Multivariate Bayesian spatial modelling of child labour and out of school children in Nigeria: mapping areas of excess risk

This research implemented a Bayesian multivariate spatial model to identify areas where the risk of children being out of school, engaging in child labour, or undertaking hazardous work differs from the risk of poverty in Nigeria. Using recently released geo-coded UNICEF MICS survey data, the results showed how the spatial distribution of the three outcomes studied differ and highlighted specific areas where the excess risk for each outcome, after accounting for poverty, is high. The model was also used to predict the risk of each outcome across the whole of Nigeria at a high spatial resolution.