M-quantile regression for multivariate longitudinal data: analysis of the Millennium Cohort Study data

Abstract: We propose a M-quantile regression model for the analysis of multivariate, continuous, longitudinal data. M-quantile regression represents an appealing alternative to standard regression models, as it combines the robustness of quantile and the efficiency of expectile regression, providing a complete picture of the response variable distribution. Discrete, individual-specific, random parameters are used to account for both dependence within the same response recorded at different times and association between different responses observed on the same sample unit at a given time. A suitable parametrisation is also introduced in the linear predictor to account for possible dependence between the individual specific random parameters and the vector of observed covariates, that is to account for endogeneity of some covariates. An extended EM algorithm is proposed to derive model parameter estimates under a maximum likelihood approach. The model is applied to the analysis of the strengths and difficulties questionnaire scores from the Millennium Cohort Study in the UK.