Background/Aims: Understanding predictors, moderators and mediators of child externalising behaviour problems could lead to improvements in engagement and outcomes for children and their families. Parental cognitions, including self-efficacy and attributions, have been proposed as mediators in the relationship between parenting and child behaviour problems. Although mediation is increasingly recognised as an effective way of examining relationships between variables, only a small proportion of studies that identify potential mediators actually conduct a mediation analysis. This thesis consists of two studies: a systematic review (Journal Article 1) and an empirical study (Journal Article 2). The systematic review investigated the ways in which mediation analysis has been used to examine the role of parental cognitions in studies of child behavioural problems, and to assess the methodological quality of these studies. The empirical study investigated the role of parental cognitions and characteristics in relation to pre-school child externalising behaviour problems. Methods: In Journal Article 1, a systematic search of three electronic databases, a quality assessment of included studies, and a subsequent narrative synthesis were conducted. In Journal Article 2, 125 parents of children aged 3-6 years old attending a community-based parent management training programme, across three local authorities, completed a battery of self-report questionnaires before and after the intervention. Correlational and mediation analyses were conducted to investigate relationships between child behaviour and parental attachment style, metacognition, dysfunctional attributions and parental stress. We also tested the possibility that parents' reported levels of stress and child behaviour problems, and their demographic variables, played a role in whether they completed the intervention. Results: In Journal Article 1, after screening, 14 studies were reviewed using an adapted quality criteria tool. The most commonly studied parental cognition was parental self-efficacy, with a small number of studies investigating parental attributions. A variety of approaches to mediation analysis had been used and caution should be exercised when interpreting the results of many of the reviewed studies. Despite a growing recognition of the limitations of some traditional methods (e.g. causal steps approach), research into mediators of child externalising behaviour could be improved by a wider adoption of more appropriate tools, in line appropriate theoretical frameworks. In Journal Article 2, as hypothesised, the results indicated significant relationships between parents' attachment insecurity and baseline levels of parental stress, parental attributions and child behaviour problems. Support was found for the hypothesis that parental attributions mediated the relationship between attachment insecurity and child externalising behaviour problems. We did not find significant that any demographic variables other than parent age predicted whether parents completed the programme. Conclusions: Taken together, the two studies provide evidence of a complex relationship between parental factors, particularly parental cognitions, and externalising child behaviour problems. The systematic review found some evidence that parental cognitions mediate how aspects of parenting (e.g. behaviour and affect) and child externalising behaviour problems are associated, and the empirical study also showed that parental attributions are important in relation to child behaviour problems. Of particular interest was the finding that they mediate the relationship between child behaviour problems and attachment insecurity. However, to advance the field both theoretically and clinically, future studies should endeavour to ensure adequate sample size and power, using optimal study designs, in conjunction with strong theoretical grounding. Exploring cognitive mediators beyond self-efficacy, such as parental attributions, will allow us to further develop our understanding of the relationship between child behaviour and parenting.

Background: Children who enter foster care usually do so because of maltreatment by their birth families. Early adversity such as this is associated with many negative outcomes, including disturbances of attachment and mental health in childhood and throughout the lifespan. Young children (under 5) are particularly at risk due to maltreatment rates being highest in this age range, and because of the vital brain development that occurs during this time. Improving the quality of existing relationships for young children is the most cost effective way to improve mental health outcomes. It is important that research investigates which relational and psychological variables that exist within the foster carer-child relationship may be protective against developing negative mental health outcomes, so as to inform carer training and future interventions for this vulnerable group. Aims: The aims of this research project were twofold. The first aim was to systematically review the existing literature on links between foster carer psychological variables (such as commitment to their foster child), and/or child psychological variables (such as their attachment style), and the mental health outcomes of children in foster care. The second aim was to investigate whether foster carer acceptance, commitment, awareness of influence and reflective functioning (RF) predict the mental health outcomes of Scottish preschool aged children who are looked after in foster care. Method: A systematic review of the existing literature was undertaken to address the first aim. The search strategy resulted in 12 quantitative studies that investigated links between child or carer psychological variables and child mental health outcomes. An empirical study of 179 pre-school aged children in foster care in Scotland was carried out to address the second aim. Participants were taking part in a wider RCT of a novel intervention to improve outcomes and permanency decisions for children in foster care. Foster carer acceptance, commitment, and awareness of influence was assessed using the This Is My Baby Interview, and scores of RF were coded from the transcripts of this interview using a computer-based algorithm. Child mental health information was gathered using the Infant Toddler Social Emotional Assessment. Data was gathered at 2 time points; baseline assessments occurred around 4 weeks after entry to care, and follow-up assessments were carried out a year later. Results: The systematic review found good evidence that foster child attachment security is linked to more positive mental health outcomes. It also found some evidence suggesting that foster carer psychological variables such as commitment and quality of caregiving also relate to child mental health outcomes, but this research is in its infancy and it is therefore difficult to draw firm conclusions around this. The results of the empirical study showed that carer commitment and awareness of influence predict child competence at baseline, and RF predicts internalising and externalising problems at follow up. No predictive relationships were found between carer variables and child mental health over time. Conclusion: The results from both studies suggest that carer psychological variables such as commitment to their foster child may relate to child mental health development. These results have implications in terms of foster carer training, and for intervention development for this vulnerable population. This research is however in its infancy, and the results suggest a complex picture with regard to carer psychological variables and child mental health. Large-scale high quality longitudinal research is needed to provide a clear understanding of these relationships.

With a significant number of African American single-parent families responsible for raising a generation of male children, the focus of this qualitative case study was on exploring the African American single mother-son dyad to identify metacommunicative signals delivered from mothers to sons. This study was grounded in a theoretical framework combining attachment theory and social learning theory. The research questions focused on identifying metacommunication messages passed from mothers to sons and how metacommunication patterns influence the youth's social identity. Four single mothers with adolescent sons and 4 unrelated adult sons of single mothers participated in semistructured interviews. Data were collected and analyzed using content analysis and coding supported with NVivo software. Key findings revealed that the metacommunication was a dominant form of communication in the African-American family construct, and affected the parenting styles. From the mother's retrospective reports, African-American mother's adapted an authoritarian or helicopter parenting styles to control and protect their sons from racism, becoming victims of crime and violence, being arrested, or incarcerated. The key finding from the sons' retrospective reports was that negative metacommunication from single mothers to sons was associated with insecure attachment, avoidance, and risky behaviors. The implications for social change are that positive metacommunication can strengthen the African American single mother-son dyad. This information may lead to intervention strategies for targeting negative metacommunication patterns from African American single mothers to sons and teaching new communication rules that foster a secure relationship.