Methods for analysing the relationship between poverty, parental work intensity, child emotional symptoms and conduct problems over time

Abstract

This article exposes the methods employed to analyse the complex associations between poverty and work intensity over time on the longitudinal trajectories of mental health wellbeing in a cohort of children. This study used data from nine waves of birth cohort 1 of the Growing Up in Scotland (GUS) study (2005/06–2017/18) to fit a bivariate multilevel non-linear growth curve model for the change in conduct problems and emotional symptoms of children over time with the trajectories of poverty and parental work intensity over time as the main covariates of interest. We explain in detail: (a) how we arrive at valid measures for our outcome of interest by testing for longitudinal measurement invariance and (b) the principled approach of growth mixture modelling undertaken to derive our main covariates of interest. Both procedures are the preamble for the main model of interest that addresses the substantive research question of how changes over time in poverty and parental employment are associated with changes over time in children's wellbeing.•We expose the rationale behind and the procedures for implementing Longitudinal Measurement Invariance testing for the repeated measures of emotional and conduct problems.•We expose the rationale behind and the procedures for implementing a growth mixture modelling approach to derive longitudinal measures of poverty and work intensity.•We provide details of the bivariate growth curve model fitted to analyse the effect of the derived longitudinal measures of poverty and work intensity on the valid longitudinal measures of emotional and conduct problems.