Family systems approach to attachment relations, war trauma, and mental health among Palestinian children and parents

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=zept20 European Journal of Psychotraumatology ISSN: 2000-8198 (Print) 2000-8066 (Online) Journal homepage: http://www.tandfonline.com/loi/zept20 Family systems approach to attachment relations, war trauma, and mental health among Palestinian children and parents Raija-Leena Punamäki, Samir R. Qouta & Kirsi Peltonen To cite this article: Raija-Leena Punamäki, Samir R. Qouta & Kirsi Peltonen (2017) Family systems approach to attachment relations, war trauma, and mental health among Palestinian children and parents, European Journal of Psychotraumatology, 8:sup7, 1439649, DOI: 10.1080/20008198.2018.1439649 To link to this article: https://doi.org/10.1080/20008198.2018.1439649 © 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. Published online: 20 Mar 2018. Submit your article to this journal Article views: 128 View related articles View Crossmark data BASIC RESEARCH ARTICLE Family systems approach to attachment relations, war trauma, and mental health among Palestinian children and parents Raija-Leena Punamäki a , Samir R. Qouta b and Kirsi Peltonen a a Faculty of Social Sciences Psychology, University of Tampere, Tampere, Finland; b Department of Education and Psychology, Islamic University Gaza, Gaza City, Palestine ABSTRACT Background: Trauma affects the family unit as a whole; however, most existing research uses individual or, at most, dyadic approaches to analyse families with histories of trauma. Objective: This study aims to identify potentially distinct family types according to attachment, parenting, and sibling relations, to analyse how these family types differ with respect to war trauma, and to explore how children’s mental health and cognitive processing differ across these family types. Method: Participants included Palestinian mothers and fathers (N= 325) and their children (one per family; 49.4% girls; 10–13 years old; mean ± SD age = 11.35 ± 0.57 years) after the Gaza War of 2008–2009. Both parents reported their exposure to war trauma, secure attachment availability, and parenting practices, as well as the target child’s internalizing and externalizing symptoms [Strengths and Difficulties Questionnaire (SDQ)]. Children reported their symptoms of post-traumatic stress disorder (on the Children’s Revised Impact Event Scale), depression (Birleson), and SDQ, as well as their post-traumatic cognitions (Children’s Post Traumatic Cognitions Inventory). Results: A cluster analysis identified four family types. The largest type reflected secure attachment and optimal relationships (security and positive family relationships, 36.2%, n= 102), and the smallest exhibited insecurity and problematic relationships (insecurity and negative family relationships, 15.6%; n= 44). Further, families with discrepant experiences (23.0%; n= 65) and moderate security and neutral relationships (25.2%; n= 71) emerged. The insecurity and negative relationships family type showed higher levels of war trauma; internalizing, externalizing, and depressive symptoms among children; and dysfunctional post-traumatic cognitions than other family types. Conclusion: The family systems approach to mental health is warranted in war conditions, and therapeutic interventions for children should, thus, also involve parents and siblings. Knowledge of unique family attachment patterns is fruitful for tailoring therapeutic treatments and preventive interventions for war-affected children and families. El abordaje de sistemas familiares para las relaciones de apego, el trauma de guerra y la salud mental entre niños y padres palestinos Objetivo: el trauma afecta a la unidad familiar como un todo; sin embargo, la mayoría de las investigaciones existentes utilizan abordajes individuales o, a lo sumo, diádicos para analizar familias con historias de trauma. El objetivo de este estudio es identificar tipos de familia potencialmente distintos en función de las relaciones de apego, de crianza y entre hermanos; analizar cómo difieren estos tipos de familia con respecto al trauma de guerra; y explorar cómo la salud mental y el procesamiento cognitivo de los niños difieren entre estos tipos de familia. Métodos: Los participantes incluyeron madres y padres palestinos (N= 325) y sus hijos (uno por familia, 49.4% niñas, 10-13 años, M = 11.35 ± 0.57) después de la Guerra de Gaza de 2008 a 2009. Ambos padres comentaros su exposición a los traumas de guerra, su disponibilidad para ofrecer un apego seguro y sus prácticas de crianza, así como a los síntomas de internalización y externalización del niño (Strengths and Difficulties Questionnaire, SDQ; Cuestionario de fortalezas y dificultades). Los niños indicaron sus síntomas de trastorno de estrés postraumático (TEPT, CRIES), depresión (Birleson) y SDQ, así como sus cogniciones postraumáticas (CPTCI). Resultados: Un análisis declusters identificó cuatro tipos de familia. El tipo más numeroso reflejaba apego seguro y relaciones óptimas (seguridad y relaciones familiares positivas, 36.2%, n= 102), y el tipo menos numeroso mostraba inseguridad y relaciones problemáticas (inseguridad y relaciones familiares negativas, 15.6%, n= 44). Además, aparecieron familias con experiencias discrepantes (23.0%, n= 65), con seguridad moderadas y relaciones neutrales (25.2%, n= 71). El tipo de familia con inseguridad y relaciones negativas mostró niveles más altos de trauma de guerra; síntomas de internalización, externalización y depresión entre los niños; y cogniciones postraumáticas disfuncionales que otros tipos de familia. ARTICLE HISTORY Received 22 September 2017 Accepted 31 January 2018 KEYWORDS War trauma; attachment styles; post-traumatic stress disorder (PTSD); depression; families; Palestinian PALABRAS CLAVE trauma de guerra; estilos de apego; trastorno por estrés postraumático (TEPT); depresión; familias; palestinos 关键词 战争创伤;依恋风格;创伤 后应激障碍（PTSD）;抑 郁;家庭;巴勒斯坦 HIGHLIGHTS •Diversity of attachment and sibling relationships characterizes families in war conditions, showing both resilience and vulnerabilities. •Children in families with secure attachment, warm sibling relationships, and optimal parenting show good mental health and functional trauma processing. •Psychosocial interventions among war-affected children should also improve family relationships, e.g. by increasing secure attachment and by decreasing conflicts and rivalry in siblingship. CONTACT Raija-Leena Punamäki [email protected] Faculty of Social Sciences/Psychology, FI-33014 University of Tampere, Kalevankatu 5, Linna 4krs, Tampere, Finland EUROPEAN JOURNAL OF PSYCHOTRAUMATOLOGY, 2018 VOL. 8, 1439649 https://doi.org/10.1080/20008198.2018.1439649 © 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Conclusión: El abordaje de sistemas familiares para la salud mental está justificado en condiciones de guerra, y las intervenciones terapéuticas para los niños deberían, por tanto, implicar también a padres y hermanos. Conocer los patrones de apego familiar particulares es fructífero para adaptar los tratamientos terapéuticos y las intervenciones preventivas para niños y familias afectados por la guerra. 对巴勒斯坦儿童和父母的依恋关系，战争创伤和精神健康使用家庭系统法 目标：创伤的影响是以家庭为单位的；但大多数现有的研究仅使用个体咨询或者最多双方 咨询来分析有创伤历史的家庭。本研究旨在根据依恋、教养和同胞关系识别潜在不同的家 庭类型，然后分析这些家庭类型如何随战争创伤不同，以及探索儿童的精神健康和认知加 工如何在这些家庭类型间不同。 方法：参加者包括经历加沙战争后2008到2009年的巴勒斯坦父母(N= 325)和子女（每个家庭 一人，49.4% 女孩;10–13 岁; M = 11.35 ± 0.57）。两位父母都报告他们的战争创伤经历，安全 依恋风格和教养方式，也包括孩子的内化和外化问题（《力量和问题问卷》，SDQ）。儿 童报告他们的创伤后应激障碍（PTSD, CRIES）症状，抑郁（Birleson问卷），SDQ和他们的 创伤后认知（CPTCI）。 结果：聚类分析识别四个家庭类型。最大的类型组反映了安全依恋和最好的关系（安全和 积极关系——36.2%的家庭，n=102）。最小的组展示了不安全和问题关系（不安全和负 面关系——15.6%家庭，n=44）。还有，不一致体验的家庭（23.0%; n=65）和中等安全/ 关系(25.2%; n=71)。不安全和负面关系家庭类型和其它类型相比，显示了更高水平的战争 创伤；儿童的内化、外化和抑郁症状；非功能性的PTC。 结论：家庭系统方法应对精神健康在战争情景中是可以担保的，所以针对儿童的治疗性干 预应该让家长或者兄弟姐妹加入。对独特的家庭依恋模式的了解成效显著地让咨询式治疗 和战争影响的儿童和家人的预防性干预变得个性化。 1. Introduction When discussing families living in conflict zones, parents commonly express worry about their children’s safety, and children frequently refer to their family members as a source of security. A Palestinian father said, ‘My 3-year-old promised to guarantee electricity to our whole family when she grows up. I felt ashamed of my inability.’‘Without the help of my bigger sister, I would be dead,’a Palestinian boy said after the Gaza War of 2008–2009. Research has confirmed that families face hardships together, forming a system in which each member takes on an ‘emotional share of work,’showing endurance, manifesting symptoms, and caring for one another (BarajasGonzalez & Brooks-Gunn, 2014; Crittenden & Dallos, 2009; Montgomery, 2004). Research on the transgenerational transition of trauma has indicated both vulnerability, often reflected in relational problems, social withdrawal (Daud, Skoglund, & Rydelius, 2005), and psychiatric symptoms, such as post-traumatic stress disorder (PTSD) (Yehuda & Bierer, 2008), and resilience, which manifests as caring and empathy towards the traumatized family members (Fossion et al., 2015). According to the family systems approach, families are composed of multiple dynamically interacting relational subsystems typically involving marital, parent– children, and sibling relationships (Coyne, Downey, & Boergers, 1992; Minuchin, 1974). Although traumatic experiences, such as war, influence all family members, few studies have focused on the whole family as the unit of research. Instead, studies on trauma-affected families have typically applied a variable-oriented approach, analysing separate characteristics, such as parenting quality or attachment styles, in associating with family members’mental health (Dekel & Monson, 2010). A person-oriented approach, by contrast, can identify unique homogeneous family patterns and depict the dynamic and multiple relationships among family members in different subsystems (Bergman & Magnusson, 1997). In the current study, we apply a person-oriented approach (cluster analysis) to identify family types with different attachment, parenting, and sibling dynamics among Palestinians who live in the midst of political and military conflict in Gaza. We also analyse the role of war trauma as a predictor of family type and explore how children’s mental health and cognitive processing differ across these types. 1.1. Family systems and attachment research System theories conceptualize families according to their structure (e.g. boundaries between subsystems, dominance hierarchies, and communication transparency) and relational context (e.g. autonomy vs intimacy; harshness vs warmth) (Blass & Blatt, 1992; Kerig, 2005; Olson, 2000). Research has delineated, for instance, enmeshed families, characterized by too thin boundaries with easily evoked emotional spillover between parents and children, and disengaged families, characterized by too thick boundaries and a lack of support and emotional closeness among members. Cohesive families, in turn, enjoy a balance between autonomy and intimacy, meaning that members have access to 2R.-L. PUNAMÄKI ET AL. both emotional support and individuality and privacy (Kerig, 2005; Lindblom et al., 2017; Minuchin, 1974). Attachment theory is highly informative for understanding family relations in cases of war and other traumatic conditions. A sense of security is a core motivator for young children, and the goodness of fit among the emotional, cognitive, and behavioural responses of parents and children is essential to survival and mental health (Bowlby, 1969/1982). In secure families, children learn to trust themselves and to seek shelter with their caregivers. They dare to express both positive and negative emotions and develop working models in which they see themselves as capable, others as reliable, and their environment as predictable (Bowlby, 1969/1982). A secure attachment relationship with a sensitive and emotionally available caregiver provides a child with a safe base and supports a balance between exploring and emotional holding. By contrast, children with an insecure–avoidant style seek protection in themselves because they fail to receive security from emotionally distant caregivers. Insecure–ambivalent children try to create a feeling of safety by clinging to caregivers and other adults, who are often emotionally ambivalent and unpredictable. Ample evidence confirms the importance of a sense of security in optimal child development and mental health (Pallini, Baiocco, Schneider, Madigan, & Atkinson, 2014). 1.2. Trauma and family dynamics Models integrating attachment and family systems theories conceive of triadic, dyadic, and individual experiences as developing within the larger attachment and family–cultural networks (Crittenden & Dallos, 2009; Masten & Monn, 2015). Research has found complex dynamics in attachment relations and identified several ways that family members show vulnerability and resilience in the face of traumatic events (Besser & Neria, 2010; Freedman, Gilad, Ankri, Roziner, & Shalev, 2015). A family’s atmosphere, values, beliefs, relational scripts, codes, histories, and emotional sharing all influence family members’responses to trauma (Riggs & Riggs, 2011; Walsh, 2007). Family systems theories provide insight into family members’dynamic social and mental health responses to stress, including compensatory, buffering, and additive dynamics (Davies & Cicchetti, 2004; Minuchin, 1974). A study of war-affected Palestinian families revealed that parents and siblings take on a compensatory emotional ‘share of the work’ when expressing mental health problems and resilience (Punamäki, Qouta, El Sarraj, & Montgomery, 2006). For instance, when mothers showed high levels of depressive symptoms, fathers reported low levels. Similarly, when one sibling was a ‘symptoms carrier’ (e.g. when one sibling suffered from severe PTSD symptoms), others exhibited better adjustment. This sharing of the emotional work was especially evident in families exposed to severe war trauma. Family systems and attachment-informed research is available on military families, especially American veterans of the Iraq and Afghanistan wars (Pemberton, Kramer,Borrego,&Owen,2013; Riggs & Riggs, 2011) and Israeli soldiers, veterans, and prisoners of war (Cohen, Zerach, & Solomon, 2011; Ein-Dor, Doron, Solomon, Mikulincer, & Shaver, 2010; Zerach, Greene, Ein-Dor, & Solomon, 2012). These studies found negative family dynamics, especially when the veteran parents had PTSD. The families with a traumatized parent typically suffered a low sense of security, biased or narrow family attachment networks, relational rigidity with cemented roles of strength and weakness, and scapegoating. Transgenerational research on Holocaust (Van Ijzendoorn, Bakermans-Kranenburg, & Sagi-Schwartz, 2003) and torture (Montgomery, 2011)survivorsprovides further information on the dynamics of families affected by trauma. Parental past trauma has been found to be associated with PTSD, depression, and somatic complaints among children (Montgomery & Foldspang, 2005; Yehuda & Bierer, 2008). Poor parenting and siblingship, as well as biased and silenced family communication, can underlie children’svulnerabilityto parental trauma (Bryant, 2016; Frewen, Brown, DePierro, D’Andrea, & Schore, 2015;Schierholz, Kruger, Barenbrugge, & Ehring, 2016). Traumatized parents can be either overprotective (owing to fears and concerns for their children’ssafety) or unable to tolerate their children’smanifestationsof fear, helplessness, and neediness (Scheeringa & Zeanah, 2001). Thus, parents may adopt intrusive and insensitive child-rearing practices, typical of persons with preoccupied attachment styles, or may easily withdraw from dyadic interactions, typical of persons with avoidant attachment styles (Flykt, Kanninen, Sinkkonen, & Punamäki, 2010; Van Ee, Kleber, Jongmans, Mooren, & Out, 2016). A qualitative study confirmed that parental refugee trauma disturbed children’s creation of secure attachments because the parental fears were overwhelming (De Haene Grieten, & Verschueren, 2010). In a follow-up setting of Israeli families with traumatic war experiences, Besser and Neria (2010) showed that parental insecure attachment predicted severe depressive symptoms and poor social support, both of which can compromise optimal parent–child relations (Frewen et al., 2015). By contrast, supportive, secure, and wise parenting practices can protect children’s mental health, optimal development, and resilience in the life-endangering conditions of war (Betancourt et al., 2011; Cummings et al., 2011; Feldman, Vengrober, Eidelman-Rothman, & Zagoory-Sharon, 2013; Qouta, Punamäki, Miller, & El Sarraj, 2008). A study of Northern Irish children and parents confirmed that secure emotional family relations predicted low levels of psychological distress EUROPEAN JOURNAL OF PSYCHOTRAUMATOLOGY 3 despite prolonged and severe sectarian and political violence (Cummings et al., 2011). Follow-up analyses revealed close and dynamic interactions among traumatic political events, children’s aggressive symptoms, and family violence and conflicts. Children’s community-evoked emotional insecurity made them more vulnerable when facing family conflicts, and vice versa (Cummings, Taylor, Merrilees, Goeke-Morey, & Shirlow, 2016). 1.3. Trauma, child development, and well-being There is ample evidence of war trauma negatively affecting children’s mental health by increasing symptoms of PTSD, depression, anxiety, and aggression (Attanayake et al., 2009;Dubowetal.,2009). Less is known about war trauma’s impacts on children’s attachment styles, although, theoretically, trauma is assumed to increase the risk of insecure attachment. Childhood traumas of neglect, deprivation, and socioeconomic hardship predict the development of insecure attachments (Scheeringa & Zeanah, 2001), and some research shows a higher level of insecure attachments among traumatized children and adolescents (Huemer et al., 2016;Turunen,Haravuori,Punamaki, Suomalainen, & Marttunen, 2014). It is generally agreed that it is not exposure to war trauma alone, but also survivors’cognitive–emotional processing of their experiences that contribute to mental health consequences (Ehlers & Clark, 2000; Schnyder et al., 2015). According to cognitive theories, thinking, attributions, appraisals, and beliefs about traumatic events are decisive for the emergence and maintenance of mental health problems like PTSD or depression (Ehlers & Clark, 2000; Meiser-Stedman, 2002). A trauma victim’s negative post-traumatic cognitions (PTCs) depict the core of his or her perceptions of himself or herself as helpless and worthless, of other people as malevolent and distrustful, and of the community as dangerous and unpredictable (Foa, Ehlers, Clark, Tolin, & Orsillo, 1999; Meiser-Stedman et al., 2009). Research has confirmed that overly negative appraisals, dysfunctional memories, supressed emotions, and feelings of guilt and anger predict PTSD in children (Trickey, Siddaway, MeiserStedman, Serpell, & Field, 2012). Many childand trauma-related factors predict dysfunctional and negative PTCs (Ehlers, Mayou, & Bryant, 2003). However, we found no studies on how attachment or other family relationships contribute to the ways in which children process their traumatic experiences, which is the topic of the current research. 1.4. Research questions First, our study aims to identify distinct family types according to mother, father, and child attachment responses; sibling subsystems; and parenting among families living in war conditions. Secondly, we examine how the family types differ in their exposure to traumatic war events, as indicated by human losses, material destruction, the witnessing of horrors, and the experiencing of life-threatening situations during the Gaza War. Thirdly, we analyse how family type is associated with children’s mental health, as assessed by PTSD, depressive, and internalizing and externalizing symptoms. Fourthly, we analyse whether children’s PTCs differ across the identified family types. 2. Method 2.1. Participants and procedure Palestinian families (N=325), each with a mother, a father, and one 10–13-year-old child (mean age = 11.35 years, SD = 0.57 years; 49% girls), participated in May 2009, 3 months after the Gaza War. This is a baseline subsample of a larger randomized intervention study (N= 482) that examined the effectiveness of a post-war psychosocial intervention programme (Qouta, Palosaari, Diab, & Punamäki, 2012). In the family subsample, both mothers (n= 337) and fathers (n=328) responded, and their children participated in either the intervention group or the control group. The analysed data are from the 325 participating families for which both parents provided information to support the analysis of their family dynamics. The return rates for the mothers’and fathers’ questionnaires were 69.9% and 68.0%, respectively. The family subsample of 325 did not differ from the families who did not participate (n= 157) in terms of fathers’or mothers’work status [respectively, χ 2 (1) = 0.38, p= ns and χ 2 (1) = 0.01, p= ns], place of residence [χ 2 (1) = 0.28, p= ns], family structure [χ 2 (2) = 2.3, p= ns], or size [χ 2 (2) = 0.11, p= ns]. However, the participating families were biased towards having more girls as the target child (56.1%) than the families that did not participate (35.2%) [(χ 2 (1) = 17.72, p< 0.0001]. The original baseline sample represented the regions of the Gaza Strip that were severely bombed during the Gaza War. First, eight schools were randomly selected from 160 potential schools in these regions, taking into consideration that girls and boys go to separate schools. Then, from each of these eight schools, one sixth-grade and one seventh-grade class were randomly chosen, resulting in 16 classes whose pupils participated in the baseline assessment that serves as the data for the current cross-sectional analysis. The ethics boards of the Palestinian Ministry of Education and the Gaza Community Mental Health Program (GCMHP) reviewed and accepted the study’s protocols and measurements, and permission 4R.-L. PUNAMÄKI ET AL. for the study was received from the schools’authorities. Information sheets were provided to children and their parents explaining the procedure of the study, but parents only gave verbal consent for their children to participate in the study and in the psychosocial intervention. Six research assistants collected the children’s data in the classroom. The children took the parents’questionnaires home to complete and returned them in closed envelopes to the research assistants. The second author (SQ) supervised the data collection through weekly sessions and school visits. 2.2. Measures 2.2.1. Parental attachment security This was measured using the 10-item Security Scale (Kerns, Klepac, & Cole, 1996), which depicts a parent’s acceptance of and willingness to serve as an attachment figure and provide a secure base for a child. The items include, for example, ‘I respect my child’s opinions and encourage him/her to express them,’‘I feel a child should be given comfort and understanding when she/he is scared or upset,’and ‘I make sure my child knows that I appreciate what she/ he tries to accomplish.’Mothers and fathers responded by noting how well each item corresponded to their attitudes and behaviours towards the target child on a six-point Likert scale (1 = not at all; 6 = very well). The resulting sum variables showed moderate reliability. The Cronbach’sαvalues were 0.69 for mothers and 0.68 for fathers. 2.2.2. Parents’war trauma The 28-event checklist for Gaza War-related traumatic experiences covered personal exposure to war trauma (e.g. detained, tortured), material destruction (e.g. shelled/bombed neighbourhood, home demolished by the military), family losses (e.g. family member killed, family separation), and witnessing horrors (e.g. witnessing a killing, seeing body parts). Mothers and fathers reported whether they had experienced these events during the war (1 = yes; 0 = no). Four sum variables were constructed for both parents by counting the positive answers relating to personal exposure to trauma (seven events), material destruction (nine events), family losses (five events), and witnessing horrors (seven events). 2.2.3. Negative parenting The 20-item Child Psychological Maltreatment questionnaire (Khamis, 2000) covers emotional abuse (seven items, e.g. ‘My parents humiliate me in front of people’), emotional neglect (seven items, e.g. ‘My parents ignore my attempts to interact with them’), and harsh parenting (six items, e.g. ‘My parents force me to do things and tasks against my will’). Children and parents evaluated how well the descriptions fitted their parents (children) or their own behaviours and rearing practices on a five-point Likert scale (1 = strongly disagree; 5 = strongly agree). Three averaged sum variables were constructed for the children (α= 0.94 for emotional abuse, α= 0.89 for emotional neglect, and α= 0.53 for harsh parenting) and parental variables by combining the responses of the mothers and fathers (α= 0.94 for emotional abuse, α= 0.89 for emotional neglect, and α= 0.73 for harsh parenting). 2.2.4. Children’sattachment style Children’sattachment style was measured using the Coping Strategies Questionnaire (CSQ) (Finnegan, Hodges, & Perry, 1996) and the Security Scale (Kerns, Tomich, Aspelmeier, & Contreras, 2000). Together, these distinguish 28 everyday situations to measure avoidant, preoccupied, and secure attachment. The children’s responses reflected their mothers as attachment figures who helped them, listened to them, and cared for them daily during stressful situations. Their answers were depicted using two-stage methods. For example, ‘avoidant attachment’was assessed by 10 everyday situations, such as, ‘One day you come home from school, and you are upset about something. Your mother asks you what the problem is.’For each situation, respondents chose between two-stage forced choices: (1) talk to her about it or (2) not talk her about it. Underneath these two choices were two more alternatives: (a) sort of true for me or (b) very true for me. ‘Preoccupied attachment’was also assessed by 10 everyday situations, such as ‘Your mother says she is thinking about going to visit a relative for a week.’Here, the two choices were (1) being upset that she was going away for so long and trying to talk her out of going and (2) not being upset and not trying to talk her out of going. Again, two further sub-choices were presented: (a) very true for me and (b) sort of true for me (reverse coding). In measuring ‘felt security attachment’,the children were instructed to answer the question ‘How do you feel about your mother?’by selecting from eight provided two-stage choices, such as: ‘Some kids worry that their mom might not be there when they need her’ (but) ‘Other kids are sure their mom will be there when they need her,’and ‘Some kids feel that their mom does not help them enough with their problems’(but) ‘Other kids think that their mom helps them enough with their problems.’Averaged sum variables were formed for felt security (α= 0.66), avoidant (α= 0.63), and preoccupied (α= 0.54) attachments, and showed low reliabilities. 2.2.5. Sibling relations The quality of siblingship was assessed using an 11-item scale describing positive (warmth and intimacy) and negative (conflict and rivalry) interactions (Dunn, Slomkowski, & Beardsall, 1994). Children estimated EUROPEAN JOURNAL OF PSYCHOTRAUMATOLOGY 5 how often the described events happen in their relations with older (11 items) and younger (11 items) siblings using a five-point scale (1 = never; 5 = always). All items for older and younger sibling relationships correlated significantly, and averaged composite variables were calculated by combining the items for both siblings: warmth in siblingship (e.g. ‘We laugh and joke together’ or ‘I miss him/her when he/she is out of the home’), intimacy (‘I tell him/her about my secrets’or ‘I play and share games with him/her’), conflicts (‘He/she annoys and teases me’or ‘At times, he/she beats me and pushes me’) and rivalry (‘I feel jealous of him/her when he/she takes all my mother’s attention’or ‘I feel unhappy or jealous when other children play with him/her and ignore me’). Averaged sum variables were constructed with reasonable reliabilities (Cronbach’sαvalues were 0.72, 0.68, 0.75, and 0.79, respectively). 2.2.6. Children’spost-traumatic stress symptoms (PTSD) PTSD symptoms were evaluated using the 13-item Children’s Revised Impact Event Scale (CRIES) (Dyregrov, Gjestad, & Raundalen, 2002). The scale covers the three core symptoms of re-experiencing (four items), avoidance (four items), and hyperarousal (five items). Children indicated on a four-point Likert scale (0 = not at all; 4 = often) how often they had experienced each symptom during the last 2 weeks. A total score was constructed, and the Cronbach’sαvalue was low at 0.61. 2.2.7. Children’sdepression The Depression Self-Rating Scale for Children (Birleson, Hudson, Grey-Buchanan, & Wolff, 1987) is an 18-item self-report assessment of the cognitive, affective, and behavioural dimensions of depression. Children estimated on a three-point scale (0 = not at all; 1 = sometimes; 2 = all the time) whether they had experienced each symptom during the last 2 weeks. A sum score of the depression symptoms was then formed with a Cronbach’sαvalue of 0.78. 2.2.8. Children’sexternalizing and internalizing symptoms The Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) uses five behavioural descriptions each to assess hyperactivity, prosocial behaviour, and emotional, behavioural, and relational problems. Both parents and children estimated on a three-point Likert scale (0 = not at all; 1 = somewhat; 2 = yes, fits well) how well the description fitted the target child (or the child her/himself). Sum scores were constructed for the parents’reports by combining the mothers’and fathers’ scores (Cronbach’sαvalues for parents’internalizing and externalizing symptoms were 0.73 and 0.79, respectively) and for the children’s reports (Cronbach’sα values for children’s internalizing and externalizing symptoms were 0.71 and 0.69, respectively). 2.2.9. Children’spost-traumatic cognitive appraisals These appraisals were measured using the 25-item Children’sPost Traumatic Cognitions Inventory (CPTCI) (Meiser-Stedman et al., 2009). The items include statements relating to negative appraisals of the world and the self. The dimensions are (1) the trauma-exposed child as a feeble person in a scary world (e.g. ‘Anybody could hurt me’or ‘I can’t stop bad things from happening to me’) and (2) the life after trauma, involving disturbing and permanent negative change (e.g. ‘My life has been destroyed by the frightening event’or ‘Not being able to get over all my fears means that I am a failure’). Children evaluated on a four-point Likert scale their agreement with each statement (1 = don’t agree at all; 2 = don’t agree a bit; 3 = agree a bit; 4 = agree a lot). A sum variable was constructed for CPTCI scores with an α value of 0.85. 2.2.10. Children’swar trauma A checklist of 28 war-related events was constructed of typical experiences during the Gaza War and military occupation (UN, 2009). The checklist covered childtargeted violence (e.g. being wounded, beaten, or burned by phosphorous bombs), family-related losses (e.g. death of father, mother, siblings, or friends; loss of home; and being separated during the war), witnessing horrors (e.g. witnessing people dying and being injured and seeing body parts), and material destruction (e.g. neighbourhood shelled/bombed, home demolished, or besieged). Children reported whether they had been exposed to each event (1 = yes; 0 = no) either during the war or earlier. War trauma is a linear sum variable reflecting the total number of all ‘yes’answers. 2.2.11. Demographic variables Mothers and fathers reported family income, parental education, work situation, family size, family structure (extended or core), and children reported their age and gender. 2.2.12. Translations The research instruments for the CRIES-13, Birleson depression, and war trauma scales were available in Arabic. The children’s and parents’attachment scales and PTCs (CPTCI) were first translated by a bilingual psychologist from English into Arabic and then translated back by the research group. 2.3. Statistical analyses Cluster analyses were used to identify distinct family type subgroups based on the variables of 6R.-L. PUNAMÄKI ET AL. siblingship, parenting, and mother, father, and child attachment. We first ran a hierarchical cluster analysis to define the number of groups and then ran a K-means cluster analysis to confirm the cluster membership. The dendrogram inspection in the hierarchical clustering helped us to choose the number of initial clusters. K-means clustering produces a cluster centre (centroid) initialization and a squared Euclidean distance measure, which are used to search for the location of each case within the defined family clusters. Before the cluster analyses were run, the variables were standardized into t-scores to avoid biases due to differences in scales (Tabachnik & Fidell, 2007). Cluster membership was tested by multivariate analyses of variance (MANOVAs) using univariate analyses and Tukeyb post-hoc tests. To analyse how the identified family types were associated with families’exposure to war trauma, we ran the MANOVA with the identified cluster family type as an independent variable and fatherand mother-reported war trauma (e.g. personal exposure to trauma, material destruction, family losses, and the witnessing of horrors) as the dependent variables. Further, to analyse the associations between family types and children’s mental health, we ran a multivariate analysis of covariance (MANCOVA) with parentand child-reported internalizing and externalizing symptoms and child-reported depressive and PTSD symptoms as independent variables. A MANCOVA was also used to analyse the family types associated with post-traumatic cognitions (CPTCI), with being a feeble person in a scary world and permanent negative change as the independent variables. Children’s gender and children’s war trauma were the covariates. The MANCOVAs were followed by univariate analyses and Tukey-b post-hoc tests. 3. Results 3.1. Descriptive statistics Table 1 shows the demographic characteristics reported by the parents and their children. About a quarter (24%) of the fathers had a university education, while less than 10% of the mothers had one. Despite their education, about half (49%) of the fathers were unemployed, and nearly all (93%) of the mothers worked at home. These statistics correspond with the economic and social situation in Gaza, which deteriorated following the Israeli military siege and international boycott of the Hamas government (UN:OCHA, 2009). Nearly a third (29%) of the participants lived in extended families, and family size was large: about a quarter (26%) had more than eight children. Figure 1 shows the occurrence of the parents’ reported war trauma. Witnessing horrors was the most common war trauma among both fathers (96.2%) and mothers (89.2%). Furthermore, about half of the mothers reported personal exposure to trauma or family losses, and about 68.4% and 59.7% of fathers reported the same, respectively. Pearson product–moment correlations among parental and child attachment, siblingship, and parenting are presented in Table 2. Parents’secure attachment availability did not correlate with their children’s attachment styles. Further, children’s secure and preoccupied styles correlated significantly and positively, indicating that, in our sample, only avoidant attachment could be considered a genuinely insecure style. Children’s secure attachment correlated positively with warmth and intimacy and negatively with conflict and rivalry in sibling relations, and avoidant attachment was negatively correlated with warm and intimate siblingship and positively correlated with sibling conflict. Preoccupied attachment correlated positively with warm sibling relations. Both parents’and children’s secure attachment were negatively correlated with emotional abuse and emotional neglect, and children’s avoidant attachment correlated negatively with these parenting practices. Intimacy among siblings was Table 1. Percentages and frequencies of demographic family factors. %n Place of living City 84.3 284 Refugee camp 3.3 11 Village 12.5 42 Status Refugee 11.3 38 Citizen 88.7 299 Mother’s education Elementary 19.6 66 Preparatory 32.4 109 Secondary 39.9 134 University 8.0 27 Father’s education Elementary 21.1 71 Preparatory 28.3 95 Secondary 26.2 88 University 24.4 82 Father’s work situation Unemployed 49.3 166 Worker 12.8 43 Public employee 24.9 84 Entrepreneur/self-employed 13.1 44 Mother’s work situation Works at home 93.2 314 Worker 3.0 10 Public employee 3.9 13 Other Family type Immediate 61.9 210 Extended 28.9 98 Tribe 9.1 31 Family size Small (1–4) 23.8 80 Medium (5–7) 50.0 168 Large (8 or more) 26.2 88 EUROPEAN JOURNAL OF PSYCHOTRAUMATOLOGY 7 correlated negatively with parentand child-reported emotional neglect, and warmth among siblings correlated negatively with parent-reported harsh parenting. Conflict and rivalry correlated positively with parentreported harsh parenting, and rivalry in siblingship correlated positively with child-reported emotional neglect. 3.2. Identified family types According to the hierarchical cluster dendrogram, a four-cluster solution was selected because it covered all the data, showed sufficient entropy, and proposed clusters with relatively high numbers of members. The K-cluster analysis confirmed the four-class solution, and Table 3 presents the means, standard errors, and analysis of variance (ANOVA) statistics for the differences among the identified family types with respect to parental attachment, child attachment, siblingship, and negative parenting. The MANOVA test showed a highly significant fit [F Wilks’lambda (45,785.057) = 24.05, p< 0.0001, η 2 = 0.58]. The four distinct family types showed unique attachment, sibling relationships, and negative parenting practices. About a third belonged to the first family type class, ‘security and positive relationships’ (n= 102; 36%), characterized by high parental and child attachment security and low avoidant attachment among children. Sibling relations showed high levels of warmth and intimacy and low levels of conflict and rivalry. Further, parentand child-perceived harsh parenting were the lowest in this family type. The second family type class was labelled ‘insecurity and negative relationships’(n= 44; 15.6%) and was characterized by low parental and child attachment security and high avoidant attachment among children. However, the children’s preoccupied attachment was lowest in these families. Sibling conflicts and rivalry were most common in this family type, and warmth was very low. Both parentand child-reported harsh parenting was very high. The third family type class was called ‘discrepant experiences’(n= 65; 23.0%) and was characterized by both negative and positive family relationships, although these were perceived differently by parents and children. Most notably, parents reported very low levels of harsh parenting, whereas children perceived high levels of harsh parenting. Children’s attachment patterns showed very high avoidance, low levels of security, and relatively high levels of preoccupied attachment, whereas parents showed high levels of secure attachment. Finally, about a quarter belonged to the fourth family type class, ‘moderate security and neutral relationships’(n= 71; 25.2%). Like the parents in the insecurity and negative relationships family type, parents in this class showed low levels of secure attachment. By contrast, children showed high secure and preoccupied attachment and low avoidant attachment, similarly to children in the security and positive relationships families. Furthermore, the quality of siblingship seemed to be comparable to that in the security and positive relationships families. Both parentand childreported harsh parenting was moderate, between the levels reported by the security and positive relationships families and the insecurity and negative relationships families. 3.3. War trauma and family types The MANOVA results revealed that family types differed significantly in terms of the severity of parental war trauma [F Wilks’lambda (24,685.073) = 2.39, p< 0.0001, η 2 = 0.08]. The significant ANOVAs and Tukey-b post-hoc tests showed that fathers in insecurity and negative relationships families reported more material destruction [F(3,243) = 5.04, p< 0.002, η 2 = 0.06] than fathers in other types of families, and more personal exposure to trauma [F (3,243) = 5.50, p< 0.001, η 2 = 0.06] than fathers in security and positive relationship and discrepant experience families. Fathers in insecurity and negative relationships families also reported high levels of family losses (F(3,243) = 3.01, p< 0.03, η 2 = 0.04], but the post-hoc test was not significant. Mothers in the insecurity and negative relationships families also reported more personal exposure to trauma [F (3,243) = 9.59, p< 0.0001, η 2 = 0.11] than mothers Figure 1. Occurrence (%) of war trauma among mothers and fathers. 8R.-L. PUNAMÄKI ET AL. peer relations: A meta-analysis of recent research. Journal of Family Psychology,28(1), 118–123. Pemberton, J. R., Kramer, T. L., Borrego, J., Jr., & Owen, R. R. (2013). Kids at the VA? A call for evidence-based parenting interventions for returning veterans. Psychological Services,10(2), 194–202. Punamäki, R. L. (2014). Mental health and development among children living in violent conditions: Underlying mechanisms for promoting peace. In P. Britto, J. F. Leckman, C. Panter–Brick, & R. Salah (Eds.), Pathways to peace: The transformative power of children and families. strüngmann forum reports (Vol. 15). J Lupp, series ed. Cambridge, MA: MIT Press. Punamäki, R.-L., Qouta, S., El Sarraj, E., & Montgomery, E. (2006). Psychological distress and resources among siblings and parents exposed to traumatic events. International JournalofBehavioralDevelopment,30(5), 385–397. Qouta, S., Punamäki, R.-L., Miller, T., & El Sarraj, E. (2008). Does war beget child aggression? Military violence, gender, age and aggressive behavior in two Palestinian samples. Aggressive Behavior,34, 231–244. Qouta, S. R., Palosaari, E., Diab, M., & Punamäki, R.-L. (2012). Intervention effectiveness among war affected children: A cluster randomized controlled trial on improving mental health. Journal of Traumatic Stress,25,288–298. Riggs, S. A., & Riggs, D. S. (2011). Risk and resilience in military families experiencing deployment: The role of the family attachment network. Journal of Family Psychology,25(5), 675–687. Scheeringa, M. S., & Zeanah, C. H. (2001). A relational perspective on PTSD in early childhood. Journal of Traumatic Stress,14(4), 799–815. Schierholz, A., Kruger, A., Barenbrugge, J., & Ehring, T. (2016). What mediates the link between childhood maltreatment and depression? The role of emotion dysregulation, attachment, and attributional style. European Journal of Psychotraumatology,7, 32652. Schnyder, U., Ehlers, A., Elbert, T., Foa, E. B., Gersons, B. P., Resick, P. A., . . . Cloitre, M. (2015). Psychotherapies for PTSD: What do they have in common? European Journal of Psychotraumatology,6,28186. Sturge-Apple, M. L., Davies, P. T., Cicchetti, D., & Fittoria, M. G. (2014). A typology of interpartner conflict and maternal parenting practices in high-risk families: Examining spillover and compensatory models and implications for child adjustment. Development and Psychopathology,26(4 Pt 1), 983–998. Tabachnick, B., & Fidell, L. (2007). Using multivariate statistics (5th ed.). Boston, MA: Allyn & Bacon. Toth, S. L., & Cicchetti, D. (2011). Frontiers in translational research on trauma. Development and Psychopathology, 23(2), 353–355. Trickey, D., Siddaway, A. P., Meiser-Stedman, R., Serpell, L., & Field, A. P. (2012). A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents. Clinical Psychology Review,32(2), 122–138. Turunen, T., Haravuori, H., Punamaki, R. L., Suomalainen, L., & Marttunen, M. (2014). The role of attachment in recovery after a school-shooting trauma. European Journal of Psychotraumatology,5, 22728. UN. (2009). Human rights in Palestine and other occupied Arab territories: Report of the United Nations fact-finding mission on the Gaza conflict. New York, NY: UN General Assembly. Retrieved from http://www2.ohchrorg/english/ bodies/hrcouncil/docs/12session/A-HRC-12-48.pdf UN-OCHA. (2009). Locked in: the humanitarian impact of the two year blockade on the gaza strip. East Jerusalem, UN: Office for the Coordination of Humanitarian Affairs occupied Palestinian territory. Van Ee, E., Kleber, R. J., Jongmans, M. J., Mooren, T. T., & Out, D. (2016). Parental PTSD, adverse parenting and child attachment in a refugee sample. Attachment & Human Development,18(3), 273–291. Van IJzendoorn, M, & Bakermans-Kranenburg, M. J. (1996). Attachment representations in mothers, fathers, adolescents, and clinical groups: a meta-analytic search for normative data. Journal Of Consulting And Clinical Psychology,64(1), 8–14. Van Ijzendoorn, M. H., Bakermans-Kranenburg, M. J., & Sagi-Schwartz, A. (2003). Are children of Holocaust survivors less well-adapted? A meta-analytic investigation of secondary traumatization. Journal of Traumatic Stress,16(5), 459–469. Walsh, F. (2007). Traumatic loss and major disasters: strengthening family and community resilience. Family Process,46(2), 207–227. Yehuda, R., & Bierer, L. M. (2008). Transgenerational transmission of cortisol and PTSD risk. Progress in Brain Research,167, 121–135. Zerach, G., Greene, T., Ein-Dor, T., & Solomon, Z. (2012). The relationship between posttraumatic stress disorder symptoms and paternal parenting of adult children among ex-prisoners of war: A longitudinal study. Journal of Family Psychology,26(2), 274–284. EUROPEAN JOURNAL OF PSYCHOTRAUMATOLOGY 15