Institute of Education London

Children's psychological adjustment and father's residence parenting traits

Summary
This research project uses data from the first two sweeps of the Millennium Cohort Study (MCS) (www.cls.ioe.ac.uk) to, in particular, explore bidirectional links between child's behaviour (measured at 9 months with the Revised Carey Infant Temperament Scale) and father's transitions in and out of residence, and it investigates the relationship between early father's parenting and later child's behaviour (measured with the Strengths and Difficulties Questionnaire at 3 years) and the relationship between early child's behaviour (at 9 months) and later father's parenting (at age 3 years) both in resident and in non-resident father families. In addition, it explores both the direct effect of father's mental health on child's behaviour at age 3 years and the effect of the interaction between father's and mother's mental health on later child's behaviour, and it models the family-level mental health effect on child's behaviour in two parent families.

Methods
To account for the complex sampling design of the MCS we considered the substantial work carried out on the estimation of multilevel models under a complex survey design. A popular approach to accounting for a complex survey design is by incorporating the survey weights in the statistical analysis. Estimating the parameters of a regression model is then achieved using a pseudo-maximum likelihood approach. An alternative approach is by including in the model the design variables, i.e. the variables used at the planning stage of the survey, as explanatory variables. The inherent assumption in this approach is that conditioning on the design variables the sampling mechanism is ignorable. For this reason we include the nine MCS stratification (or design) variables as explanatory variables in our models. Confirmatory Factor Analysis (CFA) is carried out to determine the structural validity of our measures. The research questions are answered using Structural Equation Models (SEM) in which relationships between latent variables are modelled. Additive and multiplicative effects are investigated using the latent interaction (XWITH) function, and dyad models (partners nested within couples) using the two-level function in MPLUS

References
McEwen, C. & Flouri, E. (in press). Fathers' parenting, adverse life events, and adolescents' emotional and eating disorder symptoms: The role of emotion regulation. European Child & Adolescent Psychiatry

Flouri, E. & Hall, J. (in press). Adverse life events and adolescent psychopathology: Father's and mother's parenting as mediators. In M. Malikiosi-Loizos & A. Papastylianou (Eds.) Athens: Hellinica Grammata