES variables (NASQ, CDI, STAI - trait, -state) accounted for a unique change of 11% in coping , F(11,172)=14.06, p<.001. Of the individual variables within this block, CDI and anxiety-state made a significant contribution towards predicting coping. Lower levels of depression and state anxiety were significantly associated with greater levels of coping (beta = -.17, p<.05; beta = -.17, p<.05)(Table 3). Once the final primary factor was included (step 5) only scores on the CDI remained significant as an individual predictive factor of coping within this block of variables (beta = -.16, p.<.05)(Table 3).

PTSD symptomatology. The RI accounted for an insignificant .3% (Table 3) of the unique variance in coping F(12, 171)=12.97, p<.001. Decreases in PSTD symptomatology were associated with non-significant increases in children's coping scores.

A final inspection of Table 3 indicates that as each consecutive step was carried out individual variables were impacted on. Once all the primary factors were included only perception of life threat and the CDI remained individually significant in ability to predict coping. However, with the exception of PTSD symptomatology, each of the primary factor blocks significantly added to the models overall capacity to predict coping ability at time one.

Table 3.   Heirarchical multiple regression of Exposure, Child Demographics, Home Factors, Cognitive and Emotional Style and Posttraumatic Stress on coping Ability at Time One showing standardized regression coefficients, R, R2, Adjusted R2 and R2 change (N=183).

*p<0.05, **p<0.01, ***p<0.001.
Note: Preliminary analysis incorporating posttraumatic stress within negative cognitive and emotional style factors (Step 4) did not alter R2 Change. Additionally, the posttraumatic stress contribution to the model (indicated by the beta weight) was the same as step 5 of the current analysis. Thus, it appears that the explanatory value of PTSD symptoms for children's coping post-disaster is the same regardless of other negative cognitive and emotional factors.

Prediction of Coping at Time 2
At time two, three months post-disaster, over 47% of the variance (Adjusted R2) in total coping ability was explained, F(13,113)=9.56, p<.001.

Coping at time one. Prior to the inclusion of other predictive factors children's initial coping ability (as measured by the CQ) accounted for 32% of the unique variance in coping, F(1, 125)=60.62, p<.001). Coping at time one remained a significant predictive factor once all primary factors were included (beta = .23, p<.05)(Table 4). Higher levels of initial coping were related to greater levels of coping at time two.

Exposure. The ability of exposure (location and life threat) to predict children's coping decreased over time. Three months post-disaster, exposure explained an insignificant 2% of the variance (Table 4) in coping F(3,123)=21.77, p<.001.

Preexisting child characteristics. Child demographics (age, sex, ethnic group) explained an insignificant 4% of unique variance in coping, F(6,120)=12.56, p<.001.

Home factors. Home factors accounted for a unique change of 10% in coping, F(8,118)=14.20, p<.001. Beta coefficients indicated that over time upsetting talk at home about the volcanic activity continued to be a significant contributor towards explaining variance in children's coping (beta = -.49, p<.001) while perception of how much parents were upset did not (beta = -.14, ns)(Table 4). Upsetting talk at home about the volcanic activity remained significant once all the primary factors were included (beta = -.30, p<.01)(Table 4). Lower (more adaptive) scores on the home factor variables were associated with increased coping ability scores.

Negative cognitive and emotional style. NCES variables accounted for an insignificant unique change of 3% in coping, F(12, 114)=10.21, p<.001.

PTSD symptomatology. PTSD symptomatology (as measured by the RI scale) increased in importance over time. However, the contribution remained insignificant. PTSD symptomatology accounted for an insignificant 1% of the unique variance in coping, F(13,113)=9.56., p<.001. As each consecutive step in time two's analyses was carried out individual variables were impacted on. Once all the primary factors were included only the home factor _ upsetting talk at home about the volcano and coping at time one remained significant individual contributors in ability to predict children's coping ability over time. However, it should be noted that each successive step added to the model's overall ability to predict coping at Time 2.

Table 4.   Heirarchical multiple regression of Coping Ability at Time One, Exposure, Child Demographics, Home Factors, Cognitive and Emotional Style and Posttraumatic Stress on Coping Ability at Time Two showing standardized regression coefficients, R, R2, adjusted R2 and R2 change (N=126).

*p<0.05, **p<0.01, ***p<0.001.
Note: Preliminary analysis incorporating posttraumatic stress within negative cognitive and emotional style factors (Step 5) did not alter R2 Change. Posttraumatic stress's contribution (beta weight) was the same as step 6 of the current analysis. Thus, it appears that the predictive value of PTSD symptoms for children's coping post-disaster is the same regardless of other negative cognitive and emotional factors.

Discussion

Results provide further support for the utility of integrative conceptual models for understanding children's reactions to disasters (see also La Greca et al., 1996; Vernberg et al., 1996). Unlike much of the previous research, which has tended to be unsystematic and haphazard, this model has the potential to provide a useful method of organizing researchers' and clinicians' thinking about children's coping and the factors that influence them. The conceptual model can also serve to guide future research efforts in this field.

In addition to providing overall support for the utility of the current conceptual model, results suggest that the factors in the model may also be helpful for understanding changes in children's coping over time. Factors of particular salience over time were home factors, anxiety, and initial coping ability. Factors of lesser significance within the current model included, exposure factors (location and life threat), child demographics (age, sex, ethnicity) and the NCES factors, depression negative affect and PTSD.

One-month post disaster exposure explained the greatest amount of variance in children's coping ability. Consistent with previous research (e.g., La Greca et al., 1996; Vernberg et al., 1996), self-talk related to life threat was a significant contributing factor initially. Exposure in terms of location or proximity did not explain coping ability one month or three months post-trauma indicating that other factors are of more importance. It should be noted however, that while each location studied was exposed to differing levels of volcanic fall-out due to location all where under potentially similar risk from the volcano. Children may have perceived their lives to be equally threatened due to potential risk rather than the actual consequential effects of location. The very poor relationship between children's perception of life-threat and location tend to support this train of thought. For the current research, location was possibly not indicative of exposure.

Child demographics went from contributing significantly to the model initially to becoming insignificant over time. However, the contribution to the model remained approximately the same. Additionally, child demographics showed the expected relationship with other variables included within the model. Increases in children's age or developmental level were significantly related to increased distress expressed within home factors, anxiety and PTSD symptomatology. Age findings are supported by previous research indicating aspects of children's responses to stressful events are impacted on by developmental level (Atkins, 1991; Green et al., 1991; Yule, 1994). The relationship between gender and ability to cope remains unclear due to gender's poor relationship with the other variables in the model. Very little cultural differences were found in relation to children's ability to cope suggesting that children reacted similarly regardless of their ethnic group (Asian, Maori, Pacific Islander, Pakeha/European, other). A possible explanation for a lack of cultural reactive differences for children may be that they have had less time being socialized within specific cultural norms. It is also possible that the small rural communities investigated are relatively close knit creating a common subculture.

The contribution of home factors contribution to children's ability to cope one and three months post- disaster highlights the importance of caregiver communication. The explanatory value of the extent children were upset or distressed by talk about the volcano at home was particularly salient. Results supported previous findings that the extent to which children are able to cope with stressful events is often dependent on communication from caregivers, initially and over time (Atkins, 1991; Ronan, 1997; Vernberg et al., 1996).

The impact of upsetting talk at home about the volcano on coping also highlights the fact that certain forms of communication within the home may be more detrimental than helpful. While communication about a traumatic event is often purported to be a means of social support and can provide information that promotes coping (i.e., emotional processing of the fear and anxiety associated with the traumatic event)(Joseph et al., 1995), some forms of communication are obviously detrimental. Ronan (1997) found that asthmatic children perceived their parents to be significantly more upset about the eruptions than the parents themselves reported and this appeared to be related to unclear or negative parental communication. Intervention efforts may benefit from a focus on helping parents to clearly communicate with their children. That is, parents may need immediate assistance in coping that includes adaptive forms of child-related communication.

Despite the high positive relationship, both between coping and each of the NCES factors and within the NCES factors themselves only depression and anxiety held strong explanative and or predictive value. Initially depression was a significant explanatory factor along side anxiety. Previous research indicates that the co-occurrence of depression and anxiety is common (Last, Strauss & Francis, 1987; Kendall, 1994). Consistent with a developmental focus, disorders comorbid with anxiety show trends across different age groups with older children more often comorbid with depression and dysthymia than younger children (see review by Ollendick & King, 1994). The relationship between anxiety level and age and anxiety and depression are reflected in the current research.

PTSD symptomatology had very little explanatory or predictive value after taking into consideration other NCES factors within the current conceptual model. It is acknowledged that the RI may in fact be measuring many of the same negative cognitive and emotional factors that the CDI, the NASSQ, and the STAI -S & -T inventories measure. This may explain the apparent low level of importance of posttraumatic stress within the current model. The moderate to high level of correlation between step four and step five factors appears to support this train of thought. However, when preliminary analyses are taken into consideration the afore mentioned does not hold true. Preliminary analyses indicated that when posttraumatic stress is included within the NCES block it does not increase in importance. Preliminary analyses indicated that posttraumatic stress (as measured by the RI) does not make a greater contribution to the model, both with regards to an individual contribution and to the overall contribution.

Children's initial coping was one of the largest predictive factor for children's coping over time. Once all factors within the model were considered, initial coping's predictive value was second only to how upsetting talk at home was. These findings support the idea that intervention may benefit from a focus on specific forms of self-talk and related affect (e.g., life threat, social support). Additionally, the fact that initial coping ability was of greater important than NCES symptoms (including PTSD symptoms) suggests that post trauma coping ability may be more strongly influenced over time by other factors. Thus, early intervention approaches may need to focus on active coping strategies rather than, for example, simply providing reassurance to reduce distress (Long, Ronan, & Perreira-Laird, 1997; Ronan & Johnston, 1996).

Despite the poor predictive value of NCES factors the current research supports the idea of a relationship between explanative or predictive factors of PTSD symptomatology (La Greca et al., 1996; Vernberg et al., 1996) and the explanation and or prediction of coping ability. Variables such as life threat and demographic factors previously shown to impact PTSD symptomatology (La Greca et al., 1996; Vernberg et al., 1996) also impacted children's ability to cope. However, it is also the case that other factors (home factors, depression, state-anxiety) not previously or directly used in those studies to predict PTSD symptomatology were found to be important in predicting coping.

There are several limitations that should be noted for the present study. First, the current research relied on children's self-report questionnaires. However, previous research has generally supported the idea (Loeber, Green, & Lahey, 1990; Vernberg et al., 1996) that children tend to be more reliable reporters of their internalized states than other sources like parents or teachers (e.g., La Greca et al., 1996; Silverman & Eisen, 1992). This may be particularly true for reporting after a natural disaster, as typical informants (parents and teachers) may be preoccupied and under stress themselves. Under such circumstances, teachers and parents may not be as aware of children's inner emotional states as the children are themselves (La Greca et al., 1996; Vernberg et al., 1996).Another issue worth bearing in mind when interpreting results is times of data collection. Children's reports were obtained one- and three months post-disaster. Generalizability of the study's findings are thus limited to this particular time frame. The more long-term effects of how children cope and how the above factors impact on coping are currently unknown. Future research is important to examine children's reactions to disaster over a more extended time frame, using longitudinal designs similar to those used in the current study.

The variables in the current study account for a little under half of the variance in coping at Time 1 and a little more than half of the variance at Time 2. Further investigation is needed to assess the impact of other variables such as the impact of premorbid functioning, life events, and other daily hassles that may impact on coping following stressful events (see also La Greca et al., 1996).

Future research may also want to investigate the impact of both formal and informal interventions on coping ability following disasters. For example, investigation of the impact of intervention on children's anxiety (state and trait, PTSD) and the impact on actual coping ability are two obvious research avenues. Previous research indicates that most teachers report discussing disaster related events following such an occurrence. The content and outcome of these informal interventions are unclear (La Greca et al., 1996).

Finally, findings accentuate the role of perception in helping children manage distress related to natural disasters. It may be critical to target these factors in post-disaster interventions for children. The current findings provide some foundations for developing interventions, both formal and informal. Results suggest that interventions may want to be implemented both within the classroom setting, specifically directed at developing coping strategies and accessing social support, and in the home, specifically directed at healthy versus distorted forms of coping-related communication. Teacher- or counselor-led activities, emphasizing coping with disaster-related issues and reducing cognitive distortions, may provide models of effective coping and problem solving. In addition, encouraging supportive outreach to peers and from important adults (teachers and parents) appears particularly important.

Interventions aimed at enhancing children's social support via systems within the home (and without) may prove to be particularly challenging following a natural disaster. Entire communities are often affected by natural disasters. Thus, there is a high likelihood that support providers (i.e., families, friends, teachers) are also victims (La Greca et al., 1996) and may need to be helped themselves. One area may be in helping impacted adults access or develop communication skills to be effective as coping models and support systems for children in the aftermath of a disaster. Helping parents understand that helping their children can also have a beneficial impact on themselves is one avenue worth investigating.

In conclusion, support for the utility of the conceptual model in explaining children's post-disaster coping strengthens confidence in focusing on such things as life threat, level of healthy communication provided at home, and related cognitive and emotional factors that promote adaptive coping following a traumatic event. It is suggested that systematic evaluations of post-disaster interventions, especially ones based on clear, validated conceptual models, will offer the greatest hope of improving children's post-disaster psychological adjustment and coping.

References

Atkins, F.D. (1991). Children's perspective of stress and coping: An integrative review. Issues in Mental Health Nursing, 12 (2), 171-178.

Belter, R.W., & Shannon, M.P. (1993). Impact of Natural Disasters on Children and Families. In C.F. Saylor (Ed.). Children and Disasters. New York: Plenum Press.

Buros, O.K. (1978). (Ed.). The Eighth Mental Measurements Yearbook. United States of America: The Gryphon Press.

Coddington, R.D. (1972). The significance of life events as etiologic factors in the diseases of children _ II A study of a normal population. Journal of Psychosomatic Research, 16, 205-213.

Cohen, J., & Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences (2^nd ed.). Hillsdale, NJ: Lawrence Erlbaum Association.

Compas, B.E. & Epping, J.E. (1993). Stress and coping in children and families: Implications for children coping with disaster. In C.F. Saylor (Ed.), Children and Disasters (p. 11-28). New York: Plenum Press.

Compas, B.E., Forsythe, C.J. & Wagner, B.M. (1988). Consistency and variability in causal attributions and coping with stress. Cognitive Therapy and Research, 12 (3), 305-320.

Everly, G.S. & Lating, J.M. (1995). (Ed.). Psychotraumatology: Key papers and core concepts in post-traumatic stress. New York: Plenum Press.

Gibbs, M.S. (1989). Factors in the victim that mediate between disaster and Psychopathology: A review. Journal of Traumatic Stress, 2 (4), 489-514.

Green, B.L., Korol, M., Grace, M.C., Vary, M.G., Leonard, A.C., Gleser, G.C., & Smitson-Cohen, S. (1991). Children and disaster: Age, gender, and parental effects on PTSD symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 945-951.

Hodgkinson, P.E. & Stewart, M. (1991). Coping with Catastrophe: A handbook of disaster management. New York: Routledge.

Jonhston, D.M., Neall, V.E., Houghton, B.F., Ronan, K.R. and Paton, D. (1998). The social and economic impacts of the 1945 and 1995-1996 Ruapehu eruptions, New Zealand: An example of increasing social vulnerability. Manuscript in preparation.

Joseph, S., Williams, R. & Yule, W. (1995). Psychosocial perspectives on post-traumatic stress. Clinical Psychology Review, 15 (6), 515-544.

Kendall, P.C. (1994). Treating anxiety disorders in children: Results from a randomized clinical trial. Journal of Consulting and Clinical Psychology, 62 (1), 100-110.

Kendall, P.C. & Ronan, K.R. (1990). Assessment of children's anxieties, fears, and phobias. In C.R. Reynolds & R.W. Kamphaus (eds.). Handbook of Psychological and Educational Assessment of Children (pp. 223-244). New York: Guilford Press.

Kovacs, M. (1981). Rating scales to assess depression in school aged children. Acta Paedopsychiatrica, 46, 305- 315.

Kramer, J.J. & Conoley, J.C. (1992) (eds.). The Eleventh Mental Measurements Yearbook. United States of America: Buros Institute of Mental Measurements of the University of Nebraska- Lincoln.

La Greca, A.M., Silverman, W.K., Vernberg, E.M. & Prinstein, M.J. (1996). Symptoms of posttraumatic tress in children following Hurricane Andrew: A prospective study. Journal of Consulting and Clinical Psychology, 64 (4), 712-723.

Last, C.L., Strauss, C.C., & Francis, G. (1987). Comorbidity among childhood anxiety disorders. Journal of Nervous and Mental Disease, 175, 726-730.

Loeber, R., Green, S.M., & Lahey, B.B. (1990). Mental health professionals' perception of the utility of children, mothers, and teachers as informants of child psychopathology. Journal of Clinical Child Psychology, 19, 136-143.

Long, N.R., Ronan, K.R., & Perreira-Laird, J. (1997). Victims of disaster. In N. Singh (Ed.). Comprehensive Clinical Psychology: Applications in diverse populations. New York: Pergamon: in press.

Lonigan, C.J., Shannon, M.P., Taylor, C.M., Finch, A.J. & Sallee, F.R. (1994). Children exposed to disaster: II. Risk factors for the development of post-traumatic symptomatology. Journal of the American Academy of Child and Adolescent Psychiatry, 33 (1), 94-105.

McFarlane, A.C. (1987). Posttraumatic phenomena in a longitudinal study of children following a natural disaster. Journal of American Academy of Child and Adolescent Psychiatry, 26 (5), 764-769.

Ollendick, T.H. & King, N.J. (1994). Diagnosis, assessment, and treatment of internalizing problems in children: The role of longitudinal data. Journal of Consulting and Clinical Psychology, 62, 918-927.

Pynoos, R.S. & Nader, K. (1988). Psychological first aid and treatment approach to children exposed to community violence: Research implications. Journal of Traumatic Stress, 1, 445-473.

Ronan, K.R. (1997). The effect of a series of volcanic eruptions on emotional and behavioural functioning in children with atopic disorders. The New Zealand Medical Journal: in press.

Ronan, K.R., Deane, F.P. (1997). Anxiety Disorders. In P. Graham (Ed.). Cognitive Behavior Therapy for Children and Families. Cambridge: Cambridge University Press: in press.

Ronan, K.R. & Johnson, D.M. (1996). Volcanic Eruptions and Children: Assessment and school based intervention. Paper presented at the Pan Pacific 1996 Conference, Vancouver.

Ronan, K.R., Kendall, P.C. & Rowe, M. (1994). Negative affectivity in children: Development and validation of a self-statement questionnaire. Cognitive Therapy and Research, 18 (6), 509-528.

Ryan-Wagner, N.M. (1992). A taxonomy of children's coping strategies: A step toward theory development. American Journal of Orthopsychiatry, 62 (2), 256-263.

Schreurs, K.M.G., & de Ridder, D.T.D., (1997). Integration of coping and social support perspectives: Implications for the study of adaptation to chronic diseases. Clinical Psychology Review, 17, 89-112.

Shannon, M.P., Lonigan, C.J., Finch, A.J., & Taylor, C.M. (1994). Children exposed to disaster: I. Epidemiology of post-traumatic symptoms and symptom profiles. Journal of the American Academy of Child and Adolescent Psychiatry, 33 (1), 80-93.

Silverman, W.K., & Eisen, A.R. (1992). Age differences in the reliability of parents' and children's reports of child anxious symptomatology using a structured interview. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 117-124.

Spielberger, C. (1973). Manual for the State-Trait Anxiety Inventory for Children. Palo Alto, CA: Consulting Psychologist Press.

Suls, J., & David, J.P. (1996). Coping and personality: Third time's a charm. Journal of Personality, 64 (4), 993- 1005.

Taylor, S.E. & Armour, D.A. (1996). Positive illusions and coping with adversity. Journal of Personality, 64 (4), 873-898.

Terr, L.C. (1989). Treating psychic trauma in children: A preliminary discussion. Journal of Traumatic Stress, 2, 3-20.

Titchener, J.L. (1988). Clinical intervention after natural and technological disasters. In M. Lystad (Ed.). Mental Health Response to Mass Emergencies: Theory and practice (pp. 160-180). New York: Brunner/Mazel.

Treadwell, K.R., & Kendall, P.C. (1996). Self-talk in youth with anxiety disorders, states of mind, content specificity, & treatment outcome. Journal of Consulting and Clinical Psychology, 64 (5), 941-950.

Vernberg, E.M., La Greca, A.M., Silverman, W.K., & Prinstein, M.J. (1996). Prediction of posttraumatic stress symptoms in children after Hurricane Andrew. Journal of Abnormal Psychology, 105 (2), 237-248.

Vernberg, E.M. & Vogel, J. (1993). Interventions with children after disasters. Journal of Clinical Child Psychology, 22, 484-498.

Vogel, J.M., & Vernberg, E.M. (1993). Part 1: Children's psychological responses to disasters. Journal of Clinical Child Psychology, 22 (4), 464-484.

Yule, W. (1994). Posttraumatic stress disorder. In T.H. Ollendick, N.J. King, & W. Yule (Eds.). International Handbook of Phobic and Anxiety Disorders in Children and Adolescents (pp. 223-240). New York: Plenum Press.

Author Note

This research was supported by a grant from Massey University (1-0575-67214A) awarded to the second author. Correspondence can be addressed to Kevin R. Ronan, School of Psychology, Massey University, Palmerston North, New Zealand (K.R.Ronan@massey.ac.nz).

Copyright